Brain Injury

Brain Injury

ee

rP Fo Journal:

Manuscript Type:

TBIN-2009-0116.R1 Original Paper

ie

Keywords:

Brain Injury

ev

Manuscript ID:

rR

A retropsective study of heterogeneity in neurocognitive profiles associated with traumatic brain injury

cognitive, head Injury, cluster analysis, assessment

w ly

On URL: http://mc.manuscriptcentral.com/tbin

Page 1 of 48

Brain Injury

1 Heterogeneity in TBI

November 30, 2009

Deleted: November 30, 2009 Deleted: November 25, 2009

Traumatic Brain Injury 1 Running Head: Heterogeneity in TBI A Retrospective Study of Heterogeneity in Neurocognitive Profiles Associated with Traumatic Brain Injury Deleted: November 30, 2009

November 30, 2009

rP Fo

Deleted: November 25, 2009

Gerald Goldstein1, Daniel N. Allen2 and Janelle M. Caponigro1

1

Mental Illness Research, Educational, and Clinical Center (MIRECC), Veterans Affairs Pittsburgh Healthcare System

2

Department of Psychology, University of Nevada Las Vegas

rR

ee

Corresponding Author: Gerald Goldstein VA Pittsburgh Healthcare System 7180 Highland Drive (151R) Pittsburgh, PA 15206 Phone: 412-954-5356 FAX: 412954-5371 E mail: [email protected]

w

ie

ev ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Brain Injury

Page 2 of 48

2 Heterogeneity in TBI

Abstract Primary Objective: To establish empirical subtypes, based upon cognitive test results, of individuals who had sustained traumatic brain injury. Research Design: The study was retrospective, applying cluster analyses and associated statistical tests to an established database. Methods and Procedures: Neuropsychological data from veterans with brain

rP Fo

trauma were cluster analyzed using the WAIS-R and Halstead-Reitan Battery (HRB). External validity of the cluster solutions was evaluated. Experimental interventions: The study was based upon use of an established database that contained cognitive test data and information regarding diagnosis and clinical history

rR

ee

Main Outcomes and Results: The WAIS-R clusters described subgroups with near normal function, preserved verbal but impaired problem solving abilities, or

ev

global impairment. The HRB clusters differed in level of performance with Near Normal, Moderately Impaired, and Globally Impaired clusters. Cluster

ie

membership was associated with age and employment status, and but not with

w

neurological findings.

Conclusion: The outcome of traumatic brain injury is heterogeneous, and mainly associated with demographic considerations.

On

Key Words: Brain Injury, Assessment, Cluster Analysis, Intelligence, Brain Injur

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Page 3 of 48

Brain Injury

3 Heterogeneity in TBI

Much of the cognitive research with brain injury has traditionally focused on localization of cognitive abilities, in the case of focal wounds, or alternatively identifying a pattern of cognitive deficits that is associated with the more diffuse injuries associated with closed head injuries (1). However, a number of studies have suggested that brain injury does not have any protoypical pattern of cognitive performance and outcome but may be best characterized by

rP Fo

heterogeneity both in regard to neurocognitive deficit and ultimate level of functioning. To examine this heterogeneity, empirical classification methods, notably cluster analysis, have been applied to both children and adults with traumatic brain injury (TBI) (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13). This work has

ee

primarily focused on measures of intellectual abilities and memory, and as might

rR

be expected some variability in results have been present from one study to the next. However, these studies have provided evidence for homogeneous clusters

ev

that are often differentiated by level of performance, although pattern of performance differences have been reported as well.

ie

An investigation of the WAIS III (14) in TBI conducted by van der

w

Heijden and Donders (12) demonstrated that when subjected to cluster analysis, the WAIS III index scores produced three clusters of patients. The clusters were

On

differentiated largely by level of performance with one cluster characterized by above average performance, a second by average performance, and a third by

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

below average performance. Among the three clusters there were substantially

URL: http://mc.manuscriptcentral.com/tbin

Brain Injury

Page 4 of 48

4 Heterogeneity in TBI

different mean scores on the Verbal Comprehension (VC), Perceptual Organization (PO), Working Memory (WM), and Perceptual Speed (PS) factor index scores, with no overlap of profiles or other sign of configuration differences. The authors concluded that there was no cognitive profile that was unique to TBI although a relative decrease in Processing Speed scores was present in all clusters.

rP Fo

In addition to WAIS performance, evidence for such heterogeneity comes from a number of studies that concentrated on memory, showing that there are clusters representing subtypes of memory disorder such as a subgroup with limited capacity to benefit from repetition and another subgroup that showed an

ee

unusually high rate of intrusion errors (6, 9). Chan et al. (3) described three

rR

subtypes of combinations of attentional deficits. Thus, cluster analysis has revealed the presence of heterogeneity in general intelligence, memory, and

ev

attention, suggesting the possibility of its appearance in other cognitive functions. A main objective of the current study was to examine this possibility by extending

ie

the findings with IQ, memory, and attention tests to a more comprehensive

w

evaluation of neurocognitive abilities. We were specifically interested in refining the cluster analysis based classification of patients provided by van der Heijden

On

and Donders (12) with regard to cognitive profile detail and specificity through utilizing additional tests, in this case major components of the Halstead-Reitan

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Neuropsychological Test Battery (HRB) (15) which assesses a large number of

URL: http://mc.manuscriptcentral.com/tbin

Page 5 of 48

Brain Injury

5 Heterogeneity in TBI

cognitive abilities. Based on van der Heijden and Donders (12) and much of the previous literature described in Reitan and Wolfson (15) we hypothesized that we would find a generalized pattern of performance across abilities, but substantial heterogeneity with regard to level of performance. Such heterogeneity might provide the basis for formation of a number of valid, separable clusters. Both van der Heijden and Donders (12) and Millis and Ricker (9) recommended use of

rP Fo

additional, more comprehensive tests to confirm and extend their findings. Comparisons between our WAIS-R cluster solution and that obtained using the Deleted: T

HRB were therefore of particular interest. Thus, the present study provides a classification system based upon cluster analysis of a comprehensive battery of

ee

tests assessing a broader scope of cognitive abilities than was employed in past efforts at classification.

rR

Another objective was to determine if heterogeneous patterns and levels of

ev

performance may vary as a result of recovery of cognitive abilities as the length of time since brain injury increases (9). Patients in the Millis and Ricker’s study (9)

ie

had a mean post-injury time of less than 2 years. Heijden and Donders (12)

w

participants had their injuries an average of about 3 months post-injury. Thus, the present study, while involving a wide time post-injury range, focused on

On

relatively long term outcome patterns of cognitive function in a very

heterogeneous sample. The possibility of different clusters produced in the

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

present study from those obtained by van der Heidjen and Donders (12) appeared

URL: http://mc.manuscriptcentral.com/tbin

Brain Injury

Page 6 of 48

6 Heterogeneity in TBI

to be real given that with passage of time profiles based on sparing, recovery, and persisting deficits may be expected to emerge which might not be apparent in the months immediately following injury. This view is supported by the fact that Millis and Ricker (9) obtained substantial time post-injury differences among their CVLT-based clusters.

rP Fo

This study is retrospective, based upon data obtained during the 1960s and 70s. It is derived from a group of veterans who served in the military in World War II, the Korean War and the Vietnam conflict. It was done primarily because availability of this sample provided the opportunity to consider outcome of brain

ee

injury over a lengthy range of time that cannot be assessed through exclusive

rR

study of acute patients, and provided a possible basis for comparison with brain injuries associated with recent military activity. The availability of a

ev

heterogeneous sample for classification studies is also advantageous, since it permits basing findings on distributions in a natural environment before

ie

heterogeneity is attenuated by sample selection requirements of some particular study.

w

The neuropsychological assessments were accomplished during inpatient

On

treatment unrelated to the acute trauma, such as acquisition of another medical or psychiatric illness, or a residual consequence of the trauma, such as seizures.

