BIOMECHANICS OF HEADING IN YOUTH SOCCER

by

ERIN HANLON

DISSERTATION

Submitted to the Graduate School

of Wayne State University,

Detroit, Michigan

in partial fulfillment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY

2009

MAJOR: BIOMEDICAL ENGINEERING Approved by: Advisor Date

UMI Number: 3387316

All rights reserved

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript

and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion.

UMI 3387316

Copyright 2010 by ProQuest LLC. All rights reserved. This edition of the work is protected against

unauthorized copying under Title 17, United States Code.

ProQuest LLC 789 East Eisenhower Parkway

P.O. Box 1346 Ann Arbor, MI 48106-1346

© COPYRIGHT BY

ERIN HANLON

2009

All Rights Reserved

ii

DEDICATION

For my parents, for all of their love, support, and patience

iii

ACKNOWLEDGMENTS

This work was supported by in part by the National Operating Committee

for Standards in Athletic Equipment (NOCSAE). In addition to funding provided

by NOCSAE the author would like to acknowledge partial support from the

Anthony and Joyce Danielski Kales Scholarship. The author would like to thank

her advisor, Dr. Cynthia Bir, and her committee members, Dr. John Cavanaugh,

Dr. Pamela VandeVord, and Dr. Kenneth Podell for all of their guidance, support,

and expertise.

The author would also like to thank the Sports and Ballistics group for all

of their assistance with test preparation, data collection, and insight. Specifically,

the author would like to thank Charlene Brain, Sarah Stojsih, Demario Tucker,

and Jacob Mack for assistance with data collection and to Jonathan Beckwith for

assistance with data processing. Thank you to Nathan Dau and Donald

Sherman for input on testing methods and test setup. Thank you to Amanda

Esquivel and James Kopacz for assistance with subject recruitment. Thank you

to all athletes for taking time to participate in the study.

Finally, the author would like to thank her friends and family for their

support for the duration of this process. Thank you to my Mom and Dad, Brian,

and Alison. Thank you for your support and, most importantly, your patience.

TABLE OF CONTENTS

DEDICATION ........................................................................................................ ii

ACKNOWLEDGMENTS ...................................................................................... iii

LIST OF TABLES ................................................................................................. iv

LIST OF FIGURES ............................................................................................... v

CHAPTER 1 - INTRODUCTION ........................................................................... 1

1.1 Statement of the Problem ..................................................................... 1

1.2 Background and Significance ............................................................... 2

1.3 Specific Aims ...................................................................................... 15

CHAPTER 2 - NEISS DATABASE ...................................................................... 17

2.1 Introduction ........................................................................................ 17

2.2 Methodology ....................................................................................... 23

2.3 Results ............................................................................................... 24

2.4 Discussion .......................................................................................... 29

CHAPTER 3 - HEADING FREQUENCY IN YOUTH SOCCER .......................... 33

3.1 Introduction ........................................................................................ 33

3.2 Methodology ....................................................................................... 37

3.3 Results ............................................................................................... 40

3.4 Discussion .......................................................................................... 47

CHAPTER 4 - HEADING BIOMECHANICS IN YOUTH SOCCER ..................... 50

4.1 Introduction ........................................................................................ 50

4.2 Methodology ....................................................................................... 54

4.3 Results ............................................................................................... 60

4.4 Discussion .......................................................................................... 86

CHAPTER 5 - ACCELERATION MEASUREMENT SYSTEM VALIDATION ...... 90

5.1 Introduction ........................................................................................ 90

5.2 Methodology ....................................................................................... 93

5.3 Results ............................................................................................... 99

5.4 Discussion ........................................................................................ 110

CHAPTER 6 - ON FIELD MEASUREMENT OF HEAD ACCELERATION ........ 115

6.1 Introduction ...................................................................................... 115

6.2 Methodology ..................................................................................... 120

6.3 Results ............................................................................................. 122

6.4 Discussion ........................................................................................ 132

CHAPTER 7 - CONCLUSIONS AND FUTURE RECOMMENDATIONS .......... 137

7.1 Conclusions ...................................................................................... 137

7.2 Future Recommendations ................................................................ 141

APPENDIX A – HIC APPROVALS .................................................................... 143

ABSTRACT ....................................................................................................... 160

BIOGRAPHICAL STATEMENT ........................................................................ 162

iv

LIST OF TABLES

Table 1.1: Player Symptoms Following Soccer Heading .................................... 5

Table 2.1: Head injuries by mechanism ............................................................ 26

Table 2.2: Ball to head only injuries ................................................................... 27 Table 3.1: Number of games monitored outlined by age, gender and division of

play ..................................................................................................................... 38

Table 3.2: Maximum headers by any one player for a single game ................... 40

Table 3.3: Average total headers/game/team for male population ..................... 43 Table 3.4: Average total headers/game/team for female population .................. 44 Table 3.5: Total headers in each field position for females ................................ 45 Table 3.6: Total headers in each field position for males ................................... 46 Table 4.1: Heading Scenarios ............................................................................ 59

Table 4.2: Average angles at impact for each heading task .............................. 78 Table 4.3: Mean peak RMS EMG values for each muscle and each task ......... 82

Table 4.4: Average head acceleration on impact for each heading task ........... 85

Table 4.5: Average head flexion for each task ................................................... 87 Table 5.1: Average Peak Linear Accelerations for Ball to Head Conditions ..... 104 Table 5.2: Average Peak Angular Accelerations for Ball to Head Conditions .. 105

Table 5.3: Average Peak Linear Accelerations for Head to Head Conditions .. 106 Table 5.4: Average Peak Angular Accelerations for Head to Head Conditions 106

Table 6.1: Average results for headers by location .......................................... 125

Table 6.2: Description of non header impacts and the player that impacted .... 128

v

LIST OF FIGURES Figure 2.1: NEISS database hospitals by strata ............................................... 18

Figure 2.2: Soccer head injuries including both males and females .................. 25

Figure 2.3: Male and female ball only head injuries by age ............................... 28

Figure 2.4: Ball only head injuries by diagnosis ................................................. 29

Figure 3.1: Soccer Field Diagram ..................................................................... 39

Figure 3.2: Comparison of number of headers/minute, male versus female

across age groups .............................................................................................. 41

Figure 3.3: Number of headers/minute for the male population ......................... 42

Figure 3.4: Number of headers/minute for the female population ...................... 45

Figure 3.5: Total number of headers in each field position ............................... 47

Figure 4.1: FAB System with size scale ............................................................. 53

Figure 4.2: Example of player wearing FAB sensors ......................................... 55

Figure 4.3: Side view of head and trunk body angles ........................................ 56

Figure 4.4: Top view of head rotation ................................................................. 56

Figure 4.5: Neck musculature used for EMG testing a) sternocleidomastoid; b)

trapezius ............................................................................................................. 57

Figure 4.6: Torso flexion for all male players during task 2 ............................... 61

Figure 4.7: Head flexion for all male players during task 2 ............................... 62

Figure 4.8: Head rotation for all males during task 2 .......................................... 63

Figure 4.9: Torso flexion for all females during task 2 ........................................ 64

Figure 4.10: Head flexion for all females during task 2 ..................................... 64

Figure 4.11: Head rotation for all females during task 2 .................................... 65

vi

Figure 4.12: Example of single male participant’s torso flexion for all header

tasks that lack modifications (1, 2, 3, and 7) ...................................................... 66

Figure 4.13: Example of single male participant’s head flexion for all header tasks that lack modifications (1, 2, 3, and 7) ....................................................... 66

Figure 4.14: Example of single male participant’s head rotation for all header

tasks that lack modifications (1, 2, 3, and 7) ...................................................... 67

Figure 4.15: Example of single female participant’s torso flexion for all header

tasks that lack modifications (1, 2, 3, and 7) ...................................................... 68

Figure 4.16: Example of single female participant’s head flexion for all header

tasks that lack modifications (1, 2, 3, and 7) ...................................................... 68

Figure 4.17: Example of single female participant’s head rotation for all header

tasks that lack modifications (1, 2, 3, and 7) ...................................................... 69

Figure 4.18: Example of single male participant’s torso flexion for all passing

header tasks (2, 4, 5, and 6) ............................................................................... 70

Figure 4.19: Example of single male participant’s head flexion for all passing

header tasks (2, 4, 5, and 6) ............................................................................... 71

Figure 4.20: Example of single male participant’s head rotation for all passing

header tasks (2, 4, 5, and 6) ............................................................................... 71

Figure 4.21: Example of single female participant’s torso flexion for all passing

header tasks (2, 4, 5, and 6) ............................................................................... 72

Figure 4.22: Example of single female participant’s head flexion for all passing

header tasks (2, 4, 5, and 6) .............................................................................. 73

vii

Figure 4.23: Example of single female participant’s head rotation for all passing

header tasks (2, 4, 5, and 6) ............................................................................... 73

Figure 4.24: Example of single male participant’s torso flexion for all clearing

header tasks (7 and 8) ........................................................................................ 74

Figure 4.25: Example of single male participant’s head flexion for all clearing

header tasks (7 and 8) ....................................................................................... 75

Figure 4.26: Example of single male participant’s head rotation for all clearing

header tasks (7 and 8) ....................................................................................... 75

Figure 4.27: Example of single female participant’s torso flexion for all clearing

header tasks (7 and 8) ....................................................................................... 76

Figure 4.28: Example of single female participant’s head flexion for all clearing

header tasks (7 and 8) ....................................................................................... 76

Figure 4.29: Example of single male participant’s head rotation for all clearing

header tasks (7 and 8) ....................................................................................... 77

Figure 4.30: Peak EMG for all male players for each muscle a) left

sternocleidomastoid, b) right sternocleidomastoid, c) left trapezius, d) right

trapezius ............................................................................................................ 80

Figure 4.31: Peak EMG for all female players for each muscle a) left

sternocleidomastoid, b) right sternocleidomastoid, c) left trapezius, d) right

trapezius ............................................................................................................ 81

Figure 4.32: Sample RMS EMG for one male player for each muscle a) left

sternocleidomastoid, b) right sternocleidomastoid, c) left trapezius, d) right

trapezius ............................................................................................................ 83

viii

Figure 4.33: Sample RMS EMG for one female player for each muscle a) left

sternocleidomastoid, b) right sternocleidomastoid, c) left trapezius, d) right

trapezius ............................................................................................................ 84

Figure 5.1: HIT system ...................................................................................... 91

Figure 5.2: Back of HITS headband with circles marking accelerometer

placement .......................................................................................................... 94

Figure 5.3: Air cannon with soccer barrel ........................................................... 96

Figure 5.4: Head to head impact test setup for forehead testing ...................... 97

Figure 5.5: Linear regression of linear acceleration for HIII and HITS ball to head

conditions ......................................................................................................... 100

Figure 5.6: Linear regression of angular acceleration for HIII and HITS ball to

head conditions ................................................................................................ 100

Figure 5.7: Linear regression of linear acceleration for HIII and HITS head to

head conditions ................................................................................................ 101

Figure 5.8: Linear regression of angular acceleration for HIII and HITS head to

head conditions ................................................................................................ 102

Figure 5.9: Linear regression of linear acceleration for HIII and HITS ball to head

and head to head conditions combined ........................................................... 103

Figure 5.10: Linear regression of angular acceleration for HIII and HITS ball to

head and head to head conditions combined .................................................. 103

Figure 5.11: Linear acceleration for both HIII and HITS for one ball to head

forehead impact at the 12 m/s condition .......................................................... 107

ix

Figure 5.12: Linear acceleration for both HIII and HITS for one ball to head right

side impact at the 12 m/s condition .................................................................. 108

Figure 5.13: Linear acceleration for both HIII and HITS for one ball to head left

temple impact at the 12 m/s condition .............................................................. 108

Figure 5.14: Linear acceleration for both HIII and HITS for one head to head

forehead impact at the 4.75 m/s condition ....................................................... 109

Figure 5.15: Linear acceleration for both HIII and HITS for one head to head left

side impact at the 4.75 m/s condition ............................................................... 109

Figure 6.1: Wayne State University Tolerance Curve ...................................... 118

Figure 6.2: HITS headgear fitted to HIII headform .......................................... 121

Figure 6.3: Linear head acceleration by location for each header only impacts

.......................................................................................................................... 122

Figure 6.4: Angular head acceleration by location for each header only impacts

.......................................................................................................................... 123

Figure 6.5: HIC values for headers by location with mTBI tolerance level ...... 124

Figure 6.6: Linear head acceleration for header impacts for individual players

.......................................................................................................................... 126

Figure 6.7: Angular head acceleration for all header impacts for individual

players ............................................................................................................. 126

Figure 6.8: Linear head acceleration for all non header impacts by location ... 129

Figure 6.9: Angular head acceleration for all non header impacts by location 130

Figure 6.10: HIC for all non header impacts by location ................................. 130

x

Figure 6.11: Linear head acceleration for all non header impacts for individual

players ............................................................................................................. 131

Figure 6.12: Angular head acceleration for all non header impacts for individual

players ............................................................................................................. 132

1

CHAPTER 1

INTRODUCTION

1.1 Statement of the Problem

Soccer is one of the most popular sports throughout the world; Fédération

Internationale de Football Association (FIFA) has approximately 200 million

registered players worldwide (Dvorak and Junge, 2000). A large increase in

participation has taken place recently in the United States. This is seen clearly in

the American Youth Soccer Organization (2006) which today has 50,000 youth

soccer teams and over 650,000 players registered after starting out in 1964 with

only nine teams (2006). Unfortunately, the increase in youth players has caused

an increase in injuries (Metzl, 1999).

Soccer injuries not only cause trauma to the player, they also create a large

socioeconomic cost. Dvorak et al. (2000) reviewed relevant soccer injury data

and determined that approximately $30 billion dollars are spent annually for

treatment of soccer related injuries worldwide. While this value uses an average

injury rate, it also uses a conservative estimate of cost per injury (Dvorak and

Junge, 2000). As pointed out by Dvorak et al. (2000), this cost estimate does not

take into account lost wages or anything not directly related to primary medical

costs.

Head injuries are of particular concern due to their traumatic nature and the

lack of knowledge related to these injuries and their mechanisms, specifically

mild traumatic brain injury (mTBI). Head injuries represent up to 22% of all

soccer injuries (Ruchinskas, et al., 1997). A unique aspect of soccer is that there

2

are both intentional and unintentional head impacts. Intentional impacts, or

“heading” the ball, occur when a player purposefully uses their head to redirect

the ball. Unintentional impacts include: player to player impacts, player to

ground impacts, player to goalpost impacts, and unintentional player to ball

impacts.

The purpose of this study is to investigate the effect of the intentional head

impacts that occur during soccer play. Initial steps will be taken to determine the

frequency and severity of heading episodes in the field using both field

observation and a novel head band measurement system developed for use in

soccer play. Additionally, many laboratory studies have been performed on the

adult soccer population, but very few have focused on children. Therefore, an

analysis of the biomechanics of heading in youth soccer needs to be performed.

Comparisons will be made between youth and adult heading biomechanics to

determine if there is a difference based on age. Also, a comparison between the

youth heading biomechanics and the youth heading field data will be made. This

will provide valuable information as to whether the laboratory data collected in

previous studies is representative of actual on-field situations.

1.2 Background and Significance

Injury Epidemiology

Due to the increase in soccer participation, more injuries are occurring

(Metzl, 1999). Although the majority of soccer injuries are to the lower extremity

(Agel, et al., 2007, Arnason, et al., 2004, Backous, et al., 1988, Dvorak and

Junge, 2000, Elias, 2001, Junge and Dvorak, 2007, Keller, et al., 1987, Le Gall,

3

et al., 2008, Leininger, et al., 2007, Nielsen and Yde, 1989, Peterson, et al.,

2000, Poulsen, et al., 1991, Sandelin, et al., 1985, Schmidt-Olsen, et al., 1985),

head injuries are of specific concern due to the potential for long-term debilitating

effects. The American Academy of Pediatrics has classified soccer as a contact

sport, but many still believe that soccer is a safe alternative to American Football

(Patlak, et al., 2002). This may not be true when considering the previously

established concussion rates (Green and Jordan, 1998). When studying sports

in the National Collegiate Athletic Association (NCAA) Green et al. (1998) found

similar concussion rates in American football and men’s and women’s soccer.

Using the NCAA Injury Surveillance System (ISS) to determine the incidence of

concussion in various sports (Green and Jordan, 1998). Concussions were

measured per 1000 athlete exposures, where each exposure is the equivalent of

one practice or game. They found that women’s soccer actually has a higher

rate of concussion (.33 concussions/1000 athlete exposures) than both men’s

soccer (.31 concussions/1000 athlete exposures) and football (.31

concussions/1000 athlete exposures). This demonstrates the high risk of head

injury during soccer play.

Covassin et al. (2003) found that men’s and women’s soccer are in the

group of athletes at the highest risk for concussions (Covassin, et al., 2003).

Women’s soccer had the highest number and injury rate of concussions of the 15

NCAA women’s sports included in the study indicating that further research into

the area is necessary (Covassin, et al., 2003). Small stature, a greater ball-to-

head ratio, and potentially weaker neck muscles have been suggested as the

4

potential causes (Covassin, et al., 2003). All of these issues would also be

applicable to the youth population in addition to their lower skill level and lack of

experience.

Keller et al. (1987) found that younger players generally have a higher rate

of head and face injury (Keller, et al., 1987). They attributed this fact to a lack of

heading proficiency and the increase in ball to head weight ratio. Both of these

factors would indicate heading as a potential problem in youth soccer. This is of

significant importance because it has been indicated that younger athletes may

take longer to regain baseline neuropsychological levels following a concussion

(Field, et al., 2003).

Barnes et al. (1998) interviewed 144 elite soccer players in the 1993 US

Olympic Festival. Players were asked to estimate the number of times that they

head the ball during games and during practices. They were also asked to list

any symptoms that they had experienced as a result of heading and if they had

any previous head injuries. Sixty-five women aged 17-30 and 72 men aged 17-

22 completed the surveys. Using an odds ratio of the total concussions for men

and women it was found that men had an increased risk of concussion of 2.16

times that of women. Seventy-four concussions were reported in men and 28 in

women with 27 players reporting multiple concussions, 24 men and 3 women.

That indicates that 52% of players interviewed had experienced a concussion.

Of these concussions, 18% reported the mechanism of injury as collision with the

ball. Additionally 89% of men and 43% of women had previously had some type

of acute head injury during their soccer careers. Many of the players had

5

symptoms following heading the soccer ball (Table 1.1) with headache being the

highest reported symptom (Barnes, et al., 1998). These symptoms indicate that,

at a minimum, soccer heading creates a short-term problem.

Table 1.1: Player Symptoms Following Soccer Heading (Barnes, et al., 1998)

Symptom Men (%) Women (%)

Headaches 54.0 55.0

Dazed 31.0 49.0

Dizziness 18.1 38.5

Decreased Concentration 9.7 10.8

Blurred Vision 11.1 4.6

Lost Conciousness 1.4 0.0

Numbness/Tingling 12.8 7.7

Amnesia 0.0 3.1

In another study looking at concussion in soccer, Boden et al. (1998)

found similar concussion rates as Barnes et al. (Barnes, et al., 1998, Boden, et

al., 1998). Athletic trainers for each of the 15 Atlantic Coast Conference (ACC)

men’s and women’s soccer teams were asked to fill out a questionnaire for each

concussion occurring during the 1995 and 1996 seasons. In the 1995 season,

188 women and 162 men competed in ACC soccer, and in the 1996 season

there were 188 women and 163 men. During these two seasons 29 concussions

occurred in 26 players. This resulted in a concussion rate of 0.49

concussions/1000 athlete exposures, 0.6 concussions/1000 male athlete

exposures and 0.4 concussions/1000 female athlete exposures. Contact with the

6

ball was the second most common injury mechanism (24 %), but no injuries were

attributed to intentional heading. The concussion rates in this study were

significantly higher than those found in the NCAA data. This could be due to the

different levels of play in the NCAA with this conference being in the highest

division, Division I. One problem when comparing these studies is the use of

differing definitions of concussion.

Sandelin et al. (1985) used insurance reports to determine soccer related

injuries for 1980 in Finland. After eliminating exertion injuries, researchers

determined that 2072 soccer injuries occurred that year, with 13 % of these

located in the head and neck region. Another significant finding in this study was

the lack of differences between the genders and positions played, but there were

significantly more injuries in the two highest skill divisions (Sandelin, et al., 1985).

The results also coincide with Barnes et al. (1998) that gender was not a

significant factor in head injury occurrence (Barnes, et al., 1998). This study

included youth players, but the average age of those included was 26 for the

men and 23 for women. It was not determined whether or not age was a

significant factor in injury occurrence.

Mechanism of soccer related head injury has been studied previously in

adults (Agel, et al., 2007, Andersen, et al., 2004, Boden, et al., 1998, Dick, et al.,

2007, Dvorak, et al., 2007, Fuller, et al., 2005), but many of these studies have

significant limitations when looking strictly at injuries related to heading the

soccer ball. One significant problem is the lack of a set injury definition which is

required when comparing literature. Many studies use definitions that require a

7

loss of practice or play in order to be quantified as an injury which generally

eliminates those players that are just having post-heading symptoms. This

definition creates a problem in tabulating the total number of people that have

symptoms following heading a soccer ball.

Soccer Heading Studies

Research has been performed in the area of soccer heading and its

effects on players’ neuropsychological abilities and mental imaging scans

(Guskiewicz, 2002, Janda, et al., 2002, Tysvaer and Storli, 1981, Tysvaer and

Storli, 1989, Tysvaer, et al., 1989). These previous studies have provided very

contrasting results in which no clear understanding of what is occurring during or

following heading can be made. Without understanding what happens

biomechanically during heading events on the field, there is no way to determine

injury risk. The majority of the previous research has been performed on adults,

some of which are retired soccer players. Very little focus has been on the effect

of heading in the youth population, and there has still been no biomechanical

assessment of youth soccer players heading the soccer ball in the lab.

Additionally, previous studies have not isolated heading from other potential

deficit causes such as alcohol use and previous non-soccer related head injury.

Therefore, further research needs to be done to look at the effect heading has on

the youth population.

Although much research has been devoted to the adult population,

especially the elite players, the risks to the youth population have not been

studied in great detail. In fact, the risks to children are potentially greater. This is

8

primarily due to their size versus the force being applied by the ball (Lees and

Nolan, 1998). It has been reported that ball mass, impact velocity, and size of

the individual all contribute to the potential for injury (Lees and Nolan, 1998).

