Utility of Green s Word Memory Test Free Recall Subtest as a Measure of Verbal Memory: Initial Evidence from a Temporal Lobe Epilepsy Clinical Sample

Transcription

1 Archives of Clinical Neuropsychology 31 (2016) Utility of Green s Word Memory Test Free Recall Subtest as a Measure of Verbal Memory: Initial Evidence from a Temporal Lobe Epilepsy Clinical Sample Jason R. Soble 1, Katie E. Osborn 2, Michelle L. Mattingly 2, Fernando L. Vale 3, Selim R. Benbadis 4, Nancy T. Rodgers-Neame 5, Mike R. Schoenberg 2,3,4, * 1 Psychology Service, South Texas Veterans Health Care System, San Antonio, TX, USA 2 Department of Psychiatry and Behavioral Neurosciences, Morsani College of Medicine, University of South Florida, Tampa, FL, USA 3 Department of Neurosurgery and Brain Repair, Morsani College of Medicine, University of South Florida, Tampa, FL, USA 4 Department of Neurology, Morsani College of Medicine, University of South Florida, Tampa, FL, USA 5 Private Practice, Florida Comprehensive Epilepsy and Seizure Disorders Center, Tampa, FL, USA *Corresponding author at: 3515 E. Fletcher Avenue, MDC Box 14, Tampa, FL 33613, USA. Tel.: address: mschoenb@health.usf.edu (M.R. Schoenberg). Accepted 4 November 2015 Abstract This study investigated the Word Memory Test (WMT) Free Recall (FR) subtest as a conventional memory measure. Nineteen participants with pharmacoresistant left temporal lobe epilepsy (LTLE) and 16 with right temporal lobe epilepsy (RTLE) completed the WMT, Rey Auditory Verbal Learning Test (RAVLT), and Wechsler Memory Scale-Fourth Edition Logical Memory (LM) subtest during presurgical evaluation. LTLE participants performed significantly worse on FR subtest (p,.05, h 2 p = 0.17) and RAVLT Trial 7 (p,.01, h2 p = 0.25), but not on LM subtest. Age was a significant covariate for FR (p,.01, h 2 p = 0.22). Logistic regression revealed FR plus age and RAVLT age-adjusted T-scores both yielded 77.1% classification accuracy and respective diagnostic odds ratios of and Receiver operating characteristic curves to classify seizure laterality found that RAVLT and FR were significant (area under the curve [AUC] ¼ 0.82 and 0.74), whereas LM was nonsignificant (AUC ¼ 0.67). Cut scores and positive/negative predictive values were established for improved clinical classification. Keywords: Epilepsy; Assessment; Learning and memory; Lateralization Introduction Formal assessment of performance validity is an essential component of every neuropsychological evaluation (Heilbronner, Sweet, Morgan, Larrabee, & Millis, 2009). The Word Memory Test (WMT) is one such commonly used performance validity test (PVT) designed to detect invalid task performance and contains three primary effort subtests (i.e., Immediate Recognition [IR], Delayed Recognition [DR], and Consistency [CNS]), and three conventional memory subtests, including two easy memory tasks (i.e., Multiple Choice [MC] and Paired Associates [PA]) and one hard memory task (i.e., Free Recall [FR]) (Green, 2003; Green, Allen, & Astner, 1996). In prior studies, the effort indices generally have differentiated noncredible performance from genuine cognitive impairment (e.g., Carone, 2014; Carone, Green, & Drane, 2013; Flaro, Green, & Robertson, 2007; Greiffenstein, Greve, Bianchini, & Baker, 2008; Rienstra et al., 2012). In addition to extensive data supporting the WMT as a PVT, there also have been initial data to suggest that the FR subtest may simultaneously function as a bona fide verbal memory measure (e.g., Armistead-Jehle, Green, Gervais, & Hungerford, 2015; Carone et al., 2013; Davis & Wall, 2014; Eichstaedt et al., 2014; Rienstra, Spaan, & Schmand, 2009). # The Author Published by Oxford University Press. All rights reserved. For permissions, please journals.permissions@oup.com. doi: /arclin/acv084 Advance Access publication on 12 December 2015

2 80 J.R. Soble et al. / Archives of Clinical Neuropsychology 31 (2016); WMT and Genuine Memory Impairment Past findings raised concern that the WMT may incorrectly identify some patients with neurological disease/injury and severe memory deficits as having noncredible performance (e.g., Allen, Bigler, Larsen, Goodrich-Hunsaker, & Hopkins, 2007; Allen, Wu, & Bigler, 2011; Greve, Ord, Curtis, Bianchini, & Brennan, 2008; Merten, Bossink, & Schmand, 2007). As such, an algorithm referred to as the dementia profile or genuine memory impairment profile (GMIP) was developed by comparing the means of the primary effort subtests with the conventional memory subtests (see Green, 2003, 2005; Green, Montijo, & Brockhaus, 2011; Rienstra, Twennaar, & Schmand, 2013). In addition to identifying invalid performance, emerging findings suggest that the FR subtest component of the GMIP may also be sensitive to genuine verbal memory deficits. For instance, Carone and colleagues (2013) described two cases in which patients with left anterior hippocampal resections performed above cutoff on the WMT effort subtests, but impaired on the FR subtest. Goodrich-Hunsaker and Hopkins (2009) also noted profound FR deficits despite normal primary effort performance in three amnestic patients. Eichstaedt and colleagues (2014) found that the FR subtest accurately differentiated patients with left temporal lobe epilepsy (LTLE) and right temporal lobe epilepsy (RTLE), but with a smaller effect size relative to the Rey Auditory