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Modern procedures and technologies were not available during the time when the

URL: http://mc.manuscriptcentral.com/tbin

Deleted: A final objective was to determine what influence, if any, effort might have on heterogeneity of neuropsychological test performance in patients with TBI. In recent years, there has been a extensive literature that has documented the importance of effort in neuropsychological test performance among individuals with TBI, yet cluster analytic studies that have investigated neurocognitive heterogeneity have not considered this important variable as a contributing factor to difference in performance among neurocognitive clusters. While modern methods to evaluate effort were not used in this study, embedded measures from the Category Test were available and used to investigate this issue. Specifically, Forrest, Allen and Goldstein (20) reported that the best Category Test indices of effort were more than five errors on subtests I and II, both of which are simple number identification and counting tasks, and lack of improvement on subtest VI relative to subtest V, both of which require learning of exactly the same principle. It was noted in their study that while patients with structural brain damage showed such improvement, coached malingerers showed less improvement than patients with structural brain damage. It was also found that coached malingerers produced an excessive number of errors on Subtests I and II. These indexes were used to evaluate effort in the overall sample, and as external validity criteria to determine if differences in neuropsychological test performance among the clusters might be attributable to effort considerations. Formatted: Indent: First line: 36 pt Deleted: While modern procedures and technologies were not available during the time when the patients were hospitalized, Deleted: t

Page 7 of 48

Brain Injury

7 Heterogeneity in TBI

patients were hospitalized. Assessment and documentation procedures that were available then included direct observation at or close to the time of injury, review of medical records during first and subsequent hospitalizations, taking of a detailed medical and social history, physical neurological examination, and laboratory procedures available at the time, mainly EEG and skull X-ray, with rare use of pneumoencephalography and nuclear brain scans. Often neurosurgical

rP Fo

reports were available. In each case, there was a definitive neurological diagnosis such as “trauma, bone defect right frontal parietal area” or “right frontal subdural hematoma.” The neuropsychological assessment was done as part of comprehensive general medical, psychiatric, and neurological evaluations using

ee

procedures available at the time. The advantages of utilizing this opportunity to

rR

do long term studies of TBI outcome has also been illustrated in an investigation by Teiler, Adams, Walker, and Rourke (16) who evaluated psychosocial

ev

adjustment in veterans with penetrating brain injury forty years after they were wounded during World War II. The research reported here should be understood in the context of that type of study.

Participants

w

METHODS

ie On

Demographic, psychometric and clinical data for the participants are

provided in Tables 1 and 2. As can be seen from the Table, the sample was 34.6 years old when assessed with approximately 12 years of education and was

URL: http://mc.manuscriptcentral.com/tbin

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Deleted: Assessment and documentation procedures that were available then included direct observation at or close to the time of injury, review of medical records during first and subsequent hospitalizations, taking of a detailed medical and social history, physical neurological examination, and laboratory procedures available at the time, mainly EEG and skull Xray, with rare use of pneumoencephalography and nuclear brain scans. Often neurosurgical reports were available. In each case, there was a definitive neurological diagnosis such as “trauma, bone defect right frontal parietal area” or “right frontal subdural hematoma.” Comment [DA1]: Jerry, I think we should refer to the demographic and clinical characteristics from the outset, so I have moved the references to Tables 1 and 2 here. I have also provided a little discussion of the data presented in those Tables (e.g. education, gender) \. I also moved some of the description you had the sample in other parts of the paper to this spot, as I thought they fit better here. So, for example, I moved the description of the outcome of the chart evaluation on causes of injury, surgery, etc., from the “Procedures” to here. Formatted: Indent: First line: 0 pt

Brain Injury

Page 8 of 48

8 Heterogeneity in TBI

predominantly male. Full scale IQ was in the average/low average range, although the extreme range of scores at the low end may reflect the presence of impairment acquired as a result of TBI. Individuals with IQs within much of this low range are ineligible for enlistment in military service, suggesting that scores within this range reflect acquired cognitive impairment rather than preexisting low intellectual function. Eighty months had elapsed between the time of initial injury

rP Fo

and neuropsychological assessment. By far, the major cause of injury was car accident (see Table 2). About a third of the participants received surgery, and about a quarter of them had seizures following the TBI. 71% of the participants had a period of unconsciousness reported, but reliable information concerning

ee

actual length of time unconscious was not generally available. Using the

rR

definition employed here of open head injury alternatively requiring significant skull fracture, surgical debridement or related procedures, presence of a subdural

ev

hematoma, or observation of brain tissue extrusion to the surface of the skull, almost half the sample had this condition with the others having closed head

ie

injury. The psychiatric information indicated that diagnoses used at the time, such

w

as “organic brain syndrome” was given in 46 cases, but the prevalence of other disorders was relatively low. All patients provided informed consent to allow their

On

results and clinical records to be used for research purposes in accordance with established hospital policies at that time.

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Page 9 of 48

Brain Injury

9 Heterogeneity in TBI

Data were available for the Wechsler intelligence scales in use at the time of testing (WAIS or WAIS-R), HRB and associated clinical records for veterans with histories of significant TBI hospitalized at the Topeka, KS, or Pittsburgh, PA veterans (DVA) hospitals. These hospitals are both large DVA centers for treatment of psychiatric and neurological disorders from two different regions of the United States, and are thereby likely to have produced a reasonably

rP Fo

representative and diverse sample of veterans with TBI. Cases were drawn from a database of DVA patients containing HRB test data and copies of pertinent clinical records. Patients in the database were referred for neuropsychological testing, suggesting concern by clinicians regarding cognitive function.

ee

Evaluations included thorough physical neurological examinations and clinical

rR

histories, and use of the diagnostic technologies available at the time. The CT scan only came into use during the later stages of data collection and the MRI had

ev

not been available then. However, earlier methods such as pneumoencephalograms, brain scans, or EEGs were used as indicated. The brain

ie

injuries were well documented and described, and there were often actual

w

descriptions of the injury itself or neurosurgical data. While at the VA facility, patients received thorough general medical and neurological evaluations in connection with the neuropsychological assessment.

On

Participants were hospitalized at the time they received their

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

neuropsychological assessments. They had sustained a TBI between 1942 and

URL: http://mc.manuscriptcentral.com/tbin

Comment [DA2]: I moved this statement about the completeness of the data, etc., to the Procedures section because it seemed to fit better there. Deleted: Complete data sets for the cluster analyses described below were available for 80 participants for the analysis of the HRB and 78 for the analysis of the WAIS-R. Since the study was retrospective, and clinical and historical data were obtained from clinical records, there are varying amounts of missing data among the variables studied.

Brain Injury

Page 10 of 48

10 Heterogeneity in TBI

1982. The test data for the present study were not obtained at the hospitalization for the immediate, acute consequences of the trauma but for follow-up evaluation

Comment [DA3]: Jerry, I don’t think that we need to include this statement here, because it is repeated again when we discuss Table 2 (Above). Deleted:

and treatment of residual symptoms of their TBI, or treatment of another illness. The most recently injured participant was injured 3 months prior to testing ranging to a participant who sustained his injury 33 years prior to testing. Our average participant had his injury about 6 and ½ years prior to testing, which is a

rP Fo

much longer recovery period than other studies. Only 17 (20.5%) of the patients had sustained their injuries less than one year prior to testing. Data regarding evaluation procedures conducted during the acute phase of the TBI and social and medical history were available in the clinical records and were utilized for

ee

establishment of external validity of the cluster solutions.

rR

Cases rated as having sustained brain trauma were identified, and individuals with the additional diagnoses of schizophrenia or a substance use

ev

disorder were excluded. The database contained 31 patients who had a TBI and substance abuse, and 9 patients with TBI and schizophrenia. A small number of

ie

individuals with remote histories of alcoholism not diagnosed in VA medical

w

records were retained. Individuals whose clinical records did not contain substantial amounts of information needed for the study were also excluded. All

On

participants had definitive evidence of a TBI and, with few minor exceptions,

complete data for the neuropsychological tests. The wide age range indicates that

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

the sample included older veterans from the times of the Second World and

URL: http://mc.manuscriptcentral.com/tbin

Deleted: While all of the participants were veterans, the TBIs were not necessarily combat-related but were more often sustained in civilian life or while the participant was still in the military but not engaged in combat.

Page 11 of 48

Brain Injury

11 Heterogeneity in TBI

Korean Wars and younger individuals in the military at the time of the Vietnam Comment [DA4]: I moved this statement to the beginning of the Participants section, since we are talking about the demographic, psychometric and clinical information there.

conflict. Insert table 1 about here Insert table 2 about here

Deleted: Demographic and psychometric data for these participants are provided in Table 1.

