This article was downloaded by: [La Trobe University]On: 10 January 2015, At: 16:32Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: MortimerHouse, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Clinical and ExperimentalNeuropsychologyPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/ncen20

Does the presence of posttraumatic anosmia meanthat you will be disinhibited?Simon F. Crowe a & Louise M. Crowe ba School of Psychological Science , La Trobe University , Bundoora , VIC , Australiab School of Behavioural Science, Department of Psychology , University ofMelbourne , Parkville , VIC , AustraliaPublished online: 22 Feb 2013.

To cite this article: Simon F. Crowe & Louise M. Crowe (2013) Does the presence of posttraumatic anosmiamean that you will be disinhibited?, Journal of Clinical and Experimental Neuropsychology, 35:3, 298-308, DOI:10.1080/13803395.2013.771616

To link to this article: http://dx.doi.org/10.1080/13803395.2013.771616

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”)contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensorsmake no representations or warranties whatsoever as to the accuracy, completeness, or suitabilityfor any purpose of the Content. Any opinions and views expressed in this publication are the opinionsand views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy ofthe Content should not be relied upon and should be independently verified with primary sources ofinformation. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands,costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly orindirectly in connection with, in relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any substantial orsystematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distributionin any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found athttp://www.tandfonline.com/page/terms-and-conditions

Journal of Clinical and Experimental Neuropsychology, 2013

Vol. 35, No. 3, 298–308, http://dx.doi.org/10.1080/13803395.2013.771616

Does the presence of posttraumatic anosmia meanthat you will be disinhibited?

Simon F. Crowe1 and Louise M. Crowe2

1School of Psychological Science, La Trobe University, Bundoora, VIC, Australia2School of Behavioural Science, Department of Psychology, University of Melbourne, Parkville, VIC,Australia

Dispute has surrounded the issue of whether the relationship between anosmia and executive dysfunction in trau-matic brain injury (TBI) may be artefactual due to poor ascertainment. Three groups matched for age, gender,education, Full Scale IQ, and the Wechsler Working Memory Index and showing adequate symptom validity werecompared: 30 anosmic TBIs (TBI-A) matched for posttraumatic amnesia (PTA) and working memory functioningwith 36 nonanosmic TBIs (TBI-NA) and 51 controls. The groups performed the FAS test, the Animal Fluencytest, the Stroop Neurological Screening Test (SNST), the Wisconsin Card Sorting Test–64 (WCST–64) and theTrail Making Test (TMT-B) as well as tests of emotional functioning and return to work outcome. After adjust-ing for the covariates (i.e., gender; Wechsler Test of Adult Reading, WTAR; and years of education), a significanteffect was found for items successfully completed on the SNST, the FAS task, the Animal Fluency task, and theWCST–64 categories completed. After adjusting for the covariates, a significant difference was found for num-ber of errors on the SNST and for the number and type of errors on TMT-B. The two groups did not differ interms of their affective functioning (i.e., Beck Depression Inventory or Beck Anxiety Inventory), or in terms oftheir outcome with regard to return to work. The findings support the notion that the TBI-A group demonstratedconsiderably weaker performance on executive tasks than did the nonanosmic TBIs. These patients were not, how-ever, more prone to an error-prone pattern of performance, and, if anything, their executive deficit was more likelyattributable to a reduced productivity of response.

Keywords: Anosmia; Executive; Prefrontal cortex; Disinhibition; Symptom validity.

INTRODUCTION

Damage to the ventromedial temporal lobes andthe orbital frontal structures of the brain are com-mon neuropathological sequelae of traumatic braininjury (TBI: Costanzo & Zasler, 1992; Courville,1945). The location of the olfactory nerves directlybelow the orbital and medial prefrontal cortex(OMPFC; Öngür & Price, 2000) and the proxim-ity of the uncus to the anterior clinoid process andthe lesser wing of the sphenoid bone make theseregions particularly susceptible to injury as a conse-quence of TBI. As a result, posttraumatic anosmiais a commonly observed effect of these injuries(Martzke, Swan, & Varney, 1991).

Address correspondence to Simon F. Crowe, School of Psychological Science, La Trobe University, Bundoora, Victoria, 3086,Australia (E-mail: s.crowe@latrobe.edu.au).

Estimates of the frequency of posttraumaticanosmia have been as high as 20 to 30% of survivorsof trauma (Costanzo & Becker, 1986; Levin, High,& Eisenberg, 1985), and the level of impair-ment is closely correlated with the severity of theinjury (Green, Rohling, Iverson, & Gervais, 2003;Sigurdardottir, Jerstad, Andelic, Roe, & Schanke,2010), whether this is defined by longer durationof posttraumatic amnesia, lower GCS scores, orthe presence of computed tomography (CT) abnor-malities. Costanzo and Zasler (1992) have notedthat anosmia varies in frequency from 0–16% inmild head injuries and from 15–19% in moder-ate injuries and may be as high as 24–30% inthe more severe cases, although varying levels

© 2013 Taylor & Francis

Dow

nloa

ded

by [

La

Tro

be U

nive

rsity

] at

16:

32 1

0 Ja

nuar

y 20

15

ANOSMIA, EXECUTIVE DEFICIT, AND TBI 299

of compromise have been noted by other inves-tigators (Callahan & Hinkebein, 1999: 65%; deKruijk, Leffers, Menheere, Meerhoff, Rutten, &Twijnstra, 2003: 4% anosmia, 22% hyposmia; Dotyet al., 1997: 66.8%; Ogawa & Rutka, 1999: 9.3%;Sigurdardottir et al., 2010: 10%).

Damage to the olfactory apparatus followingTBI arises from three principal causes: injury ortearing of the olfactory nerves, damage to thenose or the nasal passages, and contusion orbrain hemorrhage to the olfactory parenchyma(Costanzo & Zasler, 1992). Jennett and Teasdale(1981) have noted that shearing and abrasion of theolfactory nerves is often associated with contusionand laceration of the surrounding orbital corticalareas, a finding also reported by other researchers(Costanzo & Zasler, 1991; Malloy, Bihrle, Duffy, &Cimino, 1993).

