Epilepsy & Behavior 44 (2015) 136–142
Contents lists available at ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
The relationship between neuropsychological functioning and FDG-PET hypometabolism in intractable mesial temporal lobe epilepsy Alex A. Knopman a,b, Chong H. Wong c,d, Richard J. Stevenson a, Judi Homewood a, Armin Mohamed c,e, Ernest Somerville f,g, Stefan Eberl c,e, Lingfeng Wen c,e, Michael Fulham c,e, Andrew F. Bleasel c,d,⁎ a
Department of Psychology, Macquarie University, NSW, Australia Department of Medical Psychology, Westmead Hospital, NSW, Australia Sydney Medical School, University of Sydney, NSW, Australia d Departments of Neurology, Westmead Hospital and The Children's Hospital at Westmead, NSW, Australia e Department of Molecular Imaging, Royal Prince Alfred Hospital, NSW, Australia f Institute of Neurological Sciences, Prince of Wales Hospital, NSW, Australia g Faculty of Medicine, University of New South Wales, NSW, Australia b c
a r t i c l e
i n f o
Article history: Received 25 September 2014 Revised 20 January 2015 Accepted 21 January 2015 Available online xxxx Keywords: Focal epilepsy Episodic memory Epilepsy surgery Positron emission tomography Cognitive functioning Frontal lobes
a b s t r a c t We examined the relationship between baseline neuropsychological functioning and 18-fluorodeoxyglucose positron emission tomography (FDG-PET) in intractable mesial temporal lobe epilepsy (MTLE). We hypothesized relationships between dominant temporal lobe hypometabolism and verbal memory and between nondominant temporal lobe hypometabolism and nonverbal memory in line with the lateralized material-specific model of memory deficits in MTLE. We also hypothesized an association between performance on frontal lobe neuropsychological tests and prefrontal hypometabolism. Thirty-two patients who had undergone temporal lobectomy for treatment of MTLE and who completed both presurgical FDG-PET and comprehensive neuropsychological investigations with widely used standardized measures were included. Age-adjusted composite measures were calculated for verbal memory, nonverbal memory, relative material-specific memory, IQ, executive function, attention/working memory, and psychomotor speed. Fluorodeoxyglucose positron emission tomography was analyzed with statistical parametric mapping (SPM) to identify hypometabolism relative to healthy controls. Pearson's correlation was used to determine the relationship between regions of hypometabolism and neuropsychological functioning. Dominant temporal lobe hypometabolism was associated with relatively inferior verbal memory, while nondominant temporal lobe hypometabolism was associated with inferior nonverbal memory. No relationship was found between performance on any frontal lobe measures and prefrontal hypometabolism. Statistical parametric mapping-quantified lateralized temporal lobe hypometabolism correlates with materialspecific episodic memory impairment in MTLE. In contrast, prefrontal hypometabolism is not associated with performance on frontal lobe measures. We suggest that this is because frontal lobe neuropsychology tests may not be good measures of isolated frontal lobe functioning. © 2015 Elsevier Inc. All rights reserved.
1. Introduction Temporal lobe epilepsy (TLE) is associated not only with episodic memory and language deficits characteristic of temporal lobe dysfunction [1,2] but also with cognitive deficits that extend beyond those thought to be mediated by the temporal lobe. These include IQ, attention, visuospatial deficits, speed of processing, and higher-level executive deficits [1]. Similarly, abnormalities extend well beyond the temporal lobe in TLE on fluorodeoxyglucose positron emission tomography (FDG-PET). Specifically, in addition to ipsilateral temporal lobe hypometabolism, ⁎ Corresponding author at: Department of Neurology, Westmead Hospital, PO Box 533, Wentworthville, NSW 2145, Australia. Tel.: +61 2 9845 6753; fax: +61 2 9635 6684. E-mail address: [email protected] (A.F. Bleasel).
http://dx.doi.org/10.1016/j.yebeh.2015.01.023 1525-5050/© 2015 Elsevier Inc. All rights reserved.
