SUB-ACUTE HIPPOCAMPAL ATROPHY IN THE FIRST YEAR FOLLOWING

MODERATE TO SEVERE TRAUMATIC BRAIN INJURY

by

Danielle D. DeSouza

A thesis submitted in conformity with the requirements for the degree of Master’s of Science

Graduate Department of Rehabilitation Science

University of Toronto

© Copyright by Danielle D. DeSouza (2009)

ii

SUB-ACUTE HIPPOCAMPAL ATROPHY IN THE FIRST YEAR FOLLOWING

MODERATE TO SEVERE TRAUMATIC BRAIN INJURY

Degree of Master of Science, 2009

Danielle D. DeSouza

Graduate Department of Rehabilitation Science University of Toronto

Abstract

Rationale: Ng et al. (2008) demonstrated that sub-acute hippocampal atrophy occurred

between 4.5 and 24 months following moderate-to-severe traumatic brain injury (TBI); it

remains to be determined if atrophy occurred before 24 months. Objectives: (1) to determine

if sub-acute hippocampal atrophy occurs by the first year of injury; (2) to determine

associated clinical and demographic variables. Methods: Ten moderate-to-severe TBI

patients underwent MRI at 5 and 12 months post-injury. Glasgow Coma Scale (GCS) and

demographic variables were correlated with change. Results: Significant hippocampal

volume decreases were observed for right (P< 0.002, Cohen’s d= 0.34) and left (P< 0.036,

Cohen’s d= 0.22) sides. GCS was significantly correlated with right (r= -0.663, P< 0.037),

but not left percent hippocampal volume change (r= -0.327, P< 0.356). No significant

correlations were observed for demographic variables. Conclusion: Sub-acute hippocampal

atrophy occurs between 5 and 12 months post-injury and is associated with injury severity.

iii

Acknowledgements

First, I would like to thank my supervisor Dr. Robin Green for this wonderful learning

opportunity and for introducing me to the world of neuroimaging. I would also like to

thank my committee members Dr. David Mikulis and Dr. Nancy Lobaugh for their

expertise along the way.

Thank you to all of my GDRS friends and lab mates who have been extremely

supportive and made the past two years some of the most memorable. In particular,

thanks to Ephrem Pano, April Arundine, Diana Frasca, Yuko Koshimori, Alexandra

Arnold-Oatley, Sabrina Agnihotri, Melanie Andre, Jake Yoon and Quoc Hao Mach for all

of their support (not to mention the statistics and formatting help)!

Thank you to all of the participants from Toronto Rehab. This study would not have

been possible without your co-operation.

Last, but not least, thank you to my wonderful family for all of their patience, love

and support throughout this process. Mom, Dad, Danita and Tanya this thesis is as much

yours as it is mine. Michael, your support and understanding have meant the world to me.

iv

Table of Contents

Abstract ......................................................................................................................ii Acknowledgements...................................................................................................iii Table of Contents......................................................................................................iv List of Figures ...........................................................................................................vi List of Tables ...........................................................................................................vii List of Abbreviations ..............................................................................................viii

Chapter 1 General Introduction.............................................................................. 1

1.1 Rationale ................................................................................................................1 Chapter 2 Literature Review ..................................................................................4

2.1 Introduction............................................................................................................4 2.1.1 Definition and Classification of TBI…………………………………………………4 2.1.2 General Behavioural Characteristics of TBI……………………………………….5 2.2 Pathophysiological Manifestations of Moderate to Severe TBI ............................6

2.2.1 Overview………………………………………………………………………………...6 2.2.2 Acute Injury……………………………………………………………………………...7 2.2.2.1 Primary Injury………………………………………………………………..7

2.2.2.2 Secondary Injury...................................................................................9 2.2.3 Sub-Acute Atrophy & Behavioural Change……………………………………….14

2.2.3.1 Animal Studies of Sub-Acute Atrophy…………………………………..15 2.2.3.2 Human Studies: Global Sub-Acute Atrophy…………………………...17 2.3 Hippocampal Injury Following TBI…………………………………………………19 2.3.1 Human Studies: Sub-Acute Hippocampal Atrophy…………………….20

2.3.2 Controlling for Limitations of Previous Human Studies Examining Sub-Acute Atrophy: Ng et al. (2008) and Greenberg et al. (2008…………………….21 2.4 Correlates of Outcome from TBI......................................................................23 2.4.1 Correlates of Behavioural Outcome .................................................................23 2.4.2 Correlates of Pathophysiological Outcome......................................................25

2.4.3 Summary of Studies Examining Correlates of TBI Outcome……………………27 2.5 Summary of Literature Review .........................................................................27 2.6 Current Study....................................................................................................28

2.6.1 Objectives/Hypotheses......................................................................................29 Chapter 3 Methods............................................................................................... 31

3.1 Participants ......................................................................................................31 3.1.1 TBI Participants………………………………………………………………………31 3.1.1.1 Inclusion/Exclusion Criteria………………………………………………….31 3.1.2 Healthy Control Convenience Sample……………………………………………..32

3.2 Materials...........................................................................................................32 3.2.1 Demographic and Injury Variables………………………………………………..32

3.2.2 Structural MRI Protocol ...................................................................................33

v

3.3 Image Processing and Analysis for Hippocampal Volumes ............................33 3.3.1 Pre-Processing Steps........................................................................................33

3.3.2 Processing Steps............................................................................................34 3.3.3 Quality Control.............................................................................................35 3.3.4 Manual Segmentation of the Hippocampi.....................................................36 3.4 Inter- and Intra- Rater Reliability Assessment.................................................39

3.5 Design and Procedure......................................................................................40 3.5.1 Design ...............................................................................................................40

3.5.2 Procedure……………………………………………………………………………...40 3.6 Data Analysis ...................................................................................................41

3.6.1 Objective 1: Detection of Sub-Acute Hippocampal Atrophy by the First Year of Moderate and Severe TBI..................................................................................................41 3.6.2 Objective 2 (Exploratory): Examination of Correlates of Sub-Acute Hippocampal Atrophy........................................................................................................42 Chapter 4 Results................................................................................................. 44

4.1 Demographic and Injury Characteristics ........................................................44 4.2 Objective 1: Detection of Sub-Acute Atrophy in the First Year of Injury.…44

4.2.1 Hypothesis 1: TBI patients will show significant loss of volume across time that is not explained by age-related decline...................................................45

4.2.2 Objective 2 (Exploratory): Preliminary Examination of Correlates of Sub-Acute Hippocampal Atrophy......................................................................................50

Chapter 5 Discussion ........................................................................................... 52 5.1 Summary of Key Findings………………………………………………………….51 5.1.1 Sub-Acute Atrophy- Relationship to Prior Findings; Mechanisms…………52 5.1.2 Correlates of Sub-Acute Atrophy…………………………………………………55

5.2 Implications......................................................................................................58 5.3 Limitations........................................................................................................60 5.4 Future Direction for Research .........................................................................61 5.5 Conclusions ......................................................................................................62

References................................................................................................................63

vi

List of Figures

Figure 3.1 Tail of Left Hippocampus in Coronal View………………………………37 Figure 3.2 Body of Left Hippocampus ………………………………………………37 Figure 3.3 Left Hippocampus in Sagittal View………………………………………38 Figure 4.1 Mean Right Hippocampal Volume for the TBI and Control Groups……..45 Figure 4.2 Mean Left Hippocampal Volume for the TBI and Control Groups………46 Figure 4.3 Left and Right Percent Hippocampal Volume Changes for Individual

TBI Participants………………………………………………………………48 Figure 4.4 Left and Right Percent Hippocampal Volume Changes for Individual

TBI Participants……………………………………………………………….49 Figure 4.5 Correlation between Right Percent Hippocampal Volume Change and

GCS Score…………………………………………………………………….50

vii

List of Tables

Table 4.1 Group Data for TBI and Control Participants ...................................................45

viii

List of Abbreviations

Aβ Amyloid Beta ANOVA Analysis of Variance APOE4 Apoplipoprotein E ε4 allele APP Beta Amyloid Precursor Protein ATP Adenosine Triphosphate BVC Brain Volume Change CA Cornu Ammonis Ca2+ Calcium ion CBF Cerebral Blood Flow CCI Controlled Cortical Impact CDR Clinical Dementia Rating CIMT Constrained Induced Movement Therapy CSF Cerebrospinal Fluid CT Computerized Tomography DAI Diffuse Axonal Injury FCC Focal Cortical Contusion GCS Glasgow Coma Scale GOAT Galveston Orientation and Amnesia Test GOS Glasgow Outcome Scale GOSE Extended Glasgow Outcome Scale GM Gray Matter ICC Intraclass Correlation Coefficient ICP Intracranial Pressure K+ Potassium ion MRI Magnetic Resonance Imaging MVA Motor Vehicle Accident Na+ Sodium ion NS Not Significant SAH Subarachnoid Hemorrhage SD Standard Deviation SDH Subdural Hematoma TBI Traumatic Brain Injury VBP Volume of Brain Parenchyma VBR Ventricle-to-Brain Ratio WM White Matter

1

Chapter 1 General Introduction

1.1 Rationale

With advances made in brain injury prevention and safety promotion awareness (e.g.,

recommendations by the Canadian Medical Association) (Elford, http://www.phac-

aspc.gc.ca) as well as in medical treatment of traumatic brain injury (Hedges et al., 2009;

Lane et al., 2000; Rosner, 2003), more and more people are surviving serious brain

injuries (Chiu et al. 2007; De Silva et al., 2009; Lane et al., 2000). Indeed, traumatic brain

injury (TBI) is currently one of the leading causes of morbidity worldwide (Dikmen et

al., 2003; Yattoo and Tabish, 2008). Because TBI commonly affects young adults,

survivors can endure decades of disability (Centers for Disease Control and Prevention,

1999) caused by impairments to cognitive, motor and emotional functioning (Draper and

Ponsford, 2008; Jorge et al., 2007; Walker and Pickett, 2007). In Canada, approximately

50,000 individuals are affected annually (Ontario Brain Injury Association [OBIA],

2001), with 11.4/100,000 individuals sustaining a severe TBI (Zygun et al., 2005) and

associated costs estimated at $3 billion per year (Brain Injury Association of Canada

[BIAC], 2008).

Given the enormous impact of TBI, research into the underlying causes of disability

is critically needed in order to facilitate the development of treatments that can mitigate

its consequences. To this end, the focus of the current research is on gaining a better

understanding of a putative impediment to recovery from TBI. There is growing evidence

from the animal and human literature for a second wave of damage to the brain following

serious TBI, including the hippocampi (Ariza et al., 2006; Bigler et al., 2002; di Paola et

al al., 2008; Tomaiuolo et al., 2004). Given the importance of the hippocampi to memory

2

- and therefore to all aspects of day-to-day life - gaining a better understanding of

whether or not this sub-acute atrophy occurs and over what time course can help us to

design treatments to pre-empt atrophy, and thereby improve recovery from TBI. While

whole brain sub-acute atrophy can occur following TBI (MacKenzie et al., 2002; Trivedi

et al., 2007), the focus of the current study is on the hippocampi due to its known

vulnerability in trauma and its important contributions to memory processes.

A number of studies have demonstrated the acute loss of hippocampal volume after

TBI (Bigler et al., 1997; Bouilleret et al., 2009; Immonen et al., 2009), however very few

studies have specifically examined whether atrophy can occur sub-acutely; that is,

atrophy commencing after the acute events of injury (e.g., edema) have resolved. A

recent study conducted in our lab (Ng et al., 2008) - and one which obviated

methodological limitations of prior studies examining the possibility of sub-acute atrophy

- demonstrated significant hippocampal volume loss. Ng et al (2008) measured sub-acute

atrophy with an initial scan taken at 4.5 months post-injury and a follow-up scan taken at

24 months post-injury; however, it is possible that atrophy had occurred earlier than 24

months post-injury. This result is important in and of itself; however, an important further

question is whether atrophy occurred earlier than 24 months post-injury. Indeed,

preliminary findings from our lab suggest that sub-acute atrophy may occur by the first

year of injury.

Also, Ng et al. (2008) did not investigate the correlates of atrophy. Although a cross-

sectional study by Bigler et al. (1997) determined injury severity (as measured by the

Glasgow Coma Scale) was positively correlated with hippocampal volume 100+ days

post-TBI, a comprehensive review of the literature indicated that few longitudinal studies

have examined the relationship between clinical and demographic variables and the

3

presence and/or amount of sub-acute hippocampal atrophy following TBI (Bendlin et al.,

2008; Trivedi et al., 2007).

Therefore, the primary objective of the current study was to determine if sub-acute

hippocampal atrophy occurs in the first year of injury in a sample of moderately and

severely impaired TBI patients. A secondary, exploratory objective was to examine

correlates of sub-acute hippocampal atrophy.

Having a better understanding of when sub-acute hippocampal atrophy occurs can aid

in the development of interventions aimed at offsetting this secondary wave of atrophy by

providing information on when to introduce treatment. Furthermore, determining which

clinical and demographic variables are related to the pathophysiological outcome from

TBI will increase our understanding of which patients are at a greater risk of sub-acute

pathophysiological outcome. This understanding is important for long-term clinical and

occupational planning as well as identifying which patients should be targeted for

prophylactic treatment. Before discussing the methods of the current study, a review of

the relevant literature will be provided.

4

Chapter 2 Literature Review

This chapter will review the literature that is most germane to the objectives of this

thesis. The literature on mechanisms of damage in moderate to severe TBI will be

examined with a particular emphasis on the findings of Ng et al. (2008), which the

current study was designed to extend. With regard to a secondary exploratory objective,

the literature concerning correlates of clinical and pathophysiological outcomes from

moderate to severe TBI will be provided.

2.1 Introduction

Given the deleterious clinical and economic burden of TBI and the increasing

numbers affected, it is of value to have a thorough understanding of the pathophysiology

of TBI in order to design interventions that will optimize outcome for this group.

2.1.1 Definition and Classification of TBI

TBI refers to an externally inflicted trauma to the brain that may result in significant

impairments in physical, cognitive, social and affective functioning (NIH Consensus

Development Panel, 1999; Salmond and Sahakian, 2005), which can persist into the

longer term (Dikmen et al., 2003; Green et al., 2008; Levine et al., 2008).

TBI is typically classified as mild, moderate and severe, both in clinical practice and

in the scientific literature. These designations are commonly made using broad clinical

measures of injury severity such as the Glasgow Coma Scale (GCS), which measures

depth of coma, length of time in post-traumatic amnesia (PTA; i.e., the length of time a

5

person is disoriented and unable to encode new events following the injury) (Ruijs et al.,

1994), and acute care length of stay (ACLOS).

2.1.2 General Behavioural Characteristics of TBI

Following moderate and severe brain injuries, a wide range of impairments may be

sustained. Patients may often experience sensory deficits affecting auditory, olfactory and

visual domains (Haxel et al., 2008; Jury et al., 2001; Kapoor and Ciuffreda, 2002). Motor

function is often compromised (Walker and Pickett, 2007), including impairments in

motor speech (Wang et al., 2005), fine motor coordination (Neistadt, 1994), spastic

paralysis (Zafonte et al., 2004), balance and gait (Walker and Pickett, 2007). Changes in

mood and anxiety are common, both in reaction to lost participation in day-to-day life

(Hiott and Labbate, 2002; Jorge and Robinson, 2002; Parsons et al., 1995) and due to

direct organic changes. Emotion and behaviour regulation deficits, including anger and

antisocial behaviour are also common (Tonks et al., 2007).

Arguably the most disruptive consequences of TBI, however, are cognitive deficits.

While all domains of cognitive functioning can be affected, executive functioning

(Draper and Ponsford, 2008; McDonald et al., 2002; Ponsford et al., 2008), attention and

speed of processing (Draper and Ponsford, 2008; Madigan et al., 2000; Vitaz et al., 2003)

and memory (Levin et al., 1988; Levine, 1989; Ponsford et al., 2008; Schacter and

Crovitz, 1977; Vakil et al., 2005; Vitaz et al., 2003) have been reported, with memory

impairment being one of the most frequent complaints of patients and their family

members (Arcia and Gualtieri, 1993; Oddy et al., 1985; Vakil, 2005).

Memory impairments are due in part to damage to the hippocampi and to white

matter projections into and out of this structure. The formation of new long-term

6

memories, a process critically dependent on the hippocampi (Nadel and Moscovitch,

1997; Scoville and Milner, 1957; Squire, 1992; Winocur, 1985), as well as structures in

the diencephalon (McKee and Squire, 1992; Squire, 1987) and the frontal lobes (Fletcher

and Henson, 2001; Lee et al., 2000), is typically impaired after moderate and severe TBI.