Verbal Learning Test (RAVLT; Rey, 1941; Schmidt, 1996). Armistead-Jehle and colleagues (2015) found that the FR subtest performed similarly to the California Verbal Learning Test (CVLT; Delis, Kramer, Kaplan, & Ober, 1987) and CVLT-II (Delis, Kramer, Kaplan, & Ober, 2000) as a measure of verbal memory. In contrast, Donders and Strong (2013) reported that none of the WMT conventional memory subtests differed significantly between groups with varying traumatic brain injury severity nor correlated with coma duration. Davis and Wall (2014) also highlighted limitations of current WMT normative data as 40% of their sample had at least one impaired WMT score, but performed normally on the CVLT-II. Clearly, additional research is needed to evaluate whether the FR subtest functions as a clinically useful verbal memory measure. WMT and Association with Neuropathological Disease Individuals with TLE commonly exhibit neuropsychological deficits, particularly in verbal memory (e.g., Elger, Helmstaedter, & Kurthen, 2004; Thompson & Duncan, 2005), and reflect a unique sample with material-specific memory impairment in which the FR subtest could be investigated. In prior studies, RAVLT scores have outperformed other verbal memory tests to predict laterality of seizure onset and ultimate side of surgical resection (Loring et al., 2008; Soble et al., 2014) and shared common variance with the FR subtest (e.g., Eichstaedt et al., 2014; Rienstra et al., 2012). As such, the primary goal of this study was to examine the WMT FR subtest as verbal memory measure among those with video-eeg diagnosed TLE in comparison with the RAVLT. The Wechsler Memory Scale-Fourth Edition (WMS-IV; Wechsler, 2009) Logical Memory subtest was also included as an additional verbal memory comparison measure, given its common use in epilepsy evaluations and some evidence of sensitivity to LTLE (Barr et al., 1997; Helmstaedter, Wietzke, & Lutz, 2009; Hermann et al., 1996; Jones-Gotman et al., 2010; Lencz et al., 1992; Wilde et al., 2001). Moreover, suboptimal performance validity has been found even among clinically presenting epilepsy surgery candidates without external incentive (e.g., Loring, Lee, & Meador, 2005), which further highlighted the necessity of incorporating performance validity assessment in every evaluation (see Heilbronner et al., 2009). Thus, if the WMT FR subtest is found to be sensitive to verbal memory deficits, it may eventually decrease evaluation time burden and associated health-care expenditures by allowing the clinician to obtain data on performance validity and verbal memory functioning with a single measure while also adhering to evidence-based neuropsychological practice and professional guidelines (Board of Directors (2007) of the American Academy of Clinical Neuropsychology, 2007; Chelune, 2010; Patient Protection and Affordable Care Act, 42 U.S.C et seq., 2010). A second aim was to evaluate if adding an age covariate increased the accuracy of the WMT FR score to identify patients with TLE, which would support the need to develop age-corrected normative data for this subtest. We hypothesized that the FR subtest would demonstrate convergent validity with the RAVLT, and its overall classification accuracy for identifying patients with LTLE would increase after the age covariate was added. Method Participants Participants for this case-controlled study included 19 patients with LTLE and 16 demographically and clinically similar patients with RTLE diagnosed based on video-eeg and administered a comprehensive outpatient neuropsychological evaluation as part of a presurgical evaluation for the treatment of localization-related TLE. Participants were selected from a consecutive series of patients meeting inclusion criteria for this study, obtained from an archival database approved by the Institutional Review Board of a large academic medical center as part of a larger study. Each participant provided written informed

3 consent. All participants were diagnosed with localization-related TLE with seizure onset identified from the temporal, mesial temporal, or frontotemporal cortical areas and consistent seizure semiology. Seizure onset was localized to unilateral temporal/mesial temporal region based on video-eeg for 17 LTLE and 15 RTLE participants. For two with LTLE and one with RTLE, video-eeg localized seizure onset to the frontotemporal region, with other clinical data (seizure semiology, magnetic resonance imaging [MRI], etc.) indicating temporal/mesial temporal seizure localization. In order to meet inclusion criteria, participants were at least 18 years of age; fluent in English; absence of past or present comorbid neurological disease or systemic infection; and had video-eeg confirmed TLE, MRI study of brain without evidence of contralateral temporal lobe pathology (e.g., contralateral mesial temporal sclerosis), and a neuropsychological study without evidence of invalid performance on performance validity measures. All patients with LTLE underwent intracarotid methohexital (Wada) procedure (n ¼ 19). In addition, those RTLE patients with increased risk of cerebral reorganization secondary to left handedness and/or family history of left handedness as self-reported during clinical interview, or known or suspected seizure onset prior to age 6 years, underwent Wada procedure to assess