Neuroradiological investigation supports therelationship between traumatic anosmia andthe damage to ventral brain structures. Varney,Pinkston, and Wu (2001) report that in patientswith traumatic anosmia, hypofunction of theorbital frontal region can be noted with positronemission tomography, and Varney’s group (Varney& Bushnell, 1998) have also noted that orbitalfrontal hypometabolism, as demonstrated usingsingle-photon emission computed tomogra-phy, correlates with reports of behavior changeby significant others. Yousem, Geckle, Bilker,McKeown, and Doty (1996) have reported thatmagnetic resonance imaging of patients withposttraumatic anosmia demonstrates abnormal-ities predominantly in the olfactory bulbs andthe subfrontal regions, and those patients withcomplete posttraumatic anosmia demonstrated agreater loss of volume in the olfactory bulbs andolfactory tracts than did those with only partialanosmia. Rombaux and colleagues (2006) havenoted a similar correlation between olfactory func-tioning and olfactory bulb volume in a group of25 patients with traumatic anosmia. They furthernoted that parosmia was consistently associatedwith the presence of cerebral damage and also withless olfactory bulb volume. Costanzo and Zasler(1992) have even gone so far as to suggest that“olfactory function in and of itself seems to be arelatively good marker for associated neuropatho-logical abnormalities in the medial frontal andanterior temporal lobes” (p. 21).

Neuropsychological comparison of anosmic withnormosmic groups indicates that the anosmic par-ticipants feature longer duration of coma, moresevere deficits in complex attention tasks includ-ing the Trail Making Test, compromise in newlearning and memory including impairment for

the California Verbal Learning Test Trial 5, andproblem-solving difficulty including compromiseon the Wisconsin Card Sorting Test (Callahan &Hinkebein, 1999). Our own previous study usinga small sample of 12 anosmic TBI subjects indi-cated that the anosmic subjects produced a hightendency towards error on the Controlled OralWord Association Test in comparison to nonanos-mic TBIs matched for age, sex, education, verbalIQ, and level of memory impairment (Crowe, 1996).

Cummings (1985; Lichter & Cummings, 2001)has described three distinct syndromes arisingfrom damage to the frontal lobe: a dorsolateralprefrontal convexity syndrome featuring compro-mise in executive functions including decreasedverbal and design fluency, abnormal motor pro-gramming, impaired set shifting, reduced learningand memory retrieval, and poor problem solving;a medial frontal/anterior cingulate syndrome fea-turing apathy and diminished initiative (abulia);and an orbital and ventromedial prefrontal syn-drome featuring disinhibition, irritability, impulsiv-ity, emotional lability, poor insight, poor judgment,and distractibility. Malloy et al. (1993) have simi-larly described an orbital medial frontal syndromecharacterized by anosmia, amnesia with confabu-lation, go/no-go deficits, personality changes, poorsocial and vocational adjustment, and hypersensi-tivity to pain, and they have noted that the con-dition is distinct from the pattern that emergesfollowing dorsolateral frontal damage.

The establishment of the convergent validity oftraumatic anosmia as a sign of damage to theorbital frontal area of the brain and hence signif-icant psychosocial deficit has been attempted bya number of investigators. Varney (1988; Varney& Menefee, 1993) has reported that patients withsevere or total posttraumatic anosmia have a highrate of chronic unemployability (92%) and thateven those subjects with mild forms of dysnosmiawere chronically unemployed (64%). The latter find-ings were made even despite the fact that theseparticipants had no clear neurological, intellectual,or memory deficits that explained their unemploy-ment. Varney (1988) also noted that these subjectshad other cognitive and emotional problems includ-ing absent mindedness (100%), poor planning andanticipation (95%), indecisiveness and poor deci-sion making (93%), perplexity responses (83%),unevenness in the quality of work output (80%),unreliability (74%), inability to learn from mistakes(74%), and an inability to get along with fellowemployees and supervisors (60%).

There has, however, been considerable disputein the literature (Greiffenstein, Baker, & Gola,2002, 2003; Varney, 2002) as to whether the

Dow

nloa

ded

by [

La

Tro

be U

nive

rsity

] at

16:

32 1

0 Ja

nuar

y 20

15

300 CROWE AND CROWE

relationship between posttraumatic anosmia andlate psychosocial and neuropsychological outcomesmay be artefactual. Greiffenstein and colleagues(2002) criticized Varney’s original (1988) findingon three grounds: (a) The injuries were inconse-quential; (b) the diagnosis of anosmia was madeon the basis of self-report; and (c) hospital recordsdemonstrating the presence of anosmia at the timeof the injury were not presented. In response,Varney (2002) noted that “anosmia is not reliablyassociated with any type of test failure” (p. 852).Greiffenstein et al., (2003) responded that “orbital-frontal damage may be sufficient to cause anosmia,but it is not necessary” (p. 152) and further thathe “fails to provide a crucial contrast: employmentrates of head trauma victims without smell com-plaints” (p. 153). The aim of this study was todetermine the convergent validity of the clinicalsign of traumatic anosmia as an indicator of dis-ruption of the orbital and medial prefrontal cortexand its concomitant effect upon neurocognitive andpsychosocial functioning following TBI.

METHOD

Participants

Seventy-nine individuals with TBI (39 TBI withanosmia, TBI-A; 40 without, TBI-NA) wereentered into a database from a private practicethat conducted neuropsychological assessments formedicolegal, diagnostic, and rehabilitation pur-poses and were administered the executive functiontests as part of a larger test battery. The par-ticipants gave permission for summary data aris-ing from their assessments to be used as researchdata. All of the clinical participants had docu-mented evidence of TBI, and all were at least1 year post injury. The control group consistedof 51 noninjured volunteers from the local com-munity, and the data were gathered as a partof a normative study conducted at La TrobeUniversity in Victoria, Australia. Full medical his-tories were not taken from the controls, but all par-ticipants denied neurological or psychiatric impair-ment. The control group were administered a testbattery assessing memory, attention/concentration,and executive abilities as well as tests of symptomvalidity.

For all groups, the demographic informationgathered included age, gender, and years of educa-tion. For the TBI groups, length of posttraumaticamnesia (PTA) and details of any olfactory dis-ruption, including the level of smell loss as well aswhether this disruption had been self-reported or

was medically documented (i.e., as a part of theindependent medical examination process under-taken by a neurologist, neurosurgeon, or otolaryn-gologist), were also recorded. The specialist’s reportof olfactory functioning was drawn from the avail-able documentation of the independent medicalexaminations of each patient, which was also ver-ified by patient self-report. The details of theseexaminations were not extensively documented butwere office-based clinical neurological examina-tions of olfactory functioning based upon theexamination of 3–4 odorants (usually including cof-fee and soap). Based upon any report of anosmia,the TBI group was split into two groups: thosewith no anosmia (TBI-NA) and those with anosmia(TBI-A). The TBI-A group included 39% (14/36) ofmedically confirmed cases with anosmia and 61%(22/36) of patients who self-reported this symptom.Many of the participants (70%) were involved in lit-igation at the time of assessment, most under theworkman’s compensation legislation.