It has been well established that individuals with TBI are impaired on verbal learning,

recall and recognition tasks (Ariza et al., 2006; Baddeley et al., 1987; Ferri-Campos et al.,

2008; Millis and Ricker, 1994; Vanderploeg et al., 2001; Wiegner and Donders, 1999;

Zec et al. 2001) as well as visuo-spatial learning, recall and recognition (Ariza et al.,

2006; Brooks, 1976; Brooker and George, 1984; Lehnung et al., 2003; Shum et al., 2000).

The physiological mechanisms of injury to the hippocampi and other regions of the

brain are complex. Injury can occur both acutely and sub-acutely, and within these

periods, a variety of mechanisms may operate to produce damage that is both focal and

diffuse.

2.2 Pathophysiological Manifestations of Moderate to Severe TBI

2.2.1 Overview

During TBI, a number of mechanical forces act on the brain to cause irreversible

damage. Although injuries whereby an individual remains static may result in TBI, the

most common mechanism of damage is from acceleration-deceleration injuries. This type

of injury occurs when inertial forces from the trauma cause the brain to move within the

skull. The result is typically a combination of focal damage where the brain strikes the

inner surface of the skull (for example in the form of cortical contusions or cerebral

haemorrhaging) and diffuse damage throughout the brain (Banich, 2004; Besenski et al.,

7

2002; Lezak et al., 2004). Damage may occur not only at the moment of impact, as is the

case with primary injury, but also minutes to months post-injury as pathophysiological

changes initiated by the primary injury take place (Lezak et al., 2004; Morganti-Kossman

et al., 2007).

Focal injuries include cortical contusions, infarcts, hematomas and tissue compression

secondary to raised intracranial pressure, with midline shift and herniation secondary to

edema and haemorrhaging occurring in very severe cases (Kan et al., 2006; Valadka et

al., 2000). Diffuse damage occurs to vulnerable axons as they twist and shear during the

trauma (de la Plata et al., 2007; Smith et al., 2003) and as small blood vessels supplying

axons are disrupted via petechial hemorraging (Povlishock and Katz, 2005), causing

mechanical and biochemical injury. This widespread white matter damage is called

diffuse axonal injury (DAI), and it largely contributes to the morbidity and mortality seen

in moderate and severe TBI (Adams et al., 1991). Furthermore, DAI can also contribute

to gray matter damage to areas remote from the initial impact site as slower, chronic

mechanisms acting on the white matter may lead to eventual cell body death (Anderson et

al., 1996). Eventually, these cellular mechanisms occurring post-injury lead to gross

atrophy, whereby large brain areas can deteriorate even after the acute phase of injury.

2.2.2 Acute Injury

2.2.2.1 Primary Injury

Traumatic damage to the brain occuring at the moment of impact has been widely

studied. Biomechanically, TBI can result from a violent blow to the head, directly

deforming and damaging brain tissue or it can result from acceleration-deceleration

8

incidents (e.g., a motor vehicle accident [MVA]) resulting in DAI and petechial white

matter haemorrhaging (Povlishock and Katz, 2005). TBI can also cause contusions to the

brain as the brain collides with the inner surface of the skull (Besenski et al., 2002). In

this scenario, various forces act on the brain to cause damage without the skull being

penetrated. Commonly, the forces involved in this type of injury (e.g., linear, rotational or

angular) produce hemorrhaging in the protective meningeal layers of the brain or

pronounced focal contusions in a coup and countercoup pattern (Besenski, 2002).

During TBI, not all areas of the brain are equally susceptible to damage; the frontal

and temporal lobes are reported to be the most vulnerable on impact (Bigler, 2007;

Lundy-Ekman, 2007; Pierallini et al., 2000). Although this is in part due to the action of

the linear forces involved (Besenski et al., 2002), it is also the result of the brain’s

orientation within the skull. Compartments in the skull known as the anterior and middle

cranial fossa surround the surface of the frontal and temporal lobes, respectively (Bigler,

2007). Consequently, bony surface protuberances cause damage, particularly in these

areas, when a sudden force causes the brain to move and forcefully make contact with

these protrusions.

Gray matter damage caused during primary injury usually results in subsequent tissue

loss due to irreversible neuronal damage and cell death. Cell death can occur by way of

three mechanisms that are not mutually exclusive: apoptosis, necrosis and autophagy,

with necrosis appearing to be the dominant means in primary injury (Kovesdi et al.,

2007). Necrosis is a passive process whereby mechanical deformation of the neuron

results in membrane damage, primarily to the cell soma, and also activation of an

inflammatory response leading to secondary injury (Kovesdi et al., 2007). At the moment

of impact, a cell may undergo immediate swelling and organelle breakdown, but if the

9

cell is damaged beyond a certain point, the swelling causes it to rupture, prompting the

secondary inflammatory response that is characteristic of necrosis (Goodlett and Horn,

2001).

In addition to the focal cortical damage that occurs in primary injury, DAI can also

occur in very severe circumstances when axons immediately rupture on impact (Maxwell

et al., 1993; Povlishock and Katz, 2005). More commonly, DAI is a delayed response that

occurs days to months following the TBI.

2.2.2.2 Secondary Injury

In the minutes to days post-injury, a number of additional events may take place to

cause increased neuropathology. It is important to note that some events may also occur

months post-injury. Although these events are also considered secondary, they occur

following the acute phase of injury. These sub-acute mechanisms will be described in

another section.

As previously discussed, necrosis is initiated during primary injury. In contrast, the

inflammatory response that causes the recruitment of neighbouring cells to also undergo

cell death, contributing to atrophy beyond the initial lesion site, is a secondary process

(Goodlett and Horn, 2001). The release of intracellular components after cell membrane

damage causes the inflammatory response associated with necrosis. Unlike apoptosis,

whereby cell signals are sent to phagocytes (of the immune system) to engulf the dying

cell, necrotic cells lack a signal (Sahuquillo et al., 2001). This makes it more difficult for

the immune system cells to locate and remove cells that have died by necrosis.

Furthermore, when cells die by necrosis they can release harmful enzymes usually stored

10

within the lysosome. Once released, these enzymes can trigger a chain reaction of further

cell death to neighbouring cells.

Also, TBI causes damage to surviving cells via secondary complex cellular

mechanisms, such as glutamate excitotoxicity, triggered by the initial event (Fujimoto et

al., 2004). Additionally, several secondary neurological effects such as edema, hypoxia,

herniation and hemorrhaging may occur acutely and contribute to the morbidity of this

population (Bigler et al., 1992; Pierallini et al., 2000; Povlishock and Katz, 2005).

Acutely, several disruptions to ionic balance may take place following injury. On

impact, neurotransmitter release, mitochondrial dysfunction, and subsequent membrane

depolarization of the injured cell may occur (Buki et al., 2000; Katayama et al., 1990;

Marmarou, 2007). Although the primary mechanical deformation initiates these events,

many of these neuronal changes are delayed and classified as secondary. These changes

are delayed because it is the combination of slower molecular processes (e.g., increased

axolemma permeability, increased intracellular calcium, cytoskeletal degredation) acting

together over time that can cause further damage. An example of when several molecular

processes co-occur is during glutamate excitotoxicity, a process involving the

extracellular accumulation of glutamate, the most abundant excitatory neurotransmitter in

the human nervous system. This accumulation results in the over-stimulation of

glutamate receptors, prolonged depolarization, and ionic imbalances including increases

in intracellular calcium, which subsequently leads to further cell dysfunction. Although

elevated levels of glutamate post-TBI have been reported in both animal and human

studies (Matsushita et al., 2000; Yamamoto et al., 1999), the use of glutamate antagonists

(i.e., NMDA receptor antagonists) as a treatment to offset these chemical cascades has

yielded mixed results (Beauchamp et al., 2008). As a result, recent research has focused

11

on metabolic consequences of injury and examined impaired glucose metabolism as a

potential cause of impairment.

In a study by Bergsneider et al. (2000), decreased cerebral glucose utilization (as

determined using 18F fluorodeoxyglucose positron emission tomography) was regionally

obtained in 88% of their sample of mild to severe TBI patients (n=42). Also, when global

reduction was examined, the prevalence was higher in severe TBI patients (86%) versus

their mild to moderately impaired counterparts (67%).

Ionic imbalances can also cause abnormal water accumulation in the intercellular

spaces between cells or within the brain cells themselves resulting in edema (Yi and

Hazell, 2006). Acute, secondary edema is one of the leading causes of death following

trauma to the brain. It results from an accumulation of fluid in the brain parenchyma and

it is highly associated with elevations in intracranial pressure (ICP) (Marmarou, 2007).

Together, these changes are frequent causes of mortality in this population. Two types of

edema may take place following brain injury: vasogenic and cytotoxic (Klatzo, 1967).

Although vasogenic edema has been highly implicated as the source of rapid, acute

increases in ICP following TBI, recent literature has determined that cytotoxic edema or

fluid accumulating within cells can co-exist with the former type, and is actually the

predominant type of edema when ionic imbalances are sustained over time, usually as the

result of impaired ATP-dependent Na+/K+ pumps (Ito et al., 1996; Marmarou, 2007;

Sweeney et al., 1995).

Previous research has suggested that the hippocampus is particularly vulnerable to

these excitotoxic mechanisms (Tate and Bigler, 2000). In a study by Sun and Faden

(1995), glutamate release was prohibited using sodium channel blockers in an animal

model of TBI. The prevention of glutamate release reduced the amount of hippocampal

12

neuronal loss in the CA1 and CA3 pyramidal cell fields from 59% to 22% in the

ipsilateral side. The CA1 area of the hippocampus appears to be particularly vulnerable to

the effects of TBI. Bramlett et al. (1999) demonstrated that following a secondary

hypoxic insult in a rodent model of TBI, there was a significant increase in the frequency

of damaged neurons in the CA1 areas of the hippocampus.

Another common process following TBI is secondary axonal damage. In contrast to

the theory that the twisting and shearing of axons caused by TBI are the immediate cause

of swollen axon balls or disruption of the injured axon segment (Adams et al., 1982;

Gennarelli et al., 1982; Povlishock et al., 1983), more recent findings (e.g., Li et al.,

1999; Pettus et al., 1994; Wolf et al., 1999) have determined that focal changes at the

axon membrane, primarily following acceleration-deceleration TBI, results in progressive

secondary changes that hinder axonal transport causing swelling of the axon and eventual

detachment from its downstream segment over time (Park et al., 2008; Maxwell et al.,

1993; Povlishock et al., 1983; Smith and Meaney, 2000). Additional studies have also

provided evidence for impaired axonal transport, whereby organelle delivery via

anterograde transport is hindered as a result of permeability changes in the axon

membrane caused by the mechanical deformation of the brain at the moment of impact

(Buki et al., 1999; Smith and Meaney, 2000). These primary changes in permeability,

caused by the formation of pores in the axolemma, result in an increased secondary influx

of extracellular calcium ions (Ca2+) within the cell. This atypical increase produces local

degradation of the cytoskeleton (Buki et al., 1999; Smith and Meaney, 2000) organelle

accumulation, axonal swelling and eventual detachment (Buki and Povlishock, 2006).

Adding to this, there is evidence for the rapid accumulation of proteins following

brain injury. In particular, β-amyloid precursor protein (APP) and amyloid-β (Aβ)

13

peptides, typically known for their role in Alzheimer’s disease histopathology, have been

examined (Smith et al., 2003). Although several epidemiological studies have reported a

link between TBI and the development of AD (Heyman et al., 1984; Rasmusson et al.,

1995; Schofield et al., 1997), until recently it was not understood how Aβ peptides

(derived from APP) aggregate to form plaques rapidly following brain injury. Much of

the evidence points to damaged axons as the main source of Aβ plaque formation

following TBI. APP is typically found in large quantities in axons, as it is carried via fast

axonal transport mechanisms. Thus, when white matter areas are damaged, this protein

has been shown to accumulate in damaged axons (Sheriff et al., 1994), and also, APP and

Aβ can co-accumulate in the terminal swellings of disconnected axons following brain

trauma, as shown in a swine model of TBI (Smith et al., 1999). Moreover, identifying

APP via immunocytochemical methods has been shown to be a useful tool for identifying

DAI in humans (Sherriff et al., 1994).

With all of these secondary pathological changes occurring on a cellular and

molecular level, including excitotoxicity and white matter damage, it is not surprising

that sub-acute atrophy is a consequence of moderate to severe TBI. Although sub-acute

atrophy can occur to the brain globally, some structures, in particular the hippocampus,

are particularly vulnerable to trauma. In the next section, studies examining sub-acute

atrophy post-TBI will be discussed. These studies include those that focused on global

brain changes and those that examined changes to the hippocampi, specifically.

14

2.2.3 Sub-Acute Atrophy & Behavioural Change

In addition to the mechanisms of injury described above, there is now growing

evidence from our lab and others that atrophy takes place during the sub-acute stage of

recovery, after the acute neurological events of injury (e.g., edema, hematoma, microglial

proliferation, encephalomalacia, clearing of traumatized cells) have resolved (Di

Giovanni et al., 2005; Ng et al., 2008; Greenberg et al., 2008; Trachtenberg, 1983). For

the current purposes, sub-acute atrophy refers to progression of volume loss and tissue

damage that cannot be attributed to the resolution of acute injury (e.g., volume loss in a

region of the brain that is actually caused by resolution of edema or hematoma). No

particular mechanisms of atrophy are presumed. In the current study, the period after

which any observed atrophy is considered sub-acute is from approximately 4.5 months

onwards, when edema is presumed to have resolved. This is different from early atrophy

which occurs at the time of injury or shortly thereafter.

Following early recovery from TBI, studies have demonstrated that the brain may not

be stable, as presumed. Some studies have provided evidence that following early

recovery (i.e., within the first year of injury), some individuals may show subsequent

cognitive decline (Himanen et al., 2005; Millis et al., 2001; Till et al., 2008). For

example, Millis et al. (2001) examined neuropsychological performance at 1 and 5 years

post-injury in a sample of mildly to severely impaired TBI patients. The results indicated

that some individuals improved or remained the same; however, approximately 15%

actually declined in performance over time.

Additionally, a study conducted in our lab (Till et al., 2008) examined the

neuropsychological performance of moderate to severe TBI patients from a baseline of

12-months post-injury to a follow-up conducted on average 2 years later (SD= 0.99). It

15

was determined that statistically significant cognitive decline was apparent on at least two

neuropsychological measures in approximately 27% of the group. Although results were

variable, decline was most commonly observed for verbal fluency and delayed recall

(tests that rely on memory, executive functioning and psychomotor speed) (Spreen and

Strauss, 1998; van Beilen et al., 2004). Sub-acute atrophy is one explanation for the

cognitive declines observed in these patients.

In fact, our laboratory has found preliminary evidence suggesting that memory

recovery is affected by sub-acute atrophy. In a study examining the relationship between

hippocampal atrophy and memory change over time, our lab determined that memory

recovery from 4.5 to 12 months post TBI was negatively correlated with hippocampal

atrophy over this same time period: therefore, the greater the atrophy, the less the

memory recovery from 4.5 to 12 months post-injury. No significant correlations were

obtained for memory recovery and hippocampal atrophy from 4.5 to 24 months post-

injury. Thus, these findings indicate that sub-acute hippocampal atrophy may indeed

occur by 12 months post-injury.

2.2.3.1 Animal Studies of Sub-Acute Atrophy

Animal studies in TBI research have been critical, as they have allowed for the

controlled and reproducible examination of TBI. Specifically, the animal literature has

provided evidence for sub-acute atrophy that occurs following the acute phase of

moderate to severe injury (Colicos et al., 1996; Immonen et al., 2009; Rodriguez-Paez et

al., 2005; Smith et al., 1997). These studies have examined both gray matter and white

matter regions of the brain and have outlined potential mechanisms underlying these sub-

acute changes. It is important to examine sub-acute white matter pathology even when

16

focusing on sub-cortical gray matter structures. Over time, the secondary loss of axons, in

addition to the primary cortical damage, can lead to further widespread damage to

cortical and sub-cortical gray matter structures, as important input and output connections

may be lost (Loftus et al., 2000).

In a study by Rodriguez-Paez et al. (2005), it was determined that chronic axonal

changes were present following TBI in a population of male rats that underwent moderate

parasagittal fluid-percussion brain injury. In this study, animals were sacrificed and

examined at one of seven time points: 3 days, 15 days, 1 month, 3 months, 6 months, 9

months or 12 months post-TBI. Compared to sham animals, there were significant

decreases in the estimated number of myelinated axons in the fimbriae (portion of the

fornix), external capsules and white matter tracts of the thalami in the TBI animals.

Although not statistically significant, the number of myelinated axons continued to

decrease over time in the fimbriae (from 3 to 12 months) and external capsules (from 15

days to 12 months). In contrast, myelinated axons in the thalami significantly decreased

over time, indicating that damage to sub-acute white matter occurred in addition to initial

acute damage. This study was one of the first to quantitatively describe chronic white

matter axonal damage post-TBI.