abnormalities have been reported in the contralateral temporal lobe and frontal, parietal, and thalamic regions mostly ipsilateral to the seizure focus [3–5]. To date, little research on the relationship between these functional abnormalities identified by FDG-PET and neuropsychological impairment in TLE exists. This stands in contrast with a sizable body of literature addressing the link between cognitive impairment and structural brain abnormalities in TLE both within and beyond the temporal lobes analyzed with quantitative MRI. This research has suggested that more widespread cognitive dysfunction, including lower IQ and executive dysfunction, is related to anatomical abnormalities outside of the temporal lobe [6]. The aim of the current study was to explore the relationship between FDG-PET hypometabolism and neuropsychological test scores in the most common form of focal epilepsy — unilateral mesial temporal
A.A. Knopman et al. / Epilepsy & Behavior 44 (2015) 136–142
lobe epilepsy (MTLE). Left TLE is reliably associated with verbal learning and memory deficits [1], while right TLE is generally thought to be associated with nonverbal memory deficits [2] (but see [7]). Indeed, several authors have reported a relationship between verbal memory and left temporal lobe hypometabolism [8–10], while nonverbal memory has been linked with right temporal lobe hypometabolism [8]. However, these studies have used either nonstandardized experimental memory stimuli [8] or memory measures which have been since replaced by more valid and reliable tests [9,10] — limitations that restrict clinical utility and practical reproducibility. Two studies have not found a correlation between temporal lobe hypometabolism and episodic memory measures in TLE [11,12]. Similarly, superseded memory measures were used in one of these studies [11], while the other study restricted their analysis to hypometabolism of the left hippocampal and parahippocampal gyri only, thereby not considering the potential role of the whole temporal lobe [12]. Thus, further investigation is warranted to clarify the relationship between episodic memory measures and lateralized temporal lobe hypometabolism in MTLE. Several lines of evidence suggest a relationship between prefrontal integrity and performance on higher-level neuropsychological tests measuring IQ, attention, working memory, psychomotor speed, and executive functions like problem-solving, abstract reasoning, response inhibition, and phonemic fluency (“frontal lobe abilities”) [13–19]. Only two studies have investigated whether extratemporal hypometabolism in TLE is correlated with neuropsychological functioning [3,4]. The first study reported metabolic asymmetry of the prefrontal cortex to be strongly related to IQ [3]. The second study employed quantitative statistical parametric mapping (SPM) methodology to report an association between prefrontal hypometabolism and executive dysfunction (set-shifting) in patients with MTLE, although these authors used a nonstandardized and experimental measure of executive functioning [4]. The current study examined for the first time the relationship between lateralized temporal lobe hypometabolism and separate material-specific episodic memory abilities in one sample of patients using standardized neuropsychological instruments. Additionally, the relationship between prefrontal hypometabolism and performance on frontal lobe neuropsychological tests has not previously been examined in the same sample of patients and not at all with standardized neuropsychological measures. The following hypotheses were made: first, dominant MTLE is associated with verbal memory impairment, and nondominant MTLE is associated with nonverbal memory impairment [8], with an analogous relationship between lateralized temporal lobe hypometabolism and memory using current memory measures. That is, it is expected that dominant temporal lobe hypometabolism is associated with poor verbal memory and that nondominant temporal lobe hypometabolism is associated with poor nonverbal memory; second, a relationship between poor performance on frontal lobe neuropsychological tests and hypometabolism of the prefrontal cortex is predicted. 2. Method
137
was obtained from most patients. No patient with dual pathology or psychiatric comorbidity (major depression, major anxiety, or psychosis) was included. For the majority of patients, hemispheric language dominance was investigated with the intracarotid amytal procedure (IAP) or functional magnetic resonance imaging (fMRI). Where these data were not available, left hemispheric language dominance was assumed for right-handed patients only. Exclusions were hippocampal lesions other than HS (n = 5), IQ b 70 (n = 4) as set in similar studies [20], left-handedness but no fMRI or IAP investigation (n = 3), previous neurosurgery (n = 3), bilateral language representation on fMRI or IAP investigations (n = 2), comorbid neurological disease (Parkinson's disease, n = 1; cerebral palsy, n = 1), history of brain injury (n = 1), normal pathology (n = 1), and English as a second language (n = 1). This left a final sample of 32 patients, four of whom demonstrated reversed hemispheric language dominance. Seventeen patients had a history of febrile convulsion, while 8 patients had a family history of epilepsy. All patients were on antiepileptic polypharmacotherapy, with the majority on two types of medication, but more specific information was not available given that extant data were used. Following surgery, 17 patients were classified as Engel Class 1 [21], 13 as Engel Class 2, and two as Engel Class 3, with a mean follow-up duration of 9.08 years (SD = 3.09). Other demographic data and clinical attributes are recorded in Table 1. The group with nondominant MTLE was older than the group with dominant MTLE (p = 0.03), but this was considered to be of low importance given that neuropsychological and FDG-PET measures were calculated relative to age-matched controls. There were no significant differences between groups on IQ or other demographic and clinical characteristics (t-tests or chi-square). 2.2. Neuropsychological measures Patients underwent comprehensive presurgical neuropsychological assessment with standard measures widely used in clinical practice [22] (Table 2). Raw scores on all psychometric tests were converted to age-adjusted z-scores based on published normative tables (mean = 0, SD = 1), except for performance on the Wisconsin Card Sorting Test (WCST), for which age and education-adjusted normative data were used. Cognitive domains were created based on concurrent and divergent validity studies showing that the different component tests measure the same cognitive construct [23]. This approach has been used elsewhere [20]. Domains comprised the mean z-score of contributing tests where available and included Verbal Memory [24,25], Nonverbal Memory [24,25], IQ [26,27], Executive Functioning [26–29], Attention/Working Memory [26,27], and Psychomotor Speed [26,27, 30] (Table 2). Performance on the WCST was also included as a separate measure of executive functioning given its reported sensitivity to frontal lobe integrity [13,15]. Note that different editions of the Wechsler scales are highly correlated with each other [31]. A more sensitive measure of relative material-specific memory deficit (Relative Memory) was also calculated for each patient by subtracting the scores for Nonverbal
2.1. Patients The study was approved by the ethics committees of Western and Central Sydney Area Health Services. Patients were retrospectively selected from the epilepsy surgery databases at the Westmead and Royal Price Alfred Hospitals, Sydney, Australia. All patients at least 17 years of age with MTLE, who had undergone standard anterior temporal lobectomy, who had completed both presurgical neuropsychological and FDG-PET investigations between 1994 and 2004, who had been followed up for at least two years (median = 9.08 years; range = 3.01– 15.20 years), and who had hippocampal sclerosis (HS) proven pathologically, were included (n = 54). Mesial temporal lobe epilepsy was diagnosed by clinical history, seizure semiology, interictal and ictal scalp electroencephalography (EEG), and MRI changes consistent with HS. Ictal single-photon emission computerized tomography (SPECT)
Table 1 Demographic and clinical attributes.