for atypical hemispheric dominance for language (n ¼ 8). All patients in the study were left hemisphere dominant for language. No potential participants exhibited contralateral temporal lobe pathology or evidence of suboptimal task engagement on performance validity assessment. As such, none were excluded on either of these bases. Measures J.R. Soble et al. / Archives of Clinical Neuropsychology 31 (2016); Neuropsychological assessment procedures were completed independent of video-eeg monitoring and included administration of the WMT, RAVLT, and WMS-IV. For the purposes of this study, neuropsychological data were not used to lateralize/localize seizure onset to avoid confounding group assignment by dependent variable measures. Video-EEG, MRI study of brain with and without contrast, and, in many cases, methohexital intracarotid (Wada) procedure also were obtained in order to identify seizure lateralization and localization, presence or absence of structural lesion, and lateralization of language and memory functioning, respectively. Surface video-eeg long-term monitoring (International system) was conducted using XLTEK (Oakville, Ontario, Canada) in order to lateralize/localize seizure onset, using bilateral basilar temporal placements, such as T1 and T2 electrodes (see Benbadis et al., 2009 for greater detail), and were interpreted by neurologists with board certification in neurophysiology. MRI studies were completed using GE 3.0 Tesla scanners obtaining T1- and T2-weighted studies and read by board certified neuroradiologists to identify the presence or absence of MTS with either hippocampal atrophy and/or T2 signal change in the mesial temporal lobe structures (Sanches-Alvarez, Serrano-Castro, & Canadillas-Hidalgo, 2002). For the Wada procedure, the standard protocol outlined by Trenerry and Loring (1995) was modified from the original amobarbital protocol for methohexital use (Buchtel, Passaro, Selwa, Deveikis, & Gomez-Hassan, 2002). Rey Auditory Verbal Learning Test. The RAVLT is an orally presented, 15-word verbal learning and memory measure (Strauss, Sherman, & Spreen, 2006) found to accurately differentiate patients with LTLE from RTLE (Loring et al., 2008; Sherman et al., 2011; Soble et al., 2014). Standardized T-scores (i.e., M ¼ 50; SD ¼ 10) were derived from age-matched normative data contained in Schmidt (1996). The primary outcome measure for this variable was Trial 7 (total recall after a 30-min delay) age-corrected standardized T-score. Word Memory Test. The WMT is a computer administered PVT consisting of 20 word pairs presented at a rate of one pair per 6 s and assesses performance patterns on a series of memory tasks, enabling evaluation of task engagement (Green, 2003; Green, Iverson, & Allen, 1999; Green et al., 1996). The WMT includes three primary effort indices (IR, DR, and CNS), as well as two easy memory indices (MC and PA) and one difficult memory index (FR) with a 30-min interval between the immediate and delayed trials. Per manual guidelines, the task failure cutoff score was set at 82.5 raw percentage correct or below for each effort index (Green et al., 1996). In cases of adequate task engagement in the context of significant memory deficits, a difference score of 30 or greater is expected for the mean of the effort indices and the mean of the memory indices (Green et al., 2011), per GMIP guidelines. For this study, all participants had adequate task engagement as evidenced by either (1) primary effort index scores above established cutoffs or (2) a GMIP indicative of adequate task engagement. Wechsler Memory Scale-Fourth Edition. The WMS-IV is a battery of memory tests designed to assess verbal/auditory, visual, and working memory domains. Within the WMS-IV battery, the Logical Memory II subtest is a measure of delayed verbal/auditory memory consisting of orally presented narratives which examinees are asked to recall after a min delay. Age-adjusted Logical Memory II scores have been evaluated across multiple neurological diseases, with evidence from prior studies showing correlations with dominant hemisphere temporal lobe dysfunction as typically seen in Alzheimer s dementia and LTLE (e.g., Barr, 1997; Petersen et al., 2000). Additionally, Helmstaedter and colleagues (2009) found Logical Memory on the WMS-Revised (WMS-R; Wechsler, 1987) differentiated temporal lobe epilepsy from extra-temporal lobe epilepsy.

4 82 J.R. Soble et al. / Archives of Clinical Neuropsychology 31 (2016); Performances on the Logical Memory subtest of the original WMS have also been shown to correlate with left hippocampal MRI measurements among patients with LTLE (Lencz et al., 1992). WMS-IV Logical Memory II performances are presented as age-adjusted scaled scores (i.e., M ¼ 10; SD ¼ 3). Data Analysis Between-group differences in demographic and clinical characteristics by seizure laterality (LTLE vs. RTLE) were evaluated with analyses of variance (ANOVAs). Between-group differences in WMT FR raw percentage correct scores, RAVLT Trial 7 age-adjusted T-scores, and WMS-IV Logical Memory II age-adjusted scaled scores were compared using ANOVA. Receiver operating characteristic (ROC) curves were used for each of these three measures to classify patients by seizure laterality. As the RAVLT and WMT FR ROC curves reached statistical significance, cut scores were