TABLE 3.

Neuropsychological studies with Gulf War veterans.

References	Group (N)	Neuropsychological Tests	Strengths	Limitations	Conclusions
Goldstein et al., 1996	29 GWV, 39 non-veterans	WAIS-R Information WAIS-R Similarities WAIS-R Digit Span WAIS-R Arithmetic WAIS-R Digit Symbol WAIS-R Picture Completion WAIS-R Block Design COWAT Incidental Memory Verbal Associative Learning Short-Term Memory Symbol Digit Learning Trail Making Test CPT Grooved Pegboard	– GWVs compared to demographically matched controls on neuropsychological tests – Test battery included some measurements sensitive to GWI (i.e., Block Design)	– No case designation based on GWI – Limited measurement of mood	– Found that GWVs performed worse on impairment index in compared to controls – However, the level of impairments (1 SD) was not consistent with subjective cognitive complaints – Effect size Cohen’s d calculation∗ show small effect for Trails B (d = 0.25) and pegboard dominant (d = 0.18).
Axelrod and Milner, 1997	44 male GWV from Army Guard unit	Reitan-Indiana Aphasia Screening Test WAIS-R AVLT Stroop∗ Trail Making Test WMS-R Finger Tapping Grooved Pegboard∗ Grip Strength COWAT Category Fluency PIAT-R WCST	– Veterans compared via normative data and grouped by both objective (i.e., Grooved Pegboard, Stroop) and subjective (i.e., health complaint) measures	– Small sample size – Lack of correction for Type 1 error – Volunteer sample – No hypotheses	– No evidence of deficits – Differences in subjective complaints in psychological measures – Effect size Cohen’s d calculation∗ show large effect for Trails B (d = 1.28), and small/medium effect for motor tests (d = −0.63 to −0.48)
Hom et al., 1997	26 GWV with Haley Syndromes, 10 GWV and 10 non-deployed veteran controls	WAIS∗ Halstead Category Test∗ Tactual Performance Test Seashore Rhythm Test Speech-Sounds Perception Test Finger Oscillation Test Trail Making Test∗ Reitan-Indiana Aphasia Screening Examination, Reitan-Klove Sensory Perceptual Examination Reitean-Klove Lateral Dominance Examination Reitan Word Finding Test WMS-R∗ WRAT3∗	– Matched GWV group – GWI criteria used with a factor derived technique	– Small sample size – Limitations of those with fitting factor criteria – Multiple hypothesis testing Initial differences between control group and cases	– Differences between GWI and GWVs in global neurocognitive functioning – Cohen’s d calculation∗ showed a large effect size for Block Design (d = −1.57) and a medium effect size for Trail Making Test- Trail B (d = 0.69) – Psychological responding was consistent with other medical patients
Sillanpaa et al., 1997	49 GWV from a single Army reserve military police unit	NES-2 CPT∗ Grip Strength Grooved Pegboard∗ Neurological Screen Fingertip Number Writing perception∗ WCST∗ AVLT WAIS-R ∗	– Examiner blind to participant’s medical history – Exposure to toxins measured via self-report – Models tested to mimic GWI and psychological functioning	– Small sample size – Low variance and range in scores – Multicollinearity problems present	– Psychological model accounted for neuropsychological performance with a R2 of at least a 0.03 (at or above a small effect) for all domains – At least a small effect for exposure and symptoms seen in nearly all domains – Cohen’s d calculation∗ showed a medium effect for motor functioning (d = 0.76)
Vasterling et al., 1998	43 GWVs: 19 with PTSD and 24 without	Letter Cancelation Stroop CPT∗ WCST WAIS-R Digit Span WAIS-R Arithmetic∗ Rey-AVLT∗ CVMT∗	– Investigated GWVs with PTSD	– Small sample size – Lack of comparison sample of participants with differing mental disorders – Unable to manipulate trauma exposure	– Veterans with PTSD had deficiencies in sustained attention, mental manipulation, information acquisition, and retroactive interference
Anger et al., 1999	66 GWV with unexplained symptoms from WA and OR, 35 GWV controls	Behavioral Assessment and Research System (BARS): Simple Reaction Time∗ Selective Attention Test Digit Span∗ Symbol Digit∗ Serial Digit Learning ODTP∗	– Compared those with GWI with controls – Physician blind to participant status	– Volunteer sample Self-selection bias	– Specific problems demonstrated in processing speed – Individuals compared based on processing speed differences also found that those with slower processing speed had deficits in memory and attention