Of the 79 clinical participants, only those whomet the following criteria were included in theanalysis: (a) had sustained a documented trau-matic brain injury; (b) had sustained a documentedperiod of posttraumatic amnesia—PTA durationthat was coded into discrete categories of sever-ity: mild (<5 min to 60 min), moderate (1 hourto 24 hours), and severe (>1 day), in the man-ner proposed by Russell and Smith (1961)—mostof the patients were in the moderate to severerange; (c) were a minimum of six months postinjury; and (d) performed above the cutoff scoreson tests of symptom validity testing. Duration ofPTA was calculated based on the background med-ical reports that were provided as well as fromthe client’s self-report. Glasgow Coma Scale scoreswere unavailable for many of the patients in thesample, circumventing the use of this more accuratemeasure of injury severity.

Of the initial pool of participants who had sus-tained a TBI, 13 participants (3 TBI-A and 10 TBI-NA) did not perform above the cutoffs indicativeof symptom valid performance and were thereforeexcluded from the final data set. Thus the finalsample included 30 TBI-NA, 36 TBI-A, and 51 con-trols. Refer to Table 1 for demographic and clinicaldata regarding the samples. There was no signif-icant difference between the clinical groups (p >

.05) for duration of PTA.

Materials

All assessment measures were administered accord-ing to the respective published guidelines.

Dow

nloa

ded

by [

La

Tro

be U

nive

rsity

] at

16:

32 1

0 Ja

nuar

y 20

15

ANOSMIA, EXECUTIVE DEFICIT, AND TBI 301

TABLE 1Demographic characteristics of each group

Group

Characteristic Control TBI-NA TBI-A

Mean age, years 36.59 (14.30) 42.13 (7.78) 44.44 (12.15)Male/female ratio 20/31 28/2 32/4Years of education 11.63 (1.13) 10.53 (1.50) 10.75 (1.90)Mean WTAR FSIQ score 105 (8.98) 95.58 (12.24) 92.14 (14.11)Days of PTA N/A 7.7 (13.88) 31.49 (70.55)Mean WMI N/A 92.03 (10.94) 82.78 (17.52)BDI N/A 27.73 (14.84) 23.77 (11.56)BAI N/A 19.40 (13.65) 16.14 (10.56)

Note. TBI = traumatic brain injury. TBI-NA = TBI without anosmia. TBI-A = TBI with anosmia. WTAR =Wechsler Test of Adult Reading. FSIQ = Full Scale Intelligence Quotient. PTA = posttraumatic amnesia. WMI =Working Memory Index. BDI = Beck Depression Inventory. BAI = Beck Anxiety Inventory. Standard deviations inparentheses.

Predicted verbal IQ

As a measure of estimated cognitive function-ing, the Wechsler Test of Adult Reading (WTAR:Psychological Corporation, 2001) was administeredto all groups.

Working memory

The TBI groups were administered the WechslerWorking Memory Index (WMI) from the WechslerMemory Scale Version III (WMS–III; Wechsler,1997). The rationale for this maneuver was toensure that both clinical groups were equivalentin terms of their level of working memory, whichcould have had an independent effect on execu-tive functioning. There was no significant differencebetween the clinical groups (p > .05) for the WMI.

Symptom validity

In order to detect less than genuine effort, all partic-ipants were administered the Rey 15 Item Test (Rey,1964), the Rarely Missed Index (RMI) developedby Killgore and Della Pietra (2000), and the Test ofMemory Malingering (TOMM: Tombaugh, 1996).Participants had to meet the combined criteria ofperforming above the published cutoffs on all threeof the symptom validity measures (i.e., 15 item testof ≥9; RMI ≥136, and TOMM score of ≥45 onthe second trial of the task). Whilst using this con-verging set of measures does not guarantee that allof the patients who scored above the suggested cut-off scores were putting forth genuine effort, it doesmake the presence of compromised effort less likely.

Psychosocial functioning

The two TBI groups were administered two self-report inventories to ascertain their levels ofpsychosocial compromise. These were: the BeckDepression Inventory (BDI–II; Beck, Steer, &Brown, 1996) as a screen for the presence of depres-sive symptoms, and the Beck Anxiety Inventory(BAI; Beck & Steer, 1993) as a measure of self-reported anxiety. TBI groups were also comparedwith regard to pre- versus postinjury employmentstatus.

Executive function assessments

As part of a larger test battery, several tasks ofexecutive function were administered. The first wasthe Stroop Neurological Screening Test (SNST:Trennery, Crosson, DeBoe, & Leber, 1988), whichhas been described as a measure of attention andinhibition. In scoring of the SNST the followingvariables were calculated: success or failure to com-plete the color/word trial in the allocated time (i.e.,120 s), total errors, and total self-corrected errors.As it is common in injured participants for thetask not to be completed in the allocated time, thefirst 60 seconds of the test were analyzed further.For each 15-s segment up to 60 s (0–15 s, 15–30 s,30–45 s, 45–60 s), data were collected on the num-ber of colors named for the color/word trial withany self-corrected errors included in the total.

The second test was the Wisconsin CardSorting Test–64-card version (WCST–64; Kongs,Thompson, Iverson, & Heaton, 2000). WCST–64 involves matching stimuli according to a chang-ing sorting principle based on external feedback

Dow

nloa

ded

by [

La

Tro

be U

nive

rsity

] at

16:

32 1

0 Ja

nuar

y 20

15

302 CROWE AND CROWE

and is purported to be a measure of general exec-utive functioning. Only 64 cards are employed.The dependent variables measured included: totalerrors, perseverative errors (P errors), total oddballerrors, and number of completed categories.Perseverative errors are the number of errors madebecause the participant continues to use or perse-vere with a previous correct “rule,” although thematching criterion has changed. Oddball errorsoccurred when the participant matched a card tothe example card on a random basis.

The third test was the Trail Making Test (TMT-A & TMT-B; Army Individual Test Battery, 1944).The amount of time to complete the task and thenumber of errors were recorded for both trials.TMT-B is a measure reported to tap into visualsearch and cognitive alternation skills as well asof speed of processing (Crowe, 1998b; Ruffolo,Guilmette, & Willis, 2000).

The final executive tasks were the phonemicfluency (FAS) and the semantic category fluencytest (animal fluency form: Tombaugh, Kozak, &Rees, 1999). To assess phonemic fluency, partici-pants were asked to generate as many words aspossible in 60 s that started with the letter f , fol-lowed by words that started with the letter a, andfinally words that started with the letter s. The rulesemployed in the administration of the FAS taskincluded no proper nouns, no numbers, no differ-ent forms of the same word, and no homophones(Bittner & Crowe, 2006). Results were recorded astotals for the three letters individually, as well as forthe sum of all three letters. In addition, the totals foreach letter were recorded as the number of wordsproduced in the four 15-s time slices (Crowe, 1992,1998a). For the category fluency test, participantswere asked to recall as many animals as they couldduring a 60-s trial. Repetitions or perseverationswere scored as errors.

