freely chosen stride frequencies during walking and running are not correlated with freely chosen...

Accepted Manuscript

Title: Freely chosen stride frequencies during walking andrunning are not correlated with freely chosen pedallingfrequency and are insensitive to strength training

Author: Mahta Sardroodian Pascal Madeleine Michael VoigtErnst A. Hansen

PII: S0966-6362(15)00444-0DOI: http://dx.doi.org/doi:10.1016/j.gaitpost.2015.04.003Reference: GAIPOS 4463

To appear in: Gait & Posture

Received date: 19-11-2014Revised date: 13-4-2015Accepted date: 15-4-2015

Please cite this article as: Sardroodian M, Madeleine P, Voigt M, Hansen EA, Freelychosen stride frequencies during walking and running are not correlated with freelychosen pedalling frequency and are insensitive to strength training, Gait and Posture(2015), http://dx.doi.org/10.1016/j.gaitpost.2015.04.003

This is a PDF file of an unedited manuscript that has been accepted for publication.As a service to our customers we are providing this early version of the manuscript.The manuscript will undergo copyediting, typesetting, and review of the resulting proofbefore it is published in its final form. Please note that during the production processerrors may be discovered which could affect the content, and all legal disclaimers thatapply to the journal pertain.

http://dx.doi.org/doi:10.1016/j.gaitpost.2015.04.003

http://dx.doi.org/10.1016/j.gaitpost.2015.04.003

of 25

Accep

ted

Man

uscr

ipt

Highlights

- Walking, running, and cycling are supposedly generated by shared neural networks

- Stride rates in walking and running correlated positively

- Locomotion stride rates and pedalling rate were not correlated

- Strength training did not affect locomotion (contrasting prior results on cycling)

- Pedalling may be generated by neural networks mainly consolidated for locomotion

of 25

Accep

ted

Man

uscr

ipt

Movement frequency in walking, running, and cycling

2

Freely chosen stride frequencies during walking and running are not correlated with freely

chosen pedalling frequency and are insensitive to strength training

Mahta Sardroodian, Pascal Madeleine, Michael Voigt, and Ernst A. Hansen

Center for Sensory-Motor Interaction (SMI), Department of Health Science and Technology,

Aalborg University, Denmark

To be submitted as an original research paper to Gait & Posture

Corresponding author:

Ernst Albin Hansen, PhD, Associate Professor

Center for Sensory-Motor Interaction (SMI), Department of Health Science and Technology,

Aalborg University

Fredrik Bajers Vej 7, DK-9220 Aalborg, Denmark

Phone: (+45) 50 65 34 39

e-mail: [email protected]

Running head: Movement frequency in walking, running, and cycling

of 25

Accep

ted

Man

uscr

ipt


3

Acknowledgement

The authors express their thanks to the participating individuals for their time and effort. The

present study was supported by The Ministry of Culture Committee on Sports Research in Denmark

and The Obel Family Foundation.

Abstract

Despite biomechanical differences between walking, running, and cycling, these types of movement

are supposedly generated by shared neural networks. According to this hypothesis, we investigated

relationships between movement frequencies in these tasks as well as effects of strength training on

locomotion behaviour. The movement frequencies during walking, running, and cycling were

58.1±2.6 strides min-1, 81.3±4.4 strides min-1, and 77.2±11.5 revolutions min-1, respectively (n=27).

Stride frequencies in walking and running correlated positively (r=0.72, p<0.001) while no

significant correlations were found between stride frequencies during walking and running,

respectively, and pedalling frequency (r=0.16, p=0.219 and r=0.04, p=0.424). Potential changes in

the freely chosen stride frequencies and stride phase characteristics were also investigated during

walking and running through four weeks of (i) hip extension strength training (n=9), (ii) hip flexion

strength training (n=9), and (iii) no intervention (n=9). Results showed that stride characteristics

were unaffected by strength training. That is in contrast to previous observations of decreased

pedalling frequency following strength training. In total, these results are proposed to indicate that

walking and running movements are robustly generated due to an evolutionary consolidation of the

interaction between the musculoskeletal system and neural networks. Further, based on the present

results, and the fact that cycling is a postnatally developed task that likely results in a different

pattern of descending and afferent input to rhythm generating neural networks than walking and

of 25

Accep

ted

Man

uscr

ipt


4

running, we propose pedalling to be generated by neural networks mainly consolidated for

locomotion.