Similarly, Smith et al. (1997) examined rats following severe parasagittal fluid

percussion injury at nine different time points: 1 hour (h), 2h, 48h, 1 week, 2 weeks, 1

month, 2 months, 6 months and 1 year post-TBI. These authors determined that

substantial and progressive tissue loss occurred in several structures including the

cerebral cortex and hippocampus of these animals. Most notably, at all time points

between 48h and 1- year post-injury, including from 6- months to 1- year, significant

decreases in the pyramidal cell layer of the ipsilateral hippocampi were observed. These

17

authors suggested that a progressive, degenerative process might have been initiated by

the initial trauma.

2.2.3.2 Human Studies: Global Sub-Acute Atrophy

Using MR-based quantitative tools for analyses, several studies have reported sub-

acute decreases in brain parenchyma following TBI and associated increases in

cerebrospinal fluid (CSF) volume, indicative of atrophy. A portion of these studies have

been cross-sectional in design and have either collected control volumes for comparison

(Ariza et al., 2006; Bigler et al., 2002; Blatter et al., 1997; Jorge et al., 2007; Levine et

al., 2006; Levine et al., 2008; Tomaiuolo et al., 2003; Yount et al., 2002) or have used

normative databases to compare TBI findings, such as the one compiled by Blatter and

colleagues (1995). Collectively, these studies have found that increased injury severity, as

measured by GCS or PTA, is related to greater atrophy following trauma as demonstrated

by decreases in total brain volumes, hippocampal volumes, posterior cingulate gyrus

volumes, increases in ventricle-to-brain ratio (VBR), and increases in CSF volumes.

However, the cross-sectional design poses a limitation in determining individual

differences within this heterogeneous population. Also, the patients in these studies were

assessed at varying lengths of time post-injury. For example, Ariza et al. (2006)

examined patients that acquired MRI scans from 189-603 days post-injury. Variability in

the length of time between the injury and the MRI make it difficult to determine the time

course of these sub-acute changes.

A small number of human studies have employed longitudinal designs to examine

sub-acute atrophy post-TBI (Bendlin et al., 2008; Greenberg et al, 2008; Mackenzie et al.,

2002; Ng et al., 2008; Trivedi et al., 2007). The advantage of this strategy is that within-

18

subject factors are the same at all time points of analysis. Therefore any longitudinal

decreases in volumes of interest are due to increases in sub-acute atrophy, and not merely

differences between the groups (i.e., the TBI and control samples). As individual

decreases in volume can be tracked over time, longitudinal imaging is a particularly

valuable tool for the assessment of progressive neurological disorders (Whitwell, 2008).

Of the studies that employed longitudinal designs (Bendlin et al., 2008; Greenberg et

al, 2008; Mackenzie et al., 2002; Ng et al., 2008; Trivedi et al., 2007), differences in the

times of MR data acquisition have been observed. MacKenzie and colleagues (2002)

retrospectively assessed total volume of brain parenchyma (VBP) from MR images in

mild-moderate TBI patients at two time-points: within the first three months of injury and

again within three months of the initial scan. Although this study did not find significant

decreases in VBP as compared to controls, the longitudinal analysis did show that the rate

of atrophy was greater in the TBI group.

Trivedi and colleagues (2007) used MR-derived analyses to demonstrate decreases in

percent brain volume change (%BVC) in mild to severe TBI patients, as compared to

controls within approximately the first year of injury. This study supports the possibility

that TBI can lead to progressive neurological changes as soon as one year post-injury.

These longitudinal studies provide stronger evidence that sub-acute atrophy occurs.

However, one particular limitation was evident in all but the studies of Ng et al (2008)

and Greenberg et al (2008) from our lab. A sub-set of the initial MRI scans were taken

during the acute phase of injury. For example, in the MacKenzie et al. (2002) study,

initial scans were acquired as early as seven days post-TBI (range: 7-430 days).

Similarly, in Trivedi et al. (2007), some of the initial MRI scans were taken as early as 39

days following injury (range: 39-109 days). Therefore, because some individuals

19

acquired their first scans very early post-injury, any observed volume loss may reflect the

resolution of edema or hematoma. In other words, the brain may have looked smaller

over time because the edema or blood products that accrued acutely had now been

eliminated. As the authors themselves noted, this could have accounted for some or all of

the sub-acute atrophy observed. As previously mentioned, two types of edema may occur

following TBI: vasogenic and cytotoxic. The latter of these two commonly occurs when

ionic imbalances persist over time (Ito et al., 1996; Marmarou, 2007). The hippocampus,

and in particular the pyramidal neurons of this structure, have been shown to be

particularly vulnerable to ionic imbalance as a result of excitotoxic mechanisms (Figiel et

al., 1997). The edema that results could be a potential partial confound of these studies

(Mackenzie et al., 2002; Trivedi et al., 2007) if scans were taken too early.

2.3 Hippocampal Injury Following TBI

In addition to these studies examining global brain atrophy post-TBI, some studies

have specifically examined hippocampal volume decrease following trauma. Although

the hippocampi sit deep in the temporal lobes, several studies have demonstrated that this

structure is particularly vulnerable to sub-acute atrophy following TBI (Bigler et al.,

1996; Bigler et al., 1997; Bigler et al., 2002; Tate and Bigler, 2000). This is in part due to

excitotoxic injury and other mechanisms that act on the many white matter fibres that

project to, from and within this structure. These findings are not surprising considering

memory impairments are of the most common and persistent cognitive sequelae

20

following TBI (Auerbach, 1986). Thus, hippocampal atrophy and associated memory

impairments have been a particularly common research focus in the literature.

2.3.1 Human Studies: Sub-Acute Hippocampal Atrophy

In a cross-sectional study mentioned above, Tate and Bigler (2000) used MRI scans

to examine the hippocampi and fornices (major efferent fibres from hippocampi) of 86

TBI patients of all severities, as measured by the GCS. MRI scans were acquired at least

2- months post-injury and were compared to 46 control participants. It was determined

that sub-acute atrophy of both the fornix and hippocampi occurred following TBI, with

the degree of atrophy being related to the severity of the injury.

Additionally, Bigler et al. (1996) examined two TBI groups: an early group

comprised of patients who had MRI scans up to and including 90 days post-injury, and a

late group that included patients who acquired scans more than 90 days post-injury. Both

groups consisted of individuals with GCS scores between 3 and 15; however, the authors

noted that they were generally in the moderate to severe range of injury severity.

Compared to a sample of control participants (n=72), it was determined that hippocampal

volume for the late TBI group was significantly smaller than those of controls, but for the

left hippocampus only. The early TBI group did not have significantly smaller

hippocampi for either side.

In another study by Bigler et al. (1997), hippocampal volumes were significantly

smaller in individuals with TBI as compared to normal controls. Like the previous study,

TBI participants with GCS scores ranging from 3 to 15 were divided into two groups:

those that had MRI scans up to and including 100 days post-injury (early group, n=45)

and those that had MRI scans acquired more than 100 days post-injury (late group, n=55).

21

The results indicated that only the late group had significantly smaller hippocampi as

compared to controls and hippocampal volume was positively correlated with GCS.

Unlike the previous study by this group (Bigler et al., 1996), both left and right

hippocampal volumes were significantly smaller than those of normal controls.

Furthermore, the authors themselves noted that the decreases in hippocampal size

stabilized, “some time after 100 days following injury,” and thus atrophy was taking

place sub-acutely.

Although these studies provide evidence for sub-acute hippocampal volume loss

following TBI, the main limitation of these studies is that they were cross-sectional in

design.

2.3.2 Controlling for Limitations of Previous Human Studies Examining Sub-Acute Atrophy: Ng et al. (2008) and Greenberg et al. (2008)

In order to obviate some of the limitations of previously conducted studies, Ng et al.

(2008) conducted a longitudinal study to quantify global hippocampal volume changes,

and their first MRI acquisition was taken after acute neurological events were presumed

to have resolved based on expert opinion. In this repeated-measures design, MRI scans of

TBI patients (n=14) were taken at two sub-acute time-points: 4.5 months (SD= 0.5) and

24 months (SD= 4.4) post-injury. In the first part of this study, three expert

neuroradiologists reviewed all 14 pairs of scans and were informed that they belonged to

TBI patients and were taken at two sub-acute time-points. The early and late scans were

randomized into rows and the expert raters were asked to determine if deterioration, no

observable change or improvement occurred. The raters unanimously agreed that eight of

22

the 14 pairs of scans were indicative of progression of atrophy; for two of the scans, there

was agreement that progression of atrophy occurred by two of the three raters.

In the second part of the study, hippocampal and CSF volumes were obtained via

MRI volumetric analyses. These quantitative analyses revealed significant increases in

CSF volume and decreases in left and right hippocampal volume. Moreover, when the

annual percent volume change was compared to published normative data, there was far

greater volume loss; this provided evidence that volume loss was not attributable to age-

related decline. Most importantly, as initial scans were acquired at approximately 4.5

months post-injury, it was very unlikely that the acute effects of the trauma contributed to

the volume decreases. Therefore, this study provided strong evidence for bona-fide sub-

acute hippocampal atrophy from approximately 4.5 months to 24 months following brain

injury.

Corroborating these sub-acute findings, another study from our lab (Greenberg et al.

2008) observed sub-acute decreases in white matter integrity, as measured by MRI-

diffusion tensor imaging (DTI). Similar to the Ng et al. (2008) study, sub-acute changes

were observed from approximately 4.5 months (SD= 0.4) to 29 months (SD= 4.0) post-

injury in a sample of moderately to severely impaired TBI patients. The areas that were

significantly compromised were the deep frontal and temporal white matter fibres of both

hemispheres. Thus, not only has evidence been provided for sub-acute hippocampal

atrophy, but our lab has also provided evidence for sub-acute decreases in white matter

integrity in regions of interest within the frontal and temporal lobes.

Although our lab has found evidence for significant sub-acute hippocampal atrophy

and progressive damage to some white matter post-injury, the time course for these sub-

acute changes remains to be determined. In other words, sub-acute atrophy was observed

23

between 4.5 and 24 months post-injury, but it is possible that these changes could have

occurred earlier. As mentioned earlier, our lab has found preliminary cognitive evidence

suggesting that this might be the case.

2.4 Correlates of Outcome from TBI

Although sub-acute hippocampal atrophy may occur for some TBI patients by the

first year of injury, the degree of atrophy may not be the same for everyone. The outcome

from TBI can be quite variable from patient to patient. Therefore, if it is determined some

individuals with moderate to severe TBI demonstrate sub-acute hippocampal volume loss

by the first year of injury, it is important to have an understanding of which factors

moderate this pathophysiological outcome. In order to determine which variables may be

of most relevance, we can look to literature that has examined correlates of

pathophysiological outcome. We can also examine studies concerning behavioural

outcomes, as these are logically dependent upon what is taking place at a

neurophysiological level.

2.4.1 Correlates of Behavioural Outcome

The clinical outcome from TBI can vary markedly between patients and a large

number of studies have examined variables that put patients at risk for poor behavioural

outcome. With regard to pre-morbid factors, the results of studies have been mixed, but

older age, pre-injury unemployment, lower education, and pre-injury substance abuse,

have been frequently been found to be related to poor functional and cognitive outcomes

24

(Demetriades et al., 2004; Green et al., 2008; LeBlanc et al., 2006; Wilde et al., 2004;

Ponsford et al. 2008) including ongoing disability, for example, in a systematic review by

Willemse-van Son et al. (2007).

Injury severity indices (e.g., GCS and PTA) have also been correlated with outcome

from TBI. Perhaps the most commonly used index is the GCS (Teasdale and Jennett,

1974). GCS has been shown to be useful in the clinical management and prognosis of

TBI (Saatman et al., 2008; Sherer et al., 2007). The GCS is composed of three elements:

eyes, verbal and motor. A total score out of 15 is given, with scores ≤8 classified as

severe, scores between 9 and 12 moderate, and scores ≥13 mild. Several studies have

determined that a relationship exists between GCS and functional outcome from TBI

(e.g., Dikmen et al., 1995; Park et al., 2009; Ponsford et al., 2008; Utomo et al., 2009),

but a study by Zafonte et al. (1996) determined that GCS as a single measure has limited

value as a predictor of functional outcome, a finding that was also corroborated in a

review by Zasler et al. (2007).

Another measure of severity is post-traumatic amnesia (PTA), which captures the

length of time a person is disoriented and unable to encode new events following the

injury (Ruijs et al., 1994). PTA is collected daily in the acute period following TBI,

whereas initial GCS is only measured once (Wilde et al., 2006). Length of time in PTA is

often measured using the Galveston Orientation and Amnesia Test (GOAT) (Levin et al.,

1979). The GOAT, a test administered repeatedly following closed head trauma, has been

shown to predict long-term recovery (Levin et al., 1979); however, because it is labour-

intensive to measure PTA, it is often not collected. (I removed the paragraph describing

25

studies that found significant relationships with PTA…this should be less confusing to

the reader)

2.4.2 Correlates of Pathophysiological Outcome

Few studies have directly examined risk factors for pathophysiological outcome in the

brains of traumatically injured individuals (Bigler et al., 1997; Bigler et al., 2006; Ding et

al., 2008; Levine et al., 2008; Wilde et al., 2006). Of those that have, one study conducted

by Levine et al. (2008) determined that TBI severity, as measured by the GCS at

discharge from the trauma unit, was negatively correlated with parenchymal volume at

one year post-injury; thus, the more severe the TBI, the more sub-acute brain atrophy.

The authors noted that the frontal and temporal regions showed the most robust

correlations with this severity measure. Also, patients with moderate and severe TBI

could be differentiated from patients with mild TBI, and mild TBI participants could be

differentiated from and age and education matched control group.

A cross-sectional study by Bigler et al. (1997) also used GCS to examine the

relationship between injury severity and hippocampal volume following TBI. It was

determined that injury severity was positively correlated with hippocampal volume 100+

days post-TBI. This correlation was not evident for a group of TBI patients that acquired

MRI scans earlier than 100 days following injury. This suggests that sub-acute decreases

in hippocampal volume occurred and were related to initial injury severity.

Bendlin et al. (2008) also utilized GCS as a measure of severity to determine its

relationship with five different brain measures at approximately 1- year post-injury

(range= 252-380 days). Analyses were conducted using MR-DTI and included fractional

26

anisotropy of white matter, mean diffusivity of white matter, mean diffusivity of gray

matter, white matter volume and gray matter volume. Significant correlations were

obtained between GCS scores collected within 24 hours of injury and all five

pathophysiological measures. Of most relevance to the current study, a significant

negative correlation was obtained between GCS and mean diffusivity the fornix, the

major efferent tract from the hippocampus. Thus, lower GCS scores within 24 hours of

injury were associated with greater mean diffusivity of the fornix (indicative of

pathology) at approximately 1- year post-injury.

Wilde, Bigler, Pedroza and Ryser (2006) examined the relationship between PTA and

ventricle-to-brain ratio (VBR), a common measure of global cerebral atrophy, at least 90

days post-injury (range: 92-2,748 days). The authors noted their rationale for using PTA

to predict long-term atrophy was that PTA is collected repeatedly in the acute period

following TBI, whereas GCS is typically measured once. The results of this study

indicated that a strong relationship was evident between PTA and the probability of

abnormal VBR, with the longer durations of PTA being associated with increased

probability of abnormal VBR. These studies provide evidence that certain variables,

including measures of injury severity such as GCS scores and duration of PTA, are

related to the extent of sub-acute pathophysiological outcome in TBI patients.

2.4.3 Summary of Studies Examining Correlates of TBI Outcome

In summary, many studies have examined predictors of functional and cognitive

outcomes using demographic variables and measures of injury severity. While there

exists some evidence that demographic variables such as age and years of education are

27

related to functional and cognitive outcome (Demetriades et al., 2004; Green et al., 2008;

LeBlanc et al., 2006; Williamse-van Son et al., 2007; Zafonte et al., 1997), other studies

did not find these relationships (e.g., Bush et al., 2003; Girard et al., 1996; Novack et al.,

2001). Understanding the relationship between demographic variables and measures of

injury severity on cognitive and functional outcome from TBI is important because these

variables may also be related to pathophysiological outcome. It is presumed that

pathological brain changes resulting from the trauma underlie impairments in function

and cognition. If this is the case, then these variables may also be related to the

pathophysiological changes in structures known to be involved in commonly impaired

cognitive and functional domains. It remains to be determined if the demographic

variables of age and years of education are correlates of the degree of sub-acute

hippocampal volume change following moderate to severe TBI.

Additionally, a few studies have provided evidence that injury severity, as measured

by GCS and PTA, is significantly correlated with sub-acute pathophysiological outcome

from TBI (Bendlin et al., 2008; Bigler et al., 1997; Wilde et al., 2006). Although these

studies have observed a relationship between these measures and sub-acute measures of

atrophy, they did not specifically examine the relationship between injury severity and

sub-acute hippocampal volume change.