Additionally, the importance of proper technique may be especially true in the

youth population, since their skill level has not been well developed to control

their head motion when heading the ball. Therefore, the youth population could

be at an increased risk for sustaining head injuries due to rotational acceleration.

Given the increase in youth soccer participation over the past decade and

additional injury risk, it is necessary to focus research on the youth population.

Many of the earlier studies indicated that repeatedly heading the soccer

ball increased players’ risk for neuropsychological deficits and long-term

symptoms. Tysvaer et al. (Tysvaer and Storli, 1981, Tysvaer and Storli, 1989,

Tysvaer, et al., 1989) was one of the first researchers to investigate this

occurrence. In the first study, Tysvaer et al. (1981) studied 128 retired

Norwegian soccer players with an average age of 25 years. Questionnaires

were sent to each player asking about their previous soccer play. Of the 128

players, 64 had once had symptoms related to heading the soccer ball, some of

which required hospitalization. The study does not provide information on the

number of headers to which each player was subjected. Therefore, no

association between the number of headers and injury can be made.

Two additional studies were done using EEG and neuropsychological

testing to determine if active and retired soccer players displayed deficits related

to their soccer play (Tysvaer and Storli, 1989, Tysvaer, et al., 1989). Sixty-nine

9

active players were compared to controls while a parallel study of 37 retired

players used the same tests to determine long-term effects. Eighty-one percent

of the players showed a deficit in neuropsychological tests ranging from mild to

severe (Tysvaer, et al., 1989). For the study involving active players, it was

found that there was an increase in abnormal EEG findings in soccer players with

respect to controls. It was also determined that the highest abnormal findings

were in the younger players (Tysvaer and Storli, 1989).

Neuropsychological testing has been the basis for many of the previous

studies. In one such study Guskiewicz et al. (2002) studied soccer heading in

players playing in the NCAA. Six neurocognitive tests were performed on 91

soccer players, 96 other athletes, and 53 control subjects. The battery of tests

included the Trail Making Test, the Controlled Oral Word Association Test, the

Stroop Color Word Test, the Hopkins Verbal Learning Test, the Symbol Digit

Modalities Test, and the Wechsler Digit Span Test. These tests evaluated a wide

range of cognitive abilities including: orientation, concentration, visuospatial

capacity, problem-solving, verbal associations, cognitive flexibility, attention

span, verbal memory, visual tracking, incidental learning, concentration, and

immediate memory. Scholastic Aptitude Test (SAT) scores were also evaluated.

Researchers determined that there was no actual significant difference

(Guskiewicz, 2002). One test (the Hopkins Verbal Learning Test) approached

statistical significance, but once previous concussions and learning disabilities

were controlled for, significance was not found. This test was specific for

immediate memory recall which could indicate that bouts of heading can cause

10

immediate deficits, but further research is needed to determine why these deficits

may exist. Additionally, an attempt to correlate soccer exposure to test

performance was conducted. Correlations were only made in the Wechsler Digit

Span Test. This study indicated that soccer exposure did not affect

neuropsychological ability.

Similar neuropsychological testing was conducted on 53 active

professional soccer team members from The Netherlands (Matser, et al., 1998).

Both players and 27 controls were interviewed and tested using a battery of

neuropsychological examinations. The tests included in the battery differed

slightly from those used by Guskiewicz et al. (2002) as did the results. Fourteen

tests were administered including: Raven Progressive Matrices Test, Wisconsin

Card Sorting Task, Paced Auditory Serial Addition Task, Digit Symbol Test, Trail

Making Test, Stroop Test, Bourdon-Wiersman Test, Wechsler Memory Scale,

Complex Figure Test, 15-Word Learning Test, Benton’s Facial Recognition Test,

Figure Detection Test, Verbal Fluency Test, and the Puncture Test. The number

of headers experienced by each player was estimated based on position played

and number of games. Soccer players exhibited impaired performance in

memory, planning and visuoperceptual processing in comparison to the controls

with the level of impairment related to the position and number of headers as well

as concussions. The authors suggest that these data may indicate professional

soccer’s connection to neurocognitive impairment (Matser, et al., 1998).

In a similar study, Putukian et al. (2004) corroborated these results by

studying Division I male and female college athletes. Athletes were studied

11

prospectively during two practice sessions and served “as their own controls”.

Neuropsychological testing was conducted before and after each practice

session with the number of headers monitored during the session. A practice

effect was noted between the pre- and post-test scoring for attention and

concentration. However, there was no significant difference between the header

and non-header groups in either the pre- or post-test scores (Putukian, 2004).

Therefore, further tests would be required to determine the cause of the deficits.

Janda et al. (2002) performed one of the few studies involving soccer

heading in the youth population. Fifty-seven youth soccer players from five

teams having an average age of 11.5 years were studied for three seasons.

Neurocognitive testing was used to determine the effects of heading. Four

cognitive tests were included in this study: Verbal Learning, Digit Span, Symbol

Digits Modality Test, and Verbal Learning Delay. Symptoms following heading

and heading exposure were also observed throughout the duration of the study

by the coaching staffs. All impacts were included in the data sheet returned to

the researchers by the coaching staffs including those that were unintentional.

Symptoms following heading included headaches, blurred vision, nausea, and

ringing in the ears with headaches being the most common. It was also

determined that there is an inverse relationship between increased heading and

verbal learning (Janda, et al., 2002).

In a recent study, Witol and Webbe (2003) studied 60 male players at

varying levels of skill with the youngest being high school students. Using a

control of 12 males who had never played soccer, neuropsychological testing

12

was administered to evaluate a number of functions including: abstract

reasoning, general intellectual function, attention, mental flexibility, information

processing, verbal and nonverbal memory. These tests included: Shipley

Institute of Living Scale, Trail Making Test, Paced Auditory Serial Addition Test,

Test of Facial Recognition, Rey Osterreith Complex Figure Test, and Rey

Auditory Verbal Learning Test. Players were asked to self-report their heading

exposure which allowed for a cumulative heading measure to be developed to

estimate career heading. Also, players were asked in their history about

symptoms during or following games. A decrease in the scales measuring

attention, concentration, cognitive flexibility and general intellectual functioning

correlated with an increase in the number of headers (Witol and Webbe, 2003)

which supports the conclusions of Janda et al. (2002). Researchers also found

that those players described as “typical headers” had significantly more dizziness

indicating that there are short term effects caused by performing soccer headers.

Stephens et al. (2005) performed a study on 23 youth soccer players, 23

youth rugby ranging in age from 13-16 years, and age matched controls for both

the contact sports who were participants in only non-contact sports. The battery

of 13 tests included: Rey Complex Figure, WAIS-R Digit Symbol, WAIS-R Digit

Span, Trail Making, Stroop, WMS-R Logical Memory Immediate and Delayed,

the Alertness, Divided Attention, Covert Attention Shift, Flexibility and Working

Memory subtests of Test of Attention Performance, Wisconsin Card Sorting, 64-

item version, and Alternate Uses. The study showed no significant difference

between soccer players and their controls. There was also no correlation found

13

between the number of headers and neuropsychological test results (Stephens,

et al., 2005).

Imaging has also been looked at as a method to detect injuries in both

current and retired soccer players, although to a more limited extent. Tysvaer et

al. (1989), in one of the earliest soccer heading studies, used EEG to detect

abnormalities, which although it is not necessarily an imaging technique, was the

first use of brain scans in the field (Tysvaer and Storli, 1989, Tysvaer, et al.,

1989). Since the earlier studies, imaging advances have been achieved and

researchers have used MRI and computer tomography (CT) scans (Jordan, et

al., 1996, Rutherford, et al., 2003, Sortland and Tysvaer, 1989). As with the

neuropsychological testing, the imaging studies have also returned conflicting

results.

Sortland et al. (1989) performed CT scans on former international soccer

players. Thirty-three retired players were given scans with nine of these being

categorized as “typical headers”. No definition was provided by the authors, but

both the player and their teammates had categorized themselves as “typical

headers”. Scans were evaluated visually and by taking linear measurements.

When compared to normative scans, one third of the former players had

widening of the lateral ventricles. This was determined to be central cerebral

atrophy. This could also be a symptom of alcohol abuse, but this issue wasn’t

addressed. Significant differences did not occur between “typical headers” and

the other players (Rutherford, et al., 2003, Sortland and Tysvaer, 1989).

14

In one of the initial studies in the United States, Jordan et al. (1996)

reported on twenty males with an average age of 24.9 years from the US

National Soccer team. A cohort of 20 male elite-track athletes was identified.

Players completed questionnaires regarding positions played, history of head

injury, number of headers and number of years played. A scaling system was

used to estimate the number of headers for each player based on level and type

of play. All study participants were examined using magnetic resonance imaging

(MRI) to determine any neurological deficits. Based on this testing, no

differences were noted between the soccer players and the control group, but

nine of the US National Soccer players were found to have abnormal MRI results

with three of those players having multiple findings (Jordan, et al., 1996). These

results included cortical atrophy in three players, ventricular enlargement in three

players, focal atrophy in three players, cavum septum pellucidum in three

players, and cerebellar atrophy in one player.

MRI was also used in a study by Autti et al. (1997). Both soccer and

American football players were given MRI scans and were compared to the

scans of age-matched, non-athlete controls. High-signal foci were found in 11

soccer players, seven American football players, and in five of the controls. This

indicates axonal rarefaction or non-ischaemic demyelination. The majority of

these high-signal foci were not found in both T2-weighted imaging and proton

density-weighted imaging. High-signal foci were also seen on both the T2-

weighted imaging and the proton density-weighted imaging which indicates

microinfarts or ischaemic tissue damage has occurred. Of these foci found on

15

both scans, the majority were in soccer players. Autti et al. (1997) suggests that

due to the lack of helmet use in soccer, foci are being caused by slight brain

injuries taking place during play (Autti, et al., 1997).

One problem with all of the previous studies is that the biomechanics and the

physiological effects of heading is not understood completely and, therefore,

cannot be ruled out or assigned blame for injuries. As Kirkendall and Garrett

(2001) state, “it is difficult to blame purposeful heading for the reported cognitive

deficits when actual heading exposure and details of the nature of head-ball

impact are unknown” (Kirkendall and Garrett, 2001). In order to determine

whether or not heading is causing cumulative damage, it is necessary to further

understand what is taking place during actual heading events. The proposed

study will focus on studying the frequency and severity of headers as well as a

biomechanical evaluation of soccer heading in the youth soccer population. By

evaluating youth soccer players, both in the lab and on the field, assessments

can be made about differences between adult and youth heading biomechanics.

These differences, or lack thereof, will provide insight as to whether the youth

soccer population is at higher risk for injury due to soccer heading. Additionally,

by recreating previously performed laboratory tests, it will be possible to compare

lab and field data.

1.3 Specific Aims

Although previous research has been conducted to determine the effects of

repetitive heading in soccer, the results are very controversial. Conflicting results

have been observed in studies throughout the history of soccer heading

16

research. Many of these previous studies have had significant challenges within

the methodology, including the lack of controls or using improper control groups

and many have not taken into account other outside factors that could be

contributing to results. In order to determine the effects of the repetitive

subconcussive head impacts associated with heading, an in-depth analysis of the

biomechanics of heading needs to be performed. The specific aims of this

project include:

1) To determine the incidence of head injury in youth soccer related only

to head to ball impacts. This will be accomplished using the National

Electronic Injury Surveillance System (NEISS) database which was

created by the United States Consumer Product Safety Commission

(CPSC).

2) To determine the frequency of heading in youth soccer based on age,

gender, and skill level.

3) To validate a novel headband system to measure head impact

frequency during soccer play.

4) To measure the biomechanical response of youth soccer players

during heading events using the Functional Assessment of Biomechanics

motion capture system and compare them to adults.

5) To measure head impact frequency and severity using the Head

Impact Telemetry System (HITS) (Simbex, Lebanon, NH). By using a

wireless acceleration measurement system, actual field data may be

collected.

17

CHAPTER 2

NEISS DATABASE

2.1 Introduction

The United States Consumer Product Safety Commission (CPSC)

established a sample of hospitals which gather information on each emergency

room patient who has an injury related to a consumer product. Using this sample

data, estimations can be made for the entire population. The collective database

is called the National Electronic Injury Surveillance System (NEISS).

The NEISS database was designed in its original form in 1970 using a

sample of 119 hospitals. Updates have taken place throughout the duration of its

existence in order to maintain a statistically applicable hospital sample and

estimation technique. The current database collects data from 100 hospitals

nationwide using 5 strata, or divisions based on size. Hospital size is determined

by the number of visits that the emergency department handles yearly. The

current strata represent hospitals of four different sizes as well as children’s

hospitals from across the United States (Figure 2.1).

18

Figure 2.1: NEISS database hospitals by strata (Consumer Product Safety Commission 2000)

The database can be searched using different variables to eliminate

unwanted cases. Variables which can be looked at include: date, product, sex,

age, diagnosis, disposition, locale, and body part. Additionally, short notes are

often available along with the case information providing a more detailed

description of the injury and related circumstances. Product codes for each

consumer product are available and any injury that was related to the use of the

specific product is reported and can be searched for based on that product code.

Statistical weight is calculated for each hospital for each month and the

estimates are calculated using these values. The calculation of the statistical

weights takes into account any non-response of hospitals, merging hospitals, and

any additional alterations made within the sampling frame (CPSC 2001).

Statistical weights are calculated using the following equation (CPSC 2001):

19

������� � �� �′� �� ��

Where:

Nh = Number of hospitals in the 1995 sampling frame for sampling for stratum h

nh = Number of hospitals selected for the NEISS sample for stratum h

n’h = Number of in-scope hospitals in the NEISS sample for stratum h

rh = Number of NEISS hospitals participating in stratum h for the given month

Rh = Ratio adjustment for combined stratum h

Using the weights for each hospital, the estimates can then be calculated.

The following equation is used to calculate the estimates for each hospital

(2001):

�������� � � � ����′� ���

���

�� !�

Where:

m = Number of strata in the NEISS sample during the given time period

Nh = Number of hospitals in the NEISS sampling frame for sampling for stratum h

nh = Number of hospitals selected for the NEISS sample for stratum h

n’h = Number of in-scope hospitals in the NEISS sample for stratum h

rh = Number of NEISS hospitals participating in stratum h for the given month

Rh* = Ratio adjustment for combined stratum h

xhi = Number of cases for a specified product or type of injury reported by

hospital i in stratum h for the given month

20

Since sporting equipment is considered a consumer product, injuries

related to equipment are a part of the NEISS database. The NEISS database

has been used in previous studies related to sports injury (Conn, et al., 2006,

Hostetler, et al., 2005, Hostetler, et al., 2004, Sosin, et al., 1996, Xiang, et al.,

2005, Yard and Comstock, 2006, 2006, Yard, et al., 2007), some of which are

specific to soccer (Adams and Schiff, 2006, Delaney, 2004, Leininger, et al.,

2007), to determine the injuries relating to the sport that have presented to the

Emergency Departments over a specified time period. In addition to soccer,

sports which have been researched using the NEISS database include martial

arts, ice hockey, lacrosse, field hockey, water skiing, wakeboarding, snow skiing,

snowboarding, cycling, skateboarding, and rugby (Adams and Schiff, 2006,

Hostetler, et al., 2005, Hostetler, et al., 2004, Leininger, et al., 2007, Pickett, et

al., 2005, Sosin, et al., 1996, Xiang, et al., 2005, Yard and Comstock, 2006,

2006, Yard, et al., 2007). The majority of these studies include all injuries

reported during the specified activity, but there are also those that focus on a

specific variable such as the region of the body or level of injury. For example,

one study focused specifically on head injuries which led to fatalities during

bicycling (Sosin, et al., 1996).

Previous studies using the NEISS database to research the frequency of

soccer injuries in the pediatric population have been performed. These studies

have looked at the overall injuries sustained (Adams and Schiff, 2006, Leininger,

et al., 2007) and total head injuries in multiple sports (Delaney, 2004). Injuries

were investigated by examining any factor taking place during a soccer match

21

and the details of those injuries. Although these studies discuss body region and

injury cause, further investigating head injuries, specifically those caused by

impact with the ball, can provide specific information related to the injuries

associated with soccer heading.

Although previous research has been conducted to determine whether

soccer heading is injurious, the results are varied. Conflicting results have been

observed in studies throughout the history of soccer heading research

(Guskiewicz, 2002, Jordan, et al., 1996, Matser, et al., 1998, Putukian, 2004,

Tysvaer and Storli, 1981, Tysvaer and Lochen, 1991, Tysvaer and Storli, 1989,

Witol and Webbe, 2003). Therefore, it has yet to be determined if injury occurs

strictly from head contact with the ball. The purpose of the current study is to

gain insights into the occurrence of head injuries from contact with the ball only.

By reviewing hospital emergency room data coded through the NEISS data it can

be determined if ball only impacts can be a mechanism of injury in soccer.

Gender differences have been hypothesized as being an issue in soccer

heading related issues due to the ball to weight ratio differences (Barnes, et al.,

1998, Covassin, et al., 2003). Previous studies of the NEISS database have

found that the majority of players injured while playing soccer were boys (Adams

and Schiff, 2006, Leininger, et al., 2007). Leinenger et al. (2007) found that 58.6

% of total injuries were to boys, and they were also more likely to have head and

neck injuries. While the number of injuries was higher in boys than girls, the

concussion rates remained similar (Leininger, et al., 2007). Similarly, Adams and

Schiff (2007) found that boys had a higher number of total injuries (55.5 %).

22

While both of these studies found that boys have a higher incidence of injury, it is

necessary to determine if girls have a higher incidence of ball to head related

injuries. It has been hypothesized previously that girls have a higher rate of head

injuries, specifically concussions, due to their smaller stature, greater ball to head

ratio, and potentially weaker neck muscles (Covassin, et al., 2003).

Player age is also considered a possible factor for increased risk of soccer

head injuries specifically related to heading the ball. Previous studies have found

conflicting results as to which age group has a higher likelihood of experiencing

head injuries during soccer play (Adams and Schiff, 2006, Leininger, et al.,

2007). Leinenger et al. (2007) found that younger players were more likely to

have head/neck/face injuries, but Adams and Schiff (2006) found that head and

face injuries occurred more frequently in the oldest age group (15-19 years old).

This is most likely due to the fact that Leinenger et al. (2007) included 2-4 year

old players in their study, and they were not included in the other previous

studies. Although these studies looked at head injuries, the mechanism of injury

was not determined. This is necessary to determine if younger children are more

likely to suffer from ball to head injuries related to soccer heading.

Delaney (2004) studied head injuries in ice hockey, soccer, and football.

Using the NEISS database, head injuries to anyone participating in any of the

three sports during a ten year span, 1990 – 1999, were calculated. Results were

not limited by gender or age. The inclusion criteria were limited to sport being

played and body region injured. Total head injuries for the ten year period were

found to be the highest in football (204,802), followed by soccer (86,697) and ice

23

hockey (17,008). The total number of concussions during the ten year span also

followed a similar pattern with football (68,860) being the highest, followed by

soccer (21,714) and ice hockey (4,820) respectively. Although the total number

of injuries was higher for football and soccer than hockey, the injury rate was

found to be similar. Injury rate was determined by dividing the number of injured

athletes by the total number of athletes. Although this study investigates head

injuries during soccer play, injury mechanism, gender, and age were all

neglected as pure between sport comparisons were made (Delaney, 2004).

The previous studies performed a broad soccer injury analysis using the

NEISS database. In-depth analyses have not been completed on the types of

head injuries or the cause of these injuries. The current study will focus on head

injuries caused by intentional or unintentional head contact with the ball and their

diagnoses. This will represent the occurrence of injuries caused purely by ball

contact.

2.2 Methodology

Data were collected using the United States Consumer Product Safety

Commission (CPSC) National Electronic Injury Surveillance System (NEISS).

The database was queried using for soccer injuries that occurred to the head

using product code 1267. Injuries considered were limited based on age, body

region injured, and year taking place. The injuries that were included were head

injuries in children ranging in age from 5 – 18 years that occurred from 2002 –

2007. Using statistical weights collected as part of the data set for each case,

national estimates were made.

24

Included cases are both game and practice injuries. In order to focus on

ball-to-head injury, the injury mechanism for each case was investigated. Using

the narrative description provided with each case, the estimated number of head

injuries resulting from impact with the ball only, the ground, collision with another

player, the goal post, and unknown mechanism were calculated. Additionally,

age, gender, and diagnosis were assessed for injuries that occurred from impact

with the ball only. Age groups were broken down into 5-9 years old, 10-14 years

old, and 15-18 years old. Diagnoses included all of those that were seen after

limited other variables and include: concussion, contusion, laceration, internal

organ injury, and other. Internal organ injuries included such things as closed

head injuries.

Statistical analysis was performed using the Kruskal-Wallis test for testing

more than 2 groups in a hypothesis. Whenever this was significant, Mann-

Whitney tests were performed. It was determined by the Institutional Review

Board that this study did not require approval for the study of human subjects.

2.3 Results

A total of 62,022 soccer head injuries were estimated to take place from

2002 to 2007 in the United States. Following 2003, a significant increase in the

number of head injuries was seen in 2004. This increased from 9022 head

injuries in 2003 to 11,762 head injuries in 2004 (Figure 2.2). The increase held

relatively steady through 2007 which had a total of 10,122 head injuries.

Throughout the six years included in the study, ball to head injuries remained

25

consistent. The year 2004 was statistically higher than all other years, except

2005, which it was statistically lower than.

Figure 2.2: Soccer head injuries including both males and females

The majority of soccer head injuries from 2002 to 2007 were caused by a

collision with another player (38 %). Injuries caused by impact with the ball only

represent 16 % of soccer head injuries (Table 2.1). Additional injuries with an

unknown mechanism could cause an under prediction of the four main

mechanisms investigated, but was required due to a lack of description for these

injuries. Also, goalpost injuries comprised 5.75 % of total injuries and included

impact with the wall in indoor soccer.

26

Table 2.1: Head injuries by mechanism Selected Characteristics Actual, n Weighted

N %

Injury Mechanism n = 2081 N = 62024 100.00

Ball Only Head Injuries 309 9861 15.90

Unknown 471 12720 20.51

Ground 424 12471 20.11

Goalpost 117 3569 5.75

Collision with Player 760 23404 37.73

Although males made up the majority of soccer head injuries (53.99%),

they did not have the highest percentage of ball only head injuries. Females had

59.64% of those injuries caused by impact with the ball only (Table 2.2). The

only age group in which males had a higher percentage of ball only injuries was

the 5-9 year olds where females only consisted of 19.68% of the injuries. The

gender difference had the largest increase of females in comparison to males in

the 15-18 year old age group where males only made up 29.87% of the ball to

head injuries. Total head injuries were significantly different between genders

(p=.015), but ball to head injuries were not. Hospitalization following ball to head

injury was not common, with 99.67 % of players being treated and released.