The fluency tasks were also scored for clus-tering and switching in the manner proposed byTroyer, Moscovitch, and Winocur (1997; Troyer,Moscovitch, Winocur, Alexander, & Stuss, 1998).Troyer and colleagues (1997) in their critical anal-yses of verbal fluency output, observed a tendencyfor participants to produce words in distinct sub-categories. The authors termed these subcategories“clusters” and also referred to “switching” as theprocess of shifting between these subcategories.Clustering and switching have since been estab-lished as important components of fluency output,and it is thought that optimum fluency performanceis contingent on establishing an appropriate bal-ance between cluster size and number of switches(Hurks, 2012). Normative data for clustering andswitching (Troyer, 2000) from over 400 healthy

participants noted that, on average, participantsproduced clusters of 0.24 and 23.9 switches onphonemic tasks and larger clusters of 0.94 and23.4 switches on semantic tasks. Clustering is gen-erally thought to be a temporal lobe mediated func-tion that draws upon verbal memory and semanticknowledge. Switching is thought to be predomi-nately a frontal lobe function, as it necessitatescognitive flexibility, innovation, and search strate-gies (Hurks, 2012). Studies investigating clusteringand switching have revealed the involvement ofother cognitive functions, including cognitive shift-ing, search strategies, and lexical processes (Troyeret al., 1998). Deficits in switching have also beendocumented in patients with frontal dysfunction inaddition to impairment in other brain regions; theseinclude those with schizophrenia, Parkinson’s dis-ease, Huntington’s disease, and multiple sclerosis(Troyer, 2000).

Procedure

All participants gave informed consent to partici-pate, and all tests were administered individually.

Statistical analyses

The three groups (TBI-NA, TBI-A, controls) werecompared using chi-square and one-way analysisof variance (ANOVA) on demographics and pre-dicted verbal IQ to identify any differences thatcould influence performance. The two TBI groupswere also compared on WMI and length of PTA.Multivariate analysis of variance (MANOVA) andANOVA where appropriate were conducted tocompare scores across the groups on the SNST,WCST–64, and TMT-B. Tukey’s honestly signifi-cant difference (HSD) was used to ascertain spe-cific group differences. Chi-square was used tocompare the dichotomous variable of test com-pletion for the SNST. Differences between theTBI groups were examined using MANOVAs.Correlations between the variables were alsocalculated.

RESULTS

The neuropsychological measures were classifiedinto categories based on the distinction betweenproduction (i.e., the number of items success-fully produced) and the level of rule-breakingerrors. Crowe (1992, 1996, 1998a) found supportfor the notion of a dissociation of two frontal

Dow

nloa

ded

by [

La

Tro

be U

nive

rsity

] at

16:

32 1

0 Ja

nuar

y 20

15

ANOSMIA, EXECUTIVE DEFICIT, AND TBI 303

lobe syndromes on the basis of performance ontests of verbal fluency: putative involvement ofthe dorsolateral prefrontal cortex (DLPFC) lead-ing to impairment in the level of response gener-ation (i.e., production); and putative involvementof the orbital and ventromedial prefrontal cortex(OMPFC) leading to impairment of selectivity andrule governance of response (i.e., error). The find-ing of compromise in control of responding notedin the studies with TBI participants using rule-breaking errors on fluency has also been noted byMalloy and colleagues (1993) as well as by Tate(1999).

Production

Analysis of covariance (ANCOVA) was conductedto compare any differences in production on themeasures of interest (see Table 2). After adjustingfor the covariates (i.e., gender, WTAR, and yearsof education), a significant effect was found for theSNST, F(2, 93) = 10.37, p < .001, η2 = .18; theFAS task, F(2, 110) = 6.68, p = .002, η2 = .11;the Animal Fluency task, F(2, 107) = 8.73, p <

.001, η2 = .14; and the number of categories com-pleted on the WCST–64, F(2, 81) = 4.52, p = .01,η2 = .10.

Error

After adjustment for covariates (i.e., gender,WTAR, and years of education), ANCOVAs wereconducted, and a significant difference was foundfor number of errors on the SNST, F(2, 99) = 3.11,p = .049, η2 = .06, and for number and type oferrors on TMT- B, F(2, 107) = 5.33, p = .006,η2 = .09. No significant differences were found fornumber of errors on the FAS task, F(2, 109) =1.41, p = .248, η2 = .03; number of errors onthe Animal Fluency task, F(2, 106) = 2.14, p =.126, η2 = .04; or for number or type of errors(i.e., perseverative errors, oddball errors, or totalerrors) on the WCST–64, � = .94, F(6, 156) = 1.53,p = .53.

A comparison of the raw means (i.e., not adjustedfor the covariates) of just the anosmic and nonanos-mic TBI participants indicated that the anosmicparticipants were significantly different from thenonanosmic TBI patients in the following mea-sures: the SNST, F(1, 48) = 9.91, p = .003; η2 = .17;the FAS task, F(1, 64) = 5.02, p = .029, η2 = .07;the Animal Fluency task, F(1, 61) = 8.47, p < .005,η2 = .12; and the WCST–64 categories completed,F(1, 35) = 6.35, p = .016, η2 = .15. The com-parison of the TBI groups is further examined inTable 3.

TABLE 2Comparing the three groups: control and TBIs with and without anosmia

Group (scores corrected for covariates)

Characteristic Control TBI-NA TBI-A

Comparison (allgroups) Corrected for

covariates Partial η2

Production measuresStroop total words in 60 s 56.78 (1.75) 52.87 (2.56) 43.02 (2.25) <.001∗∗∗ .18FAS total words produced 40.60 (1.77) 35.23 (2.11) 30.08 (1.94) .002∗∗ .11Animals total words produced 20.88 (0.74) 20.25 (0.90) 16.36 (0.84) <.001∗∗∗ .14WCST–64 completed categories 3.42 (0.19) 3.19 (0.33) 2.32 (0.29) .01∗ .10