Procedure A standardized review of the clinical records was conducted using a

rP Fo

schedule developed for that purpose. The schedule supported a review of demographic information, phenomenology of the brain trauma, course, surgical and medical diagnostic procedures and psychiatric findings. Information concerning these areas was contained in the clinical records or obtained during

ee

history taking done at the time of the neuropsychological assessment with the

rR

patient and/or informants. With regard to demographic information we recorded age, age at time of injury, education, and employment status at the time of

ev

assessment, at the time of injury, and total time employed between the injury and the time of assessment. Data available indicated that 48 (56.5%) cases were

ie

employed, mainly in military service, at time of injury while 32 (37.6%) were not

Comment [DA5]: I moved this information about IQ testing to the participants section above Deleted: The IQ data suggest that while the mean scores were in the average range, the extreme range of scores at the low end may reflect the presence of impairment acquired as a result of TBI. Individuals with IQs within much of this low range are ineligible for enlistment in military service, suggesting that scores within this range reflect acquired cognitive impairment rather than preexisting low intellecutal function. Considering that the age range is compatible with employment, the employment data may reflect the presence of sufficient disability in many of the patients to prevent employment. All patients provided written informed consent to allow their results and clinical records to be used for research purposes in accordance with established hospital policies at that time. ¶ Formatted: Indent: First line: 36 pt

employed at time of injury. Under phenomenology of the injury we obtained

Deleted: for those individuals who were no longer in military service when the injury occurred

information about whether it was a closed or open head injury, its location on the

Deleted: ,

w

On

skull when applicable in regard to lobe or hemisphere, history of a subdural hematoma, existence of visible skull fractures, and cause of injury such as gunshot wound or car accident. Medical findings recorded included

URL: http://mc.manuscriptcentral.com/tbin

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Brain Injury

Page 12 of 48

12 Heterogeneity in TBI

neuroimaging, EEG, and physical neurological examination results, when available. Physical neurological examinations were considered to be abnormal when the results of such an examination were positive for cranial nerves, reflexes, sensory or motor dysfunction or gait disorders. Indications of psychiatric problems in the records were noted (e.g., suicidal behavior) and it was noted when examining clinicians used terminology indicating a neurobehavioral disorder

rP Fo

suggesting dementia, amnesia or aphasia. They all had neurological diagnoses such as post-traumatic encephalopathy indicating localization when appropriate, and other diagnostic information such as subdural hematoma or skull fracture. Individuals without such documentation were not included in the study. For many

Comment [DA6]: The following sentences were moved here, from another part of the document

patients, the terms from DSM-II “Organic Brain Syndrome” or “Chronic Brain

Deleted: The specific details are described in Table 2.

rR

ee

Syndrome” were in common use (17). This procedure resulted in complete data sets for the cluster analyses described below for 80 participants for the analysis of

ev

the HRB and 78 for the analysis of the WAIS-R. However, since the study was retrospective, and clinical and historical data were obtained from clinical records,

ie

there are varying amounts of missing data among the variables studied. Data Analysis

w

Two cluster analyses were conducted one for the WAIS-R and another for

On

the HRB. For both cluster analyses, cases were classified into clusters using

Ward’s method, which is commonly used in neuropsychological studies (18). As

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

is also customary, squared Euclidean distance was used as the similarity measure.

URL: http://mc.manuscriptcentral.com/tbin

Deleted: Insert table 2 about here¶ Deleted: Prior to conducting the cluster analyses, effort was evaluated using the Subtest I and II subtests of the Category Test and the amount of change in error score between Subtests V and VI. More than five errors on subtests I and II and lack of improvement on subtest VI relative to subtest V were examined as indicators of diminished effort. These indexes were used to evaluate effort in the overall sample, as well as among the clusters identified in the cluster analyses. ¶ Formatted: Indent: First line: 36 pt Deleted: probably Deleted: the most Deleted: method Deleted: of several hierarchical agglomerative methods

Page 13 of 48

Brain Injury

13 Heterogeneity in TBI

The number of clusters was determined by two methods. First, preliminary evaluations with varying numbers of cluster solutions aimed at avoiding trivial clusters in a manner similar to “scree testing,” and more definitively plotting clusters in discriminant function space, finding adequate separation among group centroids, as recommended by Aldenderfer and Blashfield (18) in the absence of objective methods for definitive determination of number of clusters. The use of

rP Fo

these heuristic methods does not rule out alternative cluster solutions, but does suggest the location of the point at which further clustering is not productive because of reduced distance among the cluster spaces. We also examined the stability of the cluster solution structure using various algorithms. In this regard,

ee

a frequently used procedure is to compare a hierarchical agglomerative method

rR

with an iterative partitioning method. Thus, we compared Ward’s method with the K-Means method to determine if similar clusters would be present regardless

ev

of the algorithm used to derive them. An analysis cross-tabulating cluster memberships derived from the two solutions was used for this purpose testing the

ie

hypothesis that there was a significant agreement between corresponding clusters obtained from Ward’s method and K-means solutions.

w

The WAIS-R cluster analysis used the eleven subtests of the WAIS-R in

On

order to determine if a similar solution could be found in this sample as had been

reported by van der Heijden and Donders (12). This earlier versions of the WAIS

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

was used because the WAIS-III did not exist during the data collection period. In

URL: http://mc.manuscriptcentral.com/tbin

Deleted: &

Brain Injury

Page 14 of 48

14 Heterogeneity in TBI

order to compare our findings with those of van der Heijden and Donders (12) we examined the traditional three factors derived from the WAIS and WAIS-R, including Verbal Comprehension (VC), Perceptual Organization (PO), and Freedom from Distractibility (FFD). The second cluster analysis used major components of the HRB including the Halstead Category Test (total errors), the Tactual Performance Test (Total

rP Fo

Time, Memory, and Location scores), the Speech Perception Test (errors), the Deleted: ,

Seashore Rhythm Test (errors), Trail Making B (time in seconds), and Finger Tapping (dominant hand-taps), for a total of eight variables. This analysis was accomplished to extend the results of prior studies of the Wechsler scales and

ee

primarily memory tests to a broader range of cognitive abilities using tests that

rR

have been repeatedly demonstrated to be sensitive to TBI (1). Following the completion of the cluster analysis, external validity studies

ev

were conducted for the HRB analysis in which cluster membership was used as an independent variable and inter-cluster differences among the variables obtained

ie

from the clinical records were examined. While satisfactory clusters can be

w

derived mathematically, they may not reflect actual subgroups or types of a disorder unless there has been validation against criteria consisting of relevant

On

variables not included in the cluster analysis itself. This process has been

described as establishing the external validity of a proposed cluster solution (18)

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

and is accomplished through consideration of a number of variables acting as

URL: http://mc.manuscriptcentral.com/tbin

Page 15 of 48

Brain Injury

15 Heterogeneity in TBI

validation criteria, which might be predictive of cluster membership. In cluster analytic work these external validity variables aside from not being included in the cluster analysis itself could reasonably be hypothesized to be associated with cluster membership. Typically, validity level is evaluated by testing for statistical significance among clusters on the variables under consideration. That is, if there is a large effect size for the difference among clusters, that would support the

rP Fo

presence of external validity. In the case of TBI there are clearly a substantial number of variables that would be appropriate for evaluating a proposed cluster solution. As an example, age at time of injury might differ substantially among clusters based only on neuropsychological test performance. For purposes of this

ee

study, these external variables were divided into several components including

rR

demographic and historical factors, the phenomenology of the injury itself including cause and damage inflicted, the immediate outcome, notably presence

ev

of unconsciousness, the long term outcome, notably post-injury employment or educational status, psychiatric consequences and medical/neurological

ie

consequences as assessed by psychiatric and neurological evaluations.

w

Since the HRB tests are scored in different ways, HRB raw test scores were converted to T-scores using normative data corrected for mean age and

On

education (19) for purposes of graphic profiling, but the raw scores were used in the statistical analyses. For comparisons among the clusters on the external

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

validity variables (age, education, age at time of injury etc.), one way analyses of

URL: http://mc.manuscriptcentral.com/tbin

Brain Injury

Page 16 of 48

16 Heterogeneity in TBI

variance (ANOVA) were used for continuous variables and because of small sample sizes in some cells for some variables, Fisher’s exact tests rather than χ2 tests were used for frequency data. In addition to the clinical record items presented here we also had available information about location of the brain injury, particularly in cases in which there was a penetrating wound and surgery. However, preliminary analysis of the data found no significant differences among

rP Fo

the HRB or WAIS-R clusters when left or right hemisphere or frontal, temporal, parietal, or occipital lobe localization were used as dependent measures in the ANOVAs performed for these variables. The details are therefore not provided in the presentation of the external validity results. RESULTS

ee

Cluster Analyses The WAIS/WAIS-R subtests

Deleted: several instances Deleted: Effort was also evaluated using the Subtest I and II subtests of the Category Test and the amount of change in error score between Subtests V and VI. Separate symptom validity tests were not available at the time of acquisition of the data. Forrest, Allen and Goldstein (20) reported that the best Category Test indices were more than five errors on subtests I and II, both of which are simple number identification and counting tasks, and lack of improvement on subtest VI relative to subtest V, both of which require learning of exactly the same principle. It was noted in their study that while patients with structural brain damage showed such improvement, coached malingerers showed less improvement than patients with structural brain damage. It was also found that coached malingerers produced an excessive number of errors on Subtests I and II. These indexes ... were [1] Comment [DA7]: Jerry – I moved this section to the description of the Participants, in the Method ... [2]

ev

rR

In choosing the number of clusters, we compared Ward’s method with the K-Means method and found good agreement (kappa = .64, p < .001) regarding

ie

classification of cases with a three-cluster solution, but not as satisfactory with a

w

four cluster solution. Similarly, when the clusters were plotted in

multidimensional space, separation among cluster centroids was clear for a three

On

Deleted: Effort Testing¶ Analysis of effort indicators for the entire sample indicated that the mean ... [3] Deleted: ¶ An abnormality on the physical neurological examination was noted ... [4] Formatted: Font: Italic Deleted: The psychiatric information indicated that diagnoses used at the time, such as “organic ... [5] Formatted: Font: Italic Formatted: Font: Italic

cluster solution but not as clear for a four cluster solution. We therefore adopted a