27

Table 2.2: Ball to head only injuries Selected Characteristics Actual, n Weighted

N %

Gender n = 309 N = 9861 100.00

Male 134 3980 40.36

Female 175 5881 59.64

Age n = 309 N = 9861 100.00

5-9 46 1232 12.49

10-14 130 3789 38.42

15-18 133 4840 49.08

Diagnosis n = 309 N = 9861 100.00

Concussion 100 3526 35.76

Contusion 36 1793 18.18

Laceration 3 99 1.00

Internal Organ 163 4317 43.78

Other 7 126 1.28

The lowest number of injuries took place in the 5-9 year old age group

(1231.66) and the highest was in the 15-18 year old age group (4840.38). The

15-18 year old age group had significantly higher ball to head injuries than both

other age groups (Figure 2.3). There was no statistical difference between the

two younger age groups. Although the females had an increase in injuries from

the younger age group to the older, the males did not. Males had an increase in

injury from the 5-9 year olds to the 10-14 year olds.

28

Figure 2.3: Male and female ball only head injuries by age. *p < 0.05

Both male and female soccer players had similar patterns of injury

diagnosis (Figure 2.4). Internal organ injuries represent the highest percentage

of injuries with 44 % for each gender. These injuries occurred significantly more

frequently than both concussions and contusions. The second most frequent

diagnosis for males and females was concussion with 38.27% and 34.06%

respectively. Additionally, concussions occurred significantly more than

contusions which accounted for only 18% of ball to head injuries. Lacerations

were not frequent injuries, but were seen more frequently in males versus

females, 2.18% and 0.21% respectively.

29

Figure 2.4: Ball only head injuries by diagnosis

2.4 Discussion

Although many studies have used the NEISS database to study the

incidence of soccer injuries (Adams and Schiff, 2006, Delaney, 2004, Leininger,

et al., 2007), this is the first to focus specifically on youth head injuries. Previous

studies have looked at the total incidence of soccer injury presenting to the

emergency departments, specific body areas injured, and the adult population,

but studying the incidence of head injury in the youth population is novel along

with investigating mechanism of injury specific to ball to head only head impacts.

This makes direct comparisons with previous studies challenging due to different

inclusion criteria.

During the study period, high school soccer participation steadily

increased from 339,101 boys participants and 295,265 girls participants in the

2001-2002 school year to 377,999 boys participants and 337,632 girls

participants in the 2006-2007 school year. Due to this increase in player

participation it was expected that an increase in ball to head injuries would occur

30

steadily as well. This steady increase did not occur demonstrating a non-linear

relationship between the number of participants and the number of head injuries.

It is, however, challenging to determine a total number of participants due to the

inclusion of organized and non-organized soccer injuries. Additionally, the

previous comparison was made for high school age players and did not include

of the entire study age population.

Ball only injuries comprised 15.9% of total head injuries. These injuries

were not necessarily caused by heading the soccer ball, as some of the impacts

were due to unintentional ball to head impact, but many of the cases were

described as heading related. It was challenging to delineate the mechanisms

further due to the limited case descriptions. However, these data indicate that

heading alone can result in injuries severe enough to require medical attention.

Injury severity for these types of injuries is also cause for concern. While

contusions are limited to skin bruising, lacerations and concussions are more

serious. Lacerations are skin tears that are generally caused by blunt trauma,

i.e. impact, and can require suturing. The most serious diagnosis, internal organ

injuries, was also found to be the most frequent diagnosis in the ball only head

injuries for both males and females, followed by concussion. There are a limited

number of possible diagnoses that can be input into the NEISS database, which

is why many of the injuries listed are internal organ injuries. This is an unspecific

diagnosis that includes closed head injuries, cerebral bleeding, and brain

contusion (Delaney, 2004). This is relevant in that it shows that impact with the

31

ball only can cause significant injury level. This diagnosis was given to 43.84%

of female injuries, and 43.68% of male injuries.

Concussions are serious injuries, also representing injuries to the brain, as

opposed to contusions and lacerations which are skin injuries. They were the

second most frequent diagnosis for both males and females. A total of 100

concussions were reported at the participating hospitals during the study period.

This correlates to almost 36% of the ball only impacts being treated. There is no

way to determine the severity of these concussions; therefore, they can range

from mild to a more severe injury. However, several narratives describe

scenarios where the child would complain of a headache after heading the ball

during a practice or game without any other etiology.

There are limitations when examining these severe injuries in terms of

percentage of total injuries. The more serious injuries most likely are over

represented when using the NEISS data alone due to the fact that the injuries

included in this study are only those which have reported to the emergency

department. Therefore, it is less likely that less severe injuries are taken to the

emergency department. Players could be experiencing symptoms related to ball

to head impacts, but if they did not go to the emergency department, these

injuries were not included. The current study most likely underestimates both

total injuries and ball to head injuries because many less severe injuries would

not be included. This is an inherent problem with using the NEISS database to

estimate injuries. Therefore, the current study represents the more severe

cases, and is potentially an underestimate of the total injury occurrences.

32

More ball to head only injuries occurred in the older age groups, which

contradicts Leininger et al. (2007) who found that the majority of head injuries

occurred to the youngest players. Leininger et al. (2007) included all head

injuries, not just those specific to ball impacts, and included a younger set of

players. This could be due to the introduction of heading in those two age

groups. Additionally the majority of those two age groups use the adult ball size,

a standard size 5 soccer ball weighing 450 g with a 22 cm diameter, as opposed

to the smaller size 4 soccer ball which weighs approximately 390 g and has a

diameter of 21 cm. Also, they would presumably have higher kick velocities.

Therefore, it was expected that an increase in ball to head only injuries occurred.

Although much research has been devoted to the adult population with

respect to soccer heading, especially the elite players, the risks to the youth

population have not been studied in great detail. The risks to children are

potentially greater due to their size versus the force being applied by the ball

(Lees and Nolan, 1998). It has been reported that ball mass, impact velocity,

and size of the individual all contribute to the potential for injury (Lees and Nolan,

1998). The importance of proper technique may be especially true in the youth

population, since their skill level has not been well developed to control their

head motion when heading the ball. The current study demonstrates that injuries

can occur from ball to head impact only. It is recognized that all of these injuries

may not be the result of purposeful heading, but it is challenging to differentiate

them from the unintentional ball to head impacts. The data do establish the

possibility of being injured based on an impact with the ball alone.

33

CHAPTER 3

HEADING FREQUENCY IN YOUTH SOCCER

3.1 Introduction

Soccer, one of the most popular team sports worldwide, has recently

shown considerable growth among the United States youth population. This is

seen clearly in the American Youth Soccer Organization (AYSO) which today has

50,000 youth soccer teams and over 650,000 players registered after starting out

in 1964 with only nine teams (2006). With this increase in the number of players,

an increase in injuries has also occurred. Dvorak et al. (2000) determined that

approximately 30 billion dollars are spent annually worldwide on the treatment of

soccer injuries. It has been estimated that up to 22% of all injuries in soccer are

to the head (Ruchinskas, et al., 1997). These injuries can occur from

unintentional or intentional impacts. Intentional impacts, or “heading” the soccer

ball, are a standard practice within the game that is taught at the youth level.

Heading is a technique where a player intentionally uses their head to redirect

the soccer ball. Generally, players are instructed to start with their feet

approximately shoulder width apart and knees bent in a slightly staggered

stance, body squared to the ball with their torso in an extended position. While

keeping their eyes on the ball, players move their torso into a flexed position and

impact the ball at the hairline on their forehead. Following impact, players are

instructed continue through the ball with follow-through and to decelerate their

motions following impact (Shewchenko, et al., 2005). There are still conflicting

results as to whether repetitive sub-concussive forces associated with heading

34

are a cause of long term health problems (Tysvaer and Lochen, 1991).

Several studies have been conducted to look at these effects, but one of

the main limitations has been the estimation of exposure incident. One of the

first steps in delineating the effects of repetitive heading should be to determine

an accurate exposure incidence. In most studies, the incidence of heading

exposure is reported by the players themselves; often times as broad estimates.

There are several limitations related to having players self report their exposure

rate especially in a retrospective manner.

In one of the first studies conducted (Tysvaer and Storli, 1989), EEG

results of current Norwegian First Division League Clubs players were reviewed

with results showing the highest abnormal results occurred in the youngest

players. Another study conducted in 1991 by Tysvaer and Løchen (1991) of

retired players showed neuropsychological deficits in 81% of the 37 participants.

In the two Tysvaer et al. (1989, 1991) studies, players were also asked if they

were typical headers however a definition was not provided as to the criteria for

being a “typical header”. 14% of the 69 active players interviewed and 27% of

the 37 retired players interviewed reported that they were typical headers,

however a quantitative value was not reported (Tysvaer and Lochen, 1991,

Tysvaer and Storli, 1989).

Matser et al. (1998) estimated the number of headers that each of the 53

active professional soccer team members from The Netherlands who participated

based on their position and number of games (Matser, et al., 1998). A

classification system was used and categorized midfielders and goalkeepers as

35

“non-headers” with forwards and defensive players as “headers”. Matser et al.

(1998) also reported the number of headers per match and in a season. The

number of headers per match and the number of matches were obtained through

player interviews. Players reported a range of 1 to 42 headers during games with

16 being the median number of headers in a single game. These numbers along

with the total number of games were then used to calculate headers per season.

A range of 50 to 2100 headers in a season were calculated which resulted in a

median of 800 headers in a season. The number of average headers/per game

by an individual player was then stratified into three groups: 0-10 (47%), 11-20

(36%), or >21 (17%) (Matser, et al., 1998).

In a recent study, Witol and Webbe (2003) studied 60 male players at

varying levels of skill. The number of headers each player experienced was

determined based on player reporting in an interview. Players were asked if they

considered themselves a header and how many headers they experience in a

typical game. Players were then placed in one of four categories: control (no

heading), low (0-4 headers/game), moderate (5-8 headers/game), or high (>9

headers/game). The results showed that 12 players were in the control group,

19 players were in the low group, 20 players were in the moderate group, and 21

players were in the high headers/game group (Witol and Webbe, 2003).

These studies demonstrate a reliance on estimations and player memory

with a lack of data on the actual exposure rate. As reported in Chapter 1, soccer

players may have deficits in the area of memory (Matser, et al., 1998) which

could provide for inaccurate recollections. Very few studies actually observe

36

players to determine their frequency of heading. Two studies have reported

results based on this type of observation. The first by Tysvaer et al. (1981)

reported results after following 20 games. These games broke down as follows:

10 First Division games, 6 English games, and 4 International games. Tysvaer et

al. (1981) reported average headers per game which were 117, 124, and 94

respectively, but the number of headers per specific player was not reported

(Tysvaer and Storli, 1981).

The second study which observed headers is also one of the few studies

conducted in the youth population (Janda, et al., 2002). A total of 57 players

participated with an average age of 11.5 years. All players were followed for at

least three seasons, with 18 players followed for two years. The number of

headers per player was monitored by the individual coaches. Over the period of

one year, players heading the ball an average of 185.9 times with the maximum

number for one player being 450 times/year one. For the second year, the 18

players monitored had an average of 129.6 headers with a maximum of 344 for

all three seasons (Janda, et al., 2002).

Although these studies provide some information regarding the number

of headers sustained during the play of soccer, they are focused on the specific

age and skill level studied, with only one study focusing on the youth population

(Janda, et al., 2002). The frequent use of athlete reporting in studies relating

heading frequency to findings is unreliable and a better method of estimation is

required. Therefore, the purpose of this study is to explore the frequency of

soccer heading in the youth population across age groups, skill levels and

37

gender. These data are essential to conduct the controlled laboratory

experiments needed to ultimately determine the effects of repetitive heading.

Given the increase in youth soccer participation over the past decade and the

continued expected growth, the current research focused on the youth

population.

3.2 Methodology

Males’ and females’ teams ranging in ages from U12 (under 12 years old)

to U18 (under 18 years old) were observed during the 2006 Canton Cup Soccer

Tournament, a weekend long tournament in Canton, MI. The tournament

director, along with the Wayne State University Human Investigations

Committee, granted approval prior to the event. Only teams participating in the

top two divisions of their age bracket were included in the study. The highest

division was given the designation by the tournament of blue and the next

highest level was red. These designations were maintained throughout the

study. A total of 158 games were observed throughout the tournament. Table

3.1 outlines the breakdown in terms of number of teams and number of games

monitored at each division, age and gender group. It should be noted that due to

the fact that soccer is a spring sport for high school aged females in Michigan,

the highest age division for the tournament was U14.

38

Table 3.1: Number of games monitored outlined by age, gender and division of play

Age/Gender Female Male

U12 U13 U14 U12 U13 U14 U15 U16 U17 U18

Number of

Teams

Blue 6 8 6 5 8 8 8 8 8 8

Red 6 8 7 8 6 8 8 N/A N/A N/A

Number of

Games

Blue 8 11 8 6 11 10 7 10 9 10

Red 8 10 12 9 8 10 11 N/A N/A N/A

Games were monitored by 22 individuals with a minimum of five years of

experience as a soccer player. At each game, a stopwatch was started at the

kick-off. This was then allowed to run throughout the entire game, including

halftime. Each header that took place during the game was recorded on a data

sheet that contained a grid outline of the soccer field (Figure 3.1). Both the

player number and time of occurrence were noted within the specific area of the

grid where the header took place. Player position was not noted, but defensive

and offensive position was defined by which half of the field the header took

place on. No personal identifiers were recorded. Different colors were used to

denote different teams to allow for a better representation of defensive versus

offensive position on the field.

39

Figure 3.1: Soccer Field Diagram

Data were normalized using time due to the differences in game length for

varying age divisions. This was done by dividing the total number of headers per

team by the total number of minutes that each team was monitored throughout

the tournament. This resulted in a value of headers/minute for each team for the

tournament. The data were then analyzed using a repeated measures ANOVA

using gender, age group, division, game day, game number within the day, and

game number within the tournament as the independent variables and

headers/minute as the independent variable. All analyses were stratified by

gender in order to account for different age groups within gender.

Total headers were also determined for each game by team, half, and field

area (12 total areas). The total number of headers was analyzed using the

generalized estimating equation (GEE). Poisson regression analysis was then

40

used to determine if there was a significant association between the number of

headers and age, division, offensive field placement, defensive field placement,

game day, game number within the day, game number within the tournament,

and position on the field.

3.3 Results

Maximum headers

Maximum headers in one game by a single player were monitored to

determine the highest exposure incidence. The maximum number of headers in

a single game by a player was 13 headers. This was observed in a U14 male

blue division game. The range of maximum headers in one game by one player

was from 4 to 13 headers (Table 3.2).

Table 3.2: Maximum headers by any one player for a single game.

Age/Gender Female Male

U12 U13 U14 U12 U13 U14 U15 U16 U17 U18

Maximum

Headers

Blue 7 5 7 11 8 13 9 7 7 9

Red 4 4 4 5 5 7 7 N/A N/A N/A

Gender effects

Significant differences were reported between the male and female

populations following adjustment for age and division (p<0.0001). The male

populations were observed to have a higher header/minute ratio than their

female counterparts with a mean of 0.135 ± 0.079 headers/minute for females

and 0.283 ± 0.122 for males (Figure 3.2). The total number of headers was also

significantly different between males and females. Therefore, all further analyses

41

were stratified by gender.

Figure 3.2: Comparison of number of headers/minute, male versus female

across age groups.

Males

Age (p<.0001), division (p<.0001), and game number (p=.015) all

demonstrated significant findings with regard to headers/minute within the male

population. Age groups U14 and higher were found to have significantly more

headers/minute than the U12 and U13 age groups. Also, a consistent positive

increase regression was noted in the male blue division from U12 to U15,

however from U15 to U18 there was no significant increase with age (Figure 3.3).

The number of headers/minute was also significantly higher in the blue division

when compared with the red and white divisions. Also of interest,

headers/minute was significantly higher for game 1 of the tournament when

compared to both games 2 and 3.

42

Figure 3.3: Number of headers/minute for the male population.

The total number of headers (Table 3.3) was significantly lower for both

the red and white division when compared to the blue division (p<.0001), but

there was no difference between the red and white divisions. The number of

headers increased with increasing age group, with the exception of U17/18 which

was slightly lower than U15/16. All age groups 13 or higher had significantly

higher number of headers compared to U12 (p<.0001). U13 players had lower

adjusted mean number of headers compared to all older age groups (p<.01), and

U14 players were lower than U15/16 and U17/18 (p<.001). No significant

difference was observed between U15/16 and U17/18 (p=0.307). Players on

defense had a higher adjusted mean number of headers (0.62 ± 0.02) compared

to players on offense (0.47 ± 0.02) (p<.0001). The adjusted mean number of

headers was lower on Saturday and Sunday compared to Friday (p<.01) with no

difference between Saturday and Sunday (p=0.056).

43

Table 3.3: Average total headers/game/team for male population Age Division Average Headers/Game/Team

12 R 9.8

B 11.8

13 R 9.9

B 15.6

14 R 16.3

B 16.8

15 R 19.7

B 23.3

15/16 W 17.0

16 B 23.7

17 B 23.8

17/18 R 17.1

18 B 23.4

Females

The average value of headers/minute was highest in the U14 girls (0.153

headers/minute). Headers/minute showed no significant associations between

headers/minute and any of the dependent variables. However, the total number

of headers (Table 3.4) had several parameters which were significant predictors.

Within the female population, division (p=.0009), length field position (p<.0001),

width field position (p=.0002), and age group (p=.036) were found to be

predictors of the number of headers.

44

It was determined that the blue division players had more headers than

those in the red division. Differences were also seen within age groups.

Significantly more headers took place in the U14 age group when compared with

both U12 and U13, but no differences were found between U12 and U13.

Table 3.4: Average total headers/game/team for female population

Age Division

Average

Headers/Game/Team

U 12 R 5.9

B 6.5

U 13 R 5.5

B 8.7

U 14 R 7.0

B 10.8

A positive increase in headers/minute with age trend similar to that

observed in the male population was also noted in the blue division of the female

population. Due to the lack of data in the older age groups, the stabilization effect

with age was unable to be assessed (Figure 3.4).

45

Figure 3.4: Number of headers/minute for the female population.

Field Position

Table 3.5: Total headers in each field position for females Field Position

Age 1A 2A 3A 4A 1B 2B 3B 4B 1C 2C 3C 4C Totals

U12 1 7 2 5 5 11 12 6 11 9 14 2 85

U13 2 11 16 1 9 18 19 9 4 9 13 4 115

U14 5 15 16 6 13 22 22 4 8 15 13 6 145

Totals 8 33 34 12 27 51 53 19 23 33 40 12 345

The total number of headers in each portion of the field is shown in Figure

3.5. The majority of the headers, for both male and female, took place in the

middle of the field. The occurrence of headers in the four corner regions was

significantly less than all other regions on the field. Females had fewer total

headers in each field position when compared with males of the same age group

(Table 3.5, 3.6).

46

Table 3.6: Total headers in each field position for males Field Position

Age 1A 2A 3A 4A 1B 2B 3B 4B 1C 2C 3C 4C Totals

U12 6 15 21 2 9 18 18 9 3 14 13 5 133

U13 6 19 15 2 12 37 24 19 10 23 25 6 198

U14 14 23 18 9 30 50 39 28 6 22 24 7 270

U15 6 31 32 10 52 77 57 36 15 45 33 13 407

U15/16 3 7 10 4 12 20 24 8 5 5 8 2 108

U16 5 16 18 4 28 33 33 21 3 14 18 6 199

U17 3 22 18 2 23 37 36 24 3 14 13 2 197

U17/18 0 17 11 5 9 13 17 11 4 12 15 2 116

U18 7 11 13 2 27 45 42 29 7 16 20 10 229

Totals 50 161 156 40 202 330 290 185 56 165 169 53 1857

47

Figure 3.5: Total number of headers in each field position

3.4 Discussion

Putukian et al. (2004) followed college soccer players for a season to

perform neuropsychological testing (Putukian, 2004). Heading contacts and

minutes played were also counted for all home games by team trainers and

physicians, and it was found that male college age players had an average of

0.783 headers/minute during a single season and females had an average of

0.753 headers/minute. Both of these values are much higher than those found in

the current study. This is most likely due to the fact that actual minutes played by

the player were used as opposed to game length. The average headers/minute

of for U15/16 males was the highest at 0.303 headers/minute, and for the

females 0.153 headers/minute for the U14 age group. It is expected that players

48

at the college level would have more headers/minute than those playing at the

lower skill and age level observed in the current study. Also, following teams

within the younger age group for an entire season could provide a more detailed

look at the differences noted between these two skill levels.

Matser et al. (1998) found that 47% of the players that were studied

headed the ball between 0 and 10 times per game. Looking at the maximum

number of headers by a single player in a game, only two players fell outside of

that range throughout the tournament. The first, a male, blue division player in

the U12 age bracket had 11 headers in a single game, and the second, also a

male, blue division player in the U14 age group had 13 headers in a single game.

These findings are similar to Matser et al. (1998) because the majority of players

were found to remain in the lower header per game category.

The current study is limited by the fact that data was collected over a

weekend long tournament and that a maximum of three games were observed

for each team as opposed to following teams for an entire season. Additionally, it

would have been an improvement if more female age groups had been available

for analysis, but due to high school scheduling that was not possible. Even with

these limitations, important trends were evident.

One of these trends was noted in the positive correlation between age and

headers/minute within the higher division in both the males and females. This

occurs up until the age division of U15 in the male population at which point a

plateau occurs. Although there is currently no data available for the higher age

groups within the female population, it is expected that there would be a similar

49

trend based on the data available for the U12-U14 age groups. This indicates

that there is an increase in the amount of headers players experience during the

years in which they are learning the skill, but that in the upper skill levels they

experience essentially the same amount throughout.

It was also noted that the vast majority of heading occurred in the middle

of the field. This is most likely related to this being the area where long kicks,

from both the goalie and players trying to cross the field, are targeted allowing for

players to align themselves for headers. The regions directly in front of the goal

(1B and 4B in Figure 3.5) lend themselves to areas of heading due to corner

kicks being directed to those regions and the desire to redirect the ball into the

goal. However, this can also be the most dangerous position due to the number

of players near the goal during a corner kick and proximity of the goal posts.