developed to identify scores for both measures suggestive of LTLE or RTLE at a 0.80 sensitivity level. Because the WMT FR is the only score for which age-adjusted standardized scores have not been developed, between-group FR subtest differences were reassessed via ANCOVA with age as a covariate. A logistic regression was used to evaluate the amount of variance in seizure laterality accounted for by WMT FR scores to further evaluate the clinical utility of the WMT FR subtest to lateralize temporal lobe dysfunction, with seizure laterality as the dependent variable (0 ¼ LTLE; 1 ¼ RTLE). In order to assess WMT FR both as a raw percentage score and with age adjustment, FR was entered into the equation as the first predictor, with the next step allowing age to enter the algorithm. A second logistic regression was performed with RAVLT Trial 7 age-adjusted T-score to predict side of TLE. Given the overlapping constructs of verbal measures being analyzed, tests are not orthogonal in nature and, therefore, familywise error rate corrections are not deemed appropriate in the context of this study (see Brandt, 2008). Accordingly, an a level of p,.05 was used. Results See Table 1 for participant demographics and clinical characteristics. The overall sample (n ¼ 35) ranged in age from 19 to 56 years (M ¼ 40.86, SD ¼ 11.10), was predominantly Caucasian (77.1%), and was approximately evenly split by gender (60% men). There were no significant differences between LTLE and RTLE groups in terms of age; educational attainment; epilepsy duration; age at epilepsy diagnosis; Beck Depression Inventory, Second Edition (Beck, Steer, & Brown, 1996) symptom endorsement; or Wechsler Adult Intelligence Scale, Fourth Edition (WAIS-IV; Wechsler, 2008) Perceptual Reasoning Index scores. In contrast, the LTLE group had a significantly lower Full Scale IQ, and showed a nonsignificant trend toward lower WAIS-IV Verbal Comprehension Index scores relative to the RTLE group (p ¼.06). Rey Auditory Verbal Learning Test The LTLE group performed significantly worse than the RTLE group on RAVLT Trial 7 with a large effect size, F(1, 33) ¼ 10.95, p,.01, h 2 p = 0.25 (see Table 1). ROC analysis was significant (p ¼.001) with an area under the curve (AUC) of The logistic regression model with RAVLT Trial 7 age-adjusted T-score as the predictor was significant (p,.01, Exp(B) ¼ 1.08; Table 1. Demographic and clinical characteristics by hemisphere of seizure onset Left (n ¼ 19) Right (n ¼ 16) p Age in years (10.28) (11.43).10 Years of education (2.09) (2.58).31 Age of epilepsy onset (13.14) (13.62).50 Epilepsy duration (11.47) (15.64).51 Beck Depression Inventory-II raw score (9.36) (11.83).29 Full Scale IQ a (11.64) (11.31).04, h 2 p = 0.12 Verbal Comprehension Index a (10.86) (13.76).06 Perceptual Reasoning Index a (10.31) (13.14).28 RAVLT (Trial 7) (14.73) (14.32) WMT FR (16.21) (15.41).002, h 2 p = , h 2 p = , h 2 p = 0.22 WMS-IV Logical Memory II 6.32 (3.70) 8.69 (3.94).08 WMT FR (age covaried) Notes: Values are presented as mean (SD). RAVLT ¼ Rey Auditory Verbal Learning Test (age-adjusted T-score). WMT FR ¼ Word Memory Test, Free Recall subtest (raw percentage correct). WMS-IV ¼ Wechsler Memory Scale, Fourth Edition (age-adjusted scaled score). a Wechsler Adult Intelligence Scale, Fourth Edition (WAIS-IV) prorated standard scores.

5 J.R. Soble et al. / Archives of Clinical Neuropsychology 31 (2016); x 2 ¼ 9.80, p,.01) and yielded an overall classification rate of 77.1%. The algorithm resulted in a positive and negative predictive power of 0.76 and 0.79, respectively, for LTLE, and a diagnostic odds ratio of (see Table 2). In order to develop cut scores for clinical interpretation of RAVLT performance, score coordinates of the ROC curve were plotted with lower T-scores classifying LTLE and higher T-scores classifying RTLE, and cutoffs were established to correctly classify seizure laterality that achieved a minimum sensitivity rate of T-scores at or below 36.5 classified participants as belonging to the LTLE group and yielded a minimum sensitivity rate of 0.81 with specificity of 0.79 and a positive likelihood ratio of T-score ranges between 37 and 41 were equivocal and failed to yield classification rates exceeding a sensitivity of T-scores at or above 41.5 yielded a minimum sensitivity rate of 0.84 with specificity of 0.69 and a positive likelihood ratio of 2.69 to predict RTLE. WMT FR Subtest The LTLE group performed significantly worse than the RTLE group on FR subtest with a moderate effect size, F(1, 33) ¼ 6.65, p,.05, h 2 p = 0.17 (see Table 1). ROC analysis was significant (p ¼.02) and yielded an AUC of When including age as a covariate for FR scores, ANCOVA was significant, and group differences were large, F(1, 32) ¼ 9.25, p,.01, h 2 p = 0.22 (Table 1). Using step-wise logistic regression, the model with FR was significant [p,.05, Exp(B) ¼ 1.06] and resulted in a significant improvement in laterality classification from chance (x 2 ¼ 6.33, p,.05) with an overall classification rate of 68.6%. Logistic regression with age (p,.05, Exp(B) ¼ 1.09) and FR scores (p,.05, Exp(B) ¼ 1.08) also was significant (x 2 ¼ 5.24, p,.05) and provided an overall classification rate of 77.1% as well as positive and negative predictive values of 0.79 and 0.75, respectively. The logistic regression combining age and WMT FR scores