Binder et al., 1999	100 GWV with unexplained symptoms	CPT ODTP∗	– Investigated self-report of cognitive ability and affective distress in conjunction with objective cognitive performance	– Cognitive measures may lack sensitivity	– Subjective complaints associated with psychological distress over objective cognitive performance
Bunegin et al., 2001	8 symptomatic GWV, 8 GWV controls	NES-2: Hand-Eye Coordination Simple Reaction Time Visual Digit Span Forward and Backward∗ Horizontal Addition Pattern Memory∗ Switching Attention∗	– Compared symptomatic GWVs with non-symptomatic GWVs – Investigated blood flow	– Small sample size – Less sensitive measures used	– Symptomatic GWVs had worse performance in memory and executive function tasks – Exposure to acetone also impacted cognitive performance in GWVs – Cohen’s d calculation∗ showed a small effect size for CPT, reaction time (d = −0.14)
Lange et al., 2001	48 symptomatic GWV, 39 GWV controls	NES∗ PASAT∗ WAIS-R Digit Span∗ CVLT RCFT Trails Making Test Category Test∗ Judgment of Line Orientation Test WAIS-R Block Design Grooved Pegboard	– Compared those with GWI with matched controls	– Volunteer sample of health-care seeking veterans – Small sample size CFS sample – Unequal cells comparisons	– Impairment found in attention (R2 = 0.12– 0.19) and executive functioning tasks (R2 = 0.07) even after controlling for mood – Cohen’s d calculation∗ showed a large effect size for CPT reaction time (d = 0.85).
White et al., 2001	193 GWV, 47 Germany deployed veterans	WAIS-R CPT Trail Making Test PASAT WCST Digit Span CVLT∗ WMS-R∗ Finger Tapping Purdue Pegboard POMS∗ TOMM	– Compared deployed and non-deployed veterans – Detailed account of toxin exposure – Stratified Random sample	– TOMM scores evidenced possible poor effort in some participants – Multiple comparisons	– Initially, mood was only significant with adjustment for multiple comparisons – Comparing those with and without exposure, had worse performance in short term memory (R2 = 0.315– 0.399), attention (R2 = 0.381), and mood ((R2 = 0.202– 0.315) – Cohen’s d calculation∗ showed small effect sizes for all neuropsych tests (d = −0.47 to 0.22
David et al., 2002	209 British GWV, 132 non-deployed era veterans	WAIS-R NART WAIS-III Letter Sequencing PASAT SART Stroop Trail Making Test WMS-R Purdue Pegboard	– Compared Gulf War deployment and medical status against other deployments (Bosnia) and controls – Stratified Random sample Blind raters Statistical analyses	– Cross-over effects Self-report symptoms	– Significance only found in PTSD measure – Cohen’s d calculation∗ showed a large effect size for Block Design (d = −2.53)
Lindem et al., 2003a	193 GWV, 47 Germany deployed veterans	WAIS-R Information subscale∗ WAIS-R Digit Span∗ WMS-R Digit span CPT∗ Trail Making Test WCST PASAT∗ Finger Tapping∗ Purdue Pegboard∗ WAIS-R Block Design WMS-R Verbal Paired Associate Learning CVLT∗ Visual Reproduction∗ POMS∗	– Investigated PTSD in relation to exposure to chemical agents – Investigated symptom severity – Better generalization with use of overall cohorts in Gulf War – Large sample size	– Correlational analyses – Lack of baseline performance or known preexisting conditions	– PTSD symptoms severity correlated with greater deficits in a wide array of neuropsychological measures in GW deployed veterans (Partial R2 = 0.02– 0.10) – CBW exposure and PTSD severity in GWVs associated with deficits in sustained attention (Partial R2 = 0.0004– 0.0015), motor speed/motor coordination (0.0000–0.0007)
Lindem et al., 2003b	193 GWV, 47 Germany deployed veterans	WAIS-R Information subscale WAIS-R Digit Span WMS-R Digit span CPT Trail Making Test A Trail Making Test B WCST PASAT Finger Tapping Purdue Pegboard WAIS-R Block Design WMS-R Verbal Paired Associate Learning CVLT Visual Reproduction POMS∗	– Investigated GWVs discrepancy between subject complaints and objective performance	– Multiple comparisons	– Subjective complaints more associated with mood symptoms
Lindem et al., 2003c	58 GWV and 19 Germany-deployed veterans	WAIS-R Information subscale WAIS-R Digit Span WMS-R Digit span CPT Trail Making Test∗ WCST∗ PASAT Finger Tapping Purdue Pegboard WAIS-R Block Design WMS-R Verbal Paired Associate Learning∗ CVLT∗ Visual Reproduction∗ TOMM∗ POMS	– Investigated motivation in GWVs	– Small sample size – Difficult to ascertain the reason behind lower TOMM scores – Low amount of those with low TOMM scores	– Variability was seen in those with lower TOMM scores particularly in attention, executive functioning, and memory