Gender (M/F) Agea (mean, SD) Educationa IQ Age at seizure onset a Duration of epilepsy a SGTCSs (present/absent)
Dominant MTLE (n = 14)
Nondominant MTLE (n = 18)
Total (n = 32)
2/12 29.6 (8.8) 9.9 (0.7) 85.4 (10.3) 8.0 (6.8) 21.9 (10.3) 6/8
7/11 37.7 (11.2)b 10.8 (1.7) 90.7 (15.6) 13.3 (10.2) 24.4 (11.2) 13/5
9/23 34.1 (10.9) 10.4 (1.4) 88.3 (13.6) 11.0 (9.1) 23.3 (10.8) 19/13
MTLE, mesial temporal lobe epilepsy; M/F, male/female; IQ, intelligence quotient; SGTCSs, secondary generalized tonic–clonic seizures. a Years. b p b 0.05.
A.A. Knopman et al. / Epilepsy & Behavior 44 (2015) 136–142
Table 2 Neuropsychological test battery. Domain Verbal Memory Nonverbal Memory
IQa Executive Functioninga
Attention/Working Memorya
Psychomotor Speeda
–
Test Logical Memory II from WMS-R/III Verbal Paired Associates II 9 Visual Paired Associates II > > = Visual Reproduction II from WMS-R/III > Faces II > ; Family Pictures II Full Scale IQ from WAIS-R/III Controlled Oral Word Association Test 9 Similarities = Block Design from WAIS-R/III ; Matrix Reasoning 9 Digit Span > > = LNS from WAIS-R/III > Spatial Span > ; Visual Memory Span 9 TMT-A = DSymb or DSymbC ; from WAIS-R/III Symbol Search b WCST
WMS-R/III, Wechsler Memory Scale — Revised or Third Edition; WAIS-R/III, Wechsler Adult Intelligence Scale — Revised or Third Edition; LNS, Letter Number Sequencing; TMT-A, Trail Making Test — Part A; DSymb, Digit Symbol; DSymbC, Digit Symbol Coding; WCST, Wisconsin Card Sorting Test. a The tests comprising these domains are sensitive to frontal lobe dysfunction [15,19,23]. b Number of perseverative errors.
Memory from the scores for Verbal Memory, where positive values represent relatively stronger verbal memory skills and negative values represent relatively stronger nonverbal memory skills. 2.3. FDG-PET acquisition and image processing Detailed procedures for presurgical FDG-PET acquisition and image processing for patients and healthy volunteer controls have been reported elsewhere [5]. Analysis of FDG-PET images was performed using Statistical Parametric Mapping (SPM 2, Wellcome Department of Cognitive Neurology, London, UK). Both patient and control images were spatially normalized onto the standard Montreal Neurological Institute (MNI, McGill University, Montreal, Quebec, Canada) brain PET template. Regional cerebral metabolic rate of glucose utilization (rCMRGlu) for each voxel was normalized to the whole-brain average of rCMRGlu using a proportional scaling method. Normalized images were then smoothed by convolution with FWHM Gaussian kernel (10 mm full width at half maximum). The 16 age-matched healthy volunteers who served as control subjects had no history of medical or neurological illness and did not take any medications known to affect FDG-PET studies. Fluorodeoxyglucose positron emission tomography was analyzed voxel-by-voxel with SPM 2 to identify significant regions of hypometabolism relative to the control group using a two-sample t-test at the individual voxel-level with a threshold of p ≤ 0.01 (z-score of 2.34). Only clusters with N250 contiguous voxels (volume of 2 cm3) and cluster significance of p ≤ 0.05 (uncorrected) were included. Significant regions of hypometabolism were superimposed onto a normalized, spatiallyregistered MNI brain template [32] with anatomic areas in each hemisphere including the anteromesial temporal cortex (temporal pole, hippocampus, parahippocampal gyrus, amygdala, and fusiform gyrus), whole temporal cortex (anteromesial cortex and lateral temporal cortex), prefrontal cortex (orbitofrontal, dorsolateral, and mesial frontal), central cortex, parietal cortex, occipital cortex, cingulate gyrus, insula, basal ganglia, and thalamus. The extent of hypometabolism comprised the number of significant voxels in each anatomical area.