resulted in an overall diagnostic odds ratio of for predicting LTLE (Table 3). Similar to the RAVLT, score coordinates of the FR ROC curve were plotted in order to develop cut scores for clinical interpretation, with lower raw scores classifying LTLE and higher raw scores classifying RTLE. Cutoff scores again were required to yield a minimum sensitivity rate of 0.80 for correctly classifying seizure laterality. Based on these coordinates of the curve, raw scores of 38.75% yielded a minimum sensitivity of 0.81 with a specificity of Table 2. Logistic regression analysis for hemisphere of seizure onset with Rey Auditory and Verbal Learning Test (RAVLT) Trial 7 (30-min delay) T-scores Variable B Standard error Wald Significance Exp(B) RAVLT 30 Delay Constant Log likelihood ¼ Nagelkerke R 2 ¼.33 Overall classification % ¼ 77.1 N ¼ 35 Model x 2 ¼ 9.80, p,.01 PPV ¼ 0.76 NPV ¼ 0.79 DOR ¼ LR+ ¼ 2.69 [95% CI 1.27, 5.72] LR2 ¼ 0.23 [95% CI 0.08, 0.68] Note: PPV ¼ positive predictive value for identifying left temporal lobe epilepsy (LTLE); NPV ¼ negative predictive value for identifying LTLE; DOR ¼ diagnostic odds ratio; LR+ ¼ positive likelihood ratio predicting left temporal lobe epilepsy; LR2 ¼ negative likelihood ratio. Table 3. Logistic regression analysis for hemisphere of seizure onset with Green s Word Memory Test (WMT) free recall subtest and age Variable B Standard error Wald Significance Exp(B) WMT Free Recall Age Constant Log likelihood ¼ Nagelkerke R 2 ¼.38 Overall classification % ¼ 77.1 N ¼ 35 Model x 2 ¼ 11.57, p,.01 PPV ¼ 0.79 NPV ¼ 0.75 DOR ¼ LR+ ¼ 1.96 [95% CI 0.99, 3.91] LR2 ¼ 0.42 [95% CI 0.18, 0.98] Note: PPV ¼ positive predictive value for identifying left temporal lobe epilepsy (LTLE); NPV ¼ negative predictive value for identifying LTLE; DOR ¼ diagnostic odds ratio; LR+ ¼ positive likelihood ratio predicting LTLE; LR2 ¼ negative likelihood ratio.

6 84 J.R. Soble et al. / Archives of Clinical Neuropsychology 31 (2016); and positive likelihood ratio of 1.72 for diagnosing LTLE, whereas WMT FR raw scores of 53.75% yielded a minimum sensitivity rate of 0.84 with a specificity of 0.56 and positive likelihood ratio of 1.92 to correctly classify RTLE. In order to assess whether WMT FR subtest scores improved seizure classification in comparison with the RAVLT alone, a third logistic regression was developed in which WMT FR subtest scores and age were included in a single step followed by RAVLT T-scores as the next step. This model resulted in a classification rate of 77.1%, which was equivalent to the classification rate with RAVLT as the sole predictor variable, as well as the logistic regression containing the combined WMT FR scores with age. Within this third model, the model was significant (x 2 ¼ 13.89, p,.01), but none of the predictor variables retained their significance, suggesting a high degree of overlapping variance accounted for by the RAVLT and WMT FR subtest scores. WMS-IV Logical Memory II There was no significant difference between LTLE and RTLE groups in terms of Logical Memory II scaled scores, F(1, 33) ¼ 3.36, p ¼.08 (see Table 1). The ROC curve also was nonsignificant (AUC ¼ 0.67, p ¼.09). Thus, no further analyses were performed for this measure. Discussion Findings of this study mirror previous research (e.g., Loring et al., 2008; Sherman et al., 2011; Soble et al., 2014) again supporting the sensitivity of the RAVLT to mesial temporal dysfunction of the language dominant hemisphere. Further, these findings provide evidence of convergent validity between the WMT FR subtest and RAVLT and support emerging data to suggest that the FR subtest may also be a bona fide test of verbal memory (Armistead-Jehle et al., 2015; Carone et al., 2013; Davis & Wall, 2014; Goodrich-Hunsaker & Hokins, 2009). The lack of significant difference in WMS-IV Logical Memory subtest age-adjusted scaled scores between LTLE and RTLE was consistent with findings by Barr (1997) with the WMS-R (Wechsler, 1987) Logical Memory subtest as well as with the auditory and visual memory indices of the WMS-Third Edition (WMS-III, Wechsler, 1997) by Wilde and colleagues (2001) among patients with TLE. These findings may be partly influenced by the fact that Logical Memory, unlike the RAVLT and WMT, relies on single trial learning as opposed to repeated trials, which may increase the possibility of nonmemory factors, such as poor attention and comprehension, adversely affecting performance on this task (Jones-Gotman et al., 2010). Our findings add substantively to prior case studies (i.e., Carone et al., 2013; Goodrich-Hunsaker & Hokins, 2009) and empirical results (e.g., Armistead-Jehle et al., 2015; Eichstaedt et al., 2014) suggesting that the clinical utility of the WMT extends beyond just performance validity assessment as its FR subtest may also function as a sensitive measure of verbal memory impairment. The particular advantage of this study was evaluating the WMT FR and RAVLT memory indices in a well-defined clinical sample having known lateralized dominant and nondominant hemisphere temporal lobe dysfunction. Indeed, the FR subtest was able to provide similar diagnostic accuracy as a previously established verbal memory measure (i.e., RAVLT). Moreover, the FR subtest score, both with and without the age covariate, was more accurate than the WMS-IV Logical Memory II