Key words: Central pattern generator; Motor behaviour; Resistance training; Voluntary frequency

of 25

Accep

ted

Man

uscr

ipt


5

1. Introduction

Walking, running, and cycling are common voluntary human rhythmic movements. A proper

functionality in these motor tasks is an important factor for human quality of life. Moreover, a better

understanding of the control mechanisms of such movement tasks is desirable for exercise and

rehabilitation purposes [1].

Rhythmic movements across vertebrate species are coordinated by neural networks

located in the brain and the spinal cord. The spinal components of these neural networks are termed

central pattern generators (CPGs), which generate an organized pattern of motor activity in

combination with adequate supraspinal descending and peripheral afferent influences [2-4]. The

existence of CPGs and homolog interneurons has been proven in several vertebrate species

(lampreys, mice, cats) [5]. However, it has been difficult to directly prove the existence of CPG

function in primates [6] and humans. Indirect evidence of existence of functional CPG-like spinal

neural networks has been reported in patients with spinal cord injuries [7,8] and in infants [9]. As

such, the analysis of motor behavior can be used to increase our knowledge of nervous system

organization and function [10].

According to the “common core hypothesis” [3], walking, running, and cycling may

share common central mechanisms. In support of this hypothesis, it has been shown that timing of

muscle activation in running can be described by the same basic temporal activation components

that previously had been reported for walking [11]. The authors of the latter study have suggested

that despite of distinct biomechanical differences between walking and running, these movements

are likely to be controlled by shared pattern-generating networks. Additionally, a similarity of

of 25

Accep

ted

Man

uscr

ipt


6

muscle synergies during walking and cycling has been reported. This may in turn indicate similar

synergies and modular control across different rhythmic movement patterns [12].

Freely chosen pedalling frequency has previously been studied on a cycle ergometer

[13], and it most likely reflects an innate voluntary rhythmic leg movement frequency linked to

CPG function [3,14]. It has been reported [15], and subsequently confirmed [16], that 12 weeks of

leg strength training combining hip extension and flexion exercises, caused recreationally active

individuals to reduce their freely chosen pedalling frequency. Furthermore, we recently reported

that 4 weeks of hip extension strength training reduced the freely chosen pedalling frequency in

recreationally active individuals [13]. Based on these results, it could be speculated that the freely

chosen frequency in other rhythmic movements, such as walking and running, would also be

affected by strength training.

We hypothesized that (1) freely chosen movement frequencies measured during

walking, running, and cycling correlate with each other, and that (2) freely chosen movement

frequencies during walking and running are decreased by a period of strength training as observed

previously in cycling [15,16].

2. Methods

2.1 Individuals

Twenty seven recreationally active individuals (14 men/13 women, age 24±5 years, height

1.78±0.09 m, and body mass 70.2±10.6 kg) volunteered. The study population was the same as in

our recent study [13]. The study was approved by the North Denmark Region Committee on Health

Research Ethics (N-20110025) and conformed to the standards of the Declaration of Helsinki.

of 25

Accep

ted

Man

uscr

ipt


7

2.2 Experimental design

The individuals were randomly divided into groups. A HET group (n=9) performed hip extension

strength training, while a HFT group (n=9) performed hip flexion strength training. In addition, a

CON group (n=9) was not exposed to training. The studylasted six weeks, including four weeks of

training. The individuals completed the following: Familiarization and one repetition maximum (i.e.

the maximum load that could be lifted in one repetition, 1RM) strength test, a pretest session, test

A1 after one week of training, test A2 after two weeks of training, test A3 after three weeks of

training, and finally a posttest session and a second 1RM strength test after four weeks of training.

The training consisted of two sessions per week, separated by at least one day.