FDG-PET variables (whole temporal cortex and prefrontal cortex) was necessary for normalization. Neuropsychological and FDG-PET variables were compared between the group with dominant MTLE and the group with nondominant MTLE with t-tests. Pearson's correlation was used to examine the relationship between regions of hypometabolism and neuropsychological functioning for the sample as a whole. Alpha here was set at 0.025 (i.e., 0.05/2) after the Bonferroni correction for the two interrelated temporal lobe regions (anteromesial temporal lobe and whole temporal lobe) for each hemisphere. Onetailed tests of significance were employed given the expected relationship between the material-specific memory and the corresponding temporal lobe based on similar research with FDG-PET [8–10]. Following a Bonferroni adjustment for the five frontal lobe abilities, alpha was set at 0.01 (i.e., 0.05/5) for investigation into whether each of these abilities was correlated with hypometabolism of the prefrontal cortex. Here, two-tailed tests of significance were employed given the uncertainty of the relationship between these particular brain regions and frontal lobe abilities. The only missing data were for six patients on the WCST. 3. Results 3.1. Neuropsychological functioning Fig. 1 shows the scores for Verbal Memory and Nonverbal Memory by group. There was a main effect for nonverbal memory, where nondominant MTLE was associated with poorer nonverbal memory compared with dominant MTLE (t(30) = 2.41, p = 0.02). In contrast, no main effect was found for verbal memory by group (t(30) = 1.31, p = 0.20). However, patients with dominant MTLE demonstrated weaker verbal memory than nonverbal memory (t(30) = 2.85, p = 0.01), which was not the case for patients with nondominant MTLE (t(30) = 1.04, p = 0.31). Relative Memory was also significantly different between groups (t(30) = 2.81, p = 0.009), whereby dominant MTLE was associated with worse relative verbal memory deficit compared with nondominant MTLE. Together, these results support the prediction that patients with unilateral MTLE demonstrate material-specific episodic memory deficits according to laterality. For frontal lobe abilities, there was no significant difference between groups on executive functioning (t(30) = 1.66, p = 0.11), attention/ 0 -0.2
Test performance (z-score)
138
-0.4 -0.6 -0.8 -1 -1.2 -1.4 -1.6 -1.8 -2 -2.2 Nondominant MTE
Dominant MTE
Verbal Memory Nonverbal Memory
2.4. Statistical analysis Data were analyzed with PASW Statistics (GradPack; version 18; SPSS Inc., Chicago, IL, U.S.A.). Square root transformation of some
Fig. 1. Mean (and standard error) for memory by group. Patients with nondominant MTLE demonstrated significantly poorer nonverbal memory compared with patients with dominant MTLE. Patients with dominant MTLE demonstrated significantly weaker verbal memory than nonverbal memory.
A.A. Knopman et al. / Epilepsy & Behavior 44 (2015) 136–142
working memory (t(30) = 0.37, p = 0.71), psychomotor speed (t(30) = 0.41, p = 0.69), or performance on the WCST (t(24) = 0.73, p = 0.47). Fig. 2 shows performance on frontal lobe ability measures for the whole sample. The greatest impairment was found for IQ (88.3 or z = −0.78), although a greater range was evident on WCST performance, with some patients demonstrating severe impairment. Overall, these results suggest that MTLE is associated with mildly impaired performance on frontal lobe neuropsychology tests. Education was significantly correlated with verbal memory (r = 0.41, p = 0.02, d = 0.90), IQ (r = 0.67, p b 0.001, d = 1.81), executive functioning (r = 0.54, p = 0.001, d = 1.28), attention/working memory (r = 0.48, p = 0.005, d = 1.09), and psychomotor speed (r = 0.56, p = 0.001, d = 1.35) but not with nonverbal memory (r = 0.20, p = 0.28, d = 0.41). Consequently, education was treated as a control variable for correlational analyses between regions of hypometabolism and neuropsychological functioning. Education was not held constant for analyses with WCST performance given that scores on this task were already education-adjusted. No other demographic or clinical attributes were related to neuropsychological functioning.
3.2. FDG-PET hypometabolism Ipsilateral temporal lobe hypometabolism was demonstrated in 91% of the patients (100% of the patients with dominant MTLE and 83% of the patients with nondominant MTLE). Bilateral temporal lobe hypometabolism was demonstrated in 19% of the patients (14% of the patients with dominant MTLE and 27% of the patients with nondominant MTLE). Hypometabolism beyond the ipsilateral temporal lobe was demonstrated in 69% of the patients. The most common pattern of extratemporal hypometabolism involved the ipsilateral frontal lobe and insula. Prefrontal hypometabolism was found in 47% of the patients. Patients with dominant MTLE showed greater hypometabolism of both the dominant anteromesial temporal cortex (t(30) = 8.95, p b 0.001) and the dominant whole temporal lobe (t(30) = 12.10, p b 0.001) relative to patients with nondominant MTLE. Patients with nondominant MTLE, in turn, showed greater hypometabolism of both the nondominant anteromesial temporal cortex (t(30) = 3.73, p = 0.001) and the nondominant whole temporal lobe (t(30) = 4.36, p = b0.001) relative to patients with dominant MTLE. Hypometabolism of neither the whole brain (t(30) = 0.48, p = 0.64) nor the prefrontal cortex (t(30) = 1.57, p = 0.13) differed by group.