2.5 Summary of Literature Review

The literature has revealed that sub-acute brain atrophy, and in particular,

hippocampal atrophy occurs following TBI (Bigler et al., 1996; Bigler et al., 1997; Ng et

al., 2008; Tate and Bigler, 2000). Several studies have used structural MRI volumetry as

28

a tool to capture volumetric changes in the hippocampi (Bigler et al., 1997). However, of

the studies that examined hippocampal volume post-TBI and that were longitudinal in

design, only one (Ng et al., 2008) assessed hippocampal volume at two sub-acute time

points where observed volume loss could be confidently attributed to atrophy rather than

resolution of acute neurological effects, such as edema. This study determined that sub-

acute hippocampal volume decreases occurred approximately between 4.5 months and 24

months post-TBI in moderately to severely impaired patients. However, it remains to be

determined if these sub-acute changes occurred earlier. Recent cognitive evidence from

our lab suggests that this may be the case, and that sub-acute hippocampal atrophy may

be occurring by the first year of injury.

Also, few studies have examined correlates for sub-acute brain atrophy (Bendlin et

al., 2008; Bigler et al., 1997; Wilde et al., 2006). For example, Levine et al. (2008)

determined that GCS scores at discharge were related to amount of parenchymal volume

loss observed in TBI patients approximately one year post-injury as compared to age and

sex-matched controls, and Wilde et al. (2006) found a relationship between PTA and

abnormal VBR. However, it is unclear if these factors will also be related to sub-acute

hippocampal atrophy following moderate to severe TBI.

2.6 Current Study

Although Ng et al. (2008) determined that sub-acute hippocampal atrophy occurred

following TBI, to date, there has yet to be a longitudinal study in humans that has

examined sub-acute hippocampal atrophy in the first year of injury. Thus, the aim of the

current study was to quantify hippocampal volume at two sub-acute time points in the

29

first year of injury, using structural MR-derived measures of volumetry, in a sample of

moderately to severely impaired TBI patients.

Furthermore, previous research has determined that some demographic variables are

related to functional and cognitive outcome, while clinical measures of severity, such as

GCS scores and PTA scores have been related to cognitive, functional and

pathophysiological outcome from TBI. Nonetheless, the relationship between injury

severity, demographic variables and sub-acute hippocampal atrophy following moderate

to severe TBI has yet to be examined. Therefore, a secondary, exploratory aim of the

current study was to determine if any of these variables correlate with the amount of sub-

acute hippocampal atrophy observed.

2.6.1 Objectives/Hypotheses

Objective 1: To determine if sub-acute hippocampal atrophy occurs by the first year in

moderate and severe TBI patients. This objective was addressed using MR-derived

hippocampal volumetric measurements to quantify left and right hippocampal volumes at

5 and 12 months post-injury. Volume change in patients was examined longitudinally and

was also compared to a convenience control sample tested in the current study and to

normative data from the literature.

Hypothesis 1: Based on the previous neuroimaging findings that sub-acute hippocampal

atrophy occurred between 4.5 and 24 months post-injury (Ng et al., 2008), and the

preliminary cognitive findings from our lab indicating that hippocampal atrophy may be

underlying poor memory recovery from 4.5 to 12 months post moderate to severe TBI, it

30

was hypothesized that significant sub-acute hippocampal volume loss would be observed

from 5 to 12 months post-injury.

Objective 2 (Exploratory): To determine the impact of injury severity, years of

education and age on sub-acute hippocampal atrophy. This was achieved by correlating

GCS, age and years of education with percent hippocampal volume change from 5 to 12

months post-injury for each hippocampus. This secondary objective was exploratory and

its main purpose was to generate hypotheses for future research. As previous research has

yet to specifically examine the relationship between these variables and sub-acute

hippocampal volume atrophy, a specific hypothesis was not tested. The main purpose of

this objective was to generate hypotheses for future research.

31

Chapter 3 Methods

3.1 Participants

3.1.1 TBI Participants

Ten participants with TBI were recruited from a larger, ongoing prospective study

examining cognitive and motor recovery following moderate-severe TBI. The Research

Ethics Board at the Toronto Rehabilitation Institute (TRI) approved this study. All patient

participants were recruited from the in-patient program at TRI.

3.1.1.1 Inclusion/Exclusion Criteria

Inclusion criteria for the larger recovery study are are follows: 1) acute care diagnosis

of TBI, 2) PTA of one hour or more and/or GCS score of 12 or less determined either at

Emergency or at the scene of the trauma and/or positive CT or MRI findings in acute

care, 3) mechanism of injury associated with trauma, 4) between 18 to 65 years of age, 5)

able to follow simple commands in English, and 6) able to provide informed consent to

participate in the study or availability of a legal decision maker to provide informed

consent.

Exclusion criteria for the larger study are as follows: 1) diagnosis of a disease

primarily or frequently affecting the central nervous system (CNS) (e.g., Alzheimer’s

disease, multiple sclerosis), 2) previous history of psychotic disorder, 3) TBI acquired

secondary to another brain injury, and 4) subsequent brain injury occurring over the

course of the study.

In addition to the above criteria, to be included in the current study patients had to

have acquired their first MRI already, and at the time of recruitment, they needed to be

32

within one year of injury. Thirteen patients qualified for the current study. Three were

lost to follow-up. Therefore, a retention rate of 77% was achieved, which is very high

compared to previous longitudinal studies of TBI where retention rates were under 40%

(e.g., Dikmen et al., 2003; Himanen et al., 2006).

3.1.2 Healthy Control Convenience Sample

Five healthy control participants were recruited to participate in the current study.

This convenience sample acted as a comparison group to provide evidence that any

decreases in hippocampal volume from 4.5 to 12 months post-injury observed in patients

could not be explained by normal, age-related decreases in hippocampal volume. These

individuals were either employees of TRI or were family members of employees. Control

participants were eligible for the current study if they: 1) were between the ages of 18-65

years of age, 2) had no previous history of TBI or any other disease primarily affecting

the CNS.

3.2 Materials

3.2.1 Demographic and Injury Variables

All demographic and clinical variables were obtained retrospectively from medical

records. Demographic variables acquired were: age and years of education. Injury

severity was measured by the GCS.

33

3.2.2 Structural MRI Protocol

All patient MR scans were acquired on a GE Signa-Echospeed 1.5 Tesla scanner,

located at Toronto General Hospital, part of the University Health Network (UHN), using

a standard quadrature head coil. The high-resolution 1mm isotropic T1 weighted, three-

dimensional radio-frequency spoiled-gradient recalled-echo (SPGR) images were

acquired in the axial plane (TR = 11.74 ms, TE = 5.14 ms, and flip angle = 20°, 160

slices). The scans of two of the five control participants were acquired on the same

platform as above. Three of the five were acquired on a 3.0 Tesla scanner at the Toronto

Western Hospital, also part of UHN, using a GE Signa-EXCITE 3.0 Tesla scanner (TR =

25 ms, TE = 5ms, and flip angle = 45°, 166 slices, 8-channel headcoil). Higher resolution

of the 3.0 Tesla scanner would allow for increased detection of any decline in the

controls; therefore this difference between the patients was conservative, biasing the

hypothesis against rather than in favour. All participants underwent MRI scans of the

entire head.

3.3 Image Processing and Analysis for Hippocampal Volumes

3.3.1 Pre-Processing Steps

MRI scans were obtained in Digital Imaging and Communications in Medicine

(DICOM) format, the most common way to receive scans from hospitals, and

transformed to Medical Imaging Network Common Data Form (MINC). MINC,

developed by Neelin (1993) of the McConnell Brain Imaging Centre of the Montreal

Neurological Institute at McGill University, allows for the storage and manipulation of

medical images that can be used across a variety of computer operating systems

34

(www.bic.mni.mcgill.ca/software/minc/minc2_uguide/node2.html). Once subject scans

were converted to MINC format, they were anonymized and processed.

3.3.2 Processing Steps

Processing involved the submission of the MINC files to the TBI pipeline, an

automated file processing command that converts a T1-weighted image from its native

state to a state that is ready for analysis. A number of steps are involved in this process.

First, there was a non-uniformity correction that corrected for any inhomogeneities of the

images that occurred as a result of a non-uniform signal (Sled et al., 1998). Next, the

corrected image was registered to a Montreal Neurological Institute (MNI)-Talairach

space using a linear stereotaxic transformation (Collins et al., 1994) and resampling onto

a 1mm voxel grid using a linear interpolation kernel (Mazziotta et al., 1995). The

traditional Talairach atlas (Talairach and Tournoux, 1988) is a coordinate system used to

describe the location of brain structures regardless of individual differences in size and

shape. The MNI group created a new, more representative brain atlas derived from the

traditional Talairach model. The new template was created by averaging 305 MRI scans

of normal, right-handed individuals. It is this average 305-brain template that was used to

register each image to the same space for comparison.

After the images were registered to the MNI-Talairach space, each voxel comprising

the image slices was classified into one of three tissue types: gray matter (GM), white

matter (WM) or cerebrospinal fluid (CSF). As T1-weighted images were acquired in the

current study, GM appeared darker than WM, and CSF was the darkest of the three tissue

types.

35

Next, the tissue-classified image was used to create a 3-dimensional (3D)

reconstruction of the cortical surface. The skull and scalp surrounding the cortical surface

were removed using this 3D surface reconstruction map as a mask.

3.3.3 Quality Control

After the T1-weighted scans were processed via the TBI pipeline, a series of quality

control steps were taken to visually inspect the results of each stage of processing. This

included (1) giving one of five quality ratings ranging from excellent (e.g., perfect

alignment of lobes; no over/under-scaling; proper rotation; proper translation) to terrible

(e.g., complete failure of alignment; brain is distorted, stretched or rotated in abnormal

ways) (adapted from the Centre for Addiction and Mental Health instructional laboratory

document, “Visual Quality Control Guide” (Richards, 2005), (2) flagging problematic

scans, and (3) determining if dura mater and scalp were present. Quality control steps

were required since errors could have been made during the processing stage that

rendered the scans unusable without manual alteration of the problematic areas. For

example, poor or terrible ratings on registration could cause structures to appear much

larger than they actually were, making volumetric interpretations of these structures

inaccurate. Therefore, if any of the scans received poor or terrible ratings and could not

be manually corrected, they were not included in the current study. One scan was

excluded for this reason.

36

3.3.4 Manual Segmentation of the Hippocampi

Once quality control was completed and the scans were judged to be useable, the

program DISPLAY was used to manually segment the hippocampi of each subject.

DISPLAY, an interactive software program developed at the Brain Imaging Centre at the

Montreal Neurological Institute, allowed for the simultaneous viewing of scan slices in

three orientations: coronal, sagittal, and horizontal. The majority of the manual

segmentations were completed in coronal view with the exception of the most anterior

portion or head of the hippocampus; this section was completed predominantly in sagittal

view. Hippocampi labels were created by manually tracing and filling in the voxels that

comprised the structure of interest. As each voxel comprised a space of 1mm3, DISPLAY

automatically calculated the volume of the structure being segmented by summing the

number of voxels that were labelled.

The protocol used to determine the boundaries of the hippocampus for segementation

was developed by Preussner et al. (2000). Segmentation of each hippocampus began at

the most posterior aspect of the hippocampus, where the tail bulges slightly into the

trigone of the lateral ventricle. The most medial border of the trigone of the lateral

ventricle was used as the medial border for the tail of the hippocampus. The superior

border of the tail was marked by the superior row of GM voxels (see Figure 3.1), while

one or two rows of voxels were left unlabelled at the lateral border to ensure that the tail

of the caudate nucleus was excluded. As segmentation moved anteriorly through the

brain, the tail of the hippocampus descended and changed into the body. When the body

of the hippocampus was viewed in coronal section, the fornix, the major efferent pathway

from the hippocampus could be viewed, but was not included. A portion of WM called

the fimbria was included. The fimbria is the part of the fornix that is embedded in the GM

37

of the hippocampus. To ensure that only the fimbria and not fornix was included, only

voxels that were surrounded by GM, and thus “embedded” were incorporated into the

hippocampal volume.

Figure 3.1 Tail of Left Hippocampus in Coronal View

Tail of left hippocampus circled in coronal view. Rectangle identifies most superior row of GM voxels.

Additionally, the inferior and lateral borders for labelling the hippocampal

body remained the same as for the tail. However, the medial border was created by

extending the WM below the hippocampus superior-medially at a 45 degree angle to

ensure that entorhinal cortex, located below the hippocampus, was not included

(Figure 3.2). Also, a “safety” row of voxels was left unlabelled laterally and

superiorly to ensure that the choroid plexus of the inferior horn of the lateral ventricle

was not included as hippocampus.

38

Figure 3.2 Body of Left Hippocampus in Coronal View

45-degree angle used to define medial border of hippocampus

Lastly, in the coronal view, the body of the hippocampus transitioned into the head of

with the appearance of the gyrus intralimbicus (posterior border of entorhinal cortex),

which appeared as a medial bulge. At this point, the sagittal view was also used to

identify hippocampal borders (Figure 3.3). Similar boundaries were used for this section

as were used for labelling the body, however, due to the subtle transition of the

hippocampal head to the amygdala, the sagittal view helped with identifying a thin WM

line called the alveus. The alveus separated the head of the hippocampus from the

amygdala and was included in the volume. It should be noted that for the TBI

participants, these borders were not always evident due to atrophy. In these cases, borders

that were visible were used to guide the labelling of the unclear border. For example, if

the alveus surrounding the head of the hippocampus was not visible, then the most

inferior aspect of the lateral ventricle was used to determine approximately where the

head of the hippocampus ended.

39

Figure 3.3 Left Hippocampus in Sagittal View

The alveus, which separates the head of the hippocampus from the amygdala is circled.

3.4 Inter- and Intra- Rater Reliability Assessment

As the hippocampus is a highly variable structure in both normal and TBI

populations, a series of reliability correlations were conducted on practice TBI scans prior

to working on patient scans from the current study. Reliability refers to the degree to

which multiple assessments of a subject agree or are reproducible (Bartko, 1991). The

intraclass correlation coefficient (ICC) can be used to assess both inter- and intra-rater

reliability. Also, it has been shown to be a useful tool for evaluating continuous data

(Bartko, 1966) such as hippocampal volumes. The aim of the ICC is to determine the

amount of measurement error due to subjective interpretation or human error (Shrout and

Fleiss, 1979). A high ICC for inter-rater reliability suggests that two or more raters, in

this case those conducting the hippocampal labelling, achieved similar results (i.e., little

variability) with regard to hippocampal volume acquired. On the other hand, a low ICC

on inter-rater reliability suggests that the raters achieved different results due to human

error, which could affect statistical analyses and findings. (Shrout and Fleiss, 1979).

40

Similarly, the ICC can be used to assess the amount of error produced by an

individual rater between rounds (intra-rater reliability). A high ICC on intra-rater

reliability suggests that the rater is consistent in reproducing similar hippocampal

volumes between rounds, whereas low intra-rater reliability suggests that individual rater

results between rounds were not highly reproducible.

For the current study, a minimum of five rounds of rater reliability were conducted at

least two weeks apart. Inter-rater reliability was conducted against an experienced

quantitative MRI analyst trained on the Preussner (2000) method. This individual also

trained the rater of the current study. The results of these analyses indicated that for left

side, an “almost perfect” agreement was reached (κ= .871), while a “substantial”

agreement was reached for the right side (κ= .801) (Landis and Koch, 1977). With regard

to intra-rater reliability, “almost perfect” agreements were reached for both the left (κ=

.934) and right (κ= .898) sides (Landis and Koch, 1977).

3.5 Design and Procedure

3.5.1 Design

The study employed a longitudinal design with a within subjects factor (time: 5 vs. 12

months post-injury) and a between subjects factor (group: patients vs. controls). The

dependent variables were hippocampal volume (objective 1), and age, years of education

and GCS score (objective 2).

41

3.5.2 Procedures

TBI patients who gave their informed consent were recruited to the larger study on

TBI recovery. As part of that study, they underwent neuropsychological and motor

testing while in-patients. All testing took place at the Toronto Rehabilitation Institute.

After discharge, they underwent two more neuropsychological and motor assessments

and two MRI scans. Structural MRI scans (as well as behavioural testing) took place at

two sub-acute time-points: 5-months post injury (mean= 5.23 months, SD= 0.53, range=

4.7-6.2 months) and 12-months post-injury (mean= 12.4 months, SD= 1.5, range= 11-

15.2 months). MRI scans were completed at the Toronto General Hospital.

Individuals in the control group who provided their informed constent also underwent

two MRI scans at the Toronto General Hospital. There were varying lengths of time in

between the two scans (mean= 24.6 ± 7 months, SD= 7.4, range= 14-32 months).