Proctor et al., 2003	Danish GWVs (215), comparing deployed (143) and non-deployed veterans (72)	WAIS-R Information CPT Trail-making Test Wisconsin Card Sorting Test Purdue Pegboard WAIS-R Block Designs California Verbal Learning Test WMS Visual Reproductions POMS∗ TOMM	– Blind to categorization of “higher or lower” symptom status during all phases of recruitment, testing, and interviewing – Differences in deployment missions between Danish and American groups	– Significant mean age difference between deployed (38.8 years) and non-deployed (34.8 years) – Self-report of exposure	– Evidence of increased mood complaints related to GW service – no significant domain-specific evidence of CNS dysfunction was found – No associations between reported GW Environmental exposures related to the Danish GW deployment mission and objective measures of cognitive functioning were observed
Sullivan et al., 2003	207 treatment seeking GWV (120 referred for neuropsych evaluation), 53 treatment seeking non-deployed veterans	WAIS-R Information WAIS-R Digit Span∗ Trail Making Test NES CPT Stroop Test Paced Auditory Serial Addition Test Wisconsin Card Sort Test∗ CVLT WMS-R Paired Associate Learning WMS-R Visual Reproductions∗ Hooper Visual Organization Test WAIS-R Block Design∗ Finger tapping Purdue Pegboard RCFT∗ POMS∗ TOMM	– Investigated deployment, treatment seeking, use of pyridostigmine bromide (PB) and PTSD on cognitive functioning – Matched by control group that was also treatment seeking	– Self-report of exposure – Sample size small for comparisons	– GW deployed worse than controls on attention, visuospatial skills, visual memory, and mood – PB use in GWVs worse in executive system tasks – GWVs with PTSD versus those without PTSD showed no differences – Cohen’s d calculation∗ showed a large effect sizes for block design and digit span forward (d = −2.43 to −1.00), small effect sizes for Trails A and B, digit span backward, CVLT, WMS, immediate recall, and finger tapping (d = −0.090 to 0.43), and a medium effect size for WMS, delay recall (d = −0.55).
Vasterling et al., 2003	72 GWVs deployed and 33 non-deployed GWVs	WAIS-R Digit Span WCST AVLT CVMT Purdue Pegboard WAIS-R Information	– Selection of a non-treatment seeking group of GWVs – Comparison of deployed GWVs to a group of GWVs mobilized but no deployed – Use of olfactory and neurocognitive measures with demonstrated sensitivity to neurotoxic exposures	– Sample was regionally recruited	– No evidence that performance on olfactory or neurocognitive measures were related to war-zone duty or to self-reported exposure to GW toxicants – Symptoms of emotional distress were positively correlated with self-report of health and cognitive complaints
Proctor et al., 2006	140 Army GWV with modeled estimates of nerve agent exposure	CPT Trail Making Test WAIS-R Digit Span WCST Finger Tapping Purdue Pegboard∗ WAIS-R Block Design∗ CVLT WMS-R verbal paired associate learning WMS visual reproduction	– Stratified random sampling – Examined performance by exposure to sarin and cyclosarin – Sample was unaware of sarin and cyclosarin components, analyses were conducted a prior to exposure knowledge	– Etiology undetermined given the risk of another illness between exposure and measurement (i.e., no baseline health information) – Limited objective information about exposures	– Exposure associated with poor fine psychomotor dexterity (d = 0.44) and visuospatial abilities (d = 0.43)
Barrash, 2007	301 GWV, 99 era veterans deployed elsewhere	WAIS-III Similarities∗ Block Design∗ Digit Symbol Digit Span North American Reading Test – Revised Starry Night Test COWAT AVLT∗ Benton Visual Retention Test∗ RMT-Words and Faces∗ Stroop Grooved Pegboard∗	– Study of effort and neurocognitive performance in GWVs –Grouped by credible or non-credible impairment	– Small sample of non-credible group –Decreased statistical power – Lack of measures investigating reason behind low effort	– Non-credible impairment associated with more variability in tests and worse emotional/cognitive functioning
Wallin et al., 2009	41 GWVs: 25 with GWI and 16 controls	WRAT reading Block Design Trail Making Test CVLT Pegboard	– Stratified random sampling – Used GWI criteria to divide groups	– Small sample size – Gap between deployment and time of study – Multiple analyses	– Differences only seen in mood and health measures – Cohen’s d calculation∗ showed a medium effect size for block design, Trails B, and CVLT long delay, (d = −0.73 to 0.51), and a small effect size for WRAT reading, Trails A, and Pegboard (d = −0.13 to 0.39).