0
Test performance (z-score)
-0.2
139
3.3. Association of nondominant temporal lobe hypometabolism and memory scores Fig. 3 shows the relationship between memory scores and temporal lobe regions of interest. In line with our prediction, hypometabolism of the whole nondominant temporal lobe was associated with nonverbal memory (r01.2(29) = −0.48, p = 0.003, d = 1.09). That is, poorer nonverbal memory was associated with increasing metabolic abnormality of the nondominant temporal lobe. Hypometabolism of the whole nondominant temporal lobe was also associated with Relative Memory, whereby relative nonverbal memory deficit was related to increasing metabolic abnormality of the nondominant temporal lobe (r01.2(29) = 0.47, p = 0.004, d = 1.06). Hypometabolism of the whole nondominant temporal lobe was not significantly correlated with verbal memory (r01.2(29) = 0.20, p = 0.14, d = 0.41). Similarly, hypometabolism of the more circumscribed nondominant anteromesial temporal lobe was associated with poor nonverbal memory (r01.2(29) = −0.41, p = 0.01, d = 0.90) but not with verbal memory (r01.2(29) = 0.11, p = 0.28, d = 0.22). Relative Memory was associated with nondominant anteromesial temporal lobe hypometabolism, but this relationship was not significant following the Bonferroni adjustment (r01.2(29) = 0.34, p = 0.03, d = 0.72). 3.4. Association of dominant temporal lobe hypometabolism and memory scores Hypometabolism of both the whole dominant temporal lobe and the dominant anteromesial temporal lobe was significantly associated with Relative Memory (r01.2(29) = − 0.46, p = 0.004, d = 1.04 and r01.2(29) = − 0.53, p = 0.00, d = 1.25, respectively). Specifically, relative verbal memory deficit was associated with increasing metabolic abnormality in the dominant temporal lobe. Poor verbal memory was associated with hypometabolism of the dominant anteromesial temporal lobe (r(30) = −.36, p = 0.02, d = 0.77), but this became a nonsignificant trend when controlling for education (r01.2(29) = −0.29, p = 0.06, d = 0.61). There was also a nonsignificant trend for an association between poor verbal memory and hypometabolism of the whole dominant temporal lobe (r01.2(29) = −0.17, p = 0.18, d = 0.35). Nonverbal memory was also significantly associated with hypometabolism of both these regions but in a positive direction (r01.2(29) = 0.52, p = 0.001, d = 1.22 and r01.2(29) = 0.46, p = 0.004, d = 1.04, respectively). That is, increasing hypometabolism of the whole dominant temporal lobe or the dominant anteromesial temporal lobe was associated with better nonverbal memory. We interpreted these results as being related to the fact that metabolic abnormality in the dominant temporal lobe was more likely in patients with dominant MTLE, who demonstrated intact nonverbal memory. 3.5. Association of prefrontal hypometabolism and frontal lobe neuropsychological measures
-0.4 -0.6
Hypometabolism of the prefrontal cortex was not significantly correlated with IQ (r(30) = 0.28, p = 0.12, d = 0.58), executive functioning (r(30) = 0.28, p = 0.12, d = 0.58), attention/working memory (r(30) = 0.23, p = 0.21, d = 0.47), psychomotor speed (r(30) = 0.14, p = 0.47, d = 0.28), or performance on the WCST (r(24) = 0.29, p = 0.15, d = 0.61). Thus, we could not support our second prediction that performance on frontal lobe neuropsychological tests is associated with hypometabolism of the prefrontal cortex.
-0.8 -1 -1.2 -1.4 -1.6 -1.8
4. Discussion
-2
Frontal lobe abilies Fig. 2. Mean (and standard error) for frontal lobe abilities for the group with dominant MTLE and for the group with nondominant MTLE combined. Attn/WM, attention/working memory; speed, psychomotor speed; WCST, Wisconsin Card Sorting Test.
We aimed to investigate the relationship between FDG-PET hypometabolism and widely used standardized neuropsychological measures in a well-defined, homogeneous sample of unilateral MTLE. Our first set of predictions was supported in that a relationship between
140
A.A. Knopman et al. / Epilepsy & Behavior 44 (2015) 136–142
A
B
Nondominant TL Whole
Whole
Anteromesial Pearson Correlaon
Pearson Correlaon
**
0.2 0.0 -0.2 -0.4 -0.6
Anteromesial
0.6
0.6 0.4
Dominant TL
** Relave Memory
*
0.4
**
**
0.2 0.0 -0.2 -0.4 -0.6
Verbal Memory
**
**
Nonverbal Memory
Fig. 3. A. Hypometabolism of both the whole nondominant temporal lobe and the more circumscribed anteromesial nondominant temporal lobe was correlated with Nonverbal Memory but not with Verbal Memory. The whole nondominant temporal lobe was correlated with Relative Memory. B. Hypometabolism of both the whole dominant temporal lobe and the more circumscribed anteromesial dominant temporal lobe was correlated with Relative Memory and Nonverbal Memory. Note that higher scores on Relative Memory (Nonverbal Memory subtracted from Verbal Memory) correspond to relatively better verbal memory and that lower scores correspond to relatively better nonverbal memory. *p b 0.025, **p b 0.005; TL, temporal lobe.