3.6 Data Analysis

3.6.1 Objective 1: Detection of Sub-Acute Hippocampal Atrophy by the First Year of Moderate and Severe TBI

In order to test the hypothesis that the TBI group would demonstrate greater

hippocampal volume loss over time than the control group, a repeated measures 2 (group)

x 2 (time) analysis of variance (ANOVA) was conducted for each hippocampus followed

by planned comparisons (time 1 vs. time 2 for each group) with effect sizes. Given the

small sample size, effect size permits a determination of the strength of these differences

or the degree to which the null hypothesis was believed to be false (Cohen, 1988) while

minimizing the impact of sample size.

42

3.6.2 Objective 2 (Exploratory): Examination of Correlates of Sub-Acute Hippocampal Atrophy

To examine the relationship between demographic and injury severity variables with

sub-acute hippocampal volume change, bivariate, zero-order Pearson correlations were

conducted for left and right percent hippocampal volume change with (1) injury severity,

as measured by GCS scores, (2) age and (3) years of education.

Percent change scores for the left and right hippocampi were separately calculated

using the following formula: [(hippocampus volume at time 1 – hippocampus volume at

time 2)/ average hippocampus volume at times 1 and 2] x 100, as recommended by

colleagues at the Brain Stimulation and Neuroimaging laboratory based at the Alfred

Psychiatric Research Centre (a Department of Monash University). Therefore, decreases

in hippocampal volume over time were captured by positive values, while negative values

represented increases in volume over time.

Estimated annual percent hippocampal volume change scores derived from a

published normative database (Bigler et al., 1997) were also utilized in order to compare

TBI patients to age-matched controls. The method used to derive these change scores has

previously been described by Ng et al. (2008). Briefly, Bigler et al. (1997) collected

normative hippocampal volumes of individuals between the ages of 16 and 65 years of

age. Average hippocampal volumes were grouped for each of the five decades spanning

this age range. Change scores were calculated by taking the mean hippocampal volume of

the age band that followed the age band of interest and subtracting the mean hippocampal

volume of the age band preceding the age band of interest. These scores were then

divided by 20 (as the mean span across these age bands was 20) and then annual

percentages were calculated from this value.

43

All statistical analyses, including those for the ICC, were conducted using the

Statistical Package for the Social Sciences, version 16.0 (SPSS 16.0) for Windows.

44

Chapter 4 Results

4.1 Demographic and Injury Characteristics

Table 1 presents the injury and demographic variables for the TBI and the control

groups. The convenience control group had more years of education, a difference which

was not statistically significant, but for which Cohen’s d was large (P< 0.152, Cohen’s

d= 0.83). Significant differences between groups were observed for age (P< .007,

Cohen’s d= 1.76), with the TBI group being older, and for interval between scan times,

which was shorter in the TBI group (P< .000, Cohen’s d= 3.22). While the older age of

the TBI sample biased the findings in favour of the hypothesis (i.e., because they were

older, they were more likely to show atrophy from 5 to 12 months post-injury), the

shorter time between scans in the TBI group biased the results against the hypothesis of

the study.

Consistent with the literature, the TBI group consisted of more males (60%) than

females (40%). This contrasted with the control group, which was composed of more

females (60%) than males (40%). These differences were not statistically significant (χ2

(1, n= 15) = 0.536, N.S.), although the result should be interpreted conservativley

because of acceptance of the null and the small sample size.

Table 1 also presents the hippocampal volumes for the groups for scan 1. There were

no statistically significant differences in hippocampal volume for the two groups, either

for the left (P> 0.1 Cohen’s d= 0.86) or the right (P> 0.1; Cohen’s d= 0.61) sides.

However, as expected, the hippocampal volumes of the TBI patients were smaller than

those of the controls at scan 1.

45

Table 4.1 Group Data for TBI and Control Participants

TBI Group (n=10) Control Group (n=5)

Age (mean ± SD)

50.2 ± 10.5 32.4 ± 9.7**

Months between Scans (mean ± SD)

7.3 ± 1.8 24.6 ± 7.4**

Sex (M/F)

6/4 2/3

Years of Education (mean years ± SD)

17.1 ± 2.1 19 ± 2.4

GCS (mean ± SD)

5.8 ± 4.0 N/A

Injury type (fall/MVA)

6/3§ N/A

Left Mean Hippocampus Volume, Time 1 (mm3)

2535.4 ± 394.4 2854 ± 346.6

Right Mean Hippocampus Volume, Time 1 (mm3)

2640.3 ± 415.2 2854.8 ± 267.2

Note: *indicates significance at the .05 level; **indicates significance at the .01 level § Data unavailable for one participant

4.2 Objective 1: Detection of Sub-Acute Atrophy in the First Year of Injury

4.2.1 Hypothesis 1: TBI patients will show significant loss of volume across time

that is not explained by age-related decline

Figure 4.1 presents the data for the right hippocampal volumes of patients and

convenience controls. A repeated measures ANOVA revealed a significant group x time

interaction (F1,13= 7.87; P< .05). There was a trend for a main effect of time (F1,13= 3.96,

P= .068), and no significant main effect of group, (F1,13= 1.80, N.S.).

46

Figure 4.1 Mean Right Hippocampal Volume for the TBI and Control Groups

2000

2200

2400

2600

2800

3000

3200

3400

TBI Control

Mea

n H

ippo

cam

pal V

olum

e (m

m3 )

Time 1

Time 2

*

The TBI group demonstrated a significant hippocampal volume decrease over time for the right side. This was not observed for the control group. Note: * indicates significance at P< .05 level.

Planned comparisons revealed that the TBI group demonstrated a significant decrease

in hippocampal volume from baseline to 1-year post-injury (P< 0.005, Cohen’s d= 0.34).

The control group demonstrated no significant decline or trend towards a decline; indeed,

the absolute mean volume was slightly higher at 1 year post-injury. Taken together, these

findings provide support for Hypothesis 1.

The purpose of the control group was to ensure that any decreases in hippocampal

volume in the TBI group could not be attributed to age-related volume loss. However, the

control group was significantly younger than the TBI group and therefore less vulnerable

to age-related decline. This might have exaggerated relative differences in volume change

between the groups. Therefore, published age-stratified normative data (from Bigler et

al., 1997) were additional used as a second basis for comparison. Here, percent annual

47

change scores for published healthy individuals matched for age bands (in decades) to

each of the patients were computed as in Ng et al. (2008). Annual percent change for the

right hippocampus was -0.16% (SD= 0.12), which was approximately 25 times less than

that of the TBI group, which was -4.09% (SD= 3.3), although these numbers should be

interpreted with caution given the large standard deviations.

Figure 4.2 presents the results of the repeated measures ANOVA for the left

hippocampal volume. A trend was observed for the group x time interaction (F1,13 = 3.72,

P= 0.076), while the main effects of group and time were not significant (F1,13= 3.11,

N.S.) and (F1,13= 0.508, N.S.), respectively.

Figure 4.2 Mean Left Hippocampal Volume for the TBI and Control Groups.

2000

2200

2400

2600

2800

3000

3200

3400

TBI Control

Mea

n H

ippo

cam

pal V

olum

e (m

m3 ) Time 1

Time 2

*

The TBI group demonstrated a significant hippocampal volume decrease over time for the left side. This was not observed for the control group. Note: * indicates significance at P< .05 level.

Although the interaction for the left hippocampus ANOVA showed only a trend

towards significance, planned comparisons were nonetheless undertaken because of

48

the associated directional hypothesis. Mean hippocampal volume for the TBI group

was significantly smaller at one-year post-injury compared to baseline (P< 0.05,

Cohen’s d= 0.22). For the control group, there was again no significant difference or

trend, and the absolute volume showed a slight increase from baseline to follow-up.

Therefore, in support of Hypothesis 1, the left hippocampus demonstrated sub-acute

atrophy by the first year of injury and this decrease in volume was not observed for

the control group.

Again, published normative data from Bigler et al. (1997) were used to compute

annual percent hippocampal volume changes for the left side. The TBI group

demonstrated sub-acute annual percent hippocampal atrophy that was over 7 times

that of age-matched controls: -2.59 (SD= 4.4) vs. -0.34% (SD= 0.15).

Although significance testing revealed significance, analytic stability is limited in

a small sample size. Moreover, Figures 4.1 and 4.2 reveal substantial variability in

volume change over time. Therefore, to visually inspect the reliability of these

findings, percent change is plotted for each patient individually for both hippocampi

in Figure 4.3. As can be seen from the figure, 9/10 patients showed some degree of

volume loss in both hippocampi. One patient showed volume loss on only one side,

and one patient showed increased volume for both hippocampi.

It is of interest to note that subject 7, who demonstrated the largest overall

hippocampal volume loss bilaterally (left hippocampal change= 11.93%, right

hippocampal change= 7.58%) and subject 2, who demonstrated the most favourable

outcome over time cannot be distuinguished demographically from one another. Both

were female, older and with equal years of education. Only injury severity and injury

49

mechanism differed: patient 7 had sustained a more severe injury, and an MVA as

opposed to a fall.

Figure 4.4 presents percent change for each control participant individually for

each hippocampus. None of the participants showed absolute volume loss bilaterally.

Only two demonstrated a unilateral decrease.

Figure 4.3 Left and Right Percent Hippocampal Volume Change for Individual TBI Participants.

-14-12

-10-8-6-4

-202

46

1 2 3 4 5 6 7 8 9 10

TBI Participants

Per

cen

t H

ipp

oca

mp

al V

olu

me

Ch

ang

e

Left % Decrease

Right % Decrease

Note: Volume decreases are represented by negative bars and increases are represented by positive bars for this figure. Participants 1, 3, 5, 7, 8, 9 and 10 demonstrated bilateral hippocampal volume decreases. Participants 4 and 6 demonstrated unilateral hippocampal volume decrease and participant 2 was the only one to demonstrate no hippocampal volume decrease.

50

Figure 4.4 Left and right percent hippocampal volume changes for individual control participants

-10

-8

-6

-4

-2

0

2

4

6

1 2 3 4 5

Control ParicipantsPer

cen

t H

ipp

oca

mp

al V

olu

me

Ch

ang

e

Left % ChangeRight % Change

Note: Volume decreases are represented by negative bars and increases are represented by positive bars for this figure. Only participants 4 and 5 demonstrated unilateral hippocampal volume decrease; none demonstrated bilateral volume decreases.

4.2.2 Objective 2 (Exploratory): Preliminary Examination of Correlates of Sub-Acute Hippocampal Atrophy

Figure 4.5 presents the findings of the bivariate correlation between GCS and percent

volume change for the right hippocampus. A significant negative correlation was

observed for GCS scores and right percent hippocampal change (Pearson r= -0.663, P<

0.05). Thus, greater injury severity, as indexed by lower GCS scores, was related to larger

amounts of right sub-acute hippocampal volume decrease. This relationship was not

significant for the left side, however (Pearson r= -0.327, N.S.).

51

Figure 4.5 Correlation between Right Percent Hippocampal Volume Change and GCS Score

-4

-2

0

2

4

6

8

10

0 2 4 6 8 10 12 14 16

GCS Score

Rig

ht

Per

cen

t H

ipp

oca

mp

al

Vo

lum

e C

han

ge

A significant negative correlation was obtained for GCS and right percent hippocampal volume change (Pearson r= -0.663, P< 0.05).

With regard to demographic variables, there were no significant correlations. Age did

not significantly correlate with either the left (r= 0.065, N.S.) or right (r= 0.023, N.S.)

percent hippocampal volume change scores nor did years of education for either the left

(r= 0.436, N.S.) or right (r= -0.17, N.S.).

52

Chapter 5 Discussion

5.1 Summary of key findings

The current study was designed to replicate and extend those of Ng et al. (2008) who

had found sub-acute hippocampal atrophy in moderate to severe TBI patients from 4.5

months to 24 months post-injury. No conclusions from that study could be drawn as to

whether the atrophy occurred earlier than 24 months, and preliminary cognitive findings

from our laboratory had suggested that the atrophy might be occurring earlier, as greater

atrophy from 4.5 months to 24 months post-injury correlated with poorer memory

recovery from 5 to 12 months post-TBI in an overlapping sample of patients.

Thus, the primary objective of the current study was to determine whether a group of

patients with moderate and severe TBI would demonstrate significant sub-acute

hippocampal atrophy from 5 to 12 months post-injury. The approach was to examine

atrophy within the TBI group and to compare the magnitude of atrophy to a convenience

control sample. A secondary, exploratory objective was to examine the correlates of

atrophy, examining injury severity (as measured by the GCS) and two demographic

variables (age and years of education).

Consistent with the hypothesis, TBI patients demonstrated significant left and right

sub-acute hippocampal volume loss within the first year of injury. These changes were

not observed in the control convenience sample. The findings were robust, with 7 of 10

patients showing bilateral change in absolute volume in the direction of atrophy. In

contrast, none of the controls showed bilateral absolute change in that direction. Less than

half showed even unilateral change for the worse.

53

Severity of injury, as measured by GCS, was related to the amount of sub-acute

hippocampal volume change observed for the right, but not the left hippocampus. There

were no significant associations between amount of hippocampal volume loss and either

age or years of education.

5.1.1 Sub-Acute Atrophy – Relationship to Prior Findings; Mechanisms

The current findings of sub-acute hippocampal atrophy corroborate findings in the

animal literature (Smith et al., 1997) and human literature (Ng et al., 2008). Recent

research has provided insight as to why these sub-acute decreases in hippocampal volume

may be occurring during a time when the brain is presumed to be stable (Blatter et al.,

1997; Christodoulou et al., 2001; Farne et al., 2004). Although the hippocampus is a deep

gray matter structure, it has many white matter projections that act to bring information to

and from this memory structure. Several studies have demonstrated that delayed white

matter pathology can occur following TBI (Buki and Povlishock, 2006; Greenberg et al.,

2008; Pettus et al., 1994; Rodriguez-Paez et al., 2005; Smith et al., 2003). Thus, it is

possible that some of this delayed white matter damage could underlie sub-acute gray

matter atrophy, particularly in the hippocampus. For example, destruction of axonal input

to the hippocampus from functionally related but distant areas of the brain may result in

decreased metabolism and regional cerebral blood flow to this structure (Kwakkel et al.,

2004). In the early 20th century, von Monakow developed this secondary brain lesion

effect theory, termed diaschisis (Finger et al., 2004); and today it is known as a potential

underlying factor contributing to tissue loss in areas with suppressed metabolism, blood

flow, and neural activity caused by the deafferentation of axons from distant structures.

Contrary to the original notion that diaschisis is usually reversible in early stages post-

54

injury (Finger et al., 2004; Kwakkel et al., 2004), recent case study evidence has

suggested that this may not always be the case, and individuals may present with

irreversible diaschisis and associated degeneration depending on the nature of the primary

injury (Baheti et al., 2009).

Similar to diaschisis, hippocampal atrophy can cause white matter destruction to

efferent fibres resulting in anterograde degeneration to the structures to which it projects

(e.g., the mammillary bodies) (Pierpaoli et al., 2001). This process is called Wallerian

degeneration (e.g., Stoll and Muller, 1999; Bigler, 2007). In animal models of TBI,

destruction to efferent hippocampal pathways has been observed (Hall et al., 2008) as

well as afferent loss. In the latter case, decreased neural activity and accompanying

atrophy of the structure being projected to may lead to reciprocal atrophy or dysfunction

of the hippocampus. Over time, these processes occurring at the cellular level can lead to

gross morphological changes. One might speculate that these sub-acute white matter

changes may be underlying some of the sub-acute hippocampal atrophy observed

following moderate and severe TBI.

Sub-acute hippocampal loss in the first year of injury was particularly evident for the

right hippocampus. It has been well demonstrated that there is functional specialization of

the hippocampus (Lye et al., 2004; Sass et al., 1990). Specifically, the left hippocampus

has been well associated with verbal memory (Sass et al., 1990; Sass et al., 1992), while

the right hippocampus has been associated with visual-spatial memory (Gleissner et al.,

1998). However, Lye et al. (2004) also demonstrated that there was no evidence of a clear

laterality effect for visual reproduction, a non-verbal memory task. These authors also

noted that individuals may have been using verbal encoding to help them remember the

items on this task. Therefore, it is possible that individuals have a tendency to rely on left

55

hippocampal function (due to verbal strategy use) more so than the right hippocampus.

Also, in a recent study, Leshikar et al. (2009) used functional MRI on individuals while

they performed an encoding task that required them to associate items that were either

semantically related (easy) or unrelated (difficult). It was determined that as task

difficulty increased, there was greater left hippocampal activation. Thus, as cognitive

demand increased, the left hippocampus was used more. This is important considering

TBI patients commonly have early memory impairments as a result of the initial acute

atrophy (Borgaro and Prigatano, 2002; Ferri-Campos et al., 2008). As a result they have

to use more mental effort (Kohl et al., 2009; Zasler et al., 2007) and one may postulate

that they are more actively engaging the left hippocampus as a result. Work by Verghese

et al. (2003) has shown that increased participation in cognitive activities is actually

protective against memory decline. This theory has colloquially been termed, “use it or

lose it.” Therefore, if TBI patients are using greater effort to remember things, this could

cause greater activation of the left hippocampus and thus serve as a protective mechanism

of left sub-acute hippocampal atrophy. Contrastingly, the right hippocampus could get

used less and thus be more susceptible to sub-acute atrophy.