Game number within the tournament was also determined to be a factor

in the number of headers/minute in the male population. This is of interest

because the number of headers/minute seems to decrease after the first game.

This could be due to symptoms felt following the initial game. Further research

needs to be conducted to understand the cause of this reduction following the

initial game in the tournament.

50

CHAPTER 4

HEADING BIOMECHANICS IN YOUTH SOCCER

4.1 Introduction

The overall goal of heading is redirection of the ball. Depending on the

approach of the player and the intent of the redirection, the player may move

his/her head in particular manner. Alignment of the head, neck and torso can

vary and are often dependent on the intent of the redirection i.e. clearing,

passing, or controlling (Shewchenko, et al., 2005, 2005). All of these scenarios

require a specific skill level to accomplish the intent of the redirection through the

use of correct techniques.

Proper techniques and skill level often come with age. Younger players

who are learning good techniques may not always perform the skill as taught.

This may be to a variety of reasons including improper ball size for child size, not

eliciting neck muscles, and using top of head instead of forehead. All of these

factors change the biomechanics of heading the ball. Size differences between

the player and the ball have been a recognized concern not just for heading but

for the development of foot skills as well, therefore age recommended sizes have

been developed (Lees and Nolan, 1998). The recruitment of neck muscles plays

a big part in proper heading techniques. Incorporation of entire body mass

allows for the mass of the ball to be negligible in comparison. And, finally, ball

placement affects the impact vector through the head itself.

One of the first studies to look at the biomechanics of soccer heading,

specifically head and neck motion, was performed by Ludwig (1999). Twenty-

51

four college age female soccer players performed 10 standard headers served to

them at 8 m/s. Standard camcorders were used at 120 frames per second to

capture trunk motion and various acceleration measurements during each

header. It was determined that frequent headers and non-frequent headers, as

self-described by the players, used different technique when heading the ball and

that trunk range of motion and neck motion played a part in this difference. This

indicates that players with less heading experience, i.e. non-frequent headers,

use a different technique which could be similar to the less practiced youth

population. The study did not investigate differences specific to soccer

experience level or gender. Neck muscle activation was also not investigated as

a possible contributing factor to changes between the groups.

EMG has also been studied previously during soccer heading (Bauer, et

al., 2001). Again, this study focused solely on female college soccer players.

Players were asked to perform a series of soccer headers, all at the same ball

speed, 6.8 m/s, while instrumented with EMG sensors on both their left and right

sternocleidomastoid and trapezius muscles. Three types of headers were

performed by each participant in an effort to represent the various headers seen

in the field, clearing, passing, and shooting. Each of these three types of

headers were performed while standing and also while jumping to get a more

diverse representation. It was determined that muscle activation was not

significantly different for the various header scenarios studied.

One of the more comprehensive analyses of common scenarios was

recently conducted using both head and neck motion analysis techniques and

52

EMG (Shewchenko, et al., 2005). Seven adult soccer players ranging in age

from 20 to 23 years old were asked to perform various soccer headers. Subjects

were attempting to head the ball to a target with a pre-determined level of neck

muscle activation in the sternocleidomastoid and trapezius: normal, pre-tensed,

or relaxed. Soccer balls were presented to players at two different speeds in

order to elicit a wider range of response. The authors also reported a wide

variability between subjects when looking at head angle, back angle, relative

head to back angle, and neck muscle activity. In addition, it was suggested that

additional scenarios would produce additional results (Shewchenko, et al., 2005).

This variation of biomechanical scenarios of soccer heading will only be

increased in the youth population due to the varying skill levels.

The current study aims to determine if there is a difference between adult

soccer players and youth soccer players with respect to soccer heading

technique. In order to accomplish this aim, heading scenarios that have been

previously used (Shewchenko, et al., 2005) in adults were recreated in the youth

population. Using the Functional Assessment of Biomechanics (FAB) system, a

novel motion analysis system, various head and torso body angles were

measured to provide comparisons to previous studies and to participants within

the current study.

The FAB system (Figure 4.1) uses a combination of accelerometers,

gyroscopes, and magnetometers to provide real-time kinematic and kinetic data.

Angle, force, torque, velocity, acceleration, power, and foot sole weight and

pressure are all calculated for each body segment (Biosyn Systems). The

system has a maximum data collection of 100 Hz and a battery life up to 12

hours. The current study will use the system for measuring head a

angles with an accuracy of

without the limitations of using a traditional camera system

limitations and marker occlusion

traditional motion analysis system, are both eliminated.

Figure 4.1

In addition to comparing head and torso angles between youth players

and adult players, the current study

technique between genders

will also be examined to establish their level of involvement during various

53

system has a maximum data collection of 100 Hz and a battery life up to 12

The current study will use the system for measuring head a

angles with an accuracy of ± 2 degrees. The system allows for motion analysis

without the limitations of using a traditional camera system. For example, space

limitations and marker occlusion, which are common issues with using a

motion analysis system, are both eliminated.

Figure 4.1: FAB System with size scale

In addition to comparing head and torso angles between youth players

and adult players, the current study will investigate differences soccer heading

technique between genders in the youth population. Neck muscle activity levels

will also be examined to establish their level of involvement during various

system has a maximum data collection of 100 Hz and a battery life up to 12

The current study will use the system for measuring head and torso

The system allows for motion analysis

or example, space

, which are common issues with using a

In addition to comparing head and torso angles between youth players

will investigate differences soccer heading

. Neck muscle activity levels

will also be examined to establish their level of involvement during various

54

heading scenarios in the youth population. This will be done using traditional

EMG techniques and measuring the neck muscle activity of the

sternocleidomastoid and the trapezius muscles during heading events. These

muscles are both superficial and have been used previously to determine neck

muscle activation during soccer heading (Bauer, et al., 2001, Shewchenko, et al.,

2005).

4.2 Methodology

Fifteen youth soccer players, 9 females and 6 males, participated in the

current study to investigate the biomechanics of soccer heading. Players ranged

in age from 14 to 16 years old, with an average age of 15 for the females and 16

for the males. Players performed a sequence of headers while wearing

instrumentation to assess muscle activation and body position during heading

events.

Prior to any instrumentation, anthropometric measurements were taken to

provide an accurate representation in the FAB measurements. Height, weight,

trunk length, upper arm length, forearm length, thigh length, and calf length were

measured and recorded for each participant. Additionally, player age, gender,

and skill level were recorded.

Following initial measurements, players were instrumented for data

collection. Players were fitted with the FAB (Biosyn Systems, Surrey BC,

Canada). The system uses 13 wireless sensors each of which is approximately

4 cm x 7 cm x 2 cm and is attached to the subject using an elastric strap with

velcro attachment. One sensor was placed on the head, one on the chest, one

55

around the waist, one on each upper arm, one on each wrist, one on each thigh,

one below each knee, and one at each ankle (Figure 4.2).

Figure 4.2: Example of player wearing FAB sensors

Data were sampled at 100 Hz and recorded for post-processing. Head

and trunk body angles will be collected including: cervical flexion and extension,

cervical lateral flexion, trunk flexion and extension, and trunk lateral flexion

(Figures 4.3, 4.4). Prior to data collection using the FAB, a player calibration was

performed. The player was instructed to stand facing forward with arms to the

side and feet shoulder width apart while the system performed calibration

procedures. Zero degree measurements are taken from the initial calibration

position, and the trunk and head angles measured represent a change from that

initial position.

56

Figure 4.3: Side view of head and trunk body angles a) torso flexion (+ α) and extension (- α); b) head flexion (+ β) and extension (- β)

Figure 4.4: Top view of head rotation left (+ θ) and right (- θ)

a) b)

0° 0°

Low Back Sensor

Upper Back Sensor

Head Sensor

Upper Back Sensor

Head Sensor

0°

Head Sensor

57

Players were also be instrumented with electromyography (EMG) sensors.

Surface EMG data was collected at 256 samples/second using the BioCapture

physiological monitoring system (Cleveland Medical Devices Inc., Cleveland,

OH). Sensors were placed bilaterally on the sternocleidomastoid and the

trapezius muscles (Figure 4.5) with a reference sensor behind the left elbow.

After preparing the skin using an alcohol swab, Ag/AgCl electrodes were placed

over the muscle belly approximately 2 cm apart. Tape was placed over each

sensor to maintain solid contact during all movements. Data were collected

continuously for the duration of the tests for each participant. Impact times were

marked for each heading scenario within the data collection files.

Figure 4.5: Neck musculature used for EMG testing a) sternocleidomastoid; b) trapezius (Gray, et al., 1995)

a) b)

58

Following instrumentation, players performed 8 heading scenarios (Table

4.1). The headers were performed at two speeds, 6 m/s (low) and 8 m/s (high).

Balls were launched at subjects from 6 m away, providing ample reaction time at

both test speeds. Three types of headers were evaluated, passing, clearing, and

controlling. A passing header is when the player attempts to redirect the ball to

another player at a medium distance away, a clearing header is when the player

attempts to redirect the ball as far downfield as possible, and a controlling header

is when the player attempts to head the ball a very short distance down and in

front of them self.

Additionally, three neck muscle activity levels were explored. Prior to

heading the ball players were instructed to either pre-tense their neck muscles, to

perform headers with their muscles activated as they would normally, or to have

their neck muscles completely relaxed. Along with neck muscle activation levels,

prior to impact, subjects were also instructed to try to head the ball at a certain

target. This was done to try to represent the majority of heading scenarios that

players would see in the field as well as to provide comparisons to previous adult

data (Shewchenko, et al., 2005).

59

Table 4.1: Heading Scenarios (Shewchenko, et al., 2005) Task Heading

Scenario

Ball Speed Modification Ball Target

1 Controlling Low None Front, down, 2.5 m from

player

2 Passing Low None Front, down, 5.5 m from

player

3 Clearing Low None Up and away, as far as

possible

4 Passing Low Neck tensing Front, down, 5.5 m from

player

5 Passing Low Follow through Front, down, 5.5 m from

player

6 Passing Low Torso

alignment

Front, down, 5.5 m from

player

7 Clearing High None Up and away, as far as

possible

8 Clearing High Neck tensing Up and away, as far as

possible

EMG data was processed using MyoResearch XP Master Edition 1.07

(Noroxan, Inc., Scottsdale, AZ). Data were filtered using a notch filter to remove

the noise at 60 Hz prior to any additional processing. Following filtering, EMG

data were full-wave rectified and then the RMS was calculated using a 50 ms

60

moving average. The point of impact is designated as time 0 s for all

descriptions. Generally, all header related motion took place during the 50 ms

pre and post impact.

Torso angle, head angle, and relative head to torso angle were recorded

for each impact. Each angle was measured using the FAB from the original

neutral position that the player stood in during calibration. Therefore, the 0

degree position is the original calibration stance (Figure 4.2). Additionally,

positive designation was given to flexion and a negative designation was

assigned to extension for both head and torso motion. For the twist motions,

positive was assigned to left twist and negative to right twist. Motion data were

analyzed to determine the differences between the various heading scenarios

within each subject. Additionally, each heading scenario was analyzed

individually to determine subject variability.

Head acceleration was also recorded for each impact using the FAB.

These data were not taken at the center of gravity of the head will not be used to

determine injury risk. The data will be used strictly as a comparative between

tasks.

4.3 Results

Torso flexion, head flexion, and head rotation were measured for each

task. Focus was placed on the tasks that required no modification, tasks 1, 2, 3,

and 7. Task 2 was used for the standard heading case. Extensive subject

variability occurred for all tasks. Head rotation was the most consistent between

61

subjects because it at seemed to have similar patterns although actual angles

varied.

In task 2, used as the typical header case, male torso flexion ranged from

-14.81 to 27.13 degrees (41.94 degree span) at the point of impact (Figure 4.6)

with an average of 14.90 ± 15.49 degrees (Table 4.2). Peak torso flexion took

place following impact in 83 % (5 subjects) of the male players. In the remaining

17 % (1 subject), the entire task was performed in negative flexion, or torso

extension, with the minimum extension occurring just prior to impact. Peak torso

flexion ranged from -10.13 to 30.52 degrees with an average of 21.96 ± 17.06 for

the males during task 2.

Figure 4.6: Torso flexion for all male players during standard heading

scenario (task 2)

Subject variability was also seen in head flexion. Male head flexion for

task 2 ranged from -22.54 to 34.61 degrees on impact (Figure 4.7). The average

head flexion on impact was 0.91 ± 19.56 degrees with an average peak head

flexion of 8.44 ± 21.94 degrees. Although all male participants started in head

-40

-30

-20

-10

0

10

20

30

40

50

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Male Torso Flexion

Player 1

Player 2

Player 3

Player 4

Player 5

Player 6

62

extension for task 2, half of the participants had positive head flexion during task

2 and half did not. Therefore 3 of the participants’ peak head flexion, which

ranged from -22.54 to 34.69 degrees, was their minimum head extension.

Figure 4.7: Head flexion for all male players standard heading scenario

(task 2)

Head rotation for the male population remained relatively consistent

throughout task 2 (Figure 4.8). Each subject had a stable head rotation

throughout the task, but variability still existed between subjects. Head rotation

ranged from -22.54 to 34.61 degrees on impact with an average of 9.81 ± 25.90

degrees. The maximum values for head rotation ranged from -28.06 to 43.33

degrees with an average maximum of 19.27 ± 25.91 degrees.

-60

-50

-40

-30

-20

-10

0

10

20

30

40

50

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Male Head Flexion

Player 1

Player 2

Player 3

Player 4

Player 5

Player 6

63

Figure 4.8: Head rotation for all males during standard heading scenario

(task 2)

Female torso flexion for task 2 was similar to the males with a lack of

consistency between players (Figure 4.9). A large range occurred for the impact

torso flexion of 0.10 to 35.03 degrees. All players impacted the ball with their

torso in flexion with only 1 player starting with their torso extended. This pattern

is quite similar to the males. The average torso flexion on impact was 15.96 ±

10.77 degrees. The peak torso flexion ranged from 4.78 to 37.38 degrees with

an average of 24.07 ± 9.68. Although there is no clear pattern between subjects,

67 % of participants had a peak torso flexion in the between 24.40 and 29.51.

-80

-60

-40

-20

0

20

40

60

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Male Head Rotation

Player 1

Player 2

Player 3

Player 4

Player 5

Player 6

64

Figure 4.9: Torso flexion for all females during standard heading scenario

(task 2)

Female head flexion ranged from -26.91 to 10.31 degrees upon impact

(Figure 4.10) with an average impact flexion of 1.09 ± 12.99 degrees. Maximum

head flexion ranged from -17.56 to 17.10 degrees averaging 4.58 ± 12.29

degrees. The peak head flexion generally occurred just prior to impact.

Figure 4.10: Head flexion for all females during standard heading scenario

(task 2)

-20

-10

0

10

20

30

40

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Female Torso Flexion

Player 7

Player 8

Player 9

Player 10

Player 11

Player 12

Player 13

Player 14

Player 15

-60.00

-50.00

-40.00

-30.00

-20.00

-10.00

0.00

10.00

20.00

30.00

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Female Head Flexion

Player 7

Player 8

Player 9

Player 10

Player 11

Player 12

Player 13

Player 14

Player 15

65

Female head rotation is similar to the males in that it stays relatively

consistent for each participant across the entire task (Figure 4.11). Impact head

rotation ranged from -20.84 to 44.69 degrees with an average of 2.81 ± 20.57

degrees.

Figure 4.11: Head rotation for all females during standard heading

scenario (task 2)

In general, the within subject header tasks seemed to develop a more

consistent pattern. Although the subjects did not start or finish in the same

orientation for each task, the peaks and valleys of the torso and head flexion

generally occurred at similar time points with relationship to the impact. When

this varied, it generally did so during the high speed task (task 7). This pattern is

evident in Figures 4.10 – 4.12.

The example male participant’s torso flexion ranged from 11.96 to 18.14

degrees for the four different heading tasks investigated that did not require

modification (Figure 4.12). This is a much smaller range than that seen when

comparing between participants when they were performing the same task.

-40

-30

-20

-10

0

10

20

30

40

50

60

-100 -50 0 50 100An

gle

(D

eg

ree

s)

Time (ms)

Female Head Rotation

Player 7

Player 8

Player 9

Player 10

Player 11

Player 12

Player 13

Player 14

Player 15

66

Figure 4.12: Example of single male participant’s torso flexion for all

header tasks that lack modifications (1, 2, 3, and 7)

The example male participant’s head flexion showed a very distinct

pattern (Figure 4.13). Although the flexion ranged from -22.70 to 3.93 degrees,

the maximum extension took place at the same time for all four tasks, -60 ms.

The maximum flexion also took place at the same time, -10 ms, for all four tasks.

Figure 4.13: Example of single male participant’s head flexion for all

header tasks that lack modifications (1, 2, 3, and 7)

-15

-10

-5

0

5

10

15

20

25

30

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Male Torso Flexion

Task 1

Task 2

Task 3

Task 7

-80

-70

-60

-50

-40

-30

-20

-10

0

10

20

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Male Head Flexion

Task 1

Task 2

Task 3

Task 7

67

The example male participant’s head rotation has similar angles for the

first three tasks, but the high speed header (task 7) falls further out (Figure 4.14).

The range of head rotation on impact for the first three tasks is -8.73 to 8.18

degrees while the head rotation angle for task 7 was -27.88 degrees. This

indicates a change in technique when the ball is moving at a higher speed.

Figure 4.14: Example of single male participant’s head rotation for all

header tasks that lack modifications (1, 2, 3, and 7)

The female example participant’s torso flexion, head flexion, and head

rotation are very consistent for the first 3 tasks and then they do not follow similar

patterns for task 7 (Figures 4.15 - 4.17). This shows that the female patterns are

similar to the males. It is also further evidence of a technique change for higher

speed impacts. Female torso flexion on impact ranges from 14.7 to 22.79

degrees for tasks 1 – 3, but has a torso flexion of 70.48 degrees on impact for

task 7 (Figure 4.15). Similar patterns were found in head flexion with an impact

angle range of -15.5 – 12.91 degrees for the first three tasks and an impact angle

of -59.13 degrees for task 7 (Figure 4.16). Differences between the first three

-60

-40

-20

0

20

40

60

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Male Head Rotation

Task 1

Task 2

Task 3

Task 7

68

tasks and task 7 were very apparent in the head rotation during the pre-impact

stage (Figure 4.17). The impact head rotation ranged from 1.25 – 7.21 degrees

for the first three tasks with -9.84 degrees for task 7. Similar patterns were seen

throughout the male and female participants.

Figure 4.15: Example of single female participant’s torso flexion for all

header tasks that lack modifications (1, 2, 3, and 7)

Figure 4.16: Example of single female participant’s head flexion for all

header tasks that lack modifications (1, 2, 3, and 7)

0

10

20

30

40

50

60

70

80

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Female Torso Flexion

Task 1

Task 2

Task 3

Task 7

-120

-100

-80

-60

-40

-20

0

20

40

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Female Head Flexion

Task 1

Task 2

Task 3

Task 7

69

Figure 4.17: Example of single female participant’s head rotation for all

header tasks that lack modifications (1, 2, 3, and 7)

In addition to comparing all of the un-modified tasks, comparable tasks

were also compared. The four passing tasks were compared (tasks 2, 4, 5, and

6) and the two clearing tasks were also compared (tasks 7 and 8). These tasks

were compared for both males and females. All figures are representative of an

example participant.

The first tasks that were compared were the passing tasks. Male torso

flexion upon impact for the passing was relatively consistent within participants.

For the male example participant, torso flexion ranged from 16.5 to 22.55

degrees (Figure 4.18). This within subject consistency was expected, especially

since all four tasks were at the low ball impact speed.

-100

-80

-60

-40

-20

0

20

40

60

80

100

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Female Head Rotation

Task 1

Task 2

Task 3

Task 7

70

Figure 4.18: Example of single male participant’s torso flexion for all

passing header tasks (2, 4, 5, and 6)

The male example participant’s head flexion for the passing tasks was not

as consistent on impact, but did show a similar pattern of flexion over the

duration of the task. The head flexion ranged from -22.54 to 5.53 degrees

(Figure 4.19). Two of the impacts, tasks 2 and 5, were performed with the head

in extension, while tasks 4 and 6 were performed with the head in flexion. The

maximum extension took place at very similar time points for the four tasks.

0

5

10

15

20

25

30

35

40

45

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Male Torso Flexion

Task 2

Task 4

Task 5

Task 6

71

Figure 4.19: Example of single male participant’s head flexion for all

passing header tasks (2, 4, 5, and 6)

During the passing tasks, similarly to the non modified tasks, the head

rotated very little during the activity. For the male example participant, head

rotation upon impact ranged from -4.18 to 7.78 degrees (Figure 4.20). This a

very small range, indicating that head rotation was not altered between tasks.

Figure 4.20: Example of single male participant’s head rotation for all

passing header tasks (2, 4, 5, and 6)

-70

-60

-50

-40

-30

-20

-10

0

10

20

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Male Head Flexion

Task 2

Task 4

Task 5

Task 6

-30

-20

-10

0

10

20

30

40

-100 -50 0 50 100An

gle

(D

eg

ree

s)

Time (ms)

Male Head Rotation

Task 2

Task 4

Task 5

Task 6

72

The female example participant showed similar consistency between the

four passing tasks. When inconsistencies arose, they were with task 5, which

was modified to include extended follow through. Therefore, the high torso

flexion for task 5 can be attributed to the follow through. The female example

participant had a torso flexion on impact that ranged from 14.24 to 15.99 degrees

for task 2, 4, and 6. Task 5, however, had a torso flexion on impact of 47.23

degrees (Figure 4.21).

Figure 4.21: Example of single female participant’s torso flexion for all

passing header tasks (2, 4, 5, and 6) Female head flexion showed consistency with the pattern of flexion, the

minimum flexion took place at approximately the same time point for each task,

as did the maximum flexion (Figure 4.22). For the female example participant,

head flexion ranged from -6.76 to 11.86 degrees with three of the four tasks

taking place with the head in extension. Only task 4 took place with the head in

the flexion position.

0

10

20

30

40

50

60

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Female Torso Flexion

Task 2

Task 4

Task 5

Task 6

73

Figure 4.22: Example of single female participant’s head flexion for all

passing header tasks (2, 4, 5, and 6)

Female head rotation was similar for tasks 2, 4, and 6. Task 5, the task

with the follow through modification, was performed with the head rotated right,

but to a similar degree as the other three tasks (Figure 4.23). The three tasks

with a leftward rotation ranged from 5.25 to 7.63 degrees while the task with

rotation to the right was rotated -7.35 degrees on impact.