lateralized temporal lobe hypometabolism and material-specific episodic memory impairment was found. Specifically, hypometabolism within the dominant temporal lobe was related to relatively poorer verbal memory, and nondominant temporal lobe hypometabolism was related to absolute nonverbal memory impairment. We were unable to show a relationship between performance on frontal lobe neuropsychological measures and prefrontal hypometabolism. Consequently, we could not support our second hypothesis. Our findings support the limited literature showing a relationship between lateralized temporal lobe hypometabolism and materialspecific episodic memory impairment in TLE [8–10]. However, to our knowledge, the current study is the first to show this relationship separated by verbal memory and nonverbal memory in the same sample of patients using standardized neuropsychological instruments. We also showed that patients with nondominant MTLE demonstrated poorer nonverbal memory compared with patients with dominant MTLE and that patients with dominant MTLE demonstrated poorer verbal memory than their own nonverbal memory. Together, these sets of results are in line with the lateralized material-specific memory deficit model of TLE [33]. Although its validity has been questioned in recent years [7], our data provide unique support for this model by evidencing materialspecific episodic memory correlates of SPM-quantified lateralized temporal lobe hypometabolism. In the context of the observed relationships between episodic memory and lateralized temporal lobe hypometabolism, we were unable to support a relationship between performance on frontal lobe neuropsychological tests and hypometabolism of the prefrontal cortex. This is surprising because of the widely held view that the frontal lobes are sensitive to performance on neuropsychological measures of IQ, attention, working memory, psychomotor speed, and executive abilities [13–17]. However, two other studies cast doubt on the putative ‘expected’ relationship between frontal lobe or executive cognitive abilities and frontal lobe imaging abnormalities in TLE. The first study failed to find an association between frontal lobe volume on quantitative MRI and performance on several measures of executive functioning in TLE [34]. The second study showed that frontal lobe thinning on quantitative MRI could not distinguish between varying degrees of cognitive impairment. Instead, a systematic and stepwise relationship was reported between a combination of generalized cortical thinning and abnormalities of subcortical structures and white matter with increasing levels of cognitive impairment across all cognitive domains [35]. The current results suggest that hypometabolism on FDG-PET is partially discordant with neuropsychological functioning. That is, while lateralized temporal lobe hypometabolism is associated with material-
specific episodic memory impairment both in temporal lobe epilepsy and in extratemporal epilepsy [36], the same metabolic–cognitive relationship may not apply for hypometabolism of the prefrontal cortex and frontal lobe neuropsychological test performance. One possibility is that these tests may not be good measures of frontal lobe functioning in isolation. Indeed, subcortical structures seem also to play a role in frontal lobe cognitive abilities: atrophy of the thalamus and basal ganglia has been reported to be correlated with executive functioning in patients with TLE [34,37]. Such findings are in line with our understanding of the role these structures play in relaying information processing along cortico-subcortical circuits which connect the prefrontal cortex, basal ganglia, and cerebellum via the thalamus [38]. In our sample, too few patients demonstrated metabolic abnormalities of subcortical regions to allow meaningful analysis of whether these regions correlate with frontal lobe abilities. Another viewpoint is that frontal lobe neuropsychological measures are instead sensitive to a distributed neural network of frontal and nonfrontal brain regions [13]. Indeed, emerging research on resting state connectivity using fMRI analysis points to the abnormality of various resting state networks in TLE including the default mode and executive control networks — known to be involved during higher-level cognitive tasks [39]. Our findings are in contrast with previous reports of hypometabolism of the prefrontal cortex being correlated with low IQ [3] and executive dysfunction [4] in unilateral TLE. However, there are important differences between the current methodology and that used by these authors. The first study used hand-drawn regions of interest with a liberal threshold of N10% asymmetry value to determine prefrontal metabolic abnormality in patients with TLE, only a subset of whom had HS [3]. Visual analysis of FDG-PET is affected by known biases and multiple sources of variation [40], and asymmetry measures may hide bilateral hypometabolism [41]. In contrast, we employed whole brain SPM in a welldefined sample with homogeneous MTLE, with a high threshold of 250 contiguous voxels required to detect metabolic abnormality. This precise quantitative methodology offers greater objectivity and sensitivity, particularly when compared with a healthy control group. In the second study, despite SPM methodology being used, the FDG-PET scans of the 21 patients with MTLE were not compared with those from a healthy control group, and the authors employed an experimental and unstandardized measure of executive functioning [4]. In contrast, our study utilized FDG-PET scans from a healthy control group and a range of commonly used standardized neuropsychology measures. Another technical difference between our study and that by Takaya and colleagues [4] is that they measured the mean intensity of hypometabolism in selected regions of interest, which is subject to volume-averaging effects [42].