While the current study was able to reveal that sub-acute atrophy occurs by 12

months post-injury, the findings cannot tell us whether there is continued sub-acute

atrophy, beyond the first year. Thus, it is unclear whether the observed sub-acute atrophy

is best characterized as a single, second wave of atrophy or more insidiously, as an

ongoing, neurodegenerative process. In order to determine this, a within subjects,

longitudinal study examining hippocampal volume at three time-points would be an

optimal design.

56

As a coarse comparison made for the purpose of generating hypotheses for future

research, percent volume loss in the current study was compared to percent volume loss

from 4.5 to 24 months post-injury in the Ng et al (2008) study. In order to control for

demographic variance across patients, demographically-matched groups were created.

For the left hippocampus, 10 patients (6 male, 4 female) with a mean age of 50.2 years

(SD= 10.5) and a mean GCS of 5.8 (SD= 4.0) showed a 2.6% (SD= 4.4) volume decrease

from 5 to 12 months post-injury. The comparison group from the Ng et al study (2008),

(n= 11; 8 male, 3 female) with a mean age of 40.5 years (SD= 10.7) and a mean GCS of

8.5 (SD= 4.5), showed a 6.3% (SD= 4.5) volume decrease. For the right hippocampus, 10

patients (6 male; 4 female; mean age = 50.2 years, SD= 10.5) showed a 4.1% (SD= 3.3)

volume decrease from 5 to 12 months post-injury while the comparison group from the

Ng et al study (2008), (n = 11; 8 male, 3 female; mean age of 40.5 years, SD= 10.7)

showed a 5.5% volume decrease. These descriptive statistics are compatible with the

notion of continued atrophy for the left, but not the right hippocampus. Thus for the left

side, findings are compatible with progressive neurodegeneration and for the right side,

the findings are more compatible with a second wave of atrophy that completes by

approximately the first year post-injury. It is not parsimonious to propose two different

mechanisms of change for the two hippocampi, though, and the next phase of the current

study will include a third MRI to empirically assess this important question using a

repeated measures design.

5.1.2 Correlates of Sub-Acute Atrophy

In general, the literature has found that age, initial GCS score, duration of PTA and

neuroimaging findings are correlated with functional and behavioural outcome from TBI

57

(Green et al., 2008; Levine et al., 2008; Wilde et al., 2006; Zasler, Katz and Zafonte,

2007). In the current study, GCS was significantly correlated with percent hippocampal

volume change for the right side, only: lower GCS scores (greater severity) were related

to greater right sub-acute hippocampal atrophy. This correlation was not significant for

left percent hippocampal volume change, however. These findings are partially in

accordance with the cross-sectional study by Levine et al. (2008) where GCS scores at

discharge from the trauma unit were positively related to parenchymal volume at one-

year post-injury. However, Levine et al did not examine sub-acute atrophy, they looked at

one time point only.

In contrast to the studies that have determined that demographic characteristics such

as age and years of education are related to outcome from TBI (Green et al., 2008;

Williamse-van Son et al., 2007), the current study did not find a relationship between

these variables and sub-acute hippocampal atrophy in the first year of injury. One

potential explanation for this observation is that hippocampi are highly vulnerable to sub-

acute atrophy regardless of age or years of education. Thus, the demographic and injury

variables examined may be more related to acute rather than sub-acute hippocampal

atrophy. The literature has demonstrated hippocampal vulnerability and associated

memory impairment under numerous conditions other than TBI including normal aging,

hypoxia, prolonged corticosterone administration, transient global ischemia, affective

disorders, Alzheimer’s disease and post-traumatic stress disorder (Gozal et al., 2001;

Hoschl and Hajek, 2001; Kadar et al., 1998; Ouyang et al., 2007). In fact, animal research

has determined that certain areas of the hippocampus, namely the CA1 and CA3

pyramidal cell layers, are linked to cognitive dysfunction and degeneration (Kadar et al.,

1998; Kotapka et al., 1991; Royo et al., 2007; Witgen et al., 2005). Other research has

58

determined that delayed pro-apoptotic genes did not increase in expression at 3, 6 or 12

months post-TBI indicating that apoptotic mechanisms are likely not underlying these

sub-acute changes to the hippocampus (Shimamura et al., 2005). Thus, secondary

mechanisms involving extra glutamate and calcium involvement may be underlying these

sub-acute changes, and are perhaps working in combination with other processes post-

TBI, such as decreased neurogenesis in the dentate gyrus and decreased levels of

neuroprotective genes (Shimamura et al., 2005; Yoshimura et al., 2003).

Another possible explanation for the lack of findings on significance testing is simply

the lack of power in the small sample size of the current study.

5.2 Implications

The findings from the current study are clinically important. Although one recent

study (Ng et al., 2008) determined that sub-acute hippocampal atrophy occurred from 4.5

months to 24- months post-moderate and severe TBI in a within-subjects design, the

time-course for these changes had not been determined. The current study replicated the

findings of Ng et al. (2008) by determining that sub-acute hippocampal atrophy occurs

following moderate and severe TBI and moreover, it provided evidence that atrophy

occurs by the first year of injury.

These results are highly important because many studies have provided evidence that

behavioural and functional recovery is occuring until at least 6 to 12 months post-injury

(Choi et al., 1994; Dikmen et al., 1990). Such behavioural improvements may be masking

this underlying atrophy, which may in turn be mitigating recovery. Having a better

59

understanding of whether and when sub-acute hippocampal atrophy occurs is important

for identifying windows of opportunity for treatment and rehabilitation purposes.

For example, hippocampal atrophy that occurs during a period when recovery is

occurring elsewhere in the brain may be offset by more intensive rehabilitation, which

continues throughout the first year of injury. This contrasts the focus of rehabilitation in

the 1990’s, which was aimed towards efficiency, reduced costs and shorter rehabilitation

periods (High, Sander and Struchen, 2005). Intensive rehabilitation may work to restore

connections in the brain that were damaged from the TBI. For example, there is evidence

for the effectiveness of constrained induced movement therapy (CIMT) for reduced upper

limb function following stroke (French et al., 2007; Hakkennes and Keating, 2005).

CIMT involves the restraint of the intact limb over an extended period of time, in

combination with several repetitions of task-specific training trials using the affected

limb. French et al. (2007) described that this intervention may work because the forced

use of the affected limb forms the basis of motor learning; sensorimotor coupling

contributes to the recovery or formation of neuronal pathways. Similarly, intensive

rehabilitation may work to reform or create neural pathways that were damaged by the

trauma or to prevent further deterioration of damaged pathways. For example, some

strategies may be to include verbal elaboration training or using repetitive memory drills

(Mateer et al., 1996).

Also, because sub-acute hippocampal atrophy may not be a universal phenomenon or

it may be greater in some moderate or severe TBI patients over others, it is important to

have an understanding of which factors put some individuals at a greater risk of sub-acute

hippocampal atrophy over others. In the current study, only GCS was related to sub-acute

hippocampal atrophy in the first year of injury. Thus, when injury severity, as measured

60

by the GCS is known, this may provide the patient, their families or caregivers and

rehabilitation therapists an idea of what to expect over the course of the first year of

injury.

5.3 Limitations

There were several limitations in the current study that could have affected the

results. The first limitation was that there was a small sample size of 10 TBI patients and

5 control participants. Although the repeated measures design of this study helps to

reduce within subject error, the sample size nonetheless limits both the stability of effects

and the type of analyses that could be conducted.

Also, the control group was a convenience sample and the two groups were not

matched for age and length of time between scans. Although the control group had a

longer length of time between scans, which biased the results against our hypothesis, they

were significantly younger than the TBI group. Therefore, the TBI group may have been

more vulnerable to age-related decline, thereby providing an alternative explanation for

the findings. Thus, the control group being significantly younger than the TBI group

biased the results in favour of our hypothesis.

Furthermore, all TBI patients were recruited from a single, specialized rehabilitation

hospital located in an urban area. As a result, the TBI sample in the current study may not

be representative of all individuals with moderate and severe TBI. The generalizability of

the current study may be limited and the results may not be applicable to others,

particularly those that do not have access to urban rehabilitation facilities.

61

There were some additional limitations that were innate to the materials used in the

study. Clinical data surrounding injury severity and pre-morbid characteristics were

acquired via a retrospective review of each patient’s medical record. In some cases,

patients had more than one GCS score reported (e.g., one at the scene of injury and one at

Emergency). When this was the case, the lowest reported GCS score was utilized in the

current study. GCS scores were selected over PTA scores to examine the relationship

between injury severity and sub-acute hippocampal volume decrease because all patient

data were available for this measure unlike PTA.

5.4 Future Direction for Research

The current study has provided preliminary findings that suggest sub-acute

hippocampal atrophy occurs in the first year of injury. Future studies should corroborate

these findings on a larger scale by incorporating multiple sites and a larger sample size so

that findings can be generalized to all moderate and severe TBI patients. Also, results

should be compared to a tightly matched control group composed of healthy individuals.

Future research should also address other measures of severity including PTA and

ACLOS and their relationship with sub-acute hippocampal atrophy. These measures may

be more sensitive in predicting the degree of sub-acute hippocampal atrophy from TBI.

The mechanisms underlying sub-acute hippocampal atrophy should also be a focus of

future research. For example, Colicos et al. (1996) noted that neuronal loss following TBI

can be broadly divided into three categories: (1) initial cell death from the physical

trauma, (2) necrotic cell death from excitotoxic mechanisms and (3) delayed cell death.

62

Having a better understanding of this delayed cell death may provide insight into the

outcome from TBI and may be the focus of future intervention.

Additionally, because TBI typically affects several other cognitive, behavioural and

emotional domains, future research should focus on the development of rehabilitative

methods that may offset further cognitive decline in these patients, while still being

tailored enough to fit individual patient needs. Future work in understanding secondary

mechanisms underlying delayed cell death may be useful for the development of

pharmacological intervention that can be used in combination with rehabilitation

methods.

Lastly, a future follow-up study assessing the patients in the current study at a third

time point (approximately 24 months post-injury) should be conducted. This study would

allow for hippocampal volume changes from 5 to 12 months to be directly compared to

hippocampal volume changes from 12 to 24 months, while controlling for within subject

variability. This design would help provide further insight into the time course for sub-

acute hippocampal atrophy post moderate to severe TBI.

5.5 Conclusions

In summary the results of the current study suggest that individuals with moderate

and severe TBI demonstrate sub-acute hippocampal atrophy in the first year of injury,

which is particularly evident for the right hippocampus. An additional finding is that

severity of injury, as measured by GCS, may be useful for predicting the degree of sub-

acute hippocampal atrophy an individual may experience.

63

As TBI is one of the leading causes of morbidity and mortality around the world,

future research needs to address the time course of sub-acute hippocampal atrophy

following moderate and severe injury in order to gain a better understanding of the

mechanisms underlying this outcome. A better understanding can aid in the development

of precise, mechanism-based interventions aimed at optimizing the recovery of moderate

to severe TBI patients. Furthermore, determining which clinical and demographic

variables moderate sub-acute atrophy following TBI will increase our understanding of

which patients are at a greater risk of poor pathophysiological outcome. This

understanding is important for the patient, his/her family and/or caregivers and clinicians

to know for the purposes of long-term therapeutic, family and occupational planning.

64

References Consensus conference. Rehabilitation of persons with traumatic brain injury. NIH Consensus Development Panel on Rehabilitation of Persons With Traumatic Brain Injury (1999). JAMA, 282(10), 974-983. Adams, J. H., D. I. Graham, et al. (1991). "Diffuse axonal injury in non-missile head injury." J Neurol Neurosurg Psychiatry 54(6): 481-3. Adams, J. H., D. I. Graham, et al. (1982). "Diffuse axonal injury due to nonmissile head injury in humans: an analysis of 45 cases." Ann Neurol 12(6): 557-63. Arcia, E., & Gualtieri, C.T. (1993). Association between patient report of symptoms after mild head injury and neurobehavioural performance. Brain Injury, 7(6), 481-489. Ariza, M., J. M. Serra-Grabulosa, et al. (2006). "Hippocampal head atrophy after traumatic brain injury." Neuropsychologia 44(10): 1956-61. Auerbach, S. H. (1986). "Neuroanatomical correlates of attention and memory disorders in traumatic brain injury: an application of neurobehavioural subtypes." Journal of Head Trauma Rehabilitation 1(3): 1-12. Baddeley, A.D., Vallar, G., & Wilson, B. (1987). Sentence comprehension and phonological memory: some neuropsychological evidence. Baheti, N. N., A. R. Bansal, et al. (2009). "Teaching NeuroImages: diaschisis: is it always reversible?" Neurology 72(16): e79. Banich, M.T. (2004). Cognitive neuroscience and neuropsychology, 2nd edition. New York: Houghton Mifflin Company. Bartko, J. J. (1966). "The intraclass correlation coefficient as a measure of reliability." Psychol Rep 19(1): 3-11. Bartko, J. J. (1991). "Measurement and reliability: statistical thinking considerations." Schizophr Bull 17(3): 483-9. Barzo, P., A. Marmarou, et al. (1997). "MRI diffusion-weighted spectroscopy of reversible and irreversible ischemic injury following closed head injury." Acta Neurochir Suppl 70: 115-8. Beauchamp, K., H. Mutlak, et al. (2008). "Pharmacology of traumatic brain injury: where is the "golden bullet"?" Mol Med 14(11-12): 731-40.

65

Bendlin, B. B., M. L. Ries, et al. (2008). "Longitudinal changes in patients with traumatic brain injury assessed with diffusion-tensor and volumetric imaging." Neuroimage 42(2): 503-14. Bergsneider, M., D. A. Hovda, et al. (2000). "Dissociation of cerebral glucose metabolism and level of consciousness during the period of metabolic depression following human traumatic brain injury." J Neurotrauma 17(5): 389-401. Besenski, N. (2002). "Traumatic injuries: imaging of head injuries." Eur Radiol 12(6): 1237-52. Bigler, E. D. (1996). "Brain imaging and behavioral outcome in traumatic brain injury." J Learn Disabil 29(5): 515-30. Bigler, E. D. (2007). "Anterior and middle cranial fossa in traumatic brain injury: relevant neuroanatomy and neuropathology in the study of neuropsychological outcome." Neuropsychology 21(5): 515-31. Bigler, E. D., C. V. Anderson, et al. (2002). "Temporal lobe morphology in normal aging and traumatic brain injury." AJNR Am J Neuroradiol 23(2): 255-66. Bigler, E. D., D. D. Blatter, et al. (1997). "Hippocampal volume in normal aging and traumatic brain injury." AJNR Am J Neuroradiol 18(1): 11-23. Bigler, E. D., S. M. Kurth, et al. (1992). "Degenerative changes in traumatic brain injury: post-injury magnetic resonance identified ventricular expansion compared to pre-injury levels." Brain Res Bull 28(4): 651-3. Bigler, E. D., D. K. Ryser, et al. (2006). "Day-of-injury computerized tomography, rehabilitation status, and development of cerebral atrophy in persons with traumatic brain injury." Am J Phys Med Rehabil 85(10): 793-806. Blatter, D. D., E. D. Bigler, et al. (1995). "Quantitative volumetric analysis of brain MR: normative database spanning 5 decades of life." AJNR Am J Neuroradiol 16(2): 241-51. Bouilleret, V., Cardamone, L., Liu, Y.R., Fang, K., Myers, D.E., & O’Brien, T.J. (2009). Progressive brain changes on serial manganese-enhanced MRI following traumatic brain injury in the rat. J Neurotrauma, 45(2), 267-274. Bramlett, H.M., Green, E.J., Dietrich, W.D. (1999). Exacerbation of cortical and hippocampal CA1 damage due to posttraumatic hypoxia following moderate fluid-percussion brain injury in rats. Journal of Neurosurgery, 91(4), 653-659. Brooker, A.E., & George, J.C. (1984). Visual recognition memory of severely head-injured patients. Percept Mot Skills, 59(1), 249-250.