subtest, which did not significantly distinguish between LTLE and RTLE patients. Even without the age covariate, our ROC analysis of the WMT FR in distinguishing seizure laterality obtained an AUC of 0.74, which exceeds the 0.70 AUC obtained by Wilde and colleagues (2001) in their ROC analysis of an Auditory-Visual Delayed Index difference score for the WMS-III in a similar TLE population. Diagnostic accuracy evaluation of the WMT FR subtest scores and the RAVLT also were clinically similar. The odds ratio yielded by RAVLT Trial 7 age-adjusted T-score was 11.8 and the WMT FR subtest plus age was 11.3, which fall into diagnostic tests generally considered to provide useful additional information (Fischer, Bachmann, & Jaeschke, 2003). In this small sample, the odds of having RTLE increased 8% with each increase in T-score point on the RAVLT Trial 7 score. Similarly, the odds of having RTLE increased 8% for every one-point increase in WMT FR subtest percent correct score. The significant predictor of age in the WMT FR algorithm, but not in the RAVLT T-score algorithm, was thought to reflect age as a potential moderator in FR subtest scores. From a clinical standpoint, using a cutoff T-score for the RAVLT Trial 7 of,36.5 achieved a minimum sensitivity rate of 0.81 for correctly classifying patients to the LTLE, whereas a cut score of.41.5 T-scores assured a minimum sensitivity rate of 0.84 for identifying RTLE. Cutoffs for the WMT FR subtest raw scores of,38.75% for identifying LTLE and.53.75% for identifying RTLE yielded minimum sensitivity rates of 0.81 and 0.84, respectively. The positive likelihood ratios for the RAVLT Trial 7 T-score and WMT FR score both were clinically meaningful with the 95% confidence intervals not falling below By setting separate cut scores for LTLE and RTLE group membership, we provide a conservative means for predicting seizure laterality using these measures. With this approach, it should be understood that when scores fail to meet threshold for either laterality group, but rather fall within the range between cut scores, the RAVLT and WMT FR scores become less valuable predictors of seizure laterality and are therefore considered equivocal.

7 Given the added accuracy of WMT FR subtest once the effect of age was taken into account, development of WMT FR age-adjusted norms may further increase utility of these clinical data in addition to purposes of assessing adequate task engagement/performance validity (see Armistead-Jehle et al., 2015; Davis & Wall, 2014; Rienstra et al., 2009; although see also Donders & Strong, 2013). To date, two sets of age- and education-adjusted norms for WMT indices have been published derived from Dutch and Canadian participants (Rienstra et al., 2009). These normative data are limited due to small sample sizes, limited age ranges, and derivation from a Dutch (N ¼ 115) and Canadian (N ¼ 40) samples. Consequently, the generalizability of these normative data is limited. Applying the same rigorous, age-stratified norming techniques as has been done with other commonly available memory measures(e.g., RAVLT, CVLT-II, and WMS-IV), is likely to further increase the utility of the WMT FR subtest as a verbal memory measure. From a burden reduction standpoint, using the WMT as both a PVT and a verbal memory measure would allow for maximum clinical utility and may eventually reduce the need to administer other verbal memory tests. This also can decrease costs while not compromising professional guidelines regarding performance validity assessment (Heilbronner et al., 2009) or evidence-based clinical practice, given our results suggest that the WMT FR subtest score plus age-correction is as sensitive to LTLE as the RAVLT. There are several notable limitations to this study. First, data were retrospective and limited to those who consented to be included in our archival database. Notably, participants were predominately Caucasian and spoke English. Thus, these results may not generalize to other samples with bilingual or non-english speaking persons. Although these data offer further preliminary evidence, the small sample size prohibits clinical application of these. As such, it is premature to reduce or eliminate the clinical assessment battery to lateralize and predict postsurgical outcome from proposed temporal lobectomy for the treatment of pharmacoresistant TLE. However, the observed similarity of WMT FR subtest with RAVLT Trial 7 subtest scores derived from this welldescribed clinical sample of patients with known lateralized neurological disease and memory deficits provides evidence of convergent validity between measures and strongly argues additional study of the WMT FR subtest as a declarative verbal memory test is warranted. Specifically, although the current study sample size compares favorably with the sample sizes included by Green (2005) in clinical comparison samples on the WMT interpretive output, clinical application will require evaluation with larger sample sizes. Future research may also explore the WMT FR subtest as a potential measure to predict surgical outcome from temporal lobectomy and/or identify treatment needs for individuals with epilepsy-related material-specific memory deficits. Although our data were limited to TLE, they suggest that future research of the FR subtest as a measure of verbal memory