Deleted: Description of the Total Sample ¶ A summary of the information... [6]

three cluster solution. The cluster profiles are presented in Figure 1. The first

Deleted: ¶

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Formatted: Font: Italic Deleted:

Page 17 of 48

Brain Injury

17 Heterogeneity in TBI

cluster obtains roughly average mean subtest scores with the exception of Digit Symbol, which is below average. Insert figure 1 about here Deleted: Symbol,

The second cluster has a profile characterized by average range Verbal

Deleted: which is below average.

subtests and substantially impaired Performance subtests, except that Picture

Formatted: Indent: First line: 36 pt

Completion approaches the average range. Recent considerations concerning the

Deleted: subtest scores with the exception of Digit Symbol,

rP Fo

factor analysis of the WAIS have supported a distinction within the PO factor Deleted: , 22

between tests of visual organization and visual reasoning (20, 21). This distinction has been based to a large extent on the addition of new tests such as Matrix Reasoning, but it can be supported to a limited extent in the WAIS/WAIS-

ee

R, since Block Design, Picture Arrangement and Object Assembly all require

rR

some form of conceptual reasoning, while Picture Completion only requires limited analytic reasoning ability and would appear to be more dependent upon

ev

visual organization and knowledge of general information. Since the mean score for Picture Completion is substantially higher than the mean scores for Block

ie

Design, Picture Arrangement and Object Assembly, and since the VC scores in

w

this cluster approach or are at the average range, we would suggest that the second cluster can be characterized as a “Visual Reasoning” subgroup. The third

On

cluster could be characterized as a globally impaired intellectual function cluster as impairment is present on both the Verbal and Performance scales.

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Brain Injury

Page 18 of 48

18 Heterogeneity in TBI

Thus, we have characterized the clusters with the terms Near Normal, Visual Reasoning and Globally Impaired, recognizing that level of performance predominantly separates the clusters. However, there is also evidence for differences in pattern of performance particularly for the Visual Reasoning cluster. There is also some crossing over of subtests among clusters (e.g., Block Design and Object Assembly between Clusters 2 and 3) although these

rP Fo

differences are relatively small. While the Near Normal cluster reflects several average or above intellectual abilities, there still appears to be a deficit in processing speed as evidenced by a low score on Digit Symbol. The Visual Reasoning cluster has well preserved language ability, but relatively impaired

ee

attentional and spatial-constructional skills.

rR

The profile for the WAIS-R factor index scores is also presented in Figure 1. Level of performance provides the main basis for differences among the

ev

clusters, but they have different profiles as well. The Visual Reasoning cluster does better on VCI and FFD than on POI, while POI is slightly higher in the

ie

Globally Impaired cluster. In the Near Normal cluster the mean scores for VCI and POI are higher than FFD. The Halstead-Reitan Battery

w On

Again, a three cluster solution was chosen for the HRB on the same bases as were used for the WAIS-R. The comparison between Ward’s method and K-

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

means solutions yielded a kappa of .92 (p < .001) indicating excellent agreement

URL: http://mc.manuscriptcentral.com/tbin

Page 19 of 48

Brain Injury

19 Heterogeneity in TBI

between the two algorithms. The profile is presented in Figure 2. Table 3 provides quantitative raw score information that may be useful in judging magnitude of impairment and the Russell, Neuringer, and Goldstein (20) rating equivalents for the mean raw scores. These ratings range from 0 indicating excellent performance to 5 indicating severely impaired performance. Use of these ratings is particularly appropriate since they were constructed using a

rP Fo

sample of veterans who constituted the Topeka portion of the sample used in the present study. They provide information concerning severity of impairment as compared with a sample of non-brain damaged veteran patients. Insert figure 2 about here

ee

Insert table 3 about here

rR

As with the Wechsler subtests, separation among the HRB clusters is primarily on the basis of level of performance. There is a Globally Impaired

ev

cluster that does relatively poorly on all of the tests, a cluster that reflects Normal or Close to Normal performance and an Intermediate group that appears to have

ie

moderately impaired function in some cases performing more like the Near

w

Normal cluster and in some more like the Globally Impaired cluster. The Globally Impaired cluster appears to be most impaired in processing speed as measured by

On

the Trail Making Test, and shows little more impairment than the Moderate cluster on scores obtained from the Tactual Performance Test. Comparisons between the WAIS-R and HRB clusters

URL: http://mc.manuscriptcentral.com/tbin

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Formatted: Font: Bold, Not Italic Formatted: Font: Bold

Brain Injury

Page 20 of 48

20 Heterogeneity in TBI

In order to determine to what extent the WAIS-R and HRB classified the same cases into comparable subgroups, cluster memberships were cross-tabulated. The results are presented in Table 4. The degree of agreement was low (kappa = .21, p = .006) with the HRB classifying substantially more cases as impaired than did the WAIS-R. Overall, twenty five cases (31.6%) were placed into the Near Normal cluster by the HRB while 41 cases (51.9%) were placed in

rP Fo

the Near Normal cluster in the WAIS-R cluster analysis. Thus, since all of the cases had documented brain injuries, it is apparent that the HRB has greater sensitivity to its presence than the WAIS-R. Insert table 4 about here

ee External Validity

Formatted: Font: Bold, Not Italic

rR

Given this greater sensitivity to brain damage external validity data are

ev

only presented for the HRB cluster analysis. The results contained in Table 5 present a pattern of significant and non-significant differences among clusters.

ie

Neurological factors such as type, cause, and location of injury, or abnormalities

w

found on physical examination, skull-X-Ray, or EEG were not associated with significant intercluster differences. On the other hand, significant differences were

On

found for sociodemographic variables including age at time of assessment, age at

time of injury, and employment status. The Globally Impaired cluster was older at time of injury and assessment and had a lower percentage of employed

URL: http://mc.manuscriptcentral.com/tbin

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 21 of 48

Brain Injury

21 Heterogeneity in TBI

individuals than was the case for the other clusters. Education was of borderline significance (p = .08) indicating that the Globally Impaired cluster had a lower mean educational level than did the other clusters. Deleted: With regard to differences in effort among the clusters, no significant differences were present for

Insert table 5 about here

Deleted: effort, the mean score for

Discussion

rP Fo

Deleted: Category Test Subtests 1 and 2,

A sample of veterans who sustained TBI was found to have an outcome pattern largely involving marked variation in level, with some indication of differences in pattern, of cognitive function. Three cluster solutions were identified for both analyses of WAIS/WAIS-R and HRB data with level of performance being the major consideration separating the clusters. However, in

rR

an analysis performed using the 11 WAIS-R subtests, one of the clusters had a clear performance pattern marked by relatively intact verbal abilities and impaired

ev

visual reasoning abilities. In their WAIS-III cluster analytic study, van der Heijden and Donders (12) also proposed a three-cluster solution but it was

ie

essentially entirely based upon level of performance without a dissociation

With regard to generalizeabilty of the current findings, while the current

On

sample consisted entirely of veterans, it had different characteristics from those involved in numerous studies of veterans with focal brain wounds sustained in

URL: http://mc.manuscriptcentral.com/tbin

ly

combat, notably the studies of the Teuber group (e.g., 21). Only 10 of our

Deleted: of 5.35 (SD=5.40) for the total sample again with a nonsignificant difference among the clusters. Errors on Subtests I and II were substantially lower than those obtained by Forrest et al.’s (20) coached malingering group and improvement from Subtest V to VI is also greater. Not all participants improved on Subtest VI relative to V, but only 6 participants made more than one more error on Subtest VI relative to V while 80% of the participants showed some improvement. Deleted: Thus, overall findings and intercluster differences are not readily attributable to effort related considerations.

w

between verbal abilities and visual reasoning.