2.3 Familiarization and determination of maximal strength

The individuals were familiarized with all procedures including 1RM strength testing, treadmill

walking and running, and ergometer cycling. Height, body mass, and leg length were measured.

Leg length was defined as the distance between the top of the anterior superior iliac spine to the

bottom of the lateral malleolus [17]. Next, the 1RM was determined for leg extension in HET and

for leg flexion in HFT (Fig. 1A-B). Determination of 1RM in CON was done so that five

individuals performed leg extension and four individuals performed leg flexion. Strength tests were

always preceded by 10 min warm-up on the cycle ergometer, at 100 W. Then, individuals

performed a standardized strength testing protocol. For more details see [13]. For the strength

testing and training, a Plamax Adjustable Pulley (Impulse Health Tech Ltd. Co., Jimo, Qingdao,

Shandong Province, China) was used. Walking and running were performed on a motorized

treadmill (Trimline 7200, Tyler, Texas, USA). Cycling was performed on an SRM cycle ergometer

(Schoberer Rad Messtechnik, Jülich, Germany) adjusted according to each individual’s preference

for settings of seat and handlebar [13].

of 25

Accep

ted

Man

uscr

ipt


8

2.4 Test sessions

The individuals always reported to the laboratory at the beginning of the week, at the same time of

the day. First, two pressure-sensitive sensors (SFR174, I.E.E., Contern, Luxembourg) were skin-

mounted with adhesive tape under each sole of the two feet. Positions were 1) under 2nd and

3rd metatarsal heads, and 2) under the centre of heel pad. The four sensors were connected to

custom-built amplifiers, and the signals were sampled at 2000 Hz, through a 16 bit A/D

converter, using a custom made LabVIEW-based software (LabVIEW, Austin, Texas, USA). Next,

the individuals performed 5 min of walking at 4.0 km h-1 immediately followed by 5 min of running

at 8.4 km h-1. These velocities were chosen to represent light to moderate intensities. Locomotion

was performed in a preferred way.

The LabVIEW-based software computed and saved (based on onset/offset detection)

stride duration, stance phase, and swing phase. Further, it calculated stride frequency for each foot

during the last minute of each locomotion bout. Then, the first ten error-free strides within the

recording period were selected for further analysis. Next, mean values across the two feet of each of

the stride characteristics were calculated. Afterward, a single mean of these values across the ten

strides was calculated for each stride characteristic, to be used in the further analysis.

Approximately 5 min after the running bout, the individuals performed 6 min of

ergometer cycling at freely chosen pedalling frequency, at 100 W, corresponding to light to

moderate intensity. Pedals with toe clips were used. Gear 8 and “constant Watt” operating mode

was selected on the cycle ergometer. This mode ensures a constant power output regardless of

pedalling frequency. The freely chosen pedalling frequency was noted at the end of each minute and

of 25

Accep

ted

Man

uscr

ipt


9

a mean value was calculated across the last 5 min. Data on pedalling frequencies from Test A1 and

beyond are reported elsewhere [13].

Absolute, rather than relative, values of ergometer cycling power output and

locomotion velocities were chosen since participants were relatively homogenous with respect to

daily activity level.

2.5 Strength training

All training sessions were supervised. They started with a 10 min warm up at self-selected intensity

on the cycle ergometer followed by two to three warm up sets with gradually increasing load. Both

legs were trained in an alternated way. The exercises performed in HET and HFT were identical to

those performed in the 1RM test (Fig. 1A-B). The targeted muscles in the present hip extension and

flexion exercises include the gluteus maximus muscles and the iliopsoas muscles as these muscles

contribute to propulsion and power in walking, running, and cycling [18,19]. For more details of

training protocol, see [13].