Figure 4.23: Example of single female participant’s head rotation for all

passing header tasks (2, 4, 5, and 6)

-60

-50

-40

-30

-20

-10

0

10

20

30

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Female Head Flexion

Task 2

Task 4

Task 5

Task 6

-20

-15

-10

-5

0

5

10

15

20

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Female Head Rotation

Task 2

Task 4

Task 5

Task 6

74

Tasks 7 and 8 showed slightly more inconsistency within participants than

the others previously discussed. This is mostly attributed to some abnormalities

within task 7 for both male and female participants. The differences are noted in

both the example male and female torso flexion and head rotation. The oddities

were at similar time points for both participants and took place prior to impact.

This is most likely due to task 7 being the first high speed task performed, and

the participants had yet to get used to the new ball speed.

Figure 4.24: Example of single male participant’s torso flexion for all

clearing header tasks (7 and 8)

Male torso flexion upon impact was 15.9 degrees for task 8 and 19.36 for

task 9. Although some differences did occur, they were at the very beginning of

the task and flexion upon impact remained very similar (Figure 4.24). For head

flexion, a very similar pattern occurred with minimum flexion and maximum

flexion occurring at approximately the same time for tasks 7 and 8 (Figure 4.25).

Head rotation for the males was -2.8 degrees for task 7 and -27.88 degrees for

task 8 (Figure 4.26).

-15

-10

-5

0

5

10

15

20

25

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Male Torso Flexion

Task 7

Task 8

75

Figure 4.25: Example of single male participant’s head flexion for all

clearing header tasks (7 and 8)

Figure 4.26: Example of single male participant’s head rotation for all

clearing header tasks (7 and 8) Females had similar patterns to the males for the clearing tasks. Torso

flexion was 70.48 degrees for task 7 and 23.69 degrees for task 8 upon impact

(Figure 4.27). Head flexion was also like the male head flexion in that the impact

angles were not alike, but the pattern of flexion over the event remained similar

(Figure 4.28). Head rotation was also much like the males for the clearing tasks

-60

-50

-40

-30

-20

-10

0

10

20

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Male Head Flexion

Task 7

Task 8

-60

-40

-20

0

20

40

60

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Male Head Rotation

Task 7

Task 8

76

with differences between the tasks occurring in the early stages of the event.

The impact head rotation for task 7 was -9.84 degrees and 21.83 degrees for

task 8 (Figure 4.29).

Figure 4.27: Example of single female participant’s torso flexion for all

clearing header tasks (7 and 8)

Figure 4.28: Example of single female participant’s head flexion for all

clearing header tasks (7 and 8)

0

10

20

30

40

50

60

70

80

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Female Torso Flexion

Task 7

Task 8

-120

-100

-80

-60

-40

-20

0

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Female Head Flexion

Task 7

Task 8

77

Figure 4.29: Example of single male participant’s head rotation for all

clearing header tasks (7 and 8)

Averages for each task for head flexion, head rotation, and torso flexion

are presented in Table 4.2. The large standard deviations indicate the sizeable

subject variability. Averages were not statistically compared between males and

females because of the large subject variability within their own populations.

-100

-80

-60

-40

-20

0

20

40

60

80

100

-100 -50 0 50 100

An

gle

(D

eg

ree

s)

Time (ms)

Female Head Rotation

Task 7

Task 8

78

Table 4.2: Average angles at impact for each heading task

Heading

Task

Head Flexion

(degrees)

Head Rotation

(degrees)

Torso Flexion

(degrees)

Male Female Male Female Male Female

1

4.99 ±

14.29

12.99 ±

13.74

14.94 ±

24.23

5.06 ±

24.80

18.55 ±

11.88

14.14 ±

8.11

2

0.91 ±

19.56

1.09 ±

12.99

9.81 ±

25.90

2.81 ±

20.57

14.90 ±

15.49

15.96 ±

10.77

3

-12.18 ±

22.21

-3.78 ±

13.42

17.91 ±

29.34

11.48 ±

22.34

8.53 ±

6.67

13.60 ±

10.22

4

-0.09 ±

15.92

7.83 ±

15.16

3.44 ±

27.01

6.89 ±

22.52

14.10 ±

14.35

10.97 ±

12.22

5

7.99 ±

20.90

-5.94 ±

21.56

10.54 ±

25.16

6.96 ±

26.44

17.50 ±

12.21

13.74 ±

17.67

6

12.78 ±

20.40

9.58 ±

16.33

0.95 ±

13.77

6.60 ±

22.97

11.21 ±

16.59

15.48 ±

13.69

7

-4.60 ±

15.56

-15.02 ±

19.88

11.28 ±

34.74

5.02 ±

20.15

9.12 ±

8.02

23.66 ±

19.48

8

3.88 ±

14.01

-6.05 ±

11.81

7.51 ±

35.76

4.59 ±

19.99

8.99 ±

10.02

12.60 ±

10.72

EMG values provided similarly variable results as the body position

angles. For the males, this variability was more noticeable in the trapezius

muscles (Figure 4.30). The females had a more overall inconsistency as the

variation was not limited to a specific muscle group (Figure 4.31). There was

also considerable variation within the subjects when comparing tasks. However,

79

it was not limited to a specific task as it was in the body position data. The mean

peak RMS EMG, used to provide a clearer view of muscle activation, values

(Table 4.3) indicate a much more consistent appearance of the peaks when

comparing players than the overall values provide (Figure 4.30, 4.31).

80

Figure 4.30: Peak EMG for all male players for each muscle a) left

sternocleidomastoid, b) right sternocleidomastoid, c) left trapezius, d) right trapezius

81

Figure 4.31: Peak EMG for all female players for each muscle a) left

sternocleidomastoid, b) right sternocleidomastoid, c) left trapezius, d) right trapezius

82

Table 4.3: Mean peak RMS EMG values for each muscle and each task

Muscle Gender

Task 1

Mean Peak

(mV)

Task 2

Mean Peak

(mV)

Task 3

Mean Peak

(mV)

Task 7

Mean Peak

(mV)

Left

Sternocleidomastoid

M 1.17 ± 0.46 1.63 ± 0.77 1.89 ± 0.53 2.29 ± 1.38

F 1.26 ± 1.17 1.18 ± 0.77 1.45 ± 1.25 1.33 ± 0.85

Right

Sternocleidomastoid

M 1.68 ± 1.02 1.15 ± 0.31 1.36 ± 0.91 1.48 ± 0.70

F 1.07 ± 0.71 1.32 ± 1.25 1.68 ± 1.75 1.06 ± 0.52

Left Trapezius M 1.86 ± 2.18 1.46 ± 1.28 1.95 ± 1.51 1.45 ± 0.34

F 1.15 ± 1.35 0.97 ± 0.56 1.68 ± 1.32 1.39 ± 1.51

Right Trapezius M 2.01 ± 1.43 2.40 ± 2.18 2.23 ± 1.61 2.60 ± 1.66

F 1.23 ± 1.07 1.29 ± 1.02 1.08 ± 0.71 1.13 ± 0.57

The majority of the muscle activation takes place prior to the impact event.

This is true for both the male and female participants (Figure 4.32, 4.33). Muscle

activation prior to and at impact being higher than that post-impact was seen

throughout the participants. Subjects varied as to how even their contractions

were. As seen in Figure 4.32, the male example participant had relatively

consistent contraction on both the left and right sides. In contrast, the female

example participant had much less contraction in the right trapezius versus the

left (Figure 4.33). These data also indicate subject variability.

83

Figure 4.32: Sample RMS EMG for one male player for each muscle a) left

sternocleidomastoid, b) right sternocleidomastoid, c) left trapezius, d) right trapezius

84

Figure 4.33: Sample RMS EMG for one female player for each muscle a) left sternocleidomastoid, b) right sternocleidomastoid, c) left trapezius, d) right

trapezius

85

Head acceleration was evaluated for each task for both males and

females. When looking at average head acceleration for each task, the head

acceleration appears much less variable than the biomechanics and the EMG

between tasks. There is still a wide range between players in each task which is

evident by the standard deviations (Table 4.4). Tasks 7 and 8 do not have higher

head acceleration than the low speed header tasks. Also the modifications do

not appear to have caused a change in head acceleration. The males do,

however, have higher head acceleration for each task. When evaluating

individual players for each task, there was a consistency in the head acceleration

between tasks for each player, generally with one to two tasks being lower than

the rest. These tasks did not show a pattern between subjects.

Table 4.4 : Average angular head acceleration on impact for each heading task

Heading

Task

Head Acceleration (radians/s2)

Male Female

1 89.47 ± 33.20 60.09 ± 23.08

2 100.33 ± 18.47 60.90 ± 20.05

3 97.92 ± 39.07 79.62 ± 30.30

4 95.47 ± 18.25 63.44 ± 16.25

5 95.61 ± 48.61 49.50 ± 22.19

6 75.00 ± 20.99 48.93 ± 14.07

7 109.73 ± 42.58 75.04 ± 25.60

8 85.46 ± 42.58 81.05 ± 28.47

86

4.4 Discussion

Head and back angles were found to have similar tendencies to previous

studies (Shewchenko, et al., 2005). Shewchenko et al. (2005) found that some

players tend to flex their head during contact while others extend. The current

study found this to occur as well. Overall, the torso angle was in flexion for the

majority of players at ball contact which contrasted with the previous findings

where players seemed to remain in a relatively neutral position upon impact and

move into flexion during follow through (Shewchenko, et al., 2005). Players were

found to continue with torso flexion in the follow through period in the youth

population as well. It appears that they arrive at this state earlier than the adult

players. This trend was found in all scenarios, both un-modified and modified.

Table 4.5 describes the average head flexion for the previous study

(Shewchenko, et al., 2005) in comparison to the current study.

87

Table 4.5: Average head flexion for each task

Heading

Task

Current Study

Shewchenko et al.

(2005)

Male (n = 6) Female (n = 9) Male (n = 7)

1 5 13 33

2 1 1 18

3 -12 -4 9

4 0 8 18

5 8 -6 14

6 13 10 15

7 -5 -15 4

8 4 -6 16

Due to the overall variation in player body position, it is expected that

these results would vary from previous results found in adult players. The player

variation is consistent with previous findings (Shewchenko, et al., 2005). This is

most likely due to the many possible heading scenarios. Therefore, players do

not use muscle memory to perform the task in the same manor every time, but

instead learn to adjust to whatever scenario occurs. This provides unlimited

possibilities for heading scenarios and would most likely result in additional

variation in results. The current study only looked at redirection directly at the

source of the ball launch. If additional scenarios were introduced to provide

redirection in other ways, additional distinctions would be made particularly in

head rotation and EMG activity.

88

EMG activity provided insight into the activation of the neck muscles

during the phases of heading. The majority of neck muscle activity during soccer

head was found to take place prior to impact and during impact. Overall, the

peak values were similar, but otherwise there was extensive subject variability.

In addition to inconsistency between subjects, there were also differences from

one task to another within the same subject. Although this was less noticeable in

the EMG results as in the body position results.

Angular head acceleration was compared between tasks to determine if

any of the modifications that were instituted in the study provided a decrease in

head acceleration which could lead to suggested alterations to current heading

methods. It was found that none of the modifications decreased the average

linear head acceleration for the males or the females. When looking at individual

players, decreases were seen between tasks, but no pattern was visible from

one player to the next. This just reiterates that there is extensive player

variability and what works at reducing injury risk for one player may not work for

another.

Comparisons between genders or between tasks were challenging to

make. Based on the results of the current study, it indicates that differences are

not related these variables, but that the differences occur between each player.

This also made making a comparison with the adult players unproductive as any

difference would most likely have nothing to do with age, but with the players

being different and the scenarios being slightly different.

89

One of the main limitations of the current study was the method of head

acceleration measurement. This limited the overall usefulness of the study since

we lacked the ability to make comparisons with previous studies. Also, due to

the method of head acceleration measurement, no injury criteria could be

evaluated. Initial trials were performed with a head acceleration measurement

system in place, but due to system interference the system that allowed for

angular head acceleration measurement and linear head acceleration

measurement at the center of gravity of the head was not possible. It would be

of interest in future studies to measure head acceleration in the youth population

during heading events to determine if any modifications or technique changes

can reduce linear or angular head acceleration.

90

CHAPTER 5

ACCELERATION MEASUREMENT SYSTEM VALIDATION

5.1 Introduction

Previous studies of soccer heading have lacked the ability to measure

real-time game impacts (Naunheim, et al., 2003, Naunheim, et al., 2000,

Shewchenko, et al., 2005, 2005, 2005). These previous head acceleration

measurements were done using re-creations in a laboratory or restricted setting.

By using a novel head acceleration measurement system, the Head Impact

Telemetry System (HITS) (Simbex, Lebanon, NH), linear and angular head

acceleration can be measured during actual games. The system has been

implemented and validated in both football helmets and boxing headgear

(Beckwith, et al., 2007, Duma, et al., 2005, Manoogian, et al., 2006).

Previously HITS (Figure 5.1) was implemented in football helmets (Duma,

et al., 2005), and is now commercially available for use (Duma, et al., 2005). The

data processing algorithm, previously developed by Crisco et al. (2004), allows

for calculation of both linear and angular head acceleration (Crisco, et al., 2004).

System and algorithm validation was performed using a HIII dummy

instrumented with a 3-2-2-2 accelerometer setup for both football and boxing

(Crisco, et al., 2004, Duma, et al., 2005). Correlations were found to be strong

with an R2 = .97 (Duma, et al., 2005).

This system uses six wireless accelerometers which are placed inside a

football helmet along with a wireless transceiver, data acquisition, and on-board

memory (Duma, et al., 2005). The accelerometers are spring-mounted so that

91

they are closely coupled to the head (Duma, et al., 2005). This ensures that

head acceleration is measured as opposed to helmet acceleration. Recording of

impact data occurs when any accelerometer registers above the threshold of 10

g. When the threshold is reached, 40 ms of data are recorded. This information

is then time stamped and downloaded to the sideline computer for later

processing using the algorithm (Crisco, et al., 2004).

Figure 5.1: HIT system

The HIT system has recently been modified for use in boxing headgear

(Beckwith, et al., 2007). The system is much like the football system, but the

boxing headgear has a total of twelve accelerometers as opposed to six.

Additionally, accelerometer placement was more of a concern because the

headgear has no outer shell and the accelerometers had to be placed where it

92

was least likely for impacts to occur. Therefore, the accelerometers were placed

toward the back of the headgear. The battery pack and transmitter were placed

in the back panel. This system was validated using a Hybrid III (HIII) head and

neck (Beckwith, et al., 2007). Using a 3-2-2-2 accelerometer array mounted in

the HIII head, linear and angular accelerations were calculated to compare to

those calculated using the HIT system. Facial, forehead, side, and left chin

impacts were performed using a pendulum impactor at 3 m/s, 5 m/s, and 7 m/s.

Four impacts were performed at each location and speed, with the exception of

forehead impacts not being performed at 7 m/s.

Linear head acceleration, angular head acceleration, impact location, GSI,

and HIC were calculated for both systems. High correlations, r2 = .91, for both

linear and angular head acceleration were found. Estimations made by the HITS

headgear were slightly low (2%) for linear acceleration and high (8%) for angular

acceleration. RMS error was calculated over the time series and was an average

of 5.9 ± 2.6 g for linear acceleration and 595 ± 405 rad/s2. Additionally,

correlations between the two systems were calculated for HIC and GSI. Again,

high correlations were found, r2 = .88 and r2 = .89, for HIC and GSI respectively.

It was found that a limitation of the headgear development was the need to place

accelerometers in the back of the headgear. While this potentially creates error,

it is necessary for avoiding direct accelerometer impact. This could also be a

problem during development of the soccer headgear because impacts will take

place in the forehead region and there will be no padding. Therefore,

accelerometers will need to be placed in the rear portion of the headgear.

93

Implementation of this device into a headband system that can be worn

during normal soccer play would allow for collection of real-time game head

accelerations without restricting player movement. The headband system will be

modeled after a commercially available headgear, but will provide none of the

protective effects to the players. The limitation of this system is that it does

require the player to wear some type of headband to allow for player

instrumentation to take place. The system has previously been used in sports

that require helmets or headgear of some type, but in soccer this is not the case.

Although headgear is available for soccer players, it is not a required piece of

equipment.

HITS Validation

Prior to any field testing, laboratory validation of the HITS headgear was

executed. This was done by testing various possible scenarios that could occur

during soccer games. These scenarios included head to head impacts and head

to ball impacts. These impacts were done using a modified 50th percentile HIII

head (Denton ATD, Milan, OH) instrumented with a 3-2-2-2 accelerometer setup.

The HIII head was then fitted with a HITS headgear. The scenarios were tested

using an air cannon (head to ball) and a linear impactor (head to head).

Comparisons will then be made between the HIII accelerations and the HITS

accelerations. The level of error obtained from HITS will be analyzed.

5.2 Methodology

A soccer headband HITS (Simbex Inc., Lebanon, NH), similar to those

commercially available, was instrumented with 6 (± 250 G) single-axis linear

94

accelerometers (Analog Devices, Inc.) (Figure 5.2). In order to measure forces

normally seen during the play of soccer, no padding was placed in the headband.

All accelerometers were placed in the back of the headband in order to avoid ball

contact during heading events. The battery pack, placed in the back of the

headband, is a rechargeable Nickel Metal-Hydride battery which allows for

extended use, 1 – 2 weeks depending on use, and minimal additional weight,

with the entire headband system weighing 147 g. The headband has a threshold

level of 10 g, meaning that when any accelerometer registers a reading of 10 g or

greater, the impact will be downloaded. Once an impact above the threshold is

recognized, 8 ms prior to the impact and 32 ms post impact will be recorded.

Data are downloaded to the sideline computer as long as players remain within

range, approximately 200 yards. If players are out of range, up to 100 impacts

can be stored within the headgear itself until the player returns within range.

Figure 5.2: Back of HITS headband with circles marking accelerometer

placement

95

The 50th percentile male Hybrid III (HIII) head was used as the standard of

comparison for the linear and angular head accelerations for the HITS headband.

The HIII head was instrumented with nine linear accelerometers (Endevco 7264C

and 7264D) in the 3-2-2-2 setup which was mounted inside the modified HIII

head (Padgaonkar, et al., 1975) with a tri-axial linear block placed at the center of

gravity (CG) of the head form. HIII head acceleration data were collected at

20,000 Hz while HITS acceleration data were collected at 1,000 Hz. The

headband was placed on the HIII head and Velcro straps were tightened to the

manufacturer’s specifications allowing all accelerometers to make firm contact

with the head form. Impacts occurred at the forehead, side, and temple of the

HIII head using an air cannon and a linear impactor. No impacts were performed

with an impact direction going directly through the center of gravity of the head as

this would be highly unlikely in an on-field data collection scenario. Both ball to

head and head to head contacts were simulated.

Ball to head impacts were performed using an air cannon with a barrel

fitted to accommodate a soccer ball (Figure 5.3). A standard size 5 soccer ball

with a mass of 450 g, a diameter of 22 cm, and an inflation pressure of 10 psi

was used for all testing. The ball was shot through a three screen chronograph

in order to obtain velocity readings. In order to obtain velocities representative of

soccer impacts, Helium was used in the air cannon. Impacts were performed at

8 m/s, 10 m/s, and 12 m/s based on previous research performed (Withnall, et

al., 2005). Ten impacts were performed at each velocity to the forehead (n=30),

right side (n=30), and left temple of the head (n=30). These locations were

96

chosen to represent various impacts seen during soccer play, as well as to

provide a variety of impact locations possible in soccer games while not

impacting accelerometers directly.

Figure 5.3: Air cannon with soccer barrel

Head to head impacts were conducted by mounting one HIII head and

neck face down, to a linear impactor and placing another on a trolley in front of

the impactor (Figure 5.4). By placing the head and neck on the impactor, some

flex was possible allowing for an impact more closely representative of an on-

field situation. The head on the trolley was instrumented as described above and

the HITS headgear was placed on it. Tests were run at three velocities: 2.5 m/s,

3.5 m/s, and 4.75 m/s based on previous research on head to head field impacts

(Withnall, et al., 2005). Ten impacts were performed at each of these velocities

at two locations: the forehead (n=30), and to the right side (n=30). These

97

locations represent standard impacts seen during soccer play, and provide a

more severe impact condition than the ball to head impacts.

Figure 5.4: Head to head impact test setup for forehead testing

Data analysis was conducted to determine the agreement between the

HIII and the HITS headgear. Linear head acceleration and angular head

acceleration were calculated for both systems. The HITS system data was

processed using an algorithm previously described in detail by Chu et al. (2006)

which calculates both linear and angular head acceleration based on the 6

accelerometer measurements (Chu, et al., 2006). Linear regression was used to

compare the systems for the ball to head impacts, the head to head impacts, and

then all impacts together. This was done for both linear and angular head

accelerations.

Additionally, root mean square (RMS) error was calculated using the

98

equation below for the duration of the impacts for linear head acceleration. This

will provide information about specific portions of the curve and how closely they

match up in value. Cross correlation was also calculated for linear head

acceleration. These values provide insight into how strongly the variables are

related. Cross correlation values were assessed using a scale to determine

correlation strength: >0.95 was considered “excellent”, >0.85 was considered

“good”, and >0.75 was considered “acceptable”. Due to the fact that the HIT

System has built-in data acquisition and wireless communication, the HIII and

HITS could not be linked. In order to compare HITS data and HIII data during

post-processing, data was synchronized at the point of minimum RMS error. The

two resultants were synchronized by shifting the HIII data incrementally until a 40

ms span of the HIII gave the lowest cross correlation factor.

Due to the fact that data were collected at different frequencies, HITS data

must be time matched to HIII data. In order to do this, the HIT System data was

first up sampled to match the sampling frequency of the HIII so that no HIII data

was lost and the overall numbers from the HIT System output was unaffected.

The two resultants were then synchronized by shifting the HIII data incrementally

until a 40 ms span of the HIII gave the lowest cross correlation factor.

"� ���#� � $∑ �!�,� ' !(,��()��� �

Where:

x1 = HITS measurement at single time point

x2 = HIII measurement at the same time point

99

5.3 Results

Linear regressions were performed for the ball to head impacts, head to

head impacts, and all impacts combined. All impact locations are combined in

the linear regressions. Regressions for both linear and angular accelerations are

shown below with each impact location denoted by a different shape (Figures 5.5

– 5.10).