A.A. Knopman et al. / Epilepsy & Behavior 44 (2015) 136–142
An important consideration for future research is that the precise mechanism of FDG-PET hypometabolism distant to the seizure focus in MTLE remains incompletely understood. Diaschisis seems to play a significant role with epilepsy disturbing neuronal activity along anatomical frontotemporal pathways [4]. Repeated seizure propagation along the same pathways may result in additional remote hypometabolism. The current results and others [8–10,36] suggest that regional temporal lobe hypometabolism is associated with regional cognitive dysfunction. However, the hypothesis that distant hypometabolism of the prefrontal cortex is associated with regional cognitive dysfunction is supported by some research [3,4,18] but not ours. Limitations of the current study include its retrospective nature. This did not allow control over the latency between neuropsychological assessment and FDG-PET scanning, although this was typically within a few months, and data on the latency between the last seizure and the FDG-PET scan were not collected. The fact that our sample consisted of highly selected surgical patients with MTLE for whom FDG-PET was part of the presurgical decision for surgery can be seen as another limitation. However, it is unlikely that patients were excluded from surgery because of the FDG-PET result with otherwise concordant investigations in the context of HS. We note that the heterogeneous collection of measures included as frontal lobe abilities are not perfect measures of frontal lobe functioning and that they tap into other brain areas too. However, very few neuropsychological tests measure frontal lobe functioning in isolation, and all of the measures employed in our study involve the frontal lobes to some degree. Antiepileptic medication is another factor we were unable to control, and we cannot rule out medications contributing to reductions in cognitive functioning. 5. Conclusions We employed voxel-wise quantitative analysis of FDG-PET and widely used standardized neuropsychological measures to show a correlation between lateralized temporal lobe hypometabolism and material-specific episodic memory impairment in a well-defined homogeneous sample of pathologically confirmed MTLE. In contrast, performance on frontal lobe neuropsychological measures was not related to hypometabolism of the prefrontal cortex. The results suggest that frontal lobe neuropsychology tests are not good measures of frontal lobe functioning in isolation and that clinical interpretation of FDG-PET should be made with caution. Acknowledgments This work was supported, in part, by the University of Sydney Postgraduate Award, Millennium Institute Stipend, and Pfizer Neuroscience Research Grant (03-01-07) to Dr. Chong H. Wong. Disclosures None of the authors has any conflicts of interest to disclose. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. References [1] Hermann BP, Seidenberg M, Schoenfeld J, Davies K. Neuropsychological characteristics of the syndrome of mesial temporal lobe epilepsy. Arch Neurol 1997;54:369–76. [2] Baxendale S, Paesschen W, Thompson PJ, Connelly A, Duncan JS, Harkness WF, et al. The relationship between quantitative MRI and neuropsychological functioning in temporal lobe epilepsy. Epilepsia 1998;39:158–66. [3] Jokeit H, Seitz RJ, Markowitsch HJ, Neumann N, Witte OW, Ebner A. Prefrontal asymmetric interictal glucose hypometabolism and cognitive impairment in patients with temporal lobe epilepsy. Brain 1997;120:2283–94. [4] Takaya S, Hanakawa T, Hashikawa K, Ikeda A, Sawamoto N, Nagamine T, et al. Prefrontal hypofunction in patients with intractable mesial temporal lobe epilepsy. Neurology 2006;67:1674–6.
141
[5] Wong CH, Bleasel A, Wen L, Eberl S, Byth K, Fulham M, et al. The topography and significance of extratemporal hypometabolism in refractory mesial temporal lobe epilepsy examined by FDG-PET. Epilepsia 2010;51:1365–73. [6] Bell B, Lin JJ, Seidenberg M, Hermann B. The neurobiology of cognitive disorders in temporal lobe epilepsy. Nat Rev Neurol 2011;7:154–64. [7] Glikmann-Johnston Y, Saling MM, Chen J, Cooper KA, Beare RJ, Reutens DC. Structural and functional correlates of unilateral mesial temporal lobe spatial memory impairment. Brain 2008;131:3006–18. [8] Harris IM, Fulham MJ, Miller LA. The effects of mesial temporal and cerebellar hypometabolism on learning and memory. J Int Neuropsychol Soc 2001;7: 353–62. [9] Rausch R, Henry TR, Ary CM, Engel Jr J, Mazziotta J. Asymmetric interictal glucose hypometabolism and cognitive performance in epileptic patients. Arch Neurol 1994;51:139–44. [10] Weintrob DL, Saling MM, Berkovic SF, Berlangieri SU, Reutens DC. Verbal memory in left temporal lobe epilepsy: evidence for task-related localization. Ann Neurol 2002; 