2.6 Statistical analyses

Data were tested for normal distribution using the Shapiro-Wilk test. A one-way analysis of

variance (ANOVA) was applied to test for differences between groups at pretest. For correlation

analyses on pretest data, Pearson correlation coefficients were calculated. The ratings based on the

size of r apply to both positive and negative correlations: ≤0.25 is weak, 0.26-0.50 is moderate,

0.51-0.75 is fair, and ≥0.76 is high [20]. Intra-class correlation coefficients (ICC for absolute

agreement) were calculated for evaluation of within-individual day-to-day reliability of stride

characteristics during locomotion. Student’s t-tests were applied to test for differences in pretest vs.

posttest 1RM changes between CON and each of the training groups. Two-way repeated measures

of 25

Accep

ted

Man

uscr

ipt


10

(RM) ANOVAs with time (i.e. test number) as within-individual factor and group as between-

individual factor were performed to test for differences in percentage changes of stride

characteristics from pretest to the subsequent tests between CON and each of the training groups.

Mean values, with 95% confidence intervals (CI95), were calculated for each of the individuals

across time to evaluate steadiness of the freely chosen stride frequencies during walking and

running, which has been done for freely chosen pedalling and finger tapping frequencies previously

[14]. Furthermore, Coefficient of Variation (CV) was used to measure between-individual

variability of freely chosen frequency during walking, running, and cycling at pretest. Statistics

were calculated in SPSS 21.0 (SPSS Inc., Chicago, IL, USA). Data are presented as mean±SD,

unless otherwise indicated. p<0.05 was considered statistically significant.

3. Results

3.1 Comparisons at pretest

There were no significant differences between HET, HFT, and CON at pretest with regard to age,

height, body mass, leg length, stance phase duration, swing phase duration, freely chosen stride

frequencies during walking and running, and freely chosen pedalling frequency (p=0.233-0.955)

(Stride characteristics data are presented in Tables 1 and 2).

Figure 2A shows freely chosen movement frequencies during walking (58.1±2.6

strides min-1), running (81.3±4.4 strides min-1), and cycling (77.2±11.5 revolutions min-1), for all

individuals. Since the data in the figure is ordered according to the individuals’ stride frequency

during walking, the figure clearly indicates a systematic relationship between freely chosen stride

frequency in walking and running while freely chosen pedalling frequency in cycling appears to be

unrelated to locomotion stride frequencies. Further, the figure demonstrates that there is more

between-individual variability for freely chosen pedalling frequency in cycling (CV=14.84%) than

of 25

Accep

ted

Man

uscr

ipt


11

for freely chosen stride frequencies in walking and running (CV=4.46% and CV=5.45%,

respectively). Freely chosen stride frequency during walking showed a fair correlation with the

corresponding data for running (Table 3). For additional correlations, the reader is referred to Table

3. The durations of stance phase, swing phase, and the complete stride during walking also showed

fair correlations with the corresponding data for running (Fig. 2B). The highest correlation

coefficient was found for stride durations, which directly reflect freely chosen stride frequencies.

The degree of within-individual steadiness of the freely chosen stride frequency

during walking was indicated by a mean CI95 of 1.32 strides min-1, calculated across the individuals,

with a range from 0.48-4.06 strides min-1. For comparison, the mean CI95 was 1.95 strides min-1

during running, with a range from 0.60-3.34 strides min-1.

3.2 Effects of strength training

Training adherence was 100%. 1RM increased by 33.6±13.2%, 28.1±13.0%, and 6.1±6.1% in HET,

HFT, and CON, respectively. The increases in HET and HFT were larger than the change in CON

(p=0.001 and p=0.003, respectively).

Unexpectedly, the strength training did not affect the stride frequencies and stride

phase characteristics during locomotion. Thus, for both walking and running, there was no

significant difference between each of the two training groups and CON (p=0.443-0.964). The

reader is referred to Table 1 and 2 for detailed data.

Since it was observed that strength training did not affect the primary outcome

variables of freely chosen stride frequencies and stride phase characteristics during walking and

running, the within-individual day-to-day relative reliability of these parameters was analysed

between pre-test and Test A1 (ICC=0.80-0.94, p<0.001). The reliability could be considered as

almost perfect (i.e. ICC>0.8, [21]).

of 25

Accep

ted

Man

uscr

ipt


12

4. Discussion

The present study revealed two novel findings: 1) In line with our first hypothesis, a positive

correlation was found between stride frequencies in walking and running. Though, no correlation

was found between freely chosen movement frequency in cycling and in locomotion. 2) Despite

increased maximal strength after strength training for both training groups, and contrary to our

second hypothesis, strength training did not influence freely chosen stride frequencies and the

corresponding stance phase and swing phase durations during locomotion.