Error measuresStroop total errors 0.59 (0.18) 0.79 (0.26) 1.35 (0.22) .049∗ .06FAS total errors 2.97 (0.36) 2.12 (0.42) 2.91 (0.38) ns .03Animals total errors 0.54 (0.15) 0.57 (0.17) 0.15 (0.17) ns .04WCST–64 total errors 17.85 (1.31) 18.61 (2.29) 22.02 (2.12) nsWCST–64 perseverative errors 10.65 (1.09) 10.86 (1.90) 14.30 (1.75) nsWCST oddball errors 0.57 (0.24) 0.55 (0.42) 1.26 (0.38) nsTrail Making Test-B errors 0.10 (0.35) 0.48 (0.41) 1.84 (0.39) .006∗∗ .09

Clustering and switchingFAS mean cluster size 1.60 (1.04) 1.56 (1.09) 1.82 (1.81)FAS number of clusters 8.56 (4.72) 5.07 (2.57) 4.12 (2.76) <.001∗∗∗FAS total switches 29.87 (8.93) 25.40 (9.47) 20.42 (8.50) <.001∗∗∗Animal mean cluster size 2.00 (0.96) 2.54 (1.19) 2.38 (1.68)Animal number of clusters 5.77 (1.68) 4.4 (1.16) 3.72 (1.71) <.001∗∗∗Animal total switches 11.5 (3.69) 7.90 (2.72) 7.94 (3.25) <.001∗∗∗

Note. TBI = traumatic brain injury. TBI-NA = TBI without anosmia. TBI-A = TBI with anosmia. WCST = Wisconsin Card SortingTest. All means were adjusted for the covariates (i.e., gender, WTAR IQ, and years of education, where WTAR = Wechsler Test of AdultReading). Standard errors in parentheses.∗p < .05. ∗∗p < .01. ∗∗∗p < .001.

Dow

nloa

ded

by [

La

Tro

be U

nive

rsity

] at

16:

32 1

0 Ja

nuar

y 20

15

304 CROWE AND CROWE

TABLE 3Comparing the TBI groups

Group

Characteristic TBI-NA TBI-AUncorrectedraw scores

Production measuresStroop total words in 60 s 50.19 (9.99) 40.55 (11.08) .003∗∗FAS total words produced 33.63 (11.95) 27.22 (11.26) .029∗Animals total words produced 19.20 (4.83) 15.81 (4.39) .005∗∗WCST–64 completed categories 3.13 (1.20) 2.14 (1.15) .016∗

Error measuresStroop total errors 0.81 (0.75) 1.38 (1.41) nsFAS total errors 2.17 (1.69) 3.11 (2.63) nsAnimals total errors 0.69 (1.33) 0.30 (0.59) nsWCST–64 total errors 20.13 (9.86) 24.00 (9.52) nsWCST–64 perseverative errors 11.87 (7.31) 15.61 (9.76) nsWCST–oddball errors 0.53 (0.83) 1.19 (1.99) nsTrail Making Test-B errors 0.46 (1.37) 1.71 (3.38) ns

Note. TBI = traumatic brain injury. TBI-NA = TBI without anosmia. TBI-A = TBI with anosmia. WCST =Wisconsin Card Sorting Test. Raw means, with standard deviations in parentheses.

Production of items as a function of time

A multivariate analysis of covariance (MANCOVA)was conducted to compare any differences in pro-duction on the SNST for each 15-s time slice. Afteradjusting for covariates, a significant multivariateeffect was found, Wilks’s � = .724, F(8, 180) =3.92, p < .001. Follow-up univariate analyses ofeach dependent variable found four significantresults. A significant effect was found for numberof words read aloud in the first 15 s, F(2, 93) =13.50, p < .001, η2 = .99, in the 16–30-s period,F(2, 93) = 6.47, p = .002, η2 = .90, in the 31–45-speriod, F(2, 93) = 4.12, p = .02, η2 = .72, and inthe 46–60-s period, F(2, 93) = 9.15, p < . 001, η2 =.972. The TBI-A performed more weakly than theother two groups at each time slice.

A MANCOVA was conducted to compare anydifferences in production on the FAS task for each15-s time slice for each letter. After adjustmentfor covariates, a significant multivariate effect wasfound, Wilks’s � = .558, F(24, 192) = 2.71, p <

.001. Follow-up univariate analyses of each depen-dent variable found significant results for eachtime slice.

For the letter F, there was a significant differ-ence between the groups for the number of wordsproduced in the first 15 s, F(2, 107) = 3.51, p <

.033, η2 = .64; in the 16–30-s period, F(2, 107) =4.62, p = .012, η2 = .77; in the 31–45-s period,F(2, 107) = 6.48, p = .002, η2 = .90; and in the46–60-s period, F(2, 107) = 5.05, p = . 008, η2 = .81.The TBI-A performed significantly more weaklythan did the other two groups at each time slice.

For the letter A, there was a significant differ-ence between the groups for the number of wordsproduced in the first 15 s, F(2, 107) = 4.46, p <

.014, η2 = .75; in the 16–30-s period, F(2, 107) =7.89, p = .001, η2 = .95; in the 31–45-s period,F(2, 107) = 11.97, p < .001, η2 = .99; and inthe 46–60-s period, F(2, 107) = 6.38, p = .002,η2 = .89. The TBI-A performed significantly moreweakly than did the other two groups at eachtime slice.

For the letter S, there was a significant differencebetween the groups for the number of words pro-duced in the first 15 s, F(2, 107) = 3.45, p < .034,η2 = .64; and in the 16–30-s period, F(2, 107) =9.70, p < .001, η2 = .98. For these two time slices,the two TBI groups performed significantly moreweakly than did the control group. There was asignificant difference between the groups for thenumber of words produced in the 31–45-s period,F(2, 107) = 11.88, p < .001, η2 = .99; and the num-ber produced in the 46–60-s period, F(2, 107) =16.59, p < . 001, η2 = .99. The TBI-A performedsignificantly more weakly than did the other twogroups at each time slice.

Clustering and switching

A MANCOVA was conducted to compare any dif-ferences in clustering and switching on the FAStask. After adjusting for covariates, a significantmultivariate effect was found, Wilks’s � = .736,F(3, 104) = 5.70, p < .001. Follow-up univari-ate analyses of each dependent variable found

Dow

nloa

ded

by [

La

Tro

be U

nive

rsity

] at

16:

32 1

0 Ja

nuar

y 20

15

ANOSMIA, EXECUTIVE DEFICIT, AND TBI 305

two significant results: The TBI-A group producedsignificantly fewer clusters than did the other twogroups, F(2, 105) = 11.98, p < .001, η2 = .99, andalso made significantly fewer switches, F(2, 105) =10.78, p < .001, η2 = .99. There was no differencein cluster size between the three groups, F(2, 105) =0.54, p = .59.