Deleted: for the total sample was .14 (SD=.47) and .54 (SD=.83) for Subtest II with nonsignificant differences among the clusters. There was a Deleted: or for the mean improvement on Subtest V relative to VI.

ee

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Deleted: , Deleted: proposed Deleted: ¬ Deleted: Deleted: W

Brain Injury

Page 22 of 48

22 Heterogeneity in TBI

patients sustained brain injury from gun shot or shrapnel wounds, a proportion that is probably higher than what would occur in the general population (1) but clearly not common in this sample. These results may enhance the generalizability of the findings since the proportions of various causes of brain injury would appear to be little different between what is found in veterans and non-veterans. It is also likely that the sample obtained represented a substantially

rP Fo

impaired subgroup of a larger population containing less ill or disabled Deleted: they

individuals. Furthermore, subjects were tested while hospital inpatients for an event unrelated to the acute trauma, including health problem associated with the brain jury (e.g., seizure control). Nevertheless one of the clusters obtained, for

ee

both the Wechsler and HRB analyses, identified a reasonably intact subgroup

rR

consisting of individuals functioning in the average range of general intelligence, and with adequate problem solving abilities. As a group they did show slowness

ev

of processing speed, but in isolation from other relatively intact abilities. However, the Globally Impaired clusters contain individuals with rather

ie

devastating, generalized cognitive dysfunction. Only about a third of these

w

individuals were employed at the time of testing as opposed to 72% in the Near Normal cluster.

On

Comparing the present cluster analytic findings with those of van der

Heijden and Donders (12) indicates that subgroups described shortly following brain injury appear to persist on a long-term basis. However, the WAIS-R

URL: http://mc.manuscriptcentral.com/tbin

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 23 of 48

Brain Injury

23 Heterogeneity in TBI

analysis was particularly helpful in characterizing what we have called the Visual Reasoning cluster, which did not appear in the van der Heijden and Donders study. Individuals in this cluster appeared to have suffered from substantial impairment of complex problem solving abilities, as measured by such procedures as Block Design and the Category Test. However, they seemed to have preserved or recovered their language abilities, and have average level vocabularies and

rP Fo

Deleted: really

language comprehension. Van der Heijden and Donders did not have a Visual Reasoning cluster as identified in our cluster analysis, but a subgroup that could be better characterized as having moderate generalized impairment with relatively greater impairment on the PS factor index. This cluster was similar to the one

ee

found here if one uses WM as the equivalent of FFD. That is WM was slightly

rR

better than VC and PO. However, there was not a substantial split between the VC and PO indices as was found here. Therefore, while the differences in profiles

ev

from both studies can be generally attributed to level of performance differences, there is an indication that pattern of performance also distinguishes the clusters in

ie

the case of the present study. Our Near Normal and Global profiles are quite

w

similar to those obtained by van der Heijden and Donders (12). Their highest functioning cluster had higher VCI and POI mean scores, while their lowest

On

functioning cluster had a slightly higher POI than VCI score, both being higher than the FFD equivalent Working Memory (WM) and Perceptual Speed (PS)

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

indices.

URL: http://mc.manuscriptcentral.com/tbin

Brain Injury

Page 24 of 48

24 Heterogeneity in TBI

One possibility is that the difference between the cluster patterns between the van der Heijden and Donders and present study had to do with time since injury. Shortly after injury, both language and problem solving abilities may be impaired but as time passes, it is possible that language recovers substantially in some individuals, but reasoning and problem solving abilities do not. This dissociation between relatively intact language and impaired problem solving

rP Fo

abilities was made possible by considering both the WAIS/WAIS-R and the HRB in the neuropsychological evaluation. The HRB based analysis provided information concerning status of reasoning and problem solving abilities in the three clusters, while the WAIS-R analysis identified a dissociation in a group of

ee

individuals between relatively intact and impaired abilities. The HRB data for the

rR

Moderate cluster did not reflect a dissociation between impaired and intact abilities. All of the Russell, Neuringer and Goldstein ratings reflected mild or

ev

greater impairment and none of the Heaton et al. (19) mean T scores for the Moderate cluster reached 50. However, several of the mean WAIS-R subtest

ie

standard scores fell within one standard deviation from the mean for the general population.

w

Neurological examinational methods did not appear to separate the

On

clusters, possibly because of low sensitivity of these methods. In the total sample only 32% of the EEGs and 19% of the skull X-rays were abnormal. It is possible

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

that application of advanced neuroradiological techniques may have altered

URL: http://mc.manuscriptcentral.com/tbin

Page 25 of 48

Brain Injury

25 Heterogeneity in TBI

matters, although the extensive recent literature in this area is somewhat inconclusive. For example, Lewine, Davis, Bigler, Thoma, Hill, Funke, Sloan, Hall, and Orrison (22) reported that only four of their 30 patients with TBI had abnormal MRI findings. Considering the external validity data utilized in this study, the variables that best separated the clusters were demographic in nature. As might be

rP Fo

anticipated, individuals in the globally impaired clusters were least likely to be employed, but were also older than members of the other clusters. Individuals who were relatively young at the time of injury were more likely to be in the Near Normal clusters. Education also significantly separated the clusters in the expected direction.

rR

ee

Part of the explanation for the positive findings for demographic external validity variables may be that the cognitive reserve associated with young age and

ev

advanced education may protect the brain from a poor outcome (23). The participants in the Near Normal cluster were younger and better educated than

ie

was the case for the other clusters. Thus, having a TBI in an individual who is

w

relatively young and well educated may presage a better outcome than if the reverse were true.

On

A final consideration was the performance of the overall sample and the

individual clusters on the embedded measures of effort. While the importance of

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

the evaluation of effort has been emphasized in studies of patients with TBI who

URL: http://mc.manuscriptcentral.com/tbin

Brain Injury

Page 26 of 48

26 Heterogeneity in TBI

are seeking legal remedies to cognitive disabilities associated with their brain injuries, there has not been a systematic attempt to determine whether variability in effort accounts for the neurocognitive clusters identified in prior studies. To address this issue, the current investigation examined embedded measure of effort in the Category test and found that effort had little to do with cluster membership. No differences were present among the clusters on our effort indices, suggesting

rP Fo

that the neurocognitive heterogeneity noted in TBI does not results from differing levels of motivation between the clusters. Several conclusions can be suggested from this study. Cognitive outcome following the acute phase of injury is heterogeneous, but does not appear to be

ee

associated with severity of injury. This result runs parallel with the finding that

rR

psychosocial outcome is also not associated with severity (28). Cluster analyses of extensive neuropsychological assessments separated the cases studied into

ev

three subgroups varying substantially in level of performance, but also in pattern to a lesser extent. A Globally Impaired subgroup demonstrated generalized

ie

cognitive impairment resulting in an apparent diminution of general intelligence

w

and a number of adaptive abilities assessed by the HRB. Consistent with the literature, processing speed was particularly impaired. A Moderately Impaired

On

cluster produced a higher level of performance with reasonably good preservation of language or crystallized skills and relatively poor performance on tests of

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

problem solving. The term moderate would appear to be appropriate for the HRB

URL: http://mc.manuscriptcentral.com/tbin

Page 27 of 48

Brain Injury

27 Heterogeneity in TBI

based cluster analysis, but in the case of the WAIS-R, what we have called the “Visual Reasoning” cluster performed at an average or close to average level on measures of some abilities and at a clearly impaired level at other abilities. These individuals cannot be reasonably described as moderately impaired in general since they maintain some intact abilities as well as some significantly impaired abilities. A third cluster performed at an average level on all of the WAIS-R

rP Fo

subtests with the exception of measures of attention and processing speed, particularly as reflected by the FFD Wechsler factor index score, and did not receive a worse than mildly impaired HRB rating. A number of findings that could be interpreted as indirect indicators of severity did not effectively separate

ee

the clusters. However, significant separation was obtained for age at injury, age at

rR

assessment, and employment status. Borderline significance was obtained for education. However, it might be of interest that only one of the 25 Near Normal

ev

cluster members had less than a 10th grade education (4%), 8 of the 18 Globally Impaired cluster members (44.4%) had less than a 10th grade education.

ie

This type of retrospective study provides a long-term perspective on the

w

outcome of brain injury well beyond its acute phase. A retrospective design has the obvious limitations of unavailability of modern technologies. However, the

On

opportunity to study the long term outcome of TBI among patients who have been evaluated sometimes over many years does not occur frequently, and can provide

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

insights not obtainable from acute studies since it addresses issues associated with

URL: http://mc.manuscriptcentral.com/tbin

Brain Injury

Page 28 of 48

28 Heterogeneity in TBI

persistence of TBI related symptoms over a long time period. Thus, the limitations might not outweigh the advantages of retrospective research and does not compromise the general conclusion that outcome of brain injury is diverse but may be organized into subtypes utilizing classification statistics such as cluster analysis. Nevertheless, these limitations are significant, particularly regarding the absence of consistently reliable and complete information about the acute

rP Fo

histories, particularly regarding period of unconsciousness and posttraumatic amnesia, as well as because of absence of modern diagnostic procedures. We would therefore encourage pursuit of prospective studies that systematically investigate the complexities of determination of outcomes identified here utilizing

ee

contemporary methods and technologies.