Results are discussed with an evolutionary perspective [22]. It seems reasonable to

assume that over time the human body with its physiological systems has optimized and

consolidated its ability to perform specific locomotor tasks consistently and efficiently.

Furthermore, that this has taken place at the same time as an optimization of other abilities

important for survival and reproduction. One result of the evolutionary process is the current

phenotype of the human musculo-skeletal system. Further, walking and running have become the

evolutionary consolidated forms of locomotion [23-25]. Additionally, the function of the central

neural networks most likely has been consolidated in parallel, to control walking and running

movements specifically. As a consequence, the human central nervous system should be genetically

predisposed to control the musculoskeletal system in walking and running movements.

Considering walking and running as evolutionary and genetically determined

movement patterns being generated by the same central neural networks, it may seem logical that

the preferred movement frequencies in walking and running are positively correlated, as observed

here. For comparison, cycling must be considered a movement skill typically developed from

childhood, while not subjected to the same type of evolutionary pressure and consolidation, and to

be under the control of neural networks that are tailored for locomotion. This suggestion is

of 25

Accep

ted

Man

uscr

ipt


13

supported by the observation that freely chosen stride frequencies in walking and running are rather

closely optimized with respect to movement economy [25,26]. This is not the case during cycling

where the most energetically economical pedalling frequency during submaximal cycling (50-60

revolutions min-1) is markedly lower than the freely chosen pedalling frequency (70-90 revolutions

min-1 in the present study) [14,23,27]. It follows, that the behaviour of the automated rhythmic leg

movements presumably is under CPG-influence during cycling, but in a way that the freely chosen

pedalling frequency not necessarily is strongly correlated to the freely chosen movement

frequencies in walking and running. This could be the reason for the lack of significant correlation

between the freely chosen movement frequencies during cycling and locomotion as observed here.

Additionally, this lack of correlation may also partly be a consequence of the different afferent input

that is received by the central nervous system (and the CPGs) during cycling compared to walking

and running. In cycling, the postural demands are different than in walking and running due to the

fewer movement degrees of freedom. Also, cycling has less weight bearing. Therefore, the pattern

of both descending and afferent input to the CPGs during cycling must be markedly different from

the input to the same neural networks during walking and running.

Except that stride frequencies in walking and running correlated, our hypotheses were

not confirmed. This may point towards that different neural networks generate cycling, in

comparison to walking and running. This is actually not corroborated by the existing body of

literature. Partial differences in the neural control networks may be a part of the explanation of the

observed differences in the leg frequency behaviour between walking and running, and cycling,

because additional functional layers are added to the CPGs during development from childhood [5].

Therefore, cycling may be generated in different layers of the CPGs than locomotion.

The second main observation in the present study was that the freely chosen

movement frequencies during walking and running were not sensitive to four weeks of strength

of 25

Accep

ted

Man

uscr

ipt


14

training, contrary to what was hypothesised. Strength training of lower extremity muscles most

likely has several effects on both descending and afferent input to the spinal neural networks

[28,29]. However, these neuronal changes have only been measured on inactive individuals or

during static muscle contraction where spinal pattern generating neural networks most likely are

inactive. Therefore, the specific influences of strength training on the CPGs still needs to be

clarified by relevant measurements performed on moving subjects. Strength training has been

shown to induce synaptogenesis within the spinal cord, dependent on the specific behavioural

demands of the task, and without alteration of the motor map organization [30]. Therefore, in

recreational cyclists the influence of the neuronal changes due to four weeks of strength training on

the control of CPG-influenced movement patterns seems to be strong enough to decrease the freely

chosen movement frequency during cycling. However, due to the likely evolutionary consolidation

of the movement control during walking and running, it is possible that these movements are more

robustly controlled, i.e. less sensitive to variations in descending and afferent input, than the cycling

movement. That may provide an explanation of the lack of influence of four weeks of strength

training on the freely chosen stride frequencies and stride phase characteristics.