A MANCOVA was conducted to compare anydifferences in clustering and switching on theAnimal Fluency task. After adjusting for covari-ates, a significant multivariate effect was found,Wilks’s � = .738, F(6, 168) = 4.59, p < .001.Follow-up univariate analyses of each dependentvariable found two significant results: The TBI-Agroup produced significantly fewer clusters than didthe other two groups, F(2, 86) = 7.95, p = .001,η2 = .95, and both TBI groups produced signifi-cantly fewer switches than did the control group,F(2, 86) = 8.24, p = .001, η2 = .97. There was nodifference in cluster size between the three groups,F(2, 86) = 1.69, p = .19.

Clinical comparisons between the anosmicparticipants

Comparisons between the cognitive performancesof medically confirmed cases with anosmia andthose who self-reported this symptom revealedno significant differences between the groups

(all ps > .05). Post hoc comparisons using anal-ysis of variance revealed no significant differencebetween the two anosmic groups on the productionmeasures including SNST, FAS total words pro-duced, animals other words produced, or WCST–64 number of completed categories. Similarly, whenexamining the error measures, there was no signif-icant difference for SNST errors in the overall orself-corrected, FAS errors, animal fluency errors,WCST–64 perseverative, oddball, or total errors, orTMT-B errors.

Individuals who reported themselves to be eithertotally or partially anosmic were compared; onceagain there was no significant difference betweenthe two groups using univariate analyses of vari-ance to examine production and error measures (allps > .05).

It was also interesting to consider the perfor-mance of the two TBI groups in terms of failurerates on the executive measures. Table 4 presents acomparison of the number and percentage of eachof the TBI groups performing at or below the 5thpercentile of the control group on the executivetasks.

Once again it is clear that the TBI-A groupdemonstrates a higher percentage of participantsfalling below the 5% cutoff of the controls for pro-duction measures but with no consistent tendencyfor them to show greater frequency of error-proneperformances.

TABLE 4Frequency and percentage of TBIs performing equal to or below the lowest 5% of controls

Group (scores corrected for covariates)

Characteristic Control TBI-NA TBI-A

Production measuresStroop total words in 60 s 42 4 (18.18) 13 (38.24)FAS total words produced 28 10 (33.33) 17 (47.22)Animals total words produced 13 4 (13.33) 8 (24.24)WCST–64 completed categories 1 2 (12.5) 6 (37.50)

Error measuresStroop total errors 4 0 4 (11.76)FAS total errors 7 1 (3.33) 3 (8.33)Animals total errors 2 3 (10.00) 2 (6.06)WCST–64 total errors 26 5 (31.25) 6 (28.57)WCST–64 perseverative errors 21 1 (6.25) 4 (19.04)WCST–oddball errors 2 1 (6.25) 6 (28.57)Trail Making Test-B errors 2 2 (7.14) 6 (17.14)

Clustering and switchingFAS number of clusters 3 8 (26.67) 15 (41.67)FAS total switches 18 8 (26.67) 11 (30.56)Animal number of clusters 3 2 (6.67) 16 (48.48)Animal total switches 2 0 1 (3.03)

Note. TBI = traumatic brain injury. TBI-NA = TBI without anosmia. TBI-A = TBI with anosmia. WCST =Wisconsin Card Sorting Test. Number of participants, with percentages in parentheses.

Dow

nloa

ded

by [

La

Tro

be U

nive

rsity

] at

16:

32 1

0 Ja

nuar

y 20

15

306 CROWE AND CROWE

Affective functioning

A MANCOVA was conducted to compare differ-ences between the two TBI groups on the BDI andthe BAI. After adjusting for covariates, no signif-icant multivariate effect was found, Wilks’s � =.983, F(2, 59) = 0.52, p = .59. The BDI [TBI-Amean (SD): 23.7 (11.56); TBI-NA 27.7 (14.84)] andBAI [TBI-A mean (SD): 16.1 (10.56); TBI-NA 19.4(13.65)] only correlated with each other, r = .808,p < .001, and not with any of the other cognitivevariables.

Employment

No significant difference between the two TBIgroups was noted for employment status afterinjury, χ2(2) = 0.59, p = .74 [TBI-A: prein-jury, 100% full-time employed (FTE), postinjury,16.7% FTE; 13.9% part-time employed (PTE);69.4% unemployed; TBI-NA: 96.7% FTE, 3.3%PTE; postinjury, 23.3% FTE; 10.0% PTE; 66.7%unemployed]. In this sample at least, the presenceof anosmia did not of itself appear to be predictiveof employment status.

DISCUSSION

This study compared three groups [(30 anosmicTBIs (TBI-A) matched for PTA and memory func-tioning with 36 nonanosmic TBIs (TBI-NA) and51 controls] matched for age, gender, education,Full Scale IQ, and the Wechsler Working MemoryIndex, and featuring adequate symptom validity.The findings of the study indicate that the TBI-Agroup demonstrated considerably weaker perfor-mance on executive tasks overall than did thenonanosmic TBIs, which was not attributable tothe severity of their injury, ascertainment of theiranosmia, or their lack of effort. The finding thatthe anosmic TBIs seem to be more severely affectedon executive measures than nonanosmic patientsis similar to the previous observation of Callahanand Hinkebein (1999). TBI patients with traumaticanosmia who meet criteria for symptom valid per-formance did not show consistent evidence of aheightened tendency towards impulsivity as mea-sured by errors or disinhibited responding on theneuropsychological measures administered. In fact,if anything, their performance was characterizedby the reverse situation in that they tend to bemore slow and labored in their neuropsychologicalperformances as measured by tests of productiv-ity. As a result, the clinical validity of anosmia as

a sign of compromise in orbital prefrontal func-tioning culminating in more disinhibited and lessrule-governed responding must be very cautiouslyviewed. The data also did not support a linkbetween traumatic anosmia and the presence ofdifferential occupational outcome.

This finding is consistent with a number of pre-vious findings—for example, Correia, Faust, andDoty (2001) noted that the presence of anosmiaresulting from mild to moderate closed head injuryis not in itself a reliable predictor of subsequentvocational dysfunction. Similarly, Sigurdardottirand colleagues (2010) noted that both anosmic andnonanosmic TBI patients performed in a similarway on decision-making tasks (i.e., Iowa GamblingTask and the Delis–Kaplan Executive FunctionSystem, DKEFS), with both groups failing todevelop an advantageous strategy over time.