Ball to head comparisons provided minimal correlation for both linear and

angular acceleration, R2 = 0.3403 and R2 = 0.5716 respectively (Figure 5.5, 5.6).

Correlations were also investigated for each location separately. For ball to head

testing, the forehead had the highest correlation for linear acceleration (R2 =

0.4419), followed by the right side (R2 = 0.3975) and the left temple (R2 =

0.2446). Angular acceleration had the highest correlation in the right side

impacts (R2 = 0.8022), followed by the left temple (R2 = 0.2832) and the forehead

(R2 = 0.1600). The minimal correlation is most likely due to the fact that although

various impact velocities were tested, a range of linear and angular head

accelerations were not obtained. Output accelerations for both the HIII and HITS

systems were very limited in range when impacted over the three ball to head

impact velocities. Although the correlations are not ideal due to the lack of

acceleration range, the average difference between the two acceleration

measurements is minimal. This is especially true for the linear head acceleration

which has an average difference of 2.25 g. The angular head acceleration has

an average difference of 100.58 rad/s2.

100

Figure 5.5: Linear regression of linear acceleration for HIII and HITS ball to

head conditions

Figure 5.6: Linear regression of angular acceleration for HIII and HITS ball

to head conditions

Head to head comparisons provided strong correlations for both linear and

angular acceleration, R2 = 0.8940 and R2 = 0.8998 respectively (Figure 5.7, 5.8).

The forehead location had a higher correlation in both linear acceleration (R2 =

0.9653) and angular acceleration (R2 = 0.9799) than the left side which had an

R2 = 0.8411 for linear acceleration and R2 = 0.8979 for angular acceleration. A

101

much wider range of output velocities was provided from the three impact

velocities used in the head to head impact conditions providing a much stronger

dataset for linear regression. This demonstrates that the HITS is a very good

system for measuring higher velocity impacts. The average differences between

the two systems are -2.01 g for the linear acceleration measurements and -

1721.05 rad/s2 for the angular acceleration measurements. These differences

are calculated over all three impact velocities.

Figure 5.7: Linear regression of linear acceleration for HIII and HITS head

to head conditions

102

Figure 5.8: Linear regression of angular acceleration for HIII and HITS head

to head conditions

Linear regressions were also performed for all head impacts, ball to head

and head to head, combined. This was done to determine the overall accuracy

of the system for the range of velocities over which it will be used. Very strong

correlations were found over all of the impact conditions, R2 = 0.9437 and R2 =

0.9194 for linear acceleration and angular acceleration respectively (Figure 5.9,

5.10). This provides a very strong basis for using the system for future soccer

research.

103

Figure 5.9: Linear regression of linear acceleration for HIII and HITS ball to

head and head to head conditions combined

Figure 5.10: Linear regression of angular acceleration for HIII and HITS ball

to head and head to head conditions combined

Peak linear and rotational acceleration was also calculated for each of the

impacts. These values for both the HITS and HIII systems are shown in Tables

5.1 - 5.4. All values shown are the average for the impact condition listed.

104

The HITS system slightly over predicts linear head acceleration in the ball

to head impacts. This can be seen in all impact conditions, except the forehead

condition at 12 m/s. At this condition the HITS provides a slight under prediction

(Table 5.1).

Table 5.1: Average Peak Linear Accelerations for Ball to Head Conditions

8 m/s 10 m/s 12 m/s

Impact

Location HITS (g) HIII (g) HITS (g) HIII (g) HITS (g) HIII (g)

Forehead

15.20 ±

0.49

12.14 ±

0.64

18.23 ±

1.51

16.57 ±

0.70

18.70 ±

2.54

21.09 ±

1.35

Right Side

18.03 ±

4.85

13.31 ±

0.76

19.93 ±

3.20

17.67 ±

0.87

22.98 ±

4.02

21.13 ±

0.93

Left Temple

18.03 ±

3.53

13.39 ±

0.46

18.72 ±

3.26

18.08 ±

0.46

21.64 ±

4.33

21.45 ±

1.11

Angular head acceleration for the ball to head impacts provides a different

pattern for the peak values. The HITS system over predicts angular head

acceleration in the ball to head impacts for all conditions except the forehead

impacts. Forehead impacts at all three velocities have the HITS under predicting

angular acceleration (Table 5.2).

105

Table 5.2: Average Peak Angular Accelerations for Ball to Head Conditions

8 m/s 10 m/s 12 m/s

Impact

Location

HITS

(rad/s2)

HIII

(rad/s2)

HITS

(rad/s2)

HIII

(rad/s2)

HITS

(rad/s2)

HIII

(rad/s2)

Forehead

822.07 ±

244.4

833.12 ±

74.40

857.74 ±

259.28

954.86 ±

95.32

949.57 ±

362.40

1450.81 ±

252.08

Right

Side

1416.29 ±

94.57

1343.23 ±

197.30

1828.96 ±

334.27

1509.16 ±

152.15

1882.67 ±

318.00

1715.39 ±

378.91

Left

Temple

1958.79 ±

425.65

1321.29 ±

179.61

1453.86 ±

535.28

1281.71 ±

282.16

1917.18 ±

348.35

1695.22 ±

347.64

Linear head acceleration for the head to head impacts shows a general

over prediction by the HITS for the two lower impact velocities. At the 4.75 m/s

impact condition, the HITS under predicts linear head acceleration for both the

forehead and left side conditions. Although a general over prediction occurs,

values are very similar as shown by the strong correlations between the two

systems (Table 5.3).

106

Table 5.3: Average Peak Linear Accelerations for Head to Head Conditions

2.5 m/s 3.5 m/s 4.75 m/s

Impact

Location HITS (g) HIII (g) HITS (g) HIII (g) HITS (g) HIII (g)

Forehead

33.68 ±

0.55

32.54 ±

0.69

74.26 ±

4.10

66.93 ±

1.85

115.45 ±

8.02

125.14 ±

2.80

Left Side

37.14 ±

3.46

30.58 ±

2.06

69.58 ±

10.12

63.18 ±

3.31

95.82 ±

13.02

119.65 ±

4.94

Angular head acceleration for the head to head impacts shows that the

HITS under predicts the angular head acceleration slightly for nearly all the

impact conditions. The HITS system only over predicts angular head

acceleration in the head to head impacts for one impact condition, the left side at

2.5 m/s. This is shown in Table 5.4 below.

Table 5.4: Average Peak Angular Accelerations for Head to Head Conditions

2.5 m/s 3.5 m/s 4.75 m/s

Impact

Location

HITS

(rad/s2)

HIII

(rad/s2)

HITS

(rad/s2)

HIII

(rad/s2)

HITS

(rad/s2)

HIII

(rad/s2)

Forehead

1245.63 ±

109.99

1544.50 ±

32.89

2013.79 ±

176.65

3177.79 ±

205.66

6930.20 ±

530.95

9414.49 ±

218.99

Left Side

2795.55 ±

179.72

2366.77 ±

456.59

5629.58 ±

550.05

6020.31 ±

365.15

9801.20 ±

1413.74

16218.41 ±

855.65

Root mean square (RMS) error was calculated for each impact on a point

by point basis. An example of the waveforms being compared is shown in

107

Figures 5.11 – 5.15. An average was then calculated for each impact and each

impact condition. The average RMS error of the linear accelerations for ball to

head impacts at the 8 m/s condition was 2.04 ± 0.25 for forehead impacts, 3.55 ±

0.44 for right side, and 2.20 ± 0.74 for left temple impacts. Similarly RMS error

for the 10 m/s condition was 2.35 ± 0.27 for forehead impacts, 3.86 ± 0.62 for

right side, and 1.97 ± 0.54 for left temple impacts. The 12 m/s conditions had

RMS errors of 2.55 ± 0.77 for forehead impacts, 3.08 ± 0.94 for right side, and

3.89 ± 1.21 for left temple impacts.

Figure 5.11: Linear acceleration for both HIII and HITS for one ball to head

forehead impact at the 12 m/s condition

108

Figure 5.12: Linear acceleration for both HIII and HITS for one ball to head

right side impact at the 12 m/s condition

Figure 5.13: Linear acceleration for both HIII and HITS for one ball to head

left temple impact at the 12 m/s condition

RMS errors were also calculated for the head to head conditions (Figure

5.14, 5.15). Head to head impacts RMS errors were 2.69 ± 0.32 for the forehead

and 5.83 ± 0.67 for the left side at the 2.5 m/s condition, 5.82 ± 0.65 for the

forehead and 15.19 ± 1.30 for the left side at the 3.5 m/s condition, and 9.47 ±

0.57 for the forehead and 21.89 ± 2.62 for the left side at the 4.75 m/s condition.

109

These values are higher due to the higher accelerations provided from the head

to head impact conditions.

Figure 5.14: Linear acceleration for both HIII and HITS for one head to head

forehead impact at the 4.75 m/s condition

Figure 5.15: Linear acceleration for both HIII and HITS for one head to head

left side impact at the 4.75 m/s condition

Cross correlations were performed and demonstrate a strong relationship

between the two systems. Average cross correlation (r) values were 0.95 ± 0.01

for forehead impacts, 0.88 ± 0.05 for right side, and 0.95 ± 0.04 for left temple

110

impacts for the 8 m/s condition, 0.94 ± 0.01 for forehead impacts, 0.87 ± 0.04 for

right side, and 0.96 ± 0.02 for left temple impacts for the 10 m/s condition, and

0.95 ± 0.03 for forehead impacts, 0.96 ± 0.02 for right side, and 0.92 ± 0.04 for

left temple impacts for the 12 m/s condition. Head to head impacts had cross

correlation values of 0.97 ± 0.01 for the forehead and 0.94 ± 0.01 for the left side

at the 2.5 m/s condition, 0.98 ± 0.00 for the forehead and 0.88 ± 0.04 for the left

side at the 3.5 m/s condition, and 0.94 ± 0.00 for the forehead and 0.83 ± 0.02 for

the left side at the 4.75 m/s condition. All ball to head conditions fell either

within the good or excellent range when looking at cross correlation values. Of

the nine ball to head conditions, five of them were above the 0.95 value required

for an excellent rating. Head to head conditions also provided very strong cross

correlation values, with five of six conditions falling into either the good or

excellent categories.

5.4 Discussion

Although attempts have been made to determine head acceleration during

soccer heading events, a system for on field data collection had not been

previously available. A system has now been created for research purposes,

however validation was necessary before the system could be used to collect

data during a game or scrimmage situation. The results show that the new

soccer HITS system correlates well with the standard measurement system of

the HIII 3-2-2-2 accelerometer system. Locations and impact velocities were

chosen to simulate events that take place in normal soccer play.

111

Good cross correlation values were found for the linear accelerations for

all conditions with two of the conditions having excellent correlation. The lowest

correlation value of 0.83 ± 0.02 was for the left side during head to head impact

and is not a location that is expected to be frequently impacted during soccer

play. Even as the lowest correlation, it still shows an acceptable level of

agreement. Additionally, all other linear acceleration cross correlation values

exceeds the 0.85 value and shows a very well matched system.

Strong correlations were found between the systems for both linear and

angular head acceleration for the head to head condition, 0.8940 and 0.8998

respectively. Additionally, very strong correlation was found for overall use of the

system. This was shown by performing linear regression over all conditions with

results of 0.9437 for linear acceleration and 0.9194 for angular acceleration. The

ball to head condition did not have a strong correlation, but this is due to the lack

of velocity distribution as all of the impacts were at a very low magnitude.

Although these impacts have low R2 values, they did have a very small absolute

difference, 2.25 g for linear head acceleration and 100.58 rad/s2 for angular head

acceleration. Also, all average peak values for linear acceleration were well

below 66 g which has been previously established as a 25 % risk of injury

(Zhang, et al., 2004). This indicates that a difference of ± 2.25 g would not be

clinically significant.

RMS error values showed a consistency between the two waveforms for

each of the conditions (Figures 5.11 – 5.15). Slightly higher average RMS error

values were found in some waveforms, but upon further inspection it seems as

112

though the discrepancy in the waveforms took place in the tail of the impact or in

a small secondary impact but not in the peak. Therefore, even in the impacts

with a slightly higher average RMS error the peaks were still similar.

One limitation of this system is that it requires the player to wear some

type of headband to allow for player instrumentation to take place. The system

has previously been used in sports that require helmets or headgear of some

type, but in soccer this is not the case. Although headgear is available for soccer

players, it is not a required piece of equipment. Therefore, it could be more

challenging to find players willing to wear the system during play. An additional

concern with the system is movement during play (Beckwith, et al., 2007). This

was not a problem during validation testing, but when used in the field it may be

an issue due to player hair as opposed to the HIII skin. Although this is a

concern, movement would most likely just alter the accuracy of the impact

location (Beckwith, et al., 2007). Headband slippage was not a problem in

laboratory tests, and impact location is not a primary interest for use with this

system. Therefore, headband slippage is not considered a major concern but

on-field research is warranted to assess these concerns.

During soccer play, many impact scenarios exist, and although the current

study made every effort to recreate scenarios typically seen during soccer play, it

was impossible to include all scenarios. One limitation is the limited impact

locations and velocities included. Soccer has a wide range of impact locations

and although they are not all included, the range included provided sufficient

simulation of events to validate the system. An additional limitation is the

113

simulation of on-field scenarios accurately. Using two HIII heads in head to head

impacts as opposed to just an impactor ram provided some give due to the neck

flexion on the impactor, but may not be an exact replication of on-field impacts.

The current study was not designed to recreate exact scenarios, but to provide

reasonable recreations in order to determine a correlation between the

acceleration measurement systems. Therefore, the system needs to be used in

on-field situations to determine fully its ability to accurately measure soccer head

impacts.

Rotational acceleration for left side, head-to-head impacts had the largest

difference measurement for any test method. For these impacts, the HITS under

predicted the HIII by 6417.21 rad/s2. While this discrepancy is of concern, it is

unclear how translatable these results will be to in vivo data collection. These

test conditions are intended to be representative of on-field events, however, the

complex biomechanical interactions that take place during live impacts may not

be completely captured by our simulated event as it was impossible to recreate

every possible impact scenario. Due to the high correlation found for all other

test combinations, we suggest the HITS is a viable method for recording impacts

during competition, however, while linear acceleration measures appear

acceptable, caution should be taken when evaluating rotational acceleration for

impacts similar to our head-to-head condition. As part of this ongoing work,

future studies will address this concern by identifying head to head impacts

through video analysis and comparing on-field measures with those recorded

here.

114

In conclusion, this system provides a much needed method to measure

head acceleration in soccer players during normal play. It allows for accurate

measurements to be taken which could potentially lead to an injury threshold

specific to certain soccer impacts. Additionally, this system will allow a

comparison between different types of impacts that occur during soccer play.

115

CHAPTER 6

ON FIELD MEASUREMENT OF HEAD ACCELERATION

6.1 Introduction

Head acceleration has been successfully measured during sporting events

previously (Duma, et al., 2005, Stojsih, et al., 2008), but it has been a challenge

for researchers to successfully do this during soccer games or scrimmages due

to the lack of headgear (Naunheim, et al., 2003, Naunheim, et al., 2000,

Shewchenko, et al., 2005). Attempts have been made to instrument helmets

from other sports in order to collect data from soccer heading events (Naunheim,

et al., 2000). Although these efforts provide a starting point, data do not

represent actual on-field events as they are recreations and by adding a hard

shell helmet the impact event is altered.

Naunheim et al. (2000, 2003) previously studied head acceleration in

soccer. In the first of these studies (Naunheim, et al., 2000), researchers used a

football helmet with a tri-axial accelerometer mounted on the helmet’s vertex to

measure head accelerations of high school soccer, football and hockey players.

Football and hockey impact acceleration data were collected during games, but

the soccer impacts were done in a simulated game situation. The soccer players

headed a ball kicked 30 yards while wearing the instrumented football helmet.

Significant neurological damage was not anticipated from any single impact

based on standard threshold values for Gadd Severity Index, Head Injury

Criterion, and peak linear acceleration. Soccer players did see higher values for

the three reported results than the football and hockey players. The data,

116

however, does not accurately represent the accelerations that soccer players

would see when heading the ball in a game. This is due to the fact that soccer

players would not be wearing a protective helmet, but they would instead be

heading the ball with no head protection at all.

In order to address the issues with the first study, Naunheim et al. (2003)

instrumented players with an instrumented headpiece, designed specifically for

the study, and a mouthpiece to measure linear and angular accelerations.

Subjects were asked to head a ball which was launched from a distance of six

meters. Linear accelerations of up to 199 + 27 m/s2 were measured. Angular

accelerations were reported to be 1.46 + .297 krad/s2. Although the study

provided some basiline data, the ability to measure on-field data was lacking

(Naunheim, et al., 2003).

Shewchenko et al. (2005) also performed a study measuring head

acceleration during soccer heading. This study used seven current soccer

players to measure kinematics, head acceleration, and muscle activity in the

neck. The subjects were asked to recreate heading scenarios in a laboratory

while wearing reflective targets, EMG electrodes, and a bite plate instrumented

with linear and angular accelerometers. Ten scenarios were performed using

two ball speeds, 6 m/s and 8 m/s, while high speed video, acceleration, and EMG

data were recorded. Results showed that the average peak linear acceleration

did not exceed 194 ± 40 m/s2 for any of the 10 scenarios. Average peak angular

accelerations were also calculated and did not exceed 2.41 ± 1.81 krad/s2 for the

117

scenarios recorded. Although the study is comprehensive, it still did not provide

actual field data representing what occurs in a soccer game.

Laboratory recreations have provided insight into heading impact events

(Naunheim, et al., 2003, Shewchenko, et al., 2005). These recreations provide

valuable kinematic and muscle activation data, but are only representative of the

least severe cases for head acceleration. During these recreations, players were

wearing reflective targets, EMG electrodes, and an instrumented bite plate

(Shewchenko, et al., 2005). All of this instrumentation changes the ability of the

player to move freely and, therefore, alters the dynamic of the impact. Although,

linear acceleration results from two laboratory recreation studies are similar,

there is no way to know that this is what takes place during actual game play

(Naunheim, et al., 2003, Shewchenko, et al., 2005). Ultimately, a wireless

acceleration measurement system which does not inhibit movement and

provides no head protection is needed to determine the linear and angular head

accelerations during actual soccer play.

The exact contribution of linear versus angular acceleration for a given

impact when heading the ball is related to several factors including how quickly

the neck muscles are recruited and the overall intent of the redirection

(Naunheim, et al., 2003, Shewchenko, et al., 2005, 2005). Players may

purposely rotate their head in an effort to redirect the ball towards a particular

target i.e. the goal. There is much debate as to whether linear or angular

acceleration should be investigated when studying mTBI since both have been

shown to predict injury (Gurdjian, et al., 1966, Ommaya and Hirsch, 1971). For

118

the current study, each will be evaluated independently along with various other

head injury criteria which have established thresholds for mTBI.

Two major injury criteria can be assessed along with linear and angular

head acceleration. These are the Head Injury Criterion (HIC) and the Gadd

Severity Index (GSI). HIC and GSI were developed primarily for automotive

impact. Both criteria take into account acceleration over a period of time (Gadd,

1966, Newman, et al., 2000). Each of these criteria is useful only to assess one

individual impact. There are still no suggested head acceleration limits for

multiple subconcussive impact events.

Figure 6.1: Wayne State University Tolerance Curve (Cory, et al., 2001)

The Wayne State Tolerance curve (Figure 6.1) provided the basis for both

GSI and HIC (Gurdjian, et al., 1966). Gadd (1966) developed the Severity Index

using the Wayne State Tolerance Curve plotted on a log-log scale. The slope of

119

the resulting curve was -2.5 which provided the power which is used in the

calculation. It has been suggested that life threatening injuries are increasingly

likely when GSI values are greater than 1000.

*�� � +����(.-.�/

Where:

a(t) = CG resultant translational acceleration

T = duration of acceleration

HIC is the most widely used within automotive testing. This criterion is

based on the GSI calculation and the Wayne State Tolerance curve. HIC is an

optimization of the GSI formula (Versace, 1971). As opposed to using the total

impact duration, a time interval providing the maximum value is used. Various

recommended HIC limits exist, with 1000 being the original limit for automotive

testing. This limit was based on the probability of life-threatening injury, and

represented a 16 % risk of serious brain injury or skull fracture (Prasad and

Mertz, 1985), and has now been reduced to 700 which represents a 5 % risk for

automotive impacts. Pellman et al. (2003) recommended a mTBI HIC limitation

value of 250 based on American football concussions (Pellman, et al., 2003).

0�1 � 2 1��( ' ��� + ����.��4

�56

(.-��( ' ���

Where:

t1, t2 = time points which provide maximum HIC value (generally 15 ms interval is

used)

a(t) = CG resultant translational acceleration

120

In addition to HIC and GSI, other head injury measurement values have

been previously established (Newman, et al., 2000, Ommaya, et al., 2002,

Zhang, et al., 2004) for both linear and angular head acceleration which are

described below. Zhang et al. (2004) used data from football head impact

recreations (Newman, et al., 2000) and finite element modeling to determine an

injury threshold for mTBI. Using data from these recreations as input values, the

Wayne State Brain Injury Model was used to determine probability of injury

(Zhang, et al., 2004). The model was used to calculate the brain’s mechanical

responses which were then related to an injury severity and the resultant

probability of injury. For linear head acceleration, a 25 % risk of mTBI was found

at 66 G, a 50 % risk at 82 G, and an 80 % risk at 106 G. Angular head

acceleration thresholds were also determined and were 25 % at 4600 rad/s2, 50

% at 5900 rad/s2, and 80 % at 7900 rad/s2. One limitation of using these mTBI

threshold levels in the current study is that they were developed based on

impacts that occurred between helmeted individuals and are valid for single

impact events only. The values are, however, based directly on sports injury

data and are not scaled from an animal model which provides a solid basis for

use when evaluating sports injury data.

6.2 Methodology

A total of 24 girls youth soccer players in the U14 age group agreed to

participate in the study. Prior to any testing, approval from Wayne State

University’s Human Investigation Committee was obtained. All players were

fitted with the Head Impact Telemetry System (HITS) headgear (Figure 6.2),

121

described in detail in Chapter 6, and then asked to participate in a scrimmage.

Some participants were involved in more than one scrimmage and were

processed as a new player for each. Players wore the headgear for the duration

of the scrimmages which lasted 30 to 65 minutes. Data were collected at

1000 Hz and downloaded to the sideline computer for later analysis. Games

were videotaped for later analysis in determining what type of impacts occurred

at each downloaded time point.