Further, our proposed evolutionary consolidation of movement control of walking and

running is supported by the lower between-individual variation of freely chosen frequency during

walking and running compared with cycling (Fig. 2A). This is also substantiated by the present

longitudinal data on freely chosen movement frequencies that demonstrated steady freely chosen

stride frequencies throughout the study period during walking and running. The present values of

within-individual variation were lower than what has been reported previously for freely chosen

pedalling frequency [14]. The present results give some indications of movement behaviour and

control in healthy, recreationally active, individuals. Further studies are required to elucidate the

same aspects in other populations such as in elderly or patient groups.

of 25

Accep

ted

Man

uscr

ipt


15

5. Conclusion

The present study showed that freely chosen stride frequencies and stride phase characteristics

correlated during locomotion. In addition, hip extension and hip flexion strength training did not

alter the stride characteristics during walking and running. The same central neural networks,

including CPGs, have been proposed to control the freely chosen movement frequency during

walking, running, and cycling. Considering the likely evolutionary development and consolidation

of the neural control of locomotion, the present results, and the fact that cycling is a postnatally

developed task with different descending and afferent inputs to spinal neural networks, we suggest

leg movements in cycling to be generated by neural networks evolutionary consolidated mainly for

locomotion.

Conflict of interest

None.

of 25

Accep

ted

Man

uscr

ipt


16

References

[1] Cambach W, Chadwick-Straver RV, Wagenaar RC, van Keimpema AR, Kemper HC. The

effects of a community-based pulmonary rehabilitation programme on exercise tolerance and

quality of life: A randomized controlled trial. The Eur Respir J 1997; 10: 104-13.

[2] Duysens J, Van de Crommert WAA. Neural control of locomotion; part 1: The central pattern

generator from cats to humans. Gait Posture 1998; 7: 131-41.

[3] Zehr EP. Neural control of rhythmic human movement: The common core hypothesis. Exer

Sport Sci Rev 2005; 33: 54-60.

[4] Zehr EP, Duysens J. Regulation of arm and leg movement during human locomotion.

Neuroscientist 2004; 10:347-61.

[5] Lacquaniti F, Ivaneko YP, d`Avella A, Zelik KE, Zago M. Evolutionary and developmental

modules. Front Comput Neurosci 2013; 7:1-6.

[6] Fedirchuk B, Nielsen J, Petersen N, Hultborn H. Pharmacologically evoked fictive motor

patterns in the acutely spinalized marmoset monkey (callithrix jacchus). Exp Brain Res 1998;

122: 351-61.

[7] Calancie B, Needham-Shropshire B, Jacobs P, Willer K, Zych G, Green BA. Involuntary

stepping after chronic spinal cord injury. Evidence for a central rhythm generator for

locomotion in man. J Neurol 1994;117:1143-59.

[8] Dimitrijevic MR, Gerasimenko Y, Pinter MM. Evidence for a spinal central pattern generator in

humans. Ann N Y Acad Sci 1998; 860: 360-76.

[9] Yang JF, Stephens MJ, Vishram R. Infant stepping: A method to study the sensory control of

human walking. J Physiol 1998; 507: 927-37.

of 25

Accep

ted

Man

uscr

ipt


17

[10] Goulding M. Circuits controlling vertebrate locomotion: Moving in a new direction. Nat Rev

Neurosci 2009; 10: 507-18.

[11] Cappellini G, Ivanenko YP, Poppele RE, Lacquaniti F. Motor patterns in human walking and

running. J Neurophysiol 2006; 95: 3426-37.

[12] Barroso F, Torricelli D, Moreno JC, Taylor J, Gomez-Soriano J, Esteban EB, et al. Shared

muscle synergies in human walking and cycling. J Neurophysiol 2014; 112: 1984-98.