also is warranted in other clinical samples with known memory deficits, such as Alzheimer s dementia, CNS syphilis infections, temporal lobe tumors, anterior communicating artery aneurysm hemorrhages, and other neuropathological diseases affecting Papez circuit (e.g., mesial temporal structures, anterior thalamic structures) that impair declarative verbal memory. Finally, development of age-adjusted standardized scores for the WMT FR subtest using healthy controls may further improve the utility of the WMT FR subtest as a measure of verbal declarative memory. Conflict of Interest None declared. J.R. Soble et al. / Archives of Clinical Neuropsychology 31 (2016); References Allen, M. D., Bigler, E. D., Larsen, J., Goodrich-Hunsaker, N. J., & Hopkins, R. O. (2007). Functional neuroimaging evidence for high cognitive effort on the Word Memory Test in the absence of external incentives. Brain Injury, 21, Allen, M. D., Wu, T. C., & Bigler, E. D. (2011). Traumatic brain injury alters Word Memory Test performance by slowing response time and increasing cortical activation: An fmri study of a symptom validity test. Psychological Injury and Law, 4, Armistead-Jehle, P., Green, P., Gervais, R. O., & Hungerford, L. D. (2015). An examination of the Word Memory Test as a measure of memory. Applied Neuropsychology: Adult, 15, Barr, W. B. (1997). Receiver operating characteristic curve analysis of Wechsler Memory Scale Revised scores in epilepsy surgery candidates. Psychological Assessment, 9, Barr, W. B., Chelune, G. J., Hermann, B. P., Loring, D. W., Perrine, K., Strauss, E., et al. (1997). The use of figural reproduction tests as measures of nonverbal memory in epilepsy surgery candidates. Journal of the International Neuropsychological Society, 3, Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the Beck Depression Inventory-II. San Antonio, TX: Psychological Corporation. Benbadis, S. R., LaFrance, W. C., Papandonatos, G. D., Korabathina, K., Lin, K., & Kraemer, H. C. (2009). Interrater reliability of EEG-video monitoring. Neurology, 73, Board of Directors (2007) of the American Academy of Clinical Neuropsychology. (2007). American Academy of Clinical Neuropsychology (AACN) practice guidelines for neuropsychological assessment and consultation. The Clinical Neuropsychologist, 21, Brandt, J. (2008) INS presidential address: Neuropsychological crimes and misdemeanors. The Clinical Neuropsychologist, 21, Buchtel, H. A., Passaro, E. A., Selwa, L. M., Deveikis, J., & Gomez-Hassan, D. (2002). Sodium methohexital (brevital) as an anesthetic in the Wada test. Epilepsia, 43,

8 86 J.R. Soble et al. / Archives of Clinical Neuropsychology 31 (2016); Carone, D. A. (2014). Young child with severe brain volume loss easily passes the Word Memory Test and Medical Symptom Validity Test: Implications for mild TBI. The Clinical Neuropsychologist, 28, Carone, D. A., Green, P., & Drane, D. L. (2013). Word Memory Test profiles in two cases with surgical removal of the left anterior hippocampus and parahippocampal gyrus. Applied Neuropsychology: Adult, 21, Chelune, G. J. (2010). Evidence-based research and practice in clinical neuropsychology. The Clinical Neuropsychologist, 24, Davis, J. J., & Wall, J. R. (2014). Examining verbal memory on the Word Memory Test and California Verbal Learning Test-Second Edition. Archives of Clinical Neuropsychology, 29, Delis, D. C., Kramer, J. H., Kaplan, E., & Ober, B. A. (1987). The California Verbal Learning Test. San Antonio, TX: Psychological Corporation. Delis, D. C., Kramer, J. H., Kaplan, E., & Ober, B. A. (2000). California Verbal Learning Test Second edition. San Antonio, TX: Psychological Corporation. Donders, J., & Strong, C. A. H. (2013). Does Green s Word Memory Test really measure memory? Journal of Clinical and Experimental Neuropsychology, 35, Eichstaedt, K. E., Clifton, W. E., Vale, F. L., Benbadis, S. R., Bozorg, A. M., Rodgers-Neame, N. T., et al. (2014). Sensitivityof Green s Word Memory Test genuine memory impairment profile to temporal pathology: A study in patients with temporal lobe epilepsy. The Clinical Neuropsychologist, 28, Elger, C. E., Helmstaedter, C., & Kurthen, M. (2004). Chronic epilepsy and cognition. Lancet Neurology, 3, Fischer, J. E., Bachmann, L. M., & Jaeschke, R. (2003). A reader s guide to the interpretation of diagnostic test properties: Clinical example of sepsis. Intensive Care Medicine, 29, Flaro, L., Green, P., & Robertson, E. (2007). Word Memory Test failure 23 times higher in mild brain injury than in parents seeking custody: The power of external incentives. Brain Injury, 21, Goodrich-Hunsaker, N. J., & Hokins, R. O. (2009). Word Memory Test performance in amnestic patients with hippocampal damage. Neuropsychology, 23, Green, P. (2003). Green s Word Memory Test for windows: User s manual. Edmonton, Canada: Green s Publishing. Green, P. (2005). Manual for the Word Memory Test (Revised). Edmonton, Alberta, Canada: Green s Publishing. Green, P., Allen, L., & Astner, K. (1996). The Word Memory Test: A users guide to the oral and computer-administered forms, US version 1.1. Durham, NC: Cognisyst. Green, P., Iverson, G., & Allen, L. (1999). Detecting malingering in head injury litigation with the Word Memory Test. Brain Injury, 13, Green, P., Montijo, J., & Brockhaus, R. (2011). High specificityof the Word Memory Testand