66

Brooks, D.N. (1976). Wechsler Memory Scale performance and its relationship to brain damage after severe closed head injury. J Neurol Neurosurg Psychiatry, 39(6), 593-601. Buki, A., D. O. Okonkwo, et al. (2000). "Cytochrome c release and caspase activation in traumatic axonal injury." J Neurosci 20(8): 2825-34. Buki, A. and J. T. Povlishock (2006). "All roads lead to disconnection?--Traumatic axonal injury revisited." Acta Neurochir (Wien) 148(2): 181-93; discussion 193-4. Buki, A., R. Siman, et al. (1999). "The role of calpain-mediated spectrin proteolysis in traumatically induced axonal injury." J Neuropathol Exp Neurol 58(4): 365-75. Bush, B. A., T. A. Novack, et al. (2003). "Validation of a model for evaluating outcome after traumatic brain injury." Arch Phys Med Rehabil 84(12): 1803-7. Choi, S. C., T. Y. Barnes, et al. (1994). "Temporal profile of outcomes in severe head injury." J Neurosurg 81(2): 169-73. Christensen, B. K., B. Colella, et al. (2008). "Recovery of cognitive function after traumatic brain injury: a multilevel modeling analysis of Canadian outcomes." Arch Phys Med Rehabil 89(12 Suppl): S3-15. Cifu, D. X., L. Keyser-Marcus, et al. (1997). "Acute predictors of successful return to work 1 year after traumatic brain injury: a multicenter analysis." Arch Phys Med Rehabil 78(2): 125-31. Cohen, J. (1988). Statistical power analysis for the behavoural sciences (2nd ed.). Hillsdale, NJ, Lawrence Earlbaum Associates. Colicos, M. A., C. E. Dixon, et al. (1996). "Delayed, selective neuronal death following experimental cortical impact injury in rats: possible role in memory deficits." Brain Res 739(1-2): 111-9. Collie, A., D. Darby, et al. (2001). "Computerised cognitive assessment of athletes with sports related head injury." Br J Sports Med 35(5): 297-302. Collins, D. L., P. Neelin, et al. (1994). "Automatic 3D intersubject registration of MR volumetric data in standardized Talairach space." J Comput Assist Tomogr 18(2): 192-205. de la Plata, C. M., Ardelean, A., Koovakkattu, D., Srinivasan, P., Miller, A., Phuong, V., Harper, C., Moore, C., Whittemore, A., Madden, C., Diaz-Arrastia, R., Devous Sr., M. (2007). "Magnetic resonance imaging of diffuse axonal injury: quantitative assessment of white matter lesion volume." Journal of Neurotrauma 24(4): 591-598. Demetriades, D., E. Kuncir, et al. (2004). "Outcome and prognostic factors in head injuries with an admission Glasgow Coma Scale score of 3." Arch Surg 139(10): 1066-8.

67

De Silva, M.J., Roberts, I., Perel, P., Edwards, P., Kenward, M.G., Fernandes, J., Shakur, H., & Patel, V. (2009). Patient outcome after traumatic brain injury in high-, middle- and low-income countries : analysis of data on 8927 patients in 46 countries. International Journal of Epidemiology, 38(2), 452-458. Devitt, R., A. Colantonio, et al. (2006). "Prediction of long-term occupational performance outcomes for adults after moderate to severe traumatic brain injury." Disabil Rehabil 28(9): 547-59. Di Giovanni, S., Movsesyan, V., Ahmen, F., Cernak, I., Schinelli, S., Stoica, B., & Faden, A.I. (2005). Cell cycle inhibition provides neuroprotection and reduces glial proliferation and scar formation after traumatic brain injury. PNAS, 102(23), 8333-8338. Di Paola, M., Caltagirone, C., Fadda, L., Sabatini, U., Serra, L., & Carlesimo, G. A. (2008). Hippocampal atrophy is the critical brain change in patients with hypoxic amnesia. Hippocampus, 18(7), 719-728. Dikmen, S., Machamer, J., Temkin, N., & McLean, A. (1990). Neuropsychological recovery in patients with moderate to severe head injury: 2 year follow-up. J Clin Exp Neuropsychol, 12, 507-519. Dikmen, S. S., Machamer, J. E., Powell, J. M., & Temkin, N. R. (2003). Outcome 3 to 5 years after moderate to severe traumatic brain injury. Arch Phys Med Rehabil, 84(10), 1449-1457. Ding, K., Marquez de la Plata, C., Wang, J. Y., Mumphrey, M., Moore, C., Harper, C., et al. (2008). Cerebral atrophy after traumatic white matter injury: correlation with acute neuroimaging and outcome. J Neurotrauma, 25(12), 1433-1440. Draper, K., & Ponsford, J. (2008). Cognitive functioning ten years following traumatic brain injury and rehabilitation. Neuropsychology, 22(5), 618-625. Ferri-Campos, J., Chirivella-Garrido, J., Renau-Hernandez, O., Garcia-Blazquez, M.C., Ferri-Salvador, N., Noguera-Escalera, P., & Noe-Sebastian, E. (2008). When do patients with traumatic brain injury lose verbal informationa? Implications for cognitive rehabilitation. Rev Neurol, 46(2), 109-114. Finger, S., Koehler, P. J., & Jagella, C. (2004). The Monakow concept of diaschisis: origins and perspectives. Arch Neurol, 61(2), 283-288. Fletcher, P., & Henson, R. (2001). Frontal lobes and human memory: insights from functional neuroimaging. Brain, 124(Pt 5), 849-881. French, B., Thomas, L.H., Leathley, M.J., Sutton, C.J., McAdam, J., Forster, A., Langhorne, P. et al. (2007). Repetitive task training for improving functional ability after stroke. Cochrane Database Syst Rev, 4.

68

Fujii, M., Fukazawa, K., Takayasu, S., & Sakagami, M. (2002). Olfactory dysfunction in patients with head trauma. Auris Nasus Larynx, 29(1), 35-40. Fujimoto, S. T., Longhi, L., Saatman, K. E., Conte, V., Stocchetti, N., & McIntosh, T. K. (2004). Motor and cognitive function evaluation following experimental traumatic brain injury. Neurosci Biobehav Rev, 28(4), 365-378. Gennarelli, T. A., Thibault, L. E., Adams, J. H., Graham, D. I., Thompson, C. J., & Marcincin, R. P. (1982). Diffuse axonal injury and traumatic coma in the primate. Ann Neurol, 12(6), 564-574. Girard, D., Brown, J., Burnett-Stolnack, M., Hashimoto, N., Hier-Wellmer, S., Perlman, O. Z., et al. (1996). The relationship of neuropsychological status and productive outcomes following traumatic brain injury. Brain Inj, 10(9), 663-676. Gleissner, U., C. Helmstaedter, et al. (1998). "Right hippocampal contribution to visual memory: a presurgical and postsurgical study in patients with temporal lobe epilepsy." J Neurol Neurosurg Psychiatry 65(5): 665-9. Goodlett, C. R., & Horn, K. H. (2001). Mechanisms of alcohol-induced damage to the developing nervous system. Alcohol Res Health, 25(3), 175-184. Gozal, E., Row, B. W., Schurr, A., & Gozal, D. (2001). Developmental differences in cortical and hippocampal vulnerability to intermittent hypoxia in the rat. Neurosci Lett, 305(3), 197-201. Green, R. E., Colella, B., Christensen, B., Johns, K., Frasca, D., Bayley, M., et al. (2008). Examining moderators of cognitive recovery trajectories after moderate to severe traumatic brain injury. Arch Phys Med Rehabil, 89(12 Suppl), S16-24. Green, R. E., Colella, B., Hebert, D. A., Bayley, M., Kang, H. S., Till, C., et al. (2008b). Prediction of return to productivity after severe traumatic brain injury: investigations of optimal neuropsychological tests and timing of assessment. Arch Phys Med Rehabil, 89(12 Suppl), S51-60. Greenberg, G., Mikulis, D. J., Ng, K., DeSouza, D., & Green, R. E. (2008). Use of diffusion tensor imaging to examine subacute white matter injury progression in moderate to severe traumatic brain injury. Arch Phys Med Rehabil, 89(12 Suppl), S45-50. Hakkennes, S., Keating, J.L. (2005). Constraint-induced movement therapy following stroke: a systematic review of randomised controlled trials. Aust J Physiother, 51(4), 221-231. Hall, E. D., Bryant, Y. D., Cho, W., & Sullivan, P. G. (2008). Evolution of post-traumatic neurodegeneration after controlled cortical impact traumatic brain injury in mice and rats as assessed by the de Olmos silver and fluorojade staining methods. J Neurotrauma, 25(3), 235-247.

69

Haxel, B.R., Grant, L., & Mackay-Sim, A. (2008). Olfactory dysfunction after head injury. Journal of Head Trauma Rehab, 23(6), 407-413. Hedges, J.R., Newgard, C.D., Veum-Stone, J., Selden, N.R., Adams, A.L., Diggs, B.S., Arthur, M., & Mullins, R.J. (2008). Early neurosurgical procedures enhance survival in blunt head injury: propensity score analysis. J Emerg med, 37(2), 115-123. Heyman, A., Wilkinson, W. E., Stafford, J. A., Helms, M. J., Sigmon, A. H., & Weinberg, T. (1984). Alzheimer's disease: a study of epidemiological aspects. Ann Neurol, 15(4), 335-341. High, W.M., Sander, A.M., & Struchen, M.A. (2005). Rehabilitation for traumatic brain injury. Oxford University Press. Himanen, L., Portin, R., Isoniemi, H., Helenius, H., Kurki, T., & Tenovuo, O. (2005). Cognitive functions in relation to MRI findings 30 years after traumatic brain injury. Brain Inj, 19(2), 93-100. Hiott, D. W., & Labbate, L. (2002). Anxiety disorders associated with traumatic brain injuries. NeuroRehabilitation, 17(4), 345-355. Hoschl, C., & Hajek, T. (2001). Hippocampal damage mediated by corticosteroids--a neuropsychiatric research challenge. Eur Arch Psychiatry Clin Neurosci, 251 Suppl 2, II81-88. Immonen, R. J., Kharatishvili, I., Grohn, H., Pitkanen, A., & Grohn, O. H. (2009). Quantitative MRI predicts long-term structural and functional outcome after experimental traumatic brain injury. Neuroimage, 45(1), 1-9. Ito, J., Marmarou, A., Barzo, P., Fatouros, P., & Corwin, F. (1996). Characterization of edema by diffusion-weighted imaging in experimental traumatic brain injury. J Neurosurg, 84(1), 97-103. Jorge, R., & Robinson, R. G. (2002). Mood disorders following traumatic brain injury. NeuroRehabilitation, 17(4), 311-324. Jorge, R. E., Acion, L., Starkstein, S. E., & Magnotta, V. (2007). Hippocampal volume and mood disorders after traumatic brain injury. Biol Psychiatry, 62(4), 332-338. Jury, M. A., & Flynn, M. C. (2001). Auditory and vestibular sequelae to traumatic brain injury: a pilot study. N Z Med J, 114(1134), 286-288. Kadar, T., Dachir, S., Shukitt-Hale, B., & Levy, A. (1998). Sub-regional hippocampal vulnerability in various animal models leading to cognitive dysfunction. J Neural Transm, 105(8-9), 987-1004.

70

Kan, P., Amini, A., Hansen, K., White, G. L., Jr., Brockmeyer, D. L., Walker, M. L., et al. (2006). Outcomes after decompressive craniectomy for severe traumatic brain injury in children. J Neurosurg, 105(5 Suppl), 337-342. Kapoor, N., Ciuffreda, K.J., (2002). Vision disturbances following traumatic brain injury. Curr Treat Options Neurol, 4(4), 271-280. Katayama, Y., Becker, D. P., Tamura, T., & Hovda, D. A. (1990). Massive increases in extracellular potassium and the indiscriminate release of glutamate following concussive brain injury. J Neurosurg, 73(6), 889-900. Keyser-Marcus, L. A., Bricout, J. C., Wehman, P., Campbell, L. R., Cifu, D. X., Englander, J., et al. (2002). Acute predictors of return to employment after traumatic brain injury: a longitudinal follow-up. Arch Phys Med Rehabil, 83(5), 635-641. Klatzo, I. (1967). Presidental address. Neuropathological aspects of brain edema. J Neuropathol Exp Neurol, 26(1), 1-14. Kotapka, M. J., Gennarelli, T. A., Graham, D. I., Adams, J. H., Thibault, L. E., Ross, D. T., et al. (1991). Selective vulnerability of hippocampal neurons in acceleration-induced experimental head injury. J Neurotrauma, 8(4), 247-258. Kovesdi, E., Czeiter, E., Tamas, A., Reglodi, D., Szellar, D., Pal, J., Bukovics, P., Doczi, T., Buki, A. (2007). Rescuing neurons and glia: is inhibition of apoptosis useful? Progress in Brain Research, 161, 81-95. Kwakkel, G., Kollen, B., & Lindeman, E. (2004). Understanding the pattern of functional recovery after stroke: facts and theories. Restor Neurol Neurosci, 22(3-5), 281-299. Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33(1), 159-174. Lane, P.L., Skoretz, T.G., Doig, G., & Girotti, M.J. (2000). Intracranial pressure monitoring and outcomes after traumatic brain injury. Canadian Journal of Surgery, 43(6), 442-448. Leblanc, N., Chen, S., Swank, P. R., Levin, H., & Schachar, R. (2006). Impairment and recovery in inhibitory control after traumatic brain injury in children: effect of age at injury, injury severity and lesion location. Brain Cogn, 60(2), 208-209. Lee, A., Robbins, T., & Owen, A. (2000). Episodic memory meets working memory in the frontal lobe: functional neuroimaging studies of encoding and retrieval. Crit Rev Neurobiol, 14(3-4), 165-197. Lehnung, M., Leplow, B., Ekroll, V., Benz, B., Ritz, A., Mehdorn, M., & Ferstl, R. (2003). Recovery of spatial memory and persistence of spatial orientation deficits after traumatic brain injury during childhood. Brain Injury, 17(10), 855-869.

71

Levin, H.S., High, W.M., Ewing-Cobbs, L., Fletcher, J.M., Eisenberg, H.M., Miner, M.E., & Goldstein, F.C. (1988). Memory functioning during the first year after closed head injury in children and adolescents. Neurosurgery, 22, 1043-1052. Levine, B., Fujiwara, E., O'Connor, C., Richard, N., Kovacevic, N., Mandic, M., et al. (2006). In vivo characterization of traumatic brain injury neuropathology with structural and functional neuroimaging. J Neurotrauma, 23(10), 1396-1411. Levine, B., Kovacevic, N., Nica, E. I., Cheung, G., Gao, F., Schwartz, M. L., et al. (2008). The Toronto traumatic brain injury study: injury severity and quantified MRI. Neurology, 70(10), 771-778. Levine, M.D. (1989). Attention and memory: progression and variation during the elementary school years. Pediatr Ann, 18(6), 366-372. Lezak, M., Howieson, D.B., Loring, D.W., Jannay, J.H., & Fischer, J.S.(2004). Neuropsychological Assessment. USA: Oxford University Press. Li, S., Mealing, G.A., Morley, P., Stys, P.K. (1999). Novel injury mechanism in anoxia and trauma of spinal cord white matter: glutamate release via reverse Na(+)-dependent glutamate transport. Journal of Neuroscience, 19, RC1-9. Lye, T. C., O. Piguet, et al. (2004). "Hippocampal size and memory function in the ninth and tenth decades of life: the Sydney Older Persons Study." J Neurol Neurosurg Psychiatry 75(4): 548-54. MacKenzie, J. D., Siddiqi, F., Babb, J. S., Bagley, L. J., Mannon, L. J., Sinson, G. P., et al. (2002). Brain atrophy in mild or moderate traumatic brain injury: a longitudinal quantitative analysis. AJNR Am J Neuroradiol, 23(9), 1509-1515. Madigan, N. K., DeLuca, J., Diamond, B. J., Tramontano, G., & Averill, A. (2000). Speed of information processing in traumatic brain injury: modality-specific factors. J Head Trauma Rehabil, 15(3), 943-956. Marmarou, A. (2007). A review of progress in understanding the pathophysiology and treatment of brain edema. Neurosurg Focus, 22(5), E1. Mateer, C.A., Kerns, K.A. and Eso, K.L. (1996). Management of attention and memory disorders following traumatic brain injury. Journal of Learning Disabilities, 29(6), 618-632. Matsushita, Y., Shima, K., Nawashiro, H., & Wada, K. (2000). Real-time monitoring of glutamate following fluid percussion brain injury with hypoxia in the rat. J Neurotrauma, 17(2), 143-153. Maxwell, W. L., Watt, C., Graham, D. I., & Gennarelli, T. A. (1993). Ultrastructural evidence of axonal shearing as a result of lateral acceleration of the head in non-human primates. Acta Neuropathol, 86(2), 136-144.