[13] Sardroodian M, Madeleine P, Voigt M, Hansen EA. Frequency and pattern of voluntary

pedalling is influenced after one week of heavy strength training. Hum Movement Sci 2014;

36: 58-69.

[14] Hansen EA, Ohnstad AE. Evidence for freely chosen pedalling rate during submaximal cycling

to be a robust innate voluntary motor rhythm. Exp Brain Res 2008; 186: 365-73.

[15] Hansen EA, Raastad T, Hallén J. Strength training reduces freely chosen pedal rate during

submaximal cycling. Eur J Appl Physiol 2007; 101: 419-26.

[16] Rønnestad BR, Hansen EA, Raastad T. Strength training affects tendon cross-sectional area

and freely chosen cadence differently in noncyclists and well-trained cyclists. J Strength Cond

Res 2012; 26:158-66.

[17] Terry MA, Winell JJ, Green DW, Schneider R, Peterson M, Marx RG, et al. Measurement

variance in limb length discrepancy: Clinical and radiographic assessment of interobserver and

intraobserver variability. J Pediatr Orthopaed 2005; 25: 197-201.

[18] Sasaki K, Neptune RR. Differences in muscle function during walking and running at the same

speed. J Biomech 2006; 39: 2005-13.

[19] So RC, Ng JK, Ng GY. Muscle recruitment pattern in cycling: A review. Phys Ther Sport

2005; 6: 89-96.

of 25

Accep

ted

Man

uscr

ipt


18

[20] Berg KE, Latin RW. Essentials of research methods in health, physical education, exercise

science, and recreation; 2008, p. 124-28.

[21] Kitchen L, Westmacott R, Friefeld S, MacGregor D, Curtis R, Allen A, et al. The pediatric

stroke outcome measure: A validation and reliability study. Stroke 2012; 43: 1602-08.

[22] Newman SA. Form and function remixed: Developmental physiology in the evolution of

vertebrate body plans. J Physiol 2014; 592: 2403-2412.

[23] Martin P, Sanderson D, Umberger B. Biomechanics in Sport: Performance Enhancement and

Injury Prevention; 2000, p. 142-60.

[24] McNeill AR. Energetics and optimization of human walking and running: The 2000 raymond

pearl memorial lecture. Am J Hum Biol 2002; 14: 641-48.

[25] Zarrugh M, Radcliffe C. Predicting metabolic cost of level walking. Eur J Appl Physiol 1978;

38: 215-23.

[26] Cavanagh PR, Williams KR. The effect of stride length variation on oxygen uptake during

distance running. Med Sci Sports Exer 1982; 14: 30-35.

[27] Marsh AP, Martin PE. Effect of cycling experience, aerobic power, and power output on

preferred and most economical cycling cadences. Med Sci Sports Exerc 1997; 29: 1225-32.

[28] Aagaard P, Simonsen EB, Andersen JL, Magnusson P, Dyhre-Poulsen P. Neural adaptation to

resistance training: Changes in evoked V-wave and H-reflex responses. J Appl Physiol 2002;

92: 2309-2318.

[29] Sale D, Upton A, McComas A, MacDougall J. Neuromuscular function in weight-trainers. Exp

Neurol 1983; 82: 521-31.

[30] Adkins DL, Boychuk J, Remple MS, Kleim JA. Motor training induces experience-specific

patterns of plasticity across motor cortex and spinal cord. J Appl Physiol 2006; 101: 1776-82.

of 25

Accep

ted

Man

uscr

ipt


19

Figure captions

Fig. 1. Progressive strength training was performed with both legs, two days per week, for four

weeks. A) HFT performed hip flexion training . B) HET performed hip extension training. The

figure is a modification of figure 2 in [13].

Fig. 2. A) Individual values of freely chosen movement frequencies (stride frequency during

locomotion and pedalling frequency during cycling) from the pretest. Data are ordered so that

individual 1 is the one who had the lowest value during walking whereas individual 27 had the

highest value etc. B) Duration of stance phase, swing phase, and stride during running as a function

of the corresponding values during walking. Data is from the pretest.

of 25

Accep

ted

Man

uscr

ipt


20

Tables

Table 1. Stride characteristics during walking.