The findings of this study further undermine thenotion of a direct relationship between anosmiaand cognitive functioning and expand the scope ofthe earlier observations made by Correia and col-leagues (2001) that “the main finding of this studycall into question the assertion of Varney (1988)and Martzke et al. (1991) that anosmia associatedwith TBI is a strong marker of vocational dys-function related to orbital frontal damage. Thisdoes not rule out the usefulness of anosmia asa possible indicator of orbitofrontal damage, nordoes it disconfirm a predictive link between CHI-anosmia, orbitofrontal damage and vocational dys-function in some cases” (p. 487). On the basis ofthe current series of findings, it would appear thatanosmia does not reliably predict cognitive perfor-mance, particularly that associated with measuresof disinhibition and rule breaking on standardizedneuropsychological measures, and once again thestudy indicates that olfactory impairment does notreliably predict vocational outcome.

The injuries that result in anosmia following TBIare multiply determined (i.e., nasal obstruction,shearing of the olfactory nerves, and damage to thebasal forebrain: Wise, Moonis, & Mirza, 2006), andcareful screening for the causes of the anosmia areimportant in determining the likely clinical implica-tions of the injury. Even despite this issue, however,a direct relationship between olfactory compromise,executive deficits, disinhibition, and poor employ-ment outcome continues to remain elusive.

In a previous study (Crowe, 1992), it was notedthat subjects with TBI performed with higher levelsof error but with relatively lower levels of produc-tion than did the normal controls. In the subse-quent study (Crowe, 1996), in which anosmic versusnonanosmic TBIs were compared, both TBI groupsproduced levels of production comparable to those

Dow

nloa

ded

by [

La

Tro

be U

nive

rsity

] at

16:

32 1

0 Ja

nuar

y 20

15

ANOSMIA, EXECUTIVE DEFICIT, AND TBI 307

observed in the previous study, considerably belowthose noted with the normal controls. Nonetheless,it was possible to distinguish the two groups ofTBI subjects in terms of their level of error, withthe anosmic participants producing more errorsthan did the nonanosmic TBIs. Clearly with themore comprehensive data collection afforded by thecurrent study, it is evident that this original obser-vation has not been replicated, and the observationsmade on such a small sample of patients should inretrospect be cautiously viewed.

Methodologically the manner of ascertainmentof posttraumatic anosmia employed in this studywas less than optimal, and in future studies theuse of more uniform and objective methods ofdetermining anosmia, including the use of instru-ments such as the University of Pennsylvania SmellIdentification Test, is to be highly recommended.It would also be prudent in further study to employmore fine-grained tasks of disinhibition, includinginstruments such as the Iowa Gambling Task orthe Balloon Analogue Risk Task, to ensure thatany more subtle aspects of disinhibition affectingthe performances, but which are not being notedwith the more coarse-grained instrumentation, areappropriately noted.

The results of this study indicate that the anosmicgroup demonstrated considerably weaker perfor-mance on executive tasks overall. However, TBIpatients with traumatic anosmia did not show con-sistent evidence of a heightened tendency towardsimpulsivity as measured by errors or disinhib-ited responding. If anything, their performancewas characterized by the reverse situation in thatthey tended to be more slow and labored in theirneuropsychological performances as measured bytests of productivity. It thus seems clear thatanosmia does not of itself reliably predict cogni-tive performance, particularly that associated withmeasures of disinhibition and rule breaking onstandardized neuropsychological measures, and thestudy once again indicates that olfactory impair-ment does not reliably predict vocational or emo-tional outcome.

Original manuscript received 15 October 2012Revised manuscript accepted 25 January 2013

First published online 21 February 2013

REFERENCES

Army Individual Test Battery. (1944). Manual of direc-tions and scoring. Washington, DC: War Department,Adjutant General’s Office.

Beck, A. T., & Steer, R. A. (1993). Beck Anxiety Inventorymanual. San Antonio, TX: The PsychologicalCorporation.

Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manualfor Beck Depression Inventory–II . San Antonio, TX:Psychological Corporation.

Bittner, R. M., & Crowe, S. F. (2006). The relation-ship between naming difficulty and FAS performancefollowing traumatic brain injury. Brain Injury, 20,971–980.

Callahan, C. D., & Hinkebein, J. (1999). Neuro-psychological significance of anosmia followingtraumatic brain injury. Journal of Head TraumaRehabilitation, 14, 581–587.

Correia, S., Faust, D., & Doty, R. L. (2001). Are-examination of the rate of vocational dysfunc-tion among patients with anosmia and mild tomoderate closed head injury. Archives of ClinicalNeuropsychology, 16, 477–488.

Costanzo, R. M., & Becker, D. P. (1986). Smell and tastedisorders in head injury and neurosurgery patients. InH. L. Meiselman & R. S. Rivlin (Eds.), Clinical mea-surements of taste and smell (pp. 565–578). New York,NY: Macmillan.

Costanzo, R. M., & Zasler, N. D. (1991). Head trauma.In T. Getchell, R. Doty, L. Bartoshuk, & J. Snow(Eds.), Smell and taste in health and disease (pp.711–730). New York, NY: Raven Press.

Costanzo, R. M., & Zasler, N. D. (1992). Epidemiologyand pathophysiology of olfactory and gustatory dys-function in head trauma. Journal of Head TraumaRehabilitation, 7, 15–24.

Courville, C. B. (1945). Pathology of the central nervoussystem (2nd ed.). Mountain View, CA: Pacific Press.

Crowe, S. F. (1992). Dissociation of two frontal lobesyndromes by a test of verbal fluency. Journalof Clinical and Experimental Neuropsychology, 14,327–339.

Crowe, S. F. (1996). Traumatic anosmia coincides withan organic disinhibition syndrome as assessed byperformance on a test of verbal fluency. Psychiatry,Psychology and Law, 3, 39–45.

Crowe, S. F. (1998a). Decrease in performance on theverbal fluency test as a function of time: Evaluationin a young healthy sample. Journal of Clinical andExperimental Neuropsychology, 20, 391–401.

Crowe, S. F. (1998b). The differential contribution ofmental tracking, cognitive flexibility, visual search andmotor speed to performance on Parts A and B of theTrail Making Test. Journal of Clinical Psychology, 54,585–591.

Cummings, J. L. (1985). Clinical neuropsychiatry. NewYork, NY: Grune and Stratton.

de Kruijk, J., Leffers, P., Menheere, P., Meershoff, S.,Rutten, J., & Twijnstra, A. (2003). Olfactory functionafter mild traumatic brain injury. Brain Injury, 17,73–78.

Doty, R., Yousem, D., Pham, L., Kreshak, A., Geckle,R., & Lee, W. (1997). Olfactory dysfunction inpatients with head trauma. Archives of Neurology, 54,1131–1140.

Green, P., Rohling, M., Iverson, G., & Gervais, R. (2003).Relationship between olfactory discrimination andhead injury severity. Brain Injury, 17, 479–496.