Toomey et al., 2009	1061 deployed GWV and 1128 non-deployed GWV	WAIS-III Digit Span Trail Making Test∗ PASAT CPT∗ WCST CVLT∗ RCFT∗ Finger Tapping∗ Purdue Pegboard∗ TOMM WRAT-III	– Investigated differences in deployment, toxin exposure, and GWI status – Large sample size, stratified random sampling method – Use of factor analysis – Use of Khamisiyah exposure data	– Low study participation rates – Cross-sectional design – Neuropsychology raters were not blind to condition	– Deployed veterans had worse performance on motor speed (OR = 2.35) and sustained attention (OR = 2.64) – Those with Khamisiyah exposure showed poor motor speed after controlling for mood – Cohen’s d calculation∗ showed small effect sizes for all neuropsych tests (d = −0.09 to 0.06).
Chao et al., 2010	40 GWV with a history of DOD notified sarin cyclosarin exposure risk and 40 non-exposed matched GW veteran controls	CPT Trail Making Test WAIS-III Digit Span Short Category Test COWAT Grooved Pegboard WAIS-III Digit Symbol, matching WAIS-III Block Design WAIS-III Verbal Comprehension Index CVLT-II WMS-III Logical Memory BVMT-R TOMM∗	– Used matched cohort sample – Use of Khamisiyah exposure data	– Lack of information regarding the unit and rank of veterans – Lack of information regarding symptom severity (i.e., CMI, smoking status, head injuries) – Lack of cumulative exposure for all GW veterans – Plume estimates only by unit	– No differences in cognitive measures after controlling for poor effort (i.e., failure of TOMM). – Cohen’s d calculation∗ showed a small effect sizes for all neuropsych tests (d = 0.22 to 0.26).
Chao et al., 2011	64 sarin and cyclosarin exposed GWVs and 64 “matched” unexposed GWVs	CPT∗ WAIS-III Digit Span∗ Trail Making Test Short Category Test CVLT-II∗ Grooved Pegboard TOMM	– Used matched controls to compare structural and functional differences in veterans with suspected neurotoxicant exposure – Use of more sensitive MRI (4T) – Use of some sensitive tests for neuropsychological and mood outcomes	– Lack of information regarding veteran’s unit, severity of GWI symptoms, smoking status, or history of head injury – Neurotoxicant exposure measured at unit over individual level	– Reduced gray matter and white matter in exposed veterans which was linked to neurotoxicant exposure – Exposed veterans made more omission errors and had slower responses times; omission errors was also linked to neurotoxicant exposure – Cohen’s d calculations∗ showed a medium effect size for Trails A (d = −0.64), and small effect sizes for CPT, Trails B, CVLT, and pegboard (d = −0.36 to 0.38).
Chao et al., 2016	136 GWVs: 106 who reported hearing chemical alarms sound	WAIS III Block Design∗ Digit Span∗ CVLT		– Had to rely on self-reports of deployment-related exposures – Lack of pre-GW measurements of brain structure and function – Small sample size – Lack of a non-deployed GW-era veteran control group	– Self-reported frequency of hearing chemical alarms was inversely associated with and significantly predicted performance on the Block Design visuospatial task. – This effect was partially mediated by the relationship between hearing chemical alarms and lateral occipital cortex volume.
Chao, 2017	428 deployed GWVs: 272 which met CDC criteria for CMI	CVLT-II∗	– Tested verbal memory with GWVs presenting with subjective memory complaints – Large sample size	– Only measured one domain to control for Type 1 error	– Worse performance on verbal memory associated subjective complaints over and above mood, however, there was higher endorsement of PTSD symptoms
Sullivan et al., 2018	159 GW-deployed preventative medicine personnel who had varying levels of pesticide exposure	WAIS-III information subtest Boston Naming Test Trail Making Test∗ CPT∗ WCST Finger Tapping Grooved Pegboard HVOT RCFT∗ Stanford-Binet Copying Test CVLT II POMS∗ TOMM	– Grouped veterans by exposure (low/high) to PB and pesticides – Sample had sophisticated knowledge of exposure as they were part of the medical team	– Multiple analyses – Exposures of PB and pesticide may be correlated – Classifications of groups based on self-report	– High pesticide/high PB had worse information processing speed, attention (i.e., errors), visual memory, and increased mood complaints

See original journal articles in first column for test references. ^∗Denotes significance of p < 0.05.