72

Mazaux, J. M., Masson, F., Levin, H. S., Alaoui, P., Maurette, P., & Barat, M. (1997). Long-term neuropsychological outcome and loss of social autonomy after traumatic brain injury. Arch Phys Med Rehabil, 78(12), 1316-1320. Mazziotta, J. C., A. W. Toga, et al. (1995). "A probabilistic atlas of the human brain: theory and rationale for its development. The International Consortium for Brain Mapping (ICBM)." Neuroimage 2(2): 89-101. McDonald, B. C., Flashman, L. A., & Saykin, A. J. (2002). Executive dysfunction following traumatic brain injury: neural substrates and treatment strategies. NeuroRehabilitation, 17(4), 333-344. McKee, R.D., & Squire, L.R. (1992). Equivalent forgetting rates in long-term memory for diencephalic and medial temporal lobe amnesia. J Neuroscience, 12(10), 3765-3772. Millis, S.R., Ricker, J.H. (1994). Verbal learning patterns in moderate and severe traumatic brain injury. J Clin Exp Neuropsychol, 16(4), 498-507. Millis, S. R., Rosenthal, M., Novack, T. A., Sherer, M., Nick, T. G., Kreutzer, J. S., et al. (2001). Long-term neuropsychological outcome after traumatic brain injury. J Head Trauma Rehabil, 16(4), 343-355. Morganti-Kossmann, M. C., Satgunaseelan, L., Bye, N., & Kossmann, T. (2007). Modulation of immune response by head injury. Injury, 38(12), 1392-1400. Nakase-Richardson, R., Sepehri, A., Sherer, M., Yablon, S. A., Evans, C., & Mani, T. (2009). Classification schema of posttraumatic amnesia duration-based injury severity relative to 1-year outcome: analysis of individuals with moderate and severe traumatic brain injury. Arch Phys Med Rehabil, 90(1), 17-19. Neistadt, M. E. (1994). The effects of different treatment activities on functional fine motor coordination in adults with brain injury. Am J Occup Ther, 48(10), 877-882. Ng, K., Mikulis, D. J., Glazer, J., Kabani, N., Till, C., Greenberg, G., et al. (2008). Magnetic resonance imaging evidence of progression of subacute brain atrophy in moderate to severe traumatic brain injury. Arch Phys Med Rehabil, 89(12 Suppl), S35-44. Novack, T. A., Alderson, A. L., Bush, B. A., Meythaler, J. M., & Canupp, K. (2000). Cognitive and functional recovery at 6 and 12 months post-TBI. Brain Inj, 14(11), 987-996. Novack, T. A., Bush, B. A., Meythaler, J. M., & Canupp, K. (2001). Outcome after traumatic brain injury: pathway analysis of contributions from premorbid, injury severity, and recovery variables. Arch Phys Med Rehabil, 82(3), 300-305.

73

Oddy, M., Coughlan, T., Tyerman, A., & Jenkins, D. (1985). Social adjustment after closed head injury: a further follow-up seven years after injury. J Neurol Neurosurg Psychiatry, 48(6), 564-568. Ouyang, Y. B., Voloboueva, L. A., Xu, L. J., & Giffard, R. G. (2007). Selective dysfunction of hippocampal CA1 astrocytes contributes to delayed neuronal damage after transient forebrain ischemia. J Neurosci, 27(16), 4253-4260. Park, E., Bell, J.D., & Baker, A.J. (2008). Traumatic brain injury: can the consequences be stopped? CMAJ, 178(9), 1163-1170. Pettus, E. H., Christman, C. W., Giebel, M. L., & Povlishock, J. T. (1994). Traumatically induced altered membrane permeability: its relationship to traumatically induced reactive axonal change. J Neurotrauma, 11(5), 507-522. Pierallini, A., Pantano, P., Fantozzi, L. M., Bonamini, M., Vichi, R., Zylberman, R., et al. (2000). Correlation between MRI findings and long-term outcome in patients with severe brain trauma. Neuroradiology, 42(12), 860-867. Pierpaoli, C., Barnett, A., Pajevic, S., Chen, R., Penix, L. R., Virta, A., et al. (2001). Water diffusion changes in Wallerian degeneration and their dependence on white matter architecture. Neuroimage, 13(6 Pt 1), 1174-1185. Povlishock, J. T., Becker, D. P., Cheng, C. L., & Vaughan, G. W. (1983). Axonal change in minor head injury. J Neuropathol Exp Neurol, 42(3), 225-242. Povlishock, J. T., & Katz, D. I. (2005). Update of neuropathology and neurological recovery after traumatic brain injury. J Head Trauma Rehabil, 20(1), 76-94. Preussner, J. C., Li, L.M., Serles, W., Preussner, M., Collins, D.L., Kabani, N., Lupien, S., Evans, A.C. (2000). Volumetry of hippocampus and amygdala with high-resolution MRI and three-dimensional analysis software: minimizing the discrepancies between laboratories. Cerebral Cortex, 10, 433-442. Rasmusson, D. X., Brandt, J., Martin, D. B., & Folstein, M. F. (1995). Head injury as a risk factor in Alzheimer's disease. Brain Inj, 9(3), 213-219. Richards, Blake (1995). Visual Quality Control Guide: A Document produced for the MRI lab at the Centre for Addiction and Mental Health, Toronto. Rodriguez-Paez, A. C., Brunschwig, J. P., & Bramlett, H. M. (2005). Light and electron microscopic assessment of progressive atrophy following moderate traumatic brain injury in the rat. Acta Neuropathol, 109(6), 603-616. Rosner, G. (2003). Combat hypoxia: the importance of airway management & oxygenation of the traumatic brain injury patient. JEMS, 28(3), 100-117.

74

Royo, N. C., LeBold, D., Magge, S. N., Chen, I., Hauspurg, A., Cohen, A. S., et al. (2007). Neurotrophin-mediated neuroprotection of hippocampal neurons following traumatic brain injury is not associated with acute recovery of hippocampal function. Neuroscience, 148(2), 359-370. Ruijs, M. B., Keyser, A., & Gabreels, F. J. (1994). Clinical neurological trauma parameters as predictors for neuropsychological recovery and long-term outcome in paediatric closed head injury: a review of the literature. Clin Neurol Neurosurg, 96(4), 273-283. Saatman, K. E., Duhaime, A. C., Bullock, R., Maas, A. I., Valadka, A., & Manley, G. T. (2008). Classification of traumatic brain injury for targeted therapies. J Neurotrauma, 25(7), 719-738. Salmond, C. H., & Sahakian, B. J. (2005). Cognitive outcome in traumatic brain injury survivors. Curr Opin Crit Care, 11(2), 111-116. Sass, K. J., D. D. Spencer, et al. (1990). "Verbal memory impairment correlates with hippocampal pyramidal cell density." Neurology 40(11): 1694-7. Sass, K. J., A. Sass, et al. (1992). "Specificity in the correlation of verbal memory and hippocampal neuron loss: dissociation of memory, language, and verbal intellectual ability." J Clin Exp Neuropsychol 14(5): 662-72. Schacter, D.L., & Crovitz, H.F.(1977). Memory function after closed head injury: a review of the quantitative research. Cortex, 13(2), 150-176. Schofield, P. W., Tang, M., Marder, K., Bell, K., Dooneief, G., Chun, M., et al. (1997). Alzheimer's disease after remote head injury: an incidence study. J Neurol Neurosurg Psychiatry, 62(2), 119-124. Sherer, M., Struchen, M. A., Yablon, S. A., Wang, Y., & Nick, T. G. (2008). Comparison of indices of traumatic brain injury severity: Glasgow Coma Scale, length of coma and post-traumatic amnesia. J Neurol Neurosurg Psychiatry, 79(6), 678-685. Sheriff, F. E., Bridges, L.R., Sivaloganaathan, S. (1994). Early detection of axonal injury after human head trauma using immunocytochemistry for beta-amyloid precursor protein. Acta Neuropathologica, 87, 55-62. Shimamura, M., Garcia, J. M., Prough, D. S., Dewitt, D. S., Uchida, T., Shah, S. A., et al. (2005). Analysis of long-term gene expression in neurons of the hippocampal subfields following traumatic brain injury in rats. Neuroscience, 131(1), 87-97. Shrout, P. E., & Fleiss, J. L. (1979). Intraclass correlations: uses in assessing rater reliability. Psychol Bull, 86(2), 420-428. Shum, D.H., Harris, D., & O'Gorman, J.G. (2000). Effects of severe traumatic brain injury on visual memory. J Clin Exp Neuropsychol, 22(1), 25-39.

75

Sled, J. G., Zijdenbos, A. P., & Evans, A. C. (1998). A nonparametric method for automatic correction of intensity nonuniformity in MRI data. IEEE Trans Med Imaging, 17(1), 87-97. Smith, D. H., Chen, X. H., Iwata, A., & Graham, D. I. (2003). Amyloid beta accumulation in axons after traumatic brain injury in humans. J Neurosurg, 98(5), 1072-1077. Smith, D. H., Chen, X. H., Nonaka, M., Trojanowski, J. Q., Lee, V. M., Saatman, K. E., et al. (1999). Accumulation of amyloid beta and tau and the formation of neurofilament inclusions following diffuse brain injury in the pig. J Neuropathol Exp Neurol, 58(9), 982-992. Smith, D. H., Chen, X. H., Pierce, J. E., Wolf, J. A., Trojanowski, J. Q., Graham, D. I., et al. (1997). Progressive atrophy and neuron death for one year following brain trauma in the rat. J Neurotrauma, 14(10), 715-727. Smith, D.H., & Meaney, D.F. (2000). Axonal damage in traumatic brain injury. Neuroscientist, 6(6), 483-495. Smith, D. H., Meaney, D. F., & Shull, W. H. (2003). Diffuse axonal injury in head trauma. J Head Trauma Rehabil, 18(4), 307-316. Sperry, J. L., Gentilello, L. M., Minei, J. P., Diaz-Arrastia, R. R., Friese, R. S., & Shafi, S. (2006). Waiting for the patient to "sober up": Effect of alcohol intoxication on glasgow coma scale score of brain injured patients. J Trauma, 61(6), 1305-1311. Spettell, C. M., Ellis, D. W., Ross, S. E., Sandel, M. E., O'Malley, K. F., Stein, S. C., et al. (1991). Time of rehabilitation admission and severity of trauma: effect on brain injury outcome. Arch Phys Med Rehabil, 72(5), 320-325. Spreen, O., & Strauss, E. (1998). A compedium of neuropsychological tests: administration, norms and commentary. New York: Oxford Unversity Press. Squire, L.R. (1987). Memory and Brain. USA: Oxford University Press US. Stoll, G. and H. W. Muller (1999). "Nerve injury, axonal degeneration and neural regeneration: basic insights." Brain Pathol 9(2): 313-25. Talairach, J., Tournoux, P. (1988). Co-planar sterotaxic atlas of the human brain: 3-dimensional proportional system: an approach to cerebral imaging. New York: Thieme Medical Publishers. Tate, D. F., & Bigler, E. D. (2000). Fornix and hippocampal atrophy in traumatic brain injury. Learn Mem, 7(6), 442-446. Teasdale, G., & Jennett, B. (1974). Assessment of coma and impaired consciousness. A practical scale. Lancet, 2(7872), 81-84.

76

Till, C., Colella, B., Verwegen, J., & Green, R. E. (2008). Postrecovery cognitive decline in adults with traumatic brain injury. Arch Phys Med Rehabil, 89(12 Suppl), S25-34. Tomaiuolo, F., Carlesimo, G.A., Di Paola, M., Petrides, M., Fera, F., Bonanni, R., Formisano, R., Pasqualetti, P., Caltagirone, C. (2004). Gross morphology and morphometric sequelae in the hippocampus, fornix, and corpus callosum of patients with severe non-missile traumatic brain injury without macroscopically detectable lesions: a T1 weighted MRI study. Journal of Neurology, Neurosurgery, and Psychiatry, 75, 1314-1322. Tonks, J., Williams, W.H., Frampton, I., Yates, P., & Slater, A. (2007). Reading emotions after child brain injury: a comparison between children with brain injury and non-injured controls. Brain Injury, 21(7), 731-739. Trachtenberg, M.C. (1983). Glial endocytosis of protein in the traumatized brain. Journal of Neuroscience Research, 9(4), 413-423. Trivedi, M. A., Ward, M. A., Hess, T. M., Gale, S. D., Dempsey, R. J., Rowley, H. A., et al. (2007). Longitudinal changes in global brain volume between 79 and 409 days after traumatic brain injury: relationship with duration of coma. J Neurotrauma, 24(5), 766-771. Upadhyay, D. (2008). Cognitive Functioning in TBI Patients: a Review of Literature. Middle-East Journal of Scientific Research, 3(3), 120-125. Vakil, E. (2005). The effect of moderate to severe traumatic brain injury (TBI) on different aspects of memory: a selective review. J Clin Exp Neuropsychol, 27(8), 977-1021. Valadka, A. B., Gopinath, S. P., & Robertson, C. S. (2000). Midline shift after severe head injury: pathophysiologic implications. J Trauma, 49(1), 1-8; discussion 8-10. Vanderploeg, R.D., Crowell, T.A., & Curtiss, G. (2001). Verbal learning and memory deficits in traumatic brain injury: encoding, consolidation and retrieval. J Clin Exp Neuropsychol, 23(2), 185-195. Vitaz, T. W., Jenks, J., Raque, G. H., & Shields, C. B. (2003). Outcome following moderate traumatic brain injury. Surg Neurol, 60(4), 285-291; discussion 291. Walker, W. C., & Pickett, T. C. (2007). Motor impairment after severe traumatic brain injury: A longitudinal multicenter study. J Rehabil Res Dev, 44(7), 975-982. Wang, Y. T., Kent, R. D., Duffy, J. R., & Thomas, J. E. (2005). Dysarthria associated with traumatic brain injury: speaking rate and emphatic stress. J Commun Disord, 38(3), 231-260. Whitwell, J. L. (2008). Longitudinal imaging: change and causality. Curr Opin Neurol, 21(4), 410-416. Wiegner, S., & Donders, J. (1999). Performance on the California Verbal Learning Test after Traumatic Brain Injury. J Clin Exp Neuropsychol, 21(2), 159-170.

77

Wilde, E. A., Bigler, E. D., Gandhi, P. V., Lowry, C. M., Blatter, D. D., Brooks, J., et al. (2004). Alcohol abuse and traumatic brain injury: quantitative magnetic resonance imaging and neuropsychological outcome. J Neurotrauma, 21(2), 137-147. Wilde, E. A., Bigler, E. D., Pedroza, C., & Ryser, D. K. (2006). Post-traumatic amnesia predicts long-term cerebral atrophy in traumatic brain injury. Brain Inj, 20(7), 695-699. Willemse-van Son, A. H. P., Ribbers, G.M., Verhagen, A.P., Stam, H.J. (2007). Prognostic factors of long-term functioning and productivity after traumatic brain injury: a systematic review of prospective cohort studies. Clinical Rehabilitation, 21(11), 1024-1037. Winocur, G. (1985). The hippocampus and thalamus: their roles in short- and long-term memory and the effects of interference. Behav Brain Res, 16(2-3), 135-152. Witgen, B. M., Lifshitz, J., Smith, M. L., Schwarzbach, E., Liang, S. L., Grady, M. S., et al. (2005). Regional hippocampal alteration associated with cognitive deficit following experimental brain injury: a systems, network and cellular evaluation. Neuroscience, 133(1), 1-15. Yamamoto, T., Rossi, S., Stiefel, M., Doppenberg, E., Zauner, A., Bullock, R., et al. (1999). CSF and ECF glutamate concentrations in head injured patients. Acta Neurochir Suppl, 75, 17-19. Yattoo, G., & Tabish, A. (2008). The profile of head injuries and traumatic brain injury deaths in Kashmir. J Trauma Manag Outcomes, 2(1). Yi, J. H., & Hazell, A. S. (2006). Excitotoxic mechanisms and the role of astrocytic glutamate transporters in traumatic brain injury. Neurochem Int, 48(5), 394-403. Yoshimura, S., Teramoto, T., Whalen, M. J., Irizarry, M. C., Takagi, Y., Qiu, J., et al. (2003). FGF-2 regulates neurogenesis and degeneration in the dentate gyrus after traumatic brain injury in mice. J Clin Invest, 112(8), 1202-1210. Yount, R., Raschke, K. A., Biru, M., Tate, D. F., Miller, M. J., Abildskov, T., et al. (2002). Traumatic brain injury and atrophy of the cingulate gyrus. J Neuropsychiatry Clin Neurosci, 14(4), 416-423. Zafonte, R., Elovic, E. P., & Lombard, L. (2004). Acute care management of post-TBI spasticity. J Head Trauma Rehabil, 19(2), 89-100. Zec, R.F., Zellers, D., Belman, J., Miller, J., Matthews, J., Femeau-Belman, D., & Robbs, R. (2001). Long-term consequences of severe closed head injury on episodic memory. J Clin Exp Neuropsychol, 23(5), 671-691.

78

Zygun, D. A., Laupland, K. B., Hader, W. J., Kortbeek, J. B., Findlay, C., Doig, C. J., et al. (2005). Severe traumatic brain injury in a large Canadian health region. Can J Neurol Sci, 32(1), 87-92.