Greiffenstein, M. F., Baker, W. J., & Gola, T. (2002).Brief report: Anosmia and remote outcome in closedhead injury. Journal of Clinical and ExperimentalNeuropsychology, 24, 705–709.

Greiffenstein, M. F., Baker, W. J., & Gola, T. (2003).Straw man walking: Reply to Varney (2002). Journal

Dow

nloa

ded

by [

La

Tro

be U

nive

rsity

] at

16:

32 1

0 Ja

nuar

y 20

15

308 CROWE AND CROWE

of Clinical and Experimental Neuropsychology, 25,152–154.

Hurks, P. P. M. (2012). Does instruction in semantic clus-tering and switching enhance verbal fluency in chil-dren? The Clinical Neuropsychologist, 26, 1019–1037.

Jennett, B., & Teasdale, G. (1981). Management of headinjuries. Philadelphia, PA: F. A. Davis.

Killgore, W. S., & Della Pietra, L. (2000). Using theWMS–III to detect malingering: Empirical validationof the Rarely Missed Index (RMI). Journal of Clinicaland Experimental Neuropsychology, 22, 761–771.

Kongs, S. K., Thompson, L. L., Iverson, G. L., &Heaton, R. K. (2000). WCST–64: Wisconsin CardSorting Test–64 Card Version, professional manual.Odessa, FL: Psychological Assessment Resources.

Levin, H. S., High, W. M., & Eisenberg, H. M. (1985).Impairment of olfactory recognition after closed headinjury. Brain, 108, 579–591.

Lichter, D. G., & Cummings, J. L. (2001). Frontal-subcortical circuits in psychiatric and neurological dis-orders. New York, NY: Guilford Press.

Malloy, P., Bihrle, A., Duffy, J., & Cimino, C. (1993). Theorbitomedial frontal syndrome. Archives of ClinicalNeuropsychology, 8, 185–201.

Martzke, J. S., Swan, C. M., & Varney, N. R. (1991).Neuropsychological and neuropsychiatric characteris-tics of patients with post-traumatic damage to orbitalfrontal cortex. Neuropsychology, 5, 213–226.

Ogawa, T., & Rutka, J. (1999). Olfactory dysfunc-tion in head injured workers. Acta OtolaryngolicaSupplement, 540, 50–57.

Öngür, D., & Price, J. L. (2000). The organization of net-works within the orbital and medial prefrontal cortexof rats, monkeys and humans. Cerebral Cortex, 10,206–219.

Psychological Corporation. (2001). Manual for theWechsler Test of Adult Reading (WTAR). SanAntonio, TX: Author.

Rey, A. (1964). L’examen clinique en psychologie.[Clinical examination in psychology.] Paris, France:Presses Universitaires de France.

Rombaux, P., Mouraux, A., Bertrand, B., Nicolas, G.,Duprez, T., & Hummel, T. (2006). Retronasal andorthonasal olfactory function in relation to olfactorybulb volume in patients with posttraumatic loss ofsmell. Laryngoscope, 116, 901–905.

Ruffolo, L. F., Guilmette, T. J., & Willis, W. G.(2000). Completion of time and error rates onthe Trail Making Test among patients with headinjuries, experimental malingerers, patients with sus-pect effort on testing and normal controls. TheClinical Neuropsychologist, 14, 223–230.

Russell, R., & Smith, A. (1961). Post traumatic amnesiain closed head injury. Archives of Neurology, 5, 16–29.

Sigurdardottir, S., Jerstad, T., Andelic, N., Roe, C.,& Schanke, A. K. (2010). Olfactory dysfunction,gambling task performance and intracranial lesions

after traumatic brain injury. Neuropsychology, 24,504–513.

Tate, R. L. (1999). Executive dysfunction and charactero-logical changes after traumatic brain injury: Two sidesof the same coin? Cortex, 35, 39–55.

Tombaugh, T. N. (1996). The Test of Memory Malin-gering (TOMM). Toronto, Canada: Multi-HealthSystems.

Tombaugh, T. N., Kozak, J., & Rees, L. (1999).Normative data stratified by age and education fortwo measures of verbal fluency: FAS and animalnaming. Archives of Clinical Neuropsychology, 14,167–177.

Trennery, M. R., Crosson, B., DeBoe, J., & Leber, W.R. (1988). Stroop Neurological Screening Test manual.Odessa, FL: Psychological Assessment Resources.

Troyer, A. K. (2000). Normative data for clustering andswitching on verbal fluency tests. Journal of Clinicaland Experimental Neuropsychology, 22, 370–378.

Troyer, A. K., Moscovitch, M., & Winocur, G. (1997).Clustering and switching as two components of ver-bal fluency: Evidence from younger and older healthyadults. Neuropsychology, 11, 138–146.

Troyer, A. K., Moscovitch, M., Winocur, G., Alexander,M., & Stuss, D. (1998). Clustering and switchingon verbal fluency: The effects of focal frontal—and temporal—lobe lesions. Neuropsychologia, 36,499–504.

Varney, N. R. (1988). Prognostic significance ofanosmia in patients with closed-head trauma. Journalof Clinical and Experimental Neuropsychology, 10,250–254.

Varney, N. R. (2002). Wishing doesn’t make it so: A replyto Greiffenstein, Baker and Gola. Journal of Clinicaland Experimental Neuropsychology, 24, 852–853.

Varney, N. R., & Bushnell, D. (1998). NeuroSPECTfindings in patients with posttraumatic anosmia:A quantitative analysis. Journal of Head TraumaRehabilitation, 13, 63–72.

Varney, N. R., & Menefee, L. (1993). Psychosocialand executive deficits following closed head injury:Implications for orbital frontal cortex. Journal ofHead Trauma Rehabilitation, 8, 32–41.

Varney, N., Pinkston, J., & Wu, J. (2001). QuantitativePET findings in patients with posttraumatic anosmia.Journal of Head Trauma Rehabilitation, 16, 253–259.

Wechsler, D. (1997). Wechsler Memory Scale–ThirdEdition. San Antonio, TX: The PsychologicalCorporation.

Wise, J. B., Moonis, G., & Mizra, N. (2006). Magneticresonance imaging findings in the evaluation of trau-matic anosmia. Annals of Otology, Rhinology andLaryngology, 115, 124–127.

Yousem, D. M., Geckle, R. J., Bilker, W. B., McKeown,D. A., & Doty, R. L. (1996). Posttraumatic olfactorydysfunction: MR and clinical evaluation. AmericanJournal of Neuroradiology, 17, 1171–1179.

Dow

nloa

ded

by [

La

Tro

be U

nive

rsity

] at

16:

32 1

0 Ja

nuar

y 20

15