w

ie

ev

rR ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Page 29 of 48

Brain Injury

29 Heterogeneity in TBI

References 1. Reitan, R. M., Wolfson, D. Mild head injury: Intellecutal, cognitive, and emotional consequences. Tucson, AZ: Neuropsychology Press; 2000. 2. Allen, DN, Mayfield J, Strauss G P. Neuropsychological subtypes of childhood traumatic brain injury. Presented at the meeting of the American Psychological Association, New Orleans, LA. 2006, August.

rP Fo

3. Chan, R. C., Hoosain, R., Lee, T. M., Fan, Y. W., Fong, D. Are there sub-types of attentional deficits in patients with persisting post-concussive symptoms? A cluster analytical study. Brain Injury 2003; 17:131-148. 4. Curtiss, G., Vanderploeg, R. D., Spencer, J. Patterns of verbal learning and

ee

memory in traumatic brain injury. Journal of the International Neuropsychological Society 2001; 7: 574-585.

rR

5. Demery, J. A., Pedraza, O., Hanlon, R. E. Differential profiles of verbal

ev

learning in traumatic brain injury. Journal of Clinical and Experimental Neuropsychology 2002; 24: 818-827.

ie

6. Deshpande, S. A., Millis, S. R., Reeder, K. P., Fuerst, D., Ricker, J. H. Verbal

w

learning subtypes in traumatic brain injury: a replication. Journal of Clinical and Experimental Neuropsychology 1996; 18: 836-842.

On

7. Greve, K. W., Love, J. M., Sherwin, E., Mathias, C. W., Ramzinski, P. Levy, J. Wisconsin Card Sorting Test in chronic severe traumatic brain injury: factor structure and performance subgroups Brain Injury 2002;16:29-40.

URL: http://mc.manuscriptcentral.com/tbin

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Brain Injury

Page 30 of 48

30 Heterogeneity in TBI

8. Lange, R. T., Iverson, G. L., Franzen, M. D. Neuropsychological functioning following complicated vs. uncomplicated mild traumatic brain injury. Brain Injury 2009; 23:83-91. 9. Millis, S. R., Ricker, J. H. Verbal learning patterns in moderate and severe traumatic brain injury. Journal of Clinical and Experimental Neuropsychology 1994; 16: 498-507.

rP Fo

10. Moore, A. D., Stambrook, M. Peters, L. C. Coping strategies and adjustment after closed-head injury: a cluster analytical approach Brain Injury 3;1989:171175.

11. Mottram, L., Donders, J. Cluster subtypes on the California verbal learning

ee

test-children's version after pediatric traumatic brain injury. Developmental

rR

Neuropsychology 2006; 30: 865-83.

12. van der Heijden, P., Donders, J. (2003). WASI-III factor index score patterns

ev

after traumatic brain injury. Assessment 2003; 10: 115 – 122. 13. Wiegner, S., Donders, J. Performance on the California Verbal Learning Test

ie

after traumatic brain injury. Journal of Clinical and Experimental Neuropsychology 1999; 21: 159-170.

w

14. Wechsler D. WAIS-III: Wechsler Adult Intelligence Scale-Third Edition. San Antonio TX: The Psychological Corporation; 1997.

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Page 31 of 48

Brain Injury

31 Heterogeneity in TBI

15. Reitan, R. M., Wolfson, D. The Halstead-Reitan Neuropsychological Test Battery: Theory and clinical interpretation (2nd ed.). Tucson, AZ: Neuropsychological Press; 1993. 16. Teiler, A., Adams, K. M., Walker, A. E. Rourke, B. P. Long-term effects of severe penetrating head injury on psychosocial adjustment. Journal of Consulting and Clinical Psychology, 1998; 58:531-537.

rP Fo

17. American Psychiatric Association. Diagnostic and statistical manual of mental disorders (2nd ed.) Washington DC: Author; 1952. 18. Aldenderfer, M, S., Blashfield, R.K. Cluster analysis. Beverly Hills, CA: Sage Publications; 1984.

ee

19. Heaton, R. K., Miller, S. W., Taylor, M. J., Grant I. Revised comprehensive

rR

norms for an expanded Halstead-Reitan Battery. Lutz, Fl: Psychological Assessment Resources, Inc.; 2004.

ev

20. Russell, E. W., Neuringer, C., Goldstein, G Assessment of brain damage. New York: Wiley-Interscience; 1970.

ie

21. Teuber, H.-L. (1964). The riddle of frontal lobe function in man. In Warren

w

JM, & Akert K, editors. The frontal granular cortex and behavior New York: McGraw-Hill 1964, pp 410-444.

On

22. Lewine, J. D., Davis, J. T., Bigler, E. D., Thoma, R., Hill, D., Funke, M.,

Sloan, J. H., Hall, S. Orrison, W. W. Objective documentation of traumatic brain

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Brain Injury

Page 32 of 48

32 Heterogeneity in TBI

injury subsequent to mild head trauma: Multimodal brain imagining with MEG, SPECT, and MRI. Journal of Head Trauma Rehabilitation 2007; 22: 141 – 155. 23. Stern, Y. What is cognitive reserve? Theory and research application of the reserve concept. Journal of the International Neuropsychological Society 2002; 8: 448-460.

w

ie

ev

rR

ee

rP Fo ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Page 33 of 48

Brain Injury

33 Heterogeneity in TBI

Figure Captions and Legends Figure 1. Performance on WAIS-R subtests and Index Scores as a function of cluster membership: I –Information; C = Comprehension; A = Arithmetic; D = Digit Span = V = Vocabulary; DS = Digit Symbol; PC = Picture Completion; BD

rP Fo

= Block Design; PA = Picture Arrangement; OA = Object Assembly; VCI = Verbal Comprehension Index; POI = Perceptual Organization Index; FFD = Freedom From Distractibility Index Figure 2. Performance on Halstead-Reitan Battery as a function of cluster

ee

membership: CAT = Category Test; TPT = Tactual Performance Test-Total Time; M = Tactual Performance Test-Memory; L = Tactual Performance Test-Location;

rR

SP = Speech Perception Test; RHY = Rhythm Test; TRB = Trail Making Test, Part B; TAP = Finger Tapping-Right Hand

w

ie

ev ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Brain Injury

Page 34 of 48

34 Heterogeneity in TBI

rP Fo 12 11

9

ee

8

Visual Normal Global

rR

Scaled Scores

10

7 6

4

D FF I PO I

A

VC

O

w

WAIS-R Subtests

PA

BD

PC

DS

V

D

S

A

C

ie

ev

5

I

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Page 35 of 48

Brain Injury

35 Heterogeneity in TBI

60

rP Fo

T Scores

50

Global Moderate Normal

40

20

B

Y

PR TA

TR

RH

ev

SP

L

M

T TP

rR

ee

30

AT C

HRB Tests

w

ie ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Brain Injury

Page 36 of 48

36 Heterogeneity in TBI

Table 1 Demographic and IQ Data Variable Age Years of Education Verbal IQ Performance IQ Full Scale IQ Months Between TBI and Testing

Mean 34.6 11.8 94.3 90.8 92.2 80.0

SD 12.2 2.2 14.1 15.1 13.5 98.2

Deleted: 5.0 Deleted: 3 Deleted: 9 Deleted: 0 Deleted: 9

Deleted: 1

N 77 44 57 41

% 98.7 56.4 73.1 52.6

Deleted: 4 Deleted: 78.1 Deleted: 7.7 Deleted: 2 Deleted: 86 Deleted: 9 Deleted: 8.7 Deleted: 8

ev

rR

Male Participants Employed At Time of Injury Employed Since Injury Employed at Time of Assessment