HET: Group performing progressive hip extension strength training. HFT: Group performing progressive hip flexion

Stance phase duration (s)

Swing phase duration (s)

Stride frequency

(strides min-1)

HET (n=9)

Pretest 0.60±0.03 0.44±0.04 57.6±2.8

Test A1 0.61±0.04 0.44±0.03 57.3±2.7

Test A2 0.60±0.04 0.45±0.03 57.4±2.4

Test A3 0.61±0.04 0.44±0.04 56.9±2.4

Posttest 0.62±0.04 0.43±0.03 57.2±2.3

HFT (n=9)

Pretest 0.59±0.03 0.44±0.02 58.7±2.7

Test A1 0.60±0.03 0.44±0.02 57.7±1.8

Test A2 0.61±0.03 0.43±0.02 58.1±2.0

Test A3 0.60±0.02 0.43±0.02 57.9±1.9

Posttest 0.61±0.03 0.42±0.01 58.1±1.6

CON (n=9)

Pretest 0.60±0.04 0.44±0.03 57.9±2.4

Test A1 0.61±0.03 0.44±0.03 57.0±2.4

Test A2 0.61±0.05 0.45±0.04 57.2±2.8

Test A3 0.60±0.04 0.44±0.04 57.7±2.4

Posttest 0.62±0.04 0.44±0.03 57.1±2.7

of 25

Accep

ted

Man

uscr

ipt


21

strength training. CON: Control group. Pretest: Test before training. Test A1: Test after one week of training. Test A2: Test after 2 weeks of training. Test A3: Test after 3 weeks of training. Posttest: Test after the complete training period.

of 25

Accep

ted

Man

uscr

ipt


22

Table 2. Stride characteristics during running.

For acronyms and explanations, see table 1.

Stance phase duration (s)

Swing phase duration (s)

Stride frequency

(strides min-1)

HET (n=9)

Pretest 0.26±0.02 0.49±0.05 80.0±5.0

Test A1 0.26±0.03 0.49±0.04 80.0±4.6

Test A2 0.26±0.04 0.49±0.04 80.6±4.9

Test A3 0.27±0.03 0.48±0.04 79.9±4.1

Posttest 0.27±0.03 0.48±0.04 79.8±4.5

HFT (n=9)

Pretest 0.24±0.02 0.48±0.03 83.1±3.2

Test A1 0.26±0.03 0.47±0.04 82.9±3.3

Test A2 0.26±0.03 0.46±0.03 83.3±2.9

Test A3 0.25±0.05 0.48±0.04 83.1±3.9

Posttest 0.25±0.02 0.47±0.04 82.7±2.9

CON (n=9)

Pretest 0.27±0.04 0.48±0.04 80.8±4.7

Test A1 0.27±0.04 0.49±0.04 80.1±4.3

Test A2 0.27±0.04 0.48±0.05 80.6±4.8

Test A3 0.27±0.03 0.47±0.05 80.9±4.7

Posttest 0.28±0.03 0.46±0.04 81.7±5.2

of 25

Accep

ted

Man

uscr

ipt


23

Table 3. Result of correlation analyses performed on pretest data.

Stride frequency during walking Stride frequency during running

r p r p

Pearson correlation

Age 0.08 0.704 0.09 0.646

Body height -0.46* 0.015 -0.51* 0.006

Leg length -0.38 0.053 -0.38* 0.048

Body mass -0.38 0.052 -0.57* 0.002

Pedalling frequency during cycling 0.16 0.219 0.04 0.424

Stride frequency during walking 0.72* <0.001

*Significant correlation.

of 25

Accep

ted

Man

uscr

ipt


24

Figures

Fig. 1.

of 25

Accep

ted

Man

uscr

ipt


25

A)

B)

Fig. 2.

freely chosen stride frequencies during walking and running are not correlated with freely chosen...

Documents