Figure 6.2: HITS headgear fitted to HIII headform

Following field data collection, data analysis was performed using a

validated algorithm provided by Simbex (Chu, et al., 2006, Crisco, et al., 2004).

Each individual impact was analyzed using this algorithm. Information obtained

included several parameters including HIC and resultant linear and angular

acceleration. Along with all of this information, number of headers, location of

impact, and incidence of other impact events (player collisions with other players,

player falls, collisions with goalposts, and unintentional collisions with the ball)

were also determined from the video.

122

6.3 Results

The majority of header impacts took place to the front location (n = 17)

and ranged in peak linear acceleration from 4.5 g to 34.3 g with an average peak

linear acceleration of 17.4 g (Figure 6.3). Additionally, peak angular

accelerations ranged from 493.3 rad/s2 to 3649.7 rad/s2 with an average of

1657.5 rad/s2. None of these values exceed tolerance levels previously

established.

Figure 6.3: Linear head acceleration by location for each header only

impacts

The second highest number of header impacts by location took place at

the top of the head (n = 9) and ranged in peak linear acceleration from 11.1 g to

44.4 g with an average peak linear acceleration of 19.5 g. Additionally, peak

angular accelerations ranged from 598.0 rad/s2 to 3637.2 rad/s2 with an average

of 1851.8 rad/s2 (Figure 6.4). Again, none of these values exceed tolerance

values for mTBI.

123

Figure 6.4: Angular head acceleration by location for each header only

impacts

The left and right sides had the next highest number of header impacts

with 7 and 8 respectively. The right side impacts ranged in peak linear

acceleration from 5.5 g to 62.9 g with an average peak linear acceleration of 17.4

g. Additionally, peak angular accelerations for the right side ranged from 523.7

rad/s2 to 8869.1 rad/s2 with an average of 3003.4 rad/s2. For the left side, peak

linear accelerations ranged from 11.4 g to 49.4 g with an average of 27.2 g.

Peak angular accelerations ranged from 762.1 rad/s2 to 4509.8 rad/s2 with an

average of 2586.6 rad/s2 for the left side. The back of the head had the fewest

header impacts with 6. These impacts were also low in linear acceleration with a

range of 4.9 g to 19.0 g averaging 11.9 g. Angular acceleration values were also

low with a range of 444.8 rad/s2 to 927.0 rad/s2 and an average of 723.2 rad/s2.

124

Figure 6.5: HIC values for headers by location with mTBI tolerance level

HIC values for the header impacts were generally low for all locations.

The right side had the highest HIC values with an average of 38.1 (Table 6.1)

and a peak value of 154.1 (Figure 6.5). The left side followed with an average

HIC of 27.36 and a peak of 79.40. Although the front location had the most

impacts, the HIC values remained low with an average of 7.7 and a peak of 24.0.

The top and back of the head had very low HIC values with averages of 9.5 and

2.6 respectively (Table 6.1).

125

Table 6.1: Average results for headers by location

General

Location

Peak Linear

Acceleration

(g)

Peak Angular

Acceleration

(rad/s2) HIC 15

GSI

L Side 27.2 2586.6 27.4 36.7

R Side 28.1 3003.4 38.1 51.2

Top 19.5 1851.8 9.5 15.5

Front 17.4 1657.5 7.7 13.6

Back 11.9 723.2 2.6 5.1

The impacts that each individual player saw during each scrimmage was

also investigated. The maximum number of header impacts a single player

experienced in a scrimmage was four with players 3 and 18 both having that total

(Figures 6.6, 6.7). In addition to these header impacts, player 18 also had a non

header impact, a collision with the goal post, which was above the angular

acceleration 25 % threshold for injury (Figure 6.12). Players 5 and 19 were the

only two players that had header impacts with angular accelerations exceeding

head injury tolerance limits (Zhang, et al., 2004). Both of these players had

multiple impacts during their scrimmages, not just the single tolerance exceeding

blow (Figures 6.6 – 6.7).

126

Figure 6.6: Linear head acceleration for all header impacts for individual

players

Figure 6.7: Angular head acceleration for all header impacts for individual

players

In addition to the header impacts, there were also various other impacts.

There were 21 non-header impacts recorded (Table 6.2). These included

collision with the goal, player collisions, player falls, and unintentional ball to

head impacts (Table 6.2). The majority of players participating had only one non-

header event occurring during their scrimmage and many had no occurrences.

127

Three players, however, had multiple instances. These were players 12, 13, and

29 (Table 6.2). Player 12 also had multiple header events during their

scrimmage (Figures 6.6, 6.7). All impacts to player 12 were well under the

current tolerance limits for mTBI.

128

Table 6.2: Description of non header impacts and the player that impacted

Player Description

Peak Linear

Acceleration

(g)

Peak Angular

Acceleration

(rad/s2)

HIC

15

GSI

5 Player fell 23.7 3739.3 13.7 17.9

7 Unintentional ball to head 20.4 1749.9 5.4 11.4

9 Player fell 15.1 1332.3 6.4 8.1

12 Player fell 7.7 628.5 0.5 1.2

12 Player fell 11.1 881.8 1.0 1.6

13 Player collision 10.7 811.3 1.7 2.2

13 Player collision 12 828.4 1.9 2.8

14 Player hit ground 18 823.6 4.1 7.2

15 Ball hit back of head 32.2 1090.7 25.4 31.6

17 Player collision 11.6 1247.1 4.0 7.5

19 Player fell 16 1315 6.7 10.9

21 Player collision 24 2831.8 15.4 31.7

22 Collision with goalpost 27.1 5179.5 16.5 20.4

24 Player collision 25.7 2847 17.3 20.8

26 Player fell 18.5 753.6 8.7 10.3

27 Player hit ground 5 497.5 0.2 0.3

28 Player collision 56.7 2910.3 90.8 114.7

29 Player collision 18.9 987 2.9 5.4

29 Player collision 18.9 1982.3 8.7 10.3

29 Unintentional ball to head 15 899.9 1.9 2.7

129

The majority of the non-header impacts took place to the front and to the

top of the head, with six impacts each (Figures 6.8, 6.9). None of the impacts at

either of these locations reached any of the threshold levels for either linear or

angular acceleration. Following these two locations, five impacts occurred to the

left side. Of these five impacts, one had an angular head acceleration (5179.5

rad/s2) that exceeded the 25 % risk of injury tolerance level. This was the only

non-header impact to exceed any of the previously established tolerance levels.

The right side of the head and the back of the head had limited non-header

impacts with two and one respectively.

Figure 6.8: Linear head acceleration for all non header impacts by location

130

Figure 6.9: Angular head acceleration for all non header impacts by

location

HIC was also evaluated by location of the impact for the non-header

impacts (Figure 6.10). The highest HIC value for non-header events was 90.8

and occurred to the top of the head. This value is not considered to be at an

injurious level as it is well under the tolerance level of 250 established for mTBI.

Figure 6.10: HIC for all non header impacts by location

The highest peak linear acceleration of 56.7 g, to player 28, was found to

131

occur from a player collision (Figure 6.11). This was the only impact that player

28 experienced that was high enough to trigger the HIT system. With a

corresponding angular acceleration of 2910.3 rad/s2, it is unlikely that injury

would occur as none of these values reach any of the injury tolerances currently

established for mTBI.

A separate incident, a collision with the goal, resulted in the peak angular

acceleration of 5179.5 rad/s2 which occurred to player 22 (Figure 6.12). In

addition to this non-header impact, player 22 had four header impacts during

their scrimmage. Although none of the header impacts exceeded tolerance

values for linear head acceleration, one of the headers had an angular

acceleration of 4509.1 rad/s2. This nearly exceeds the tolerance level proposed

by Zhang et al. (2004) and could increase the risk of additional impacts.

Figure 6.11: Linear head acceleration for all non header impacts for

individual players

132

Figure 6.12: Angular head acceleration for all non header impacts for

individual players 6.4 Discussion

There have not been any previous studies recording on-field

measurements of head acceleration during soccer play. Therefore, the current

study cannot be compared to previous on-field data collections. It can, however,

be compared to previous laboratory studies. The maximum linear and angular

accelerations found in the current study greatly exceed those found by Naunheim

et al. (2003). In the laboratory study, Naunheim et al. (2003) did not have linear

or angular accelerations exceeding 199 m/s2 and 1.46 krad/s2 in contrast to a

maximum linear acceleration of 617 m/s2 and a maximum angular acceleration of

8.87 krad/s2 determined in the on-field study. This is a 418 m/s2 difference in the

maximum linear accelerations seen between the two studies and a 7.41 krad/s2

difference in the angular acceleration measurements.

The current study also exceeds the acceleration measurements

determined by Shewchenko et al. (2005). Shewchenko et al. (2005) determined

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28

An

gu

lar

Acc

ele

rati

on

(ra

d/

s2)

Player

Angular Head Acceleration

p = 0.50

p = 0.25

p = 0.80

133

that the average peak linear acceleration over the different heading scenarios

participants performed to not exceed 194 m/s2 which was found to be much

lower than the peak average linear acceleration of 276 m/s2 from the current

study. A peak average angular acceleration of 2.41 krad/s2 was found for the

laboratory study in contrast to the 3.00 krad/s2 found in the current on-field study.

These values are much more similar to the on-field collections than the study

performed by Naunheim et al. (2003). This is due to the use of averages as

opposed to maximums and it does not take into account worst case scenarios.

Additionally the laboratory studies looked only at a redirection directly back

towards where the ball came from. This reduces the amount of angular

acceleration that participants will experience and only represents a small portion

of the type of headers players experienced during actual play.

Data were compared to mTBI head injury tolerance values proposed

previously (Ommaya, et al., 2002, Pellman, et al., 2003, Zhang, et al., 2004).

None of the impacts, heading events or otherwise, exceeded the 66 g threshold

which was the 25% risk of injury tolerance level (Zhang, et al., 2004) or the HIC

value of 250 (Pellman, et al., 2003). Based on these observations, it seems as

though the linear acceleration contribution of heading is not causing head injury

based on a single impacts.

Angular accelerations, however, did exceed the suggested limits. Three

angular acceleration measurements for heading events (4509.8 rad/s2, 5298.3

rad/s2, 8869.1 rad/s2) exceeded the 4500 rad/s2 limit which has been suggested

as the limit required to produce concussion in adults (Ommaya, et al., 2002). Of

134

these three impacts, one also exceeded the 25% risk of injury threshold of 4600

rad/s2 and one exceeded the 7900 rad/s2 limit which correlates to an 80% risk of

head injury (Zhang, et al., 2004). In addition to the three heading events

exceeding the 4500 rad/s2 concussion injury tolerance level, an impact caused

by a collision with the goal resulted in an angular acceleration of 5179 rad/s2

which also exceeds the 25 % risk of injury value proposed by Zhang et al. (2004).

Although single impacts exceeded the suggested mTBI tolerance levels,

there was no stoppage of play during any of the scrimmages due to injury. Funk

et al. (2007) suggest a 10 % risk of injury with a linear acceleration of 165 g, a

HIC of 400, and a peak angular acceleration of 9000 rad/s2 to produce mTBI.

These values are much higher and indicate a lower risk of injury at the values

seen in the current study. Tolerance values required to induce mTBI could be

higher than Zhang et al. (2004) as suggested (Funk, et al., 2007), or the lack of

stoppage of play could be due to other causes. This could be due to the fact that

they fell within the percentage of the population that would not be injured at the

suggested tolerances, or it is possible that these tolerance levels are not

representative of the types of impacts that occur during soccer heading. It is also

possible that angular acceleration alone is not the best predictor of injury in

soccer heading impacts. Since these were the only values that exceeding injury

tolerances, it is possible that these values alone are not representative of injury

causation.

The high levels observed could be due to a limitation of the

instrumentation itself, but it when looking at the validation of the HITS higher

135

level impacts had a much better correlation than the lower level impacts. This

indicates that the system responds accurately at the impact level at which injury

is presumed to occur. However, data seem to indicate that angular acceleration

during heading events could potentially pose a problem; especially as single

impacts are exceeding 80% risk of injury tolerances. Additionally, all heading

impacts with angular accelerations exceeding suggested limits took place to

either the right or left side of the head. This indicates an unusual heading

method or inaccurate technique. When impacts took place to the front or top of

the head, no limits were exceeded. This further emphasizes the importance of

teaching proper technique.

All of the suggested tolerance levels which current data were compared to

were developed for single impact events. Many of the players in the current

study experienced multiple impacts, not all of which were header impacts. All of

the players who experienced tolerance exceeding impacts had other impacts as

well. Player 22 actually experienced two of the four impacts that had angular

accelerations above the recommended tolerance levels, one a header and one a

collision with a goalpost. In addition to those two impacts, player 22 also had

three other headers that fell below recommendations. Based on standard

tolerance levels it is challenging to determine if the combination of all of those

impacts in a single scrimmage increases the likelihood of injury. Further

research is necessary to determine if the level of multiple impacts has an effect

on mTBI probability. Additionally, further research is needed to determine if

136

symptoms are occurring after scrimmages where head accelerations are known.

This could provide a possible correlation.

Some limitations of the current study include the limited study population.

The study was limited to a single age group and to female soccer players.

Further research would be necessary to determine if differences exist in head

acceleration based on age or gender. Additionally, the current study investigated

soccer scrimmages and not actual games. This is due to the challenge of getting

players to wear equipment that is not required during competition. Higher head

accelerations may be seen during a more competitive, and likely more

aggressive, game scenario.

137

CHAPTER 7

CONCLUSIONS AND FUTURE RECOMMENDATIONS

7.1 Conclusions

Soccer is one of the most popular sports throughout the world. With a recent

increase in youth players in the United States, an increase in injuries has also

been reported. In addition to the unintentional head impacts that occur during

soccer play, similar to those of other sporting activities, soccer players also

intentional use their head to redirect the soccer ball, an act known as “heading”.

The effect of these intentional impacts has been studied, the majority of research

being conducted on adults, with conflicting outcomes. The risks to children are

potentially greater due to their size versus the force being applied by the ball. It

has been reported that ball mass, impact velocity, and size of the individual all

contribute to the potential for injury. The importance of proper technique may be

especially true in the youth population, since their skill level has not been well

developed to control their head motion when heading the ball.

Although previous research has been conducted to determine the effects of

repetitive heading in soccer, the results are very controversial. Conflicting results

have been observed in studies throughout the history of soccer heading

research. Many of these previous studies have had significant challenges within

the methodology, including the lack of controls or using improper control groups

and many have not taken into account other outside factors that could be

contributing to results. In order to determine the effects of the repetitive

138

subconcussive head impacts associated with heading, an in-depth analysis of the

biomechanics of heading needed to be performed.

The current study investigated the effect of the intentional head impacts that

occur during soccer play. Initial steps were taken to determine the frequency and

severity of heading episodes in the field using both field observation and to

determine the possibility of head injury caused by impact with the ball only. The

NEISS database was used to determine injury occurrence from ball to head only

impacts. Ball only injuries comprised 15.9% of total head injuries. These injuries

were not necessarily caused by heading the soccer ball, as some of the impacts

were due to unintentional ball to head impact, but many of the cases were

described as heading related. It is, however, challenging to determine a total

number of participants due to the inclusion of organized and non-organized

soccer injuries. These data indicate that heading alone can result in injuries

severe enough to require medical attention. However, the current study most

likely underestimates both total injuries and ball to head injuries because many

less severe injuries would not be included. This is an inherent problem with

using the NEISS database to estimate injuries. Therefore, the current study

represents the more severe cases, and is potentially an underestimate of the

total injury occurrences.

After establishing that youth soccer players have reported to the Emergency

Department complaining of injuries that occurred from impact with the ball only,

an analysis of the biomechanics of heading in youth soccer needed to be

performed. In order to assess the biomechanics of heading in youth soccer, a

139

laboratory study was performed to determine head and back angles and neck

muscle activation during a variety of heading tasks. Heading tasks were also

included that had players modify there traditional technique in order to compare

linear head acceleration and to see if any of the tasks were effective in reducing

head acceleration. It was found that heading techniques are quite variable

between players. An overall inconsistency was found for head flexion, torso

flexion, and head rotation. It was also noted that this variability existed in EMG

data as well. This is very similar to what has been previously observed in adult

players. This is most likely due to the many possible heading scenarios.

Therefore, players do not use muscle memory to perform the task in the same

manor every time, but instead learn to adjust to whatever scenario occurs. This

provides unlimited possibilities for heading scenarios and would most likely result

in additional variation in results. The current study only looked at redirection

directly at the source of the ball launch. If additional scenarios were introduced

to provide redirection in other ways, additional distinctions would be made.

Comparisons between genders or between heading tasks were

challenging to make. Based on the results of the current study, it indicates that

differences are not related these variables, but that the differences occur

between each player. This also made making a comparison with the adult

players unproductive as any difference would most likely have nothing to do with

age, but with the players being different and the scenarios being slightly different.

Although attempts have been made to determine head acceleration during

soccer heading events, a system for on field data collection had not been

140

previously available. This system was created for use in soccer research, but

had to be validated prior to use. A validation was conducted using various

scenarios that typically occur in soccer play. The results show that the new

soccer HITS system correlates well with the standard measurement system of

the HIII 3-2-2-2 accelerometer system. Although the system requires the use of

a headband, it provides a much needed method to measure head acceleration in

soccer players during normal play. It allows for accurate measurements to be

taken which could potentially lead to an injury threshold specific to certain soccer

impacts. Additionally, this system will allow a comparison between different

types of impacts that occur during soccer play.

Once validation occurred of the HITS head acceleration system, this

system was implemented in soccer scrimmages to determine actual on-field

head acceleration. The current study found both linear and angular head

accelerations that exceeded the acceleration measurements determined

previously in laboratory studies. This could be due to the fact that laboratory

studies using players have to use a ball impact speed that is on the low end of

what would be seen in the field. Additionally the laboratory studies looked only at

a redirection directly back towards where the ball came from. This reduces the

amount of angular acceleration that participants will experience and only

represents a small portion of the type of headers players experienced during

actual play.

Data were compared to mTBI head injury tolerance values proposed

previously, and none of the impacts exceeded the injury tolerance levels for

141

linear head acceleration or HIC. Based on these observations, it seems as

though the linear acceleration contribution of heading is not causing head injury

based on a single impacts. There were, however, impacts that exceeded

suggested values for angular head acceleration. Although single impacts

exceeded the suggested mTBI tolerance levels, there was no stoppage of

scrimmage play due to injury. This could be due to the fact that they fell within

the percentage of the population that would not be injured at the suggested

tolerances, or it is possible that these tolerance levels are not representative of

the types of impacts that occur during soccer heading. It is also possible that

angular acceleration alone is not the best predictor of injury in soccer heading

impacts. Additionally, all heading impacts with angular accelerations exceeding

suggested limits took place to either the right or left side of the head. This

indicates an unusual heading method or inaccurate technique. When impacts

took place to the front or top of the head, no limits were exceeded. This further

emphasizes the importance of teaching proper technique.

7.2 Future Recommendations

One of the major challenges in the current study was comparing head

acceleration measurements to injury tolerance levels that were established for

single impacts of a different nature than those that occur during soccer play. It

would be of interest in future research to determine a threshold for multiple

impacts. This would be of specific interest in the sporting community where

many players are at risk for multiple low level impacts.

142

One of the first steps to determining this threshold occurred in the current

study, where a greater knowledge of what occurs during soccer heading events

was gained. It is, however, necessary to further investigate the head

acceleration values that are obtained in other populations. The current study

lacks any on-field data collection during soccer scrimmages involving male

players. Additionally, different age groups should be investigated to determine if

there is an difference as players age. Although it would be of interest to

determine head acceleration in younger players, it is unlikely that there would be

enough heading events to warrant this investigation.

Soccer scrimmages and not actual games were investigated due to the

challenge of getting players to wear equipment that is not required during

competition. Higher head accelerations may be seen during a more competitive,

and likely more aggressive, game scenario. Future studies would be required to

determine if these differences actually occur.

In addition to these on-field measurements, further research should be

performed in the area of soccer biomechanics within the laboratory setting. The

current study, along with previous studies, only looked at redirection directly at

the source of the ball launch. If additional scenarios were introduced to provide

redirection in other ways, additional variation would most likely be noted. It

would be of interest to determine if head acceleration changes with the addition

of alternate redirection scenarios.

143

APPENDIX A – HIC APPROVALS

144

145

146

147

148

REFERENCES AYSO, 2006, www.soccer.org.

ADAMS, A. L., AND SCHIFF, M. A., 2006. CHILDHOOD SOCCER INJURIES

TREATED IN U.S. EMERGENCY DEPARTMENTS. ACAD EMERG MED

13(5), PP. 571-574.

AGEL, J., ET AL., 2007. DESCRIPTIVE EPIDEMIOLOGY OF COLLEGIATE

MEN'S SOCCER INJURIES: NATIONAL COLLEGIATE ATHLETIC

ASSOCIATION INJURY SURVEILLANCE SYSTEM, 1988-1989

THROUGH 2002-2003. J ATHL TRAIN 42(2), PP. 270-277.

ANDERSEN, T. E., ET AL., 2004. MECHANISMS OF HEAD INJURIES IN ELITE

FOOTBALL. BR J SPORTS MED 38(6), PP. 690-696.

ARNASON, A., ET AL., 2004. A PROSPECTIVE VIDEO-BASED ANALYSIS OF

INJURY SITUATIONS IN ELITE MALE FOOTBALL: FOOTBALL

INCIDENT ANALYSIS. AM J SPORTS MED 32(6), PP. 1459-1465.

NATIONAL FEDERATION OF STATE HIGH SCHOOL ASSOCIATIONS,

http://www.nfhs.org/custom/participation_figures/default.aspx.

AUTTI, T., ET AL., 1997. BRAIN LESIONS IN PLAYERS OF CONTACT

SPORTS. LANCET 349(9059), P. 1144.

BACKOUS, D. D., ET AL., 1988. SOCCER INJURIES AND THEIR RELATION

TO PHYSICAL MATURITY. AM J DIS CHILD 142(8), PP. 839-842.

BARNES, B. C., ET AL., 1998. CONCUSSION HISTORY IN ELITE MALE AND

FEMALE SOCCER PLAYERS. AM J SPORTS MED 26(3), PP. 433-438.

149

BAUER, J. A., ET AL., 2001. IMPACT FORCES AND NECK MUSCLE ACTIVITY

IN HEADING BY COLLEGIATE FEMALE SOCCER PLAYERS. J

SPORTS SCI 19(3), PP. 171-179.