Range 19-63 3-16 46-121 53-129 60-120 3-400

Deleted: 0

ee

rP Fo

Deleted: 7.3 Deleted: 2 Deleted: 4.5

w

ie ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Page 37 of 48

Brain Injury

37 Heterogeneity in TBI

Table 2 Clinical variables %

36 44 11 24

45.0 55.0 13.8 30.4

9 3 2 47 3 4 7 5

11.4 3.8 2.6 60.3 3.8 5.1 8.9 6.3

16 25 21 22 14 29 61

20.3 31.3 26.3 25.9 17.5 36.6 70.1

46 10 6 2 6 8 6

57.5 12.5 17.1 2.5 7.5 9.2 7.6

w

ie

ev

rR

ee

N

ly

On

Variable Type of Injury Open Head Closed Head Subdural Hematoma Multiple Fractures Cause of Injury Gun Shot Wound Blast Assault Vehicle Accident Home Accident Industrial Accident Fall Unknown Positive Diagnostic Procedures Abnormal Skull X-Ray Abnormal EEG Abnormal Physical Neurological Examination History of Seizures Visible Skull Marking Received Surgery Reported to Be Unconscious Positive Psychiatric Findings Diagnosis of Organic Mental Disorder Depression Anxiety Suicidality Disinhibition Alcohol or Drug Abuse History Psychosis History

rP Fo

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Brain Injury

Page 38 of 48

38 Heterogeneity in TBI

Table 3 Comparisons of HRB Raw Score Results Among the Three Clusters Test

Cluster

Mean1

SD

RNG Rating2

T-Scores3

CAT Errors

Moderate Normal Severe Total Moderate Normal Severe Total Moderate Normal Severe Total Moderate Normal Severe Total Moderate Normal Severe Total Moderate Normal Severe Total Moderate Normal Severe Total Moderate Normal Severe Total

74.2 60.0 103.9 76.5 24.7 11.3 27.2 21.1 5.8 8.1 5.2 6.4 1.9 4.5 1.1 2.5 13.1 8.7 22.6 13.9 7.5 6.2 11.8 8.1 118.7 73.4 252.3 134.6 42.8 47.8 41.9 44.2

28.2 29.6 25.5 32.0 48.2 41.0 37.0 79.8 2.0 1.2 1.6 2.1 2.1 2.6 1.4 2.5 8.8 5.0 6.8 8.9 3.7 3.8 3.9 4.3 38.0 19.2 45.3 75.3 10.4 7.3 10.6 9.8

2 2 3 3 3 1 3 3 1 1 2 1 3 2 3 2 2 2 3 2 2 2 3 2 2 1 4 3 2 2 3 2

35 39 27 35 32 48 32 36 37 53 33 37 8-41 52 37 45 32 40 24 32 37 40 31 37 36 50 23 32 35 44 31 35

TPT Minutes

M Correct

L Correct

w

ly

On

TAPD Taps

ie

TRB Second

ev

RHY Errors

rR

SP Errors

ee

rP Fo

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Page 39 of 48

Brain Injury

39 Heterogeneity in TBI

(table 3 continued) 1

All intercluster differences were statistically significant (p>001) with the exception of Tapping (p = .054) 2 Russell, Neuringer and Goldstein (1970) Ratings: 1=Normal; 2=Mildly Impaired; 3=Moderately Impaired 3 Heaton, Miller, Taylor, & Grant I. (2004). Note: CAT = Category Test; TPT=Tactual Performance Test (Time); M=TPT (Memory); L= TPT (Location); SP= Speech Perception; RHY=Rhythm Test; TRB=Trail Making B; TAPD= Finger Tapping (Dominant Hand).

w

ie

ev

rR

ee

rP Fo

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Brain Injury

Page 40 of 48

40 Heterogeneity in TBI

Table 4 Relations Between WAIS-R and HRB Based Cluster Analyses HRB Based Clusters Normal Moderate Severe WAIS Based Normal 23 17 1 Clusters Visual 0 9 5 Global 2 10 10 Total 25 36 16 (χ2 (4) = 31.06, p<.001)

Total 41 14 22 77

Deleted: 2 Deleted: 4 Deleted: 8 Deleted: 9

rP Fo

Deleted: 3.88

w

ie

ev

rR

ee ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Page 41 of 48

Brain Injury

41 Heterogeneity in TBI

Table 5 External Validity Data for HRB Cluster Solution Cluster Formatted Table

Moderate N=37

Normal N=25

Interval Data Age Years of Education Age at Injury Months Between Injury and Assessment Nominal Data

12.00+2.07 26.0+9.6 81.92+102.7

... [7]

Deleted: .08 6…2

... [8]

.05 4…6…1.2 Deleted:

... [9]

.25 98 Deleted: Formatted: Underline

51.4

72.0

33.3

79.4 55.9

76.0 76.0

78.6 42.9

.20 4.61

43.2

44.0

56.3

.65+133.3

... [10]

χ2

Deleted: .93 5.0

... [11]

.11 37.5 Deleted:

... [12]

Deleted: 0.0

.86

.67 24 Deleted:

... [13]

56.8

56.0

43.8

13.5 36.1

12.0 20

18.8 25.0

.56 1.91

Deleted: .83 16.7

12.0 44.0 12.0

18.8 60.0 6.3

1.36 3.91 .51

Deleted: .47 16.7

... [15]

.15 Deleted: .88 Deleted:

4.7

... [16]

5.6

... [17]

8.3 69.4 8.3

ie

Gun Shot Wound Vehicle Accident Fall

2.61 5.14 .94

p Deleted: .001 3…5…0

Exact P Test Deleted: 5.85 Deleted: .06

ev

Subdural Hematoma Multiple Fractures Cause of Injury

12.40+1.4 10.8+3.3 24.9+10.9 32.7+10.5 56.58+62.12 111.53 +137.2

rR

Closed Injury

F 7.85

42.8+10.4

%

ee

Employed at Assessment Employed Since Injury Employed at Injury Type of Injury Open Injury

Global n=18

Mean and SD 29.6+10.1

32.2+11.3

rP Fo

Deleted: 50.0 ... [14]

.39 33.3 Deleted:

Positive Diagnostic Procedures and Observations

24.3 24.3 21.6

28.0 34.8 16.0

18.8 18.8

Exact Test .27 1.37

31.3 25.0 12.5

URL: http://mc.manuscriptcentral.com/tbin

.41 2.07 ..63

ly

Abnormal Examination History of Seizures Visible Skull Marking

21.6 35.1

% 16.7 32.0

On

Nominal Data Abnormal Skull X-Ray Abnormal EEG

w

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Deleted: p

…χ2 …

... [18]

.93 22.2 Deleted:

... [19]

.59 22.2 Deleted:

... [20]

Deleted: .85 27.8

... [21]

.88 22.2 Deleted:

... [22]

.80 11.1 Deleted:

... [23]

Brain Injury

Page 42 of 48

42 Heterogeneity in TBI

Received Surgery (table continues)

35.1

36.0

43.8

% Reported Unconscious Positive Psychiatric Findings Organic Mental Disorder Depression Anxiety Disinhibition Alcohol Abuse Drug Abuse Psychosis History

rP Fo

67.6

76.0

69.2

62.2 8.1 15.4 5.4 5.4 0.0 8.1

52.0 16.0 23.1 12.0 12.0 4.0 4.2

62.5 18.8 12.5 6.3 6.3 6.3 12.5

.08 Exact Test .63 .77 1.72 .54 1.1 1.1 2.48 1.1

Deleted: .96 38.9 Deleted: p χ2

.73 Formatted: Left Deleted:

.73 .45 Deleted: 1.0 Deleted: .75 Deleted: .75 .27 Deleted: .75 Deleted: Deleted:

66.7 55.6 .67 .72 6.7

Deleted: 1.22 Deleted: .54 Deleted: 1.1 Deleted: .58 Deleted: .75

ee

Deleted: 5.6 Deleted: 1.06 Deleted: .59

rR

Deleted: or Drug Abuse Deleted: 11.1 Deleted: .98 Deleted: .61

ev

Deleted: 1.1 Deleted: .73 Deleted: .69

w

ie ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Page 43 of 48

Brain Injury

43 Heterogeneity in TBI

Acknowledgements: This research was supported by the DVA VISN IV Mental Illness Research, Education and Clinical Center, VA Pittsburgh Healthcare System, Pittsburgh, PA and the Medical Research Service, Department of Veterans Affairs. Declaration of Interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

w

ie

ev

rR

ee

rP Fo

ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

URL: http://mc.manuscriptcentral.com/tbin

Brain Injury

Page 16: [1] Deleted

DAniel Allen

11/24/2009 5:43:00 PM

Effort was also evaluated using the Subtest I and II subtests of the Category Test and the amount of change in error score between Subtests V and VI. Separate symptom validity tests were not available at the time of acquisition of the data. Forrest, Allen and Goldstein (20) reported that the best Category Test indices were more than five errors on subtests I and II, both of which are simple number identification and counting tasks, and

rP Fo

lack of improvement on subtest VI relative to subtest V, both of which require learning of exactly the same principle. It was noted in their study that while patients with structural brain damage showed such improvement, coached malingerers showed less improvement

ee

than patients with structural brain damage. It was also found that coached malingerers produced an excessive number of errors on Subtests I and II. These indexes were used to

rR

evaluate effort in the present data set and were treated as external validity criteria. Page 16: [2] Comment [DA7]

DAniel Allen

11/24/2009 10:34:00 AM

ev

Jerry – I moved this section to the description of the Participants, in the Method section, as that is what the reviewers seemed to want. Page 16: [3] Deleted

Effort Testing

Gerald Goldstein

11/30/2009 11:36:00 AM

w

ie

Analysis of effort indicators for the entire sample indicated that the mean score for Category Test Subtest 1 for the total sample was .14 (SD = .47) and .54 (SD = .83) for

On

Subtest II. There was a mean improvement on Subtest V relative to VI of 5.35 (SD = 5.40) for the total sample. In comparison, errors on Subtests I and II were substantially

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 44 of 48

lower than those obtained by Forrest et al.’s (20) coached malingering group and improvement from Subtest V to VI is also greater. Not all participants improved on Subtest VI relative to V, but only 6 participants made more than one error on Subtest VI relative to V while 80% of the participants showed some improvement. Thus, the findings suggest that overall the sample exhibited adequate effort.