Medical SymptomValidity Test in groups with severe verbal memory impairment. Applied Neuropsychology, 18, Greiffenstein, M. F., Greve, K. W., Bianchini, K. J., & Baker, W. J. (2008). Test of Memory Malingering and Word Memory Test: A new comparison of failure concordance rates. Archives of Clinical Neuropsychology, 23, Greve, K. W., Ord, J., Curtis, K. L., Bianchini, K. J., & Brennan, A. (2008). Detecting malingering in traumatic brain injury and chronic pain: A comparison of three forced-choice symptom validity tests. The Clinical Neuropsychologist, 22, Heilbronner, R. L., Sweet, J. J., Morgan, J. E., Larrabee, G. J., & Millis, S. R. (2009). American Academy of Clinical Neuropsychology consensus conference statement on the neuropsychological assessment of effort, response bias, and malingering. The Clinical Neuropsychologist, 23, Helmstaedter, C., Wietzke, J., & Lutz, M. T. (2009). Unique and shared validity of the Wechsler logical memory test, the California verbal learning test, and the verbal learning and memory test in patients with epilepsy. Epilepsy Research, 87, Hermann, B. P., Seidenberg, M., Schoenfeld, J., Peterson, J., Leveroni, C., & Wyler, A. R. (1996). Empirical techniques for determining the reliability, magnitude, and pattern of neuropsychological change after epilepsy surgery. Epilepsia, 37, Jones-Gotman, M., Smith, M. L., Risse, G. L., Westerveld, M., Swanson, S. J., Giovagnoli, A., et al. (2010). The contribution of neuropsychology to diagnostic assessment in epilepsy. Epilepsy & Behavior, 18, Lencz, T., McCarthy, G., Bronen, R. A., Scott, T. M., Inserni, J. A., Sass, K. J., et al. (1992). Quantitative magnetic resonance imaging in temporal love epilepsy: Relationship to neuropathology and neuropsychological function. Annals of Neurology, 31, Loring, D. W., Lee, G. P., & Meador, K. J. (2005). Victoria Symptom Validity Test performance in non-litigating epilepsy surgery candidates. Journal of Clinical and Experimental Neuropsychology, 27, Loring, D. W., Strauss, E., Hermann, B. P., Barr, W. B., Perrine, K., Trenerry, M. R., et al. (2008). Differential neuropsychological test sensitivity to left temporal lobe epilepsy. Journal of the International Neuropsychological Society, 14, Merten, T., Bossink, L., & Schmand, B. (2007). On the limits of effort testing: Symptom validity tests and severity of neurocognitive symptoms in nonlitigant patients. Journal of Clinical and Experimental Neuropsychology, 29, Patient Protection and Affordable Care Act, 42 U.S.C et seq. (2010). Petersen, R. C., Jack, C. R., Xu, Y. C., Waring, S. C., O Brien, P. C., Smith, G. E., et al. (2000). Memory and MRI-based hippocampal volumes in aging and AD. Neurology, 54, Rey, A. (1941). L examen psychologique dans les cas d encéphalopathie traumatique. Archives de Psychologie, 28, 21. Rienstra, A., Groot, P. F. C., Spaan, P. E. J., Mahole, C. B. L., Nederveen, A. J., Walstra, G. J., et al. (2012). Symptom validity testing in memory clinics: Hippocampal-memory associations and relevance for diagnosing mild cognitive impairment. Journal of Clinical and Experimental Neuropsychology, 35, Rienstra, A., Spaan, P. E. J., & Schmand, B. (2009). Reference data for the Word Memory Test. Archives of Clinical Neuropsychology, 24, Rienstra, A., Twennaar, M. K., & Schmand, B. (2013). Neuropsychological characterization of patients with the WMT dementia profile. Archives of Clinical Neuropsychology, 28, Sanchez-Alvares, J. C., Serrano-Castro, P., & Canadillas-Hidalgo, F. (2002). Refractory epilepsy in adults. Revue Neurologique, 35, Schmidt, M. (1996). Rey Auditory Verbal Learning Test: A handbook. Los Angeles, CA: Western Psychological Services. Sherman, E. M. S., Wiebe, S., Fay-McClymont, T. B., Tellez-Zenteno, J., Metcalfe, A., & Hernandez-Ronquillo, L., et al. (2011). Neuropsychological outcomes after epilepsy surgery: Systematic review and pooled estimates. Epilepsia, 52, Soble, J. R., Eichstaedt, K. E., Waseem, H., Mattingly, M. L., Benbadis, S. R., Bozorg, A. M., et al. (2014). Clinical utility of the Wechsler Memory Scale-Fourth Edition (WMS-IV) in predicting laterality of temporal lobe epilepsy among surgical candidates. Epilepsy & Behavior, 41, Strauss, E., Sherman, E. M. S., & Spreen, O. (2006). A compendium of neuropsychological tests. New York City: Oxford University Press.

9 J.R. Soble et al. / Archives of Clinical Neuropsychology 31 (2016); Thompson, P. J., & Duncan, J. S. (2005). Cognitive decline in severe intractable epilepsy. Epilepsia, 46, Trenerry, M. R., & Loring, D. W. (1995). Intracarotid amobarbital procedure: The Wada test. Neuroimaging Clinics of North America, 5, Wechsler, D. (1987). Wechsler Memory Scale-revised manual. San Antonio, TX: The Psychological Corporation. Wechsler, D. (1997). Wechsler Memory Scale-third edition manual. San Antonio, TX: The Psychological Corporation. Wechsler, D. (2008). Wechsler Adult Intelligence Scale-fourth edition (WAIS-IV) manual. San Antonio, TX: Pearson. Wechsler, D. (2009). Administration and scoring manual for the Wechsler Memory Scale (4th ed.). San Antonio, TX: Pearson. Wilde, N., Strauss, E., Chelune, G. J., Loring, D. W., Martin, R. C., Hermann, B. P., et al. (2001). WMS-III performance in patients with temporal lobe epilepsy: Group differences and individual classification. Journal of the International Neuropsychological Society, 7,