BECKWITH, J. G., ET AL., 2007. VALIDATION OF A NONINVASIVE SYSTEM

FOR MEASURING HEAD ACCELERATION FOR USE DURING BOXING

COMPETITION. J APPL BIOMECH 23(3), PP. 238-244.

BODEN, B. P., ET AL., 1998. CONCUSSION INCIDENCE IN ELITE COLLEGE

SOCCER PLAYERS. AM J SPORTS MED 26(2), PP. 238-241.

CHU, J. J., ET AL., 2006, A NOVEL ALGORITHM TO MEASURE LINEAR AND

ROTATIONAL HEAD ACCELERATION USING SINGLE-AXIS

ACCELEROMETERS, 5TH WORLD CONGRESS OF

BIOMECHANICSMUNICH, GERMANY.

US CONSUMER PRODUCT SAFETY COMMISSION, 2000, NEISS THE

NATIONAL ELECTRONIC INJURY SURVEILLANCE SYSTEM: A TOOL

FOR RESEARCHERS, http://www.cpsc.gov/neiss/2000d015.pdf.

US CONSUMER PRODUCT SAFETY COMMISSION, 2001, THE NEISS

SAMPLE (DESIGN AND IMPLEMENTATION) 1997 TO PRESENT,

http://www.cpsc.gov/neiss/2001d011-6b6.pdf.

CONN, J. M., ET AL., 2006. NON-FATAL SPORTS AND RECREATIONAL

VIOLENT INJURIES AMONG CHILDREN AND TEENAGERS, UNITED

STATES, 2001-2003. J SCI MED SPORT 9(6), PP. 479-489.

CORY, C. Z., ET AL., 2001. THE POTENTIAL AND LIMITATIONS OF

UTILISING HEAD IMPACT INJURY MODELS TO ASSESS THE

150

LIKELIHOOD OF SIGNIFICANT HEAD INJURY IN INFANTS AFTER A

FALL. FORENSIC SCI INT 123(2-3), PP. 89-106.

COVASSIN, T., ET AL., 2003. EPIDEMIOLOGICAL CONSIDERATIONS OF

CONCUSSIONS AMONG INTERCOLLEGIATE ATHLETES. APPL

NEUROPSYCHOL 10(1), PP. 12-22.

COVASSIN, T., ET AL., 2003. SEX DIFFERENCES AND THE INCIDENCE OF

CONCUSSIONS AMONG COLLEGIATE ATHLETES. J ATHL TRAIN

38(3), PP. 238-244.

CRISCO, J. J., ET AL., 2004. AN ALGORITHM FOR ESTIMATING

ACCELERATION MAGNITUDE AND IMPACT LOCATION USING

MULTIPLE NONORTHOGONAL SINGLE-AXIS ACCELEROMETERS. J

BIOMECH ENG 126(6), PP. 849-854.

DELANEY, J. S., 2004. HEAD INJURIES PRESENTING TO EMERGENCY

DEPARTMENTS IN THE UNITED STATES FROM 1990 TO 1999 FOR

ICE HOCKEY, SOCCER, AND FOOTBALL. CLIN J SPORT MED 14(2),

PP. 80-87.

DICK, R., ET AL., 2007. DESCRIPTIVE EPIDEMIOLOGY OF COLLEGIATE

WOMEN'S SOCCER INJURIES: NATIONAL COLLEGIATE ATHLETIC

ASSOCIATION INJURY SURVEILLANCE SYSTEM, 1988-1989

THROUGH 2002-2003. J ATHL TRAIN 42(2), PP. 278-285.

DUMA, S. M., ET AL., 2005. ANALYSIS OF REAL-TIME HEAD

ACCELERATIONS IN COLLEGIATE FOOTBALL PLAYERS. CLIN J

SPORT MED 15(1), PP. 3-8.

151

DVORAK, J., AND JUNGE, A., 2000. FOOTBALL INJURIES AND PHYSICAL

SYMPTOMS. A REVIEW OF THE LITERATURE. AM J SPORTS MED

28(5 SUPPL), PP. S3-9.

DVORAK, J., ET AL., 2007. HEAD INJURIES IN THE FEMALE FOOTBALL

PLAYER: INCIDENCE, MECHANISMS, RISK FACTORS AND

MANAGEMENT. BR J SPORTS MED 41 SUPPL 1, PP. I44-46.

ELIAS, S. R., 2001. 10-YEAR TREND IN USA CUP SOCCER INJURIES: 1988-

1997. MED SCI SPORTS EXERC 33(3), PP. 359-367.

FIELD, M., ET AL., 2003. DOES AGE PLAY A ROLE IN RECOVERY FROM

SPORTS-RELATED CONCUSSION? A COMPARISON OF HIGH

SCHOOL AND COLLEGIATE ATHLETES. J PEDIATR 142(5), PP. 546-

553.

FULLER, C. W., ET AL., 2005. A SIX YEAR PROSPECTIVE STUDY OF THE

INCIDENCE AND CAUSES OF HEAD AND NECK INJURIES IN

INTERNATIONAL FOOTBALL. BR J SPORTS MED 39 SUPPL 1, PP. I3-

9.

FUNK, J. R., ET AL., 2007. BIOMECHANICAL RISK ESTIMATES FOR MILD

TRAUMATIC BRAIN INJURY. ANNU PROC ASSOC ADV AUTOMOT

MED 51, PP. 343-361.

GADD, C. W., 1966, "USE OF WEIGHTED-IMPULSE CRITERION FOR

ESTIMATING INJURY HAZARD," TENTH STAPP CAR CRASH

CONFERENCE, SOCIETY OF AUTOMOTIVE ENGINEERS, INC., NEW

YORK.

152

GRAY, H., ET AL., 1995. GRAY'S ANATOMY. CHURCHILL LIVINGSTONE,

EDINBURGH ; NEW YORK, P.^PP. PAGES

GREEN, G. A., AND JORDAN, S. E., 1998. ARE BRAIN INJURIES A

SIGNIFICANT PROBLEM IN SOCCER? CLIN SPORTS MED 17(4), PP.

795-809, VIII.

GURDJIAN, E. S., ET AL., 1966. TOLERANCE CURVES OF ACCELERATION

AND INTRACRANIAL PRESSURE AND PROTECTIVE INDEX IN

EXPERIMENTAL HEAD INJURY. J TRAUMA 6(5), PP. 600-604.

GUSKIEWICZ, K. M., 2002. NO EVIDENCE OF IMPAIRED NEUROCOGNITIVE

PERFORMANCE IN COLLEGIATE SOCCER PLAYERS. AM J SPORTS

MED 30(4), P. 630.

HOSTETLER, S. G., ET AL., 2005. CHARACTERISTICS OF WATER SKIING-

RELATED AND WAKEBOARDING-RELATED INJURIES TREATED IN

EMERGENCY DEPARTMENTS IN THE UNITED STATES, 2001-2003.

AM J SPORTS MED 33(7), PP. 1065-1070.

HOSTETLER, S. G., ET AL., 2004. CHARACTERISTICS OF ICE HOCKEY-

RELATED INJURIES TREATED IN US EMERGENCY DEPARTMENTS,

2001-2002. PEDIATRICS 114(6), PP. E661-666.

JANDA, D. H., ET AL., 2002. AN EVALUATION OF THE CUMULATIVE

CONCUSSIVE EFFECT OF SOCCER HEADING IN THE YOUTH

POPULATION. INJ CONTROL SAF PROMOT 9(1), PP. 25-31.

153

JORDAN, S. E., ET AL., 1996. ACUTE AND CHRONIC BRAIN INJURY IN

UNITED STATES NATIONAL TEAM SOCCER PLAYERS. AM J SPORTS

MED 24(2), PP. 205-210.

JUNGE, A., AND DVORAK, J., 2007. INJURIES IN FEMALE FOOTBALL

PLAYERS IN TOP-LEVEL INTERNATIONAL TOURNAMENTS. BR J

SPORTS MED 41 SUPPL 1, PP. I3-7.

KELLER, C. S., ET AL., 1987. THE MEDICAL ASPECTS OF SOCCER INJURY

EPIDEMIOLOGY. AM J SPORTS MED 15(3), PP. 230-237.

KIRKENDALL, D. T., AND GARRETT, W. E., JR., 2001. HEADING IN SOCCER:

INTEGRAL SKILL OR GROUNDS FOR COGNITIVE DYSFUNCTION? J

ATHL TRAIN 36(3), PP. 328-333.

LE GALL, F., ET AL., 2008. INJURIES IN YOUNG ELITE FEMALE SOCCER

PLAYERS: AN 8-SEASON PROSPECTIVE STUDY. AM J SPORTS MED

36(2), PP. 276-284.

LEES, A., AND NOLAN, L., 1998. THE BIOMECHANICS OF SOCCER: A

REVIEW. J SPORTS SCI 16(3), PP. 211-234.

LEININGER, R. E., ET AL., 2007. EPIDEMIOLOGY OF 1.6 MILLION

PEDIATRIC SOCCER-RELATED INJURIES PRESENTING TO US

EMERGENCY DEPARTMENTS FROM 1990 TO 2003. AM J SPORTS

MED 35(2), PP. 288-293.

MANOOGIAN, S., ET AL., 2006. HEAD ACCELERATION IS LESS THAN 10

PERCENT OF HELMET ACCELERATION IN FOOTBALL IMPACTS.

BIOMED SCI INSTRUM 42, PP. 383-388.

154

MATSER, J. T., ET AL., 1998. CHRONIC TRAUMATIC BRAIN INJURY IN

PROFESSIONAL SOCCER PLAYERS. NEUROLOGY 51(3), PP. 791-

796.

METZL, J. D., 1999. SPORTS-SPECIFIC CONCERNS IN THE YOUNG

ATHLETE: SOCCER. PEDIATR EMERG CARE 15(2), PP. 130-134.

NAUNHEIM, R. S., ET AL., 2003. LINEAR AND ANGULAR HEAD

ACCELERATIONS DURING HEADING OF A SOCCER BALL. MED SCI

SPORTS EXERC 35(8), PP. 1406-1412.

NAUNHEIM, R. S., ET AL., 2000. COMPARISON OF IMPACT DATA IN

HOCKEY, FOOTBALL, AND SOCCER. J TRAUMA 48(5), PP. 938-941.

NEWMAN, J. A., ET AL., 2000. A PROPOSED NEW BIOMECHANICAL HEAD

INJURY ASSESSMENT FUNCTION - THE MAXIMUM POWER INDEX.

STAPP CAR CRASH J 44, PP. 215-247.

NIELSEN, A. B., AND YDE, J., 1989. EPIDEMIOLOGY AND TRAUMATOLOGY

OF INJURIES IN SOCCER. AM J SPORTS MED 17(6), PP. 803-807.

OMMAYA, A. K., ET AL., 2002. BIOMECHANICS AND NEUROPATHOLOGY

OF ADULT AND PAEDIATRIC HEAD INJURY. BR J NEUROSURG 16(3),

PP. 220-242.

OMMAYA, A. K., AND HIRSCH, A. E., 1971. TOLERANCES FOR CEREBRAL

CONCUSSION FROM HEAD IMPACT AND WHIPLASH IN PRIMATES. J

BIOMECH 4(1), PP. 13-21.

155

PADGAONKAR, A., ET AL., 1975. MEASUREMENT OF ANGULAR

ACCELERATION OF A RIGID BODY USING LINEAR

ACCELEROMETERS. ASME J. APPL. MECH. 42, PP. 552-558.

PATLAK, M., ET AL., 2002, "IS SOCCER BAD FOR CHILDREN'S HEADS?

SUMMARY OF THE IOM WORKSHOP ON NEUROPSYCHOLOGICAL

CONSEQUENCES OF HEAD IMPACT IN YOUTH SOCCER," NATIONAL

ACADEMY PRESS, WASHINGTON, D.C., P. 26 P.

PELLMAN, E. J., ET AL., 2003. CONCUSSION IN PROFESSIONAL

FOOTBALL: RECONSTRUCTION OF GAME IMPACTS AND INJURIES.

NEUROSURGERY 53(4), PP. 799-812; DISCUSSION 812-794.

PETERSON, L., ET AL., 2000. INCIDENCE OF FOOTBALL INJURIES AND

COMPLAINTS IN DIFFERENT AGE GROUPS AND SKILL-LEVEL

GROUPS. AM J SPORTS MED 28(5 SUPPL), PP. S51-57.

PICKETT, W., ET AL., 2005. HEAD INJURIES IN YOUTH SOCCER PLAYERS

PRESENTING TO THE EMERGENCY DEPARTMENT. BR J SPORTS

MED 39(4), PP. 226-231; DISCUSSION 226-231.

POULSEN, T. D., ET AL., 1991. INJURIES IN HIGH-SKILLED AND LOW-

SKILLED SOCCER: A PROSPECTIVE STUDY. BR J SPORTS MED

25(3), PP. 151-153.

PRASAD, P., AND MERTZ, H. J., 1985. THE POSITION OF THE UNITED

STATES DELEGATES TO THE ISO WORKING GROUP 6 ON THE USE

OF HIC IN THE AUTOMOTIVE ENVIRONMENT. SOCIETY OF

AUTOMOTIVE ENGINEERS PAPER 851246.

156

PUTUKIAN, M., 2004. HEADING IN SOCCER: IS IT SAFE? CURR SPORTS

MED REP 3(1), PP. 9-14.

RUCHINSKAS, R. A., ET AL., 1997. MILD HEAD INJURY IN SPORTS. APPL

NEUROPSYCHOL 4(1), PP. 43-49.

RUTHERFORD, A., ET AL., 2003. THE NEUROPSYCHOLOGY OF HEADING

AND HEAD TRAUMA IN ASSOCIATION FOOTBALL (SOCCER): A

REVIEW. NEUROPSYCHOL REV 13(3), PP. 153-179.

SANDELIN, J., ET AL., 1985. ACUTE SOCCER INJURIES IN FINLAND IN 1980.

BR J SPORTS MED 19(1), PP. 30-33.

SCHMIDT-OLSEN, S., ET AL., 1985. SOCCER INJURIES OF YOUTH. BR J

SPORTS MED 19(3), PP. 161-164.

SHEWCHENKO, N., ET AL., 2005. HEADING IN FOOTBALL. PART 1:

DEVELOPMENT OF BIOMECHANICAL METHODS TO INVESTIGATE

HEAD RESPONSE. BR J SPORTS MED 39 SUPPL 1, PP. I10-25.

SHEWCHENKO, N., ET AL., 2005. HEADING IN FOOTBALL. PART 2:

BIOMECHANICS OF BALL HEADING AND HEAD RESPONSE. BR J

SPORTS MED 39 SUPPL 1, PP. I26-32.

SHEWCHENKO, N., ET AL., 2005. HEADING IN FOOTBALL. PART 3: EFFECT

OF BALL PROPERTIES ON HEAD RESPONSE. BR J SPORTS MED 39

SUPPL 1, PP. I33-39.

SORTLAND, O., AND TYSVAER, A. T., 1989. BRAIN DAMAGE IN FORMER

ASSOCIATION FOOTBALL PLAYERS. AN EVALUATION BY

157

CEREBRAL COMPUTED TOMOGRAPHY. NEURORADIOLOGY 31(1),

PP. 44-48.

SOSIN, D. M., ET AL., 1996. PEDIATRIC HEAD INJURIES AND DEATHS

FROM BICYCLING IN THE UNITED STATES. PEDIATRICS 98(5), PP.

868-870.

STEPHENS, R., ET AL., 2005. NEUROPSYCHOLOGICAL IMPAIRMENT AS A

CONSEQUENCE OF FOOTBALL (SOCCER) PLAY AND FOOTBALL

HEADING: A PRELIMINARY ANALYSIS AND REPORT ON SCHOOL

STUDENTS (13-16 YEARS). CHILD NEUROPSYCHOL 11(6), PP. 513-

526.

STOJSIH, S., ET AL., 2008. A PROSPECTIVE STUDY OF PUNCH

BIOMECHANICS AND COGNITIVE FUNCTION FOR AMATEUR

BOXERS. BR J SPORTS MED.

BIOSYN SYSTEMS, 3D FULL BODY PORTABLE KINEMATIC SYSTEM,

http://www.noraxon.com/downloads/documents/partners/Biosyn%20FAB

%20System.pdf.

TYSVAER, A., AND STORLI, O., 1981. ASSOCIATION FOOTBALL INJURIES

TO THE BRAIN. A PRELIMINARY REPORT. BR J SPORTS MED 15(3),

PP. 163-166.

TYSVAER, A. T., AND LOCHEN, E. A., 1991. SOCCER INJURIES TO THE

BRAIN. A NEUROPSYCHOLOGIC STUDY OF FORMER SOCCER

PLAYERS. AM J SPORTS MED 19(1), PP. 56-60.

158

TYSVAER, A. T., AND STORLI, O. V., 1989. SOCCER INJURIES TO THE

BRAIN. A NEUROLOGIC AND ELECTROENCEPHALOGRAPHIC

STUDY OF ACTIVE FOOTBALL PLAYERS. AM J SPORTS MED 17(4),

PP. 573-578.

TYSVAER, A. T., ET AL., 1989. SOCCER INJURIES TO THE BRAIN. A

NEUROLOGIC AND ELECTROENCEPHALOGRAPHIC STUDY OF

FORMER PLAYERS. ACTA NEUROL SCAND 80(2), PP. 151-156.

VERSACE, J., 1971. A REVIEW OF THE SEVERITY INDEX. SOCIETY OF

AUTOMOTIVE ENGINEERS PAPER 710881.

WITHNALL, C., ET AL., 2005. EFFECTIVENESS OF HEADGEAR IN

FOOTBALL. BR J SPORTS MED 39 SUPPL 1, PP. I40-48; DISCUSSION

I48.

WITOL, A. D., AND WEBBE, F. M., 2003. SOCCER HEADING FREQUENCY

PREDICTS NEUROPSYCHOLOGICAL DEFICITS. ARCH CLIN

NEUROPSYCHOL 18(4), PP. 397-417.

XIANG, H., ET AL., 2005. SKIING- AND SNOWBOARDING-RELATED

INJURIES TREATED IN U.S. EMERGENCY DEPARTMENTS, 2002. J

TRAUMA 58(1), PP. 112-118.

YARD, E. E., AND COMSTOCK, R. D., 2006. INJURIES SUSTAINED BY

PEDIATRIC ICE HOCKEY, LACROSSE, AND FIELD HOCKEY

ATHLETES PRESENTING TO UNITED STATES EMERGENCY

DEPARTMENTS, 1990-2003. J ATHL TRAIN 41(4), PP. 441-449.

159

YARD, E. E., AND COMSTOCK, R. D., 2006. INJURIES SUSTAINED BY

RUGBY PLAYERS PRESENTING TO UNITED STATES EMERGENCY

DEPARTMENTS, 1978 THROUGH 2004. J ATHL TRAIN 41(3), PP. 325-

331.

YARD, E. E., ET AL., 2007. PEDIATRIC MARTIAL ARTS INJURIES

PRESENTING TO EMERGENCY DEPARTMENTS, UNITED STATES

1990-2003. J SCI MED SPORT 10(4), PP. 219-226.

ZHANG, L., ET AL., 2004. A PROPOSED INJURY THRESHOLD FOR MILD

TRAUMATIC BRAIN INJURY. J BIOMECH ENG 126(2), PP. 226-236.

160

ABSTRACT

EVALUATION OF REPETITIVE HEADING IN YOUTH SOCCER

BY

ERIN HANLON

December 2009

Advisor: Cynthia Bir, Ph.D.

Major: Biomedical Engineering

Degree: Doctor of Philosophy

The specific aims of this project were: 1) to determine the incidence of

head injury in youth soccer related only to head to ball impacts using the National

Electronic Injury Surveillance System database; 2) determine the frequency of

heading in youth soccer based on age, gender, and skill level; 3) validate a novel

headband system to measure head impact frequency during soccer play; 4)

measure the biomechanical response of youth soccer players during heading

events using the Functional Assessment of Biomechanics motion capture system

and; 5) measure head impact frequency and severity using the Head Impact

Telemetry System (HITS). A total of 62,021.55 soccer head injuries were

estimated to occur from 2002 to 2007 in the United States with 15.9 % of these

injuries being caused by impact with the ball only. When observing the

frequency of heading occurrences, males had significantly more headers/minute

and total headers than females, but females incurred 59.64% of the injuries that

were caused by impact with the ball only. A novel head acceleration

measurement system, HITS, was found to provide a much needed method to

161

measure head acceleration in soccer players during normal play. It allows for

accurate measurements to be taken which could potentially lead to an injury

threshold specific to certain soccer impacts. The system was used in during

soccer scrimmages and measured head acceleration during impact events that

took place during these scrimmages. None of the impacts, heading or otherwise,

exceeded the tolerance levels for mild traumatic brain injury (mTBI) that were

previously established for linear head acceleration. Angular accelerations for

three heading events and one non heading even, however, did exceed the

suggested limits. Based on these observations, it seems as though the linear

acceleration contribution of heading is not causing head injury based on a single

impacts. Angular acceleration has potential problem and should be investigated

further in conjunction with determining tolerance values for multiple impacts.

162

BIOGRAPHICAL STATEMENT

ERIN HANLON

PLACE OF BIRTH: Georgetown, OH, USA EDUCATION: 2009 Ph.D Biomedical Engineering Wayne State University 2006 MS Biomedical Engineering Wayne State University 2004 BS Biomedical Engineering Wright State University ACADEMIC EXPERIENCE: 2005 to date Graduate Research Assistant Wayne State University 2007 to 2009 Mentor, Freshman Engineering Wayne State University 2005 Graduate Student Assistant Wayne State University SELECTED PUBLICATIONS: Hanlon, E., Bir, C. (2009). Validation of a Wireless Head Acceleration Measurement System for Use in Soccer Play. Journal of Applied Biomechanics. Accepted with Revisions. Hanlon, E., Bir, C. (2009). Real Time Measurement of Head Acceleration During Youth Soccer Play. Poster Presentation at the Summer Bioengineering Conference, June 2009, Lake Tahoe, California. Hanlon, E., Bir, C. (2008). A Model to Determine the Effect of Multiple Subconcussive Impacts in the Rat. Podium Presentation at the American Society of Biomechanics Annual Conference, August 2008, Ann Arbor, Michigan. Hanlon, E., Bir, C. (2007). The Determination of Heading Frequency in Youth Soccer. Podium Presentation at the American Society of Biomechanics Annual Conference, August 2007, Palo Alto, California.