URL: http://mc.manuscriptcentral.com/tbin

Page 45 of 48

Page 16: [4] Deleted

Gerald Goldstein

11/30/2009 11:36:00 AM

An abnormality on the physical neurological examination was noted when the results of such an examination appeared in the records and were positive for cranial nerves, reflexes, sensory or motor dysfunction or gait disorders. Presence of scars on the head from the injury itself or surgery was also noted. Other available diagnostic

rP Fo

procedures (e.g. pneumoencephalogram) were rarely used. As can be seen in Table 2, these variables were positive in a minority of cases, with having received surgery being the most common positive sign (36.6%). Page 16: [5] Deleted

ee

Gerald Goldstein

11/30/2009 11:36:00 AM

The psychiatric information indicated that diagnoses used at the time, such as “organic

rR

brain syndrome” was given in 46 cases, but the prevalence of other disorders was relatively low. Page 16: [6] Deleted

ev

Gerald Goldstein

Description of the Total Sample

11/30/2009 11:36:00 AM

ie

A summary of the information obtained from the clinical records are contained in

w

Table 2. Using the definition employed here of open head injury alternatively requiring

On

significant skull fracture, surgical debridement or related procedures, presence of a subdural hematoma, or observation of brain tissue extrusion to the surface of the skull,

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Brain Injury

almost half the sample had this condition with the others having closed head injury. By far, the major cause of injury was car accident. About a third of the participants received surgery, and about a quarter of them had seizures following the TBI. 71% of the participants had a period of unconsciousness reported, but reliable information concerning actual length of time unconsciousness was not generally available. The

URL: http://mc.manuscriptcentral.com/tbin

Brain Injury

Glasgow Coma Scale or related procedures were not in common use at the time patients in the sample sustained their injuries. Page 41: [7] Deleted

Gerald Goldstein

11/19/2009 12:58:00 PM

3 Page 41: [7] Deleted

Gerald Goldstein

11/19/2009 12:58:00 PM

5 Page 41: [7] Deleted

Gerald Goldstein

rP Fo

Page 41: [7] Deleted Page 41: [7] Deleted Page 41: [8] Deleted Page 41: [8] Deleted

Gerald Goldstein

11/19/2009 1:02:00 PM

9.81 Gerald Goldstein

11/19/2009 1:03:00 PM

0 Gerald Goldstein

11/19/2009 12:58:00 PM

6

Gerald Goldstein

11/19/2009 12:59:00 PM

2

rR

Page 41: [8] Deleted

11/19/2009 12:58:00 PM

0

ee

Gerald Goldstein

11/19/2009 1:03:00 PM

3.92

Page 41: [8] Deleted

Gerald Goldstein

11/19/2009 1:03:00 PM

.024

Page 41: [9] Deleted

ev

Gerald Goldstein

11/19/2009 12:59:00 PM

4

Page 41: [9] Deleted

Gerald Goldstein

6

Gerald Goldstein

1.2 Page 41: [9] Deleted

w

Page 41: [9] Deleted

Gerald Goldstein Gerald Goldstein

.65+133.3 Gerald Goldstein

11/19/2009 1:03:00 PM 11/19/2009 1:00:00 PM 11/19/2009 1:03:00 PM

ly

Page 41: [10] Deleted

11/19/2009 12:59:00 PM

On

08 Page 41: [10] Deleted

11/19/2009 12:59:00 PM

ie

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 46 of 48

.40 Page 41: [11] Deleted

Gerald Goldstein

11/19/2009 11:23:00 AM

5.0 Page 41: [11] Deleted

Gerald Goldstein

11/19/2009 10:23:00 AM

16 Page 41: [11] Deleted

Gerald Goldstein

11/19/2009 10:23:00 AM

2 Page 41: [12] Deleted

Gerald Goldstein

11/19/2009 11:27:00 AM

37.5 Page 41: [12] Deleted

Gerald Goldstein

11/19/2009 10:23:00 AM

URL: http://mc.manuscriptcentral.com/tbin

Page 47 of 48

6.16 Page 41: [12] Deleted

Gerald Goldstein

11/19/2009 10:23:00 AM

05 Page 41: [13] Deleted

Gerald Goldstein

11/19/2009 10:24:00 AM

24 Page 41: [13] Deleted

Gerald Goldstein

11/19/2009 10:24:00 AM

.89 Page 41: [14] Deleted

Gerald Goldstein

11/19/2009 11:32:00 AM

16.7

rP Fo

Page 41: [14] Deleted Page 41: [14] Deleted Page 41: [15] Deleted Page 41: [15] Deleted

Gerald Goldstein

Page 41: [15] Deleted

Gerald Goldstein

11/19/2009 10:26:00 AM

91 Gerald Goldstein

11/19/2009 11:34:00 AM

16.7 Gerald Goldstein

11/19/2009 10:27:00 AM

.84

Gerald Goldstein

11/19/2009 10:27:00 AM

.66

rR

Page 41: [16] Deleted

11/19/2009 10:26:00 AM

.20

ee

Gerald Goldstein

11/19/2009 11:36:00 AM

4.7

Page 41: [16] Deleted

Gerald Goldstein

11/19/2009 10:29:00 AM

ev 4.17

Page 41: [16] Deleted

Gerald Goldstein

2

Page 41: [17] Deleted

Gerald Goldstein Gerald Goldstein

6

Gerald Goldstein

76

11/19/2009 10:30:00 AM

On

Page 41: [17] Deleted

11/19/2009 11:36:00 AM

w

5.6 Page 41: [17] Deleted

11/19/2009 10:29:00 AM

ie

11/19/2009 10:30:00 AM

Page 41: [18] Deleted

Gerald Goldstein

11/19/2009 10:31:00 AM

Page 41: [18] Deleted

Gerald Goldstein

11/19/2009 10:30:00 AM

Page 41: [18] Deleted

Gerald Goldstein

11/19/2009 10:31:00 AM

Page 41: [19] Deleted

Gerald Goldstein

11/19/2009 11:37:00 AM

ly

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Brain Injury

χ2

22.2 Page 41: [19] Deleted

Gerald Goldstein

11/19/2009 10:31:00 AM

8 Page 41: [19] Deleted

Gerald Goldstein

11/19/2009 10:32:00 AM

87

URL: http://mc.manuscriptcentral.com/tbin

Brain Injury

Page 41: [20] Deleted

Gerald Goldstein

11/19/2009 11:38:00 AM

22.2 Page 41: [20] Deleted

Gerald Goldstein

11/19/2009 10:32:00 AM

.95 Page 41: [20] Deleted

Gerald Goldstein

11/19/2009 10:32:00 AM

62 Page 41: [21] Deleted

Gerald Goldstein

11/19/2009 11:53:00 AM

27.8 Page 41: [21] Deleted

rP Fo

Page 41: [21] Deleted Page 41: [22] Deleted Page 41: [22] Deleted

Gerald Goldstein

Page 41: [22] Deleted

Gerald Goldstein

11/19/2009 10:33:00 AM

94 Gerald Goldstein

11/19/2009 11:16:00 AM

22.2 Gerald Goldstein

11/19/2009 10:34:00 AM

2.1

Gerald Goldstein

11/19/2009 10:35:00 AM

.72

Gerald Goldstein

11/19/2009 11:39:00 AM

rR

Page 41: [23] Deleted

11/19/2009 10:33:00 AM

13

ee

11.1

Page 41: [23] Deleted

Gerald Goldstein

11/19/2009 10:38:00 AM

98

Page 41: [23] Deleted

ev

Gerald Goldstein

11/19/2009 10:39:00 AM

61

w

ie ly

On

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 48 of 48

URL: http://mc.manuscriptcentral.com/tbin