51:442–7. [11] Arnold S, Schlaug G, Niemann H, Ebner A, Luders H, Witte OW, et al. Topography of interictal glucose hypometabolism in unilateral mesiotemporal epilepsy. Neurology 1996;46:1422–30. [12] Griffith HR, Pyzalski RW, Seidenberg M, Hermann BP. Memory relationships between MRI volumes and resting PET metabolism of medial temporal lobe structures. Epilepsy Behav 2004;5:669–76. [13] Alvarez JA, Emory E. Executive function and the frontal lobes: a meta-analytic review. Neuropsychol Rev 2006;16:17–42. [14] Baldo JV, Shimamura AP, Delis DC, Kramer J, Kaplan E. Verbal and design fluency in patients with frontal lobe lesions. J Int Neuropsych Soc 2001;7:586–96. [15] Demakis GJ. A meta-analytic review of the sensitivity of the Wisconsin Card Sorting Test to frontal and lateralized frontal brain damage. Neuropsychology 2003;17: 255–64. [16] Keller SS, Baker GA, Downes JJ, Roberts N. Quantitative MRI of the prefrontal cortex and executive function in patients with temporal lobe epilepsy. Epilepsy Behav 2009;15:186–95. [17] Stuss DT, Alexander MP, Shallice T, Picton TW, Binns MA, Macdonald R, et al. Multiple frontal systems controlling response speed. Neuropsychologia 2005;43:396–417. [18] Takaya S, Mikuni N, Mitsueda T, Satow T, Taki J, Kinoshita M, et al. Improved cerebral function in mesial temporal lobe epilepsy after subtemporal amygdalohippocampectomy. Brain 2009;132:185–94. [19] Glascher J, Tranel D, Paul LK, Rudrauf D, Rorden C, Hornaday A, et al. Lesion mapping of cognitive abilities linked to intelligence. Neuron 2009;61:681–91. [20] Hermann BP, Seidenberg M, Lee EJ, Chan F, Rutecki P. Cognitive phenotypes in temporal lobe epilepsy. J Int Neuropsychol Soc 2007;13:12–20. [21] Engel Jr J, Van Ness PC, Rasmussen TB, Ojemann LM. Outcome with respect to epileptic seizures. In: Engel JJ, editor. Surgical treatment of the epilepsies. Second ed. New York: Raven Press; 1993. p. 609–21. [22] Djordjevic J, Bellerose J, Rudy L, Jones-Gotman M. Neuropsychological assessment of patients with epilepsy: an international survey. American Epilepsy Society 63rd Annual Meeting, December 4–8 Boston, MA; 2009. [23] Strauss E, Sherman E, Spreen O. A compendium of neuropsychological tests: administration, norms, and commentary. Third ed. New York: Oxford University Press; 2006. [24] Wechsler D. Wechsler Memory Scale — Revised. San Antonio, TX: The Psychological Corporation; 1987. [25] Wechsler D. Wechsler Memory Scale — Third Edition. San Antonio, TX: The Psychological Corporation; 1997. [26] Wechsler D. Wechsler Adult Intelligence Scale — Revised. San Antonio, TX: The Psychological Corporation; 1981. [27] Wechsler D. Wechsler Adult Intelligence Scale — Third Edition. San Antonio, TX: The Psychological Corporation; 1997. [28] Heaton RK, Chelune GJ, Talley JL, Kay GG, Curtis G. Wisconsin Card Sorting Test (WCST) manual, revised and expanded. Odessa, FL: Psychological Assessment Resources; 1993. [29] Tombaugh TN, Kozak J, Rees L. Normative data stratified by age and education for two measures of verbal fluency: FAS and animal naming. Arch Clin Neuropsychol 1999;14:167–77. [30] Tombaugh TN. Trail Making Test A and B: normative data stratified by age and education. 2004;19:203–14. [31] Tulsky D, Zhu J, Ledbetter M. WAIS-III WMS-III Technical Manual. San Antonio, TX: The Psychological Corporation; 1997. [32] Tzourio-Mazoyer N, Landeau B, Papathanassiou D, Crivello F, Etard O, Delcroix N, et al. Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage 2002;15:273–89. [33] Milner B. Interhemispheric differences in the localization of psychological processes in man. Brit Med Bull 1971;27:272–7. [34] Tuchscherer V, Seidenberg M, Pulsipher D, Lancaster M, Guidotti L, Hermann BP. Extrahippocampal integrity in temporal lobe epilepsy and cognition: thalamus and executive functioning. Epilepsy Behav 2010;17:478–82. [35] Dabbs K, Jones J, Seidenberg M, Hermann BP. Neuroanatomical correlates of cognitive phenotypes in temporal lobe epilepsy. Epilepsy Behav 2009;15:445–51. [36] Knopman AA, Wong CH, Stevenson RJ, Homewood J, Mohamed A, Somerville E, et al. The cognitive profile of occipital lobe epilepsy and the selective association of left temporal lobe hypometabolism with verbal memory impairment. Epilepsia 2014; 55:e80–4. [37] Riley JD, Moore S, Cramer SC, Lin JJ. Caudate atrophy and impaired frontostriatal connections are linked to executive dysfunction in temporal lobe epilepsy. Epilepsy Behav 2011;21:80–7.
142
A.A. Knopman et al. / Epilepsy & Behavior 44 (2015) 136–142
[38] Alexander GE, Crutcher MD, DeLong MR. Basal ganglia-thalamocortical circuits: parallel substrates for motor, oculomotor, “prefrontal” and “limbic” functions. Prog Brain Res 1990;85:119–46. [39] Cataldi M, Avoli M, de Villers-Sidani E. Resting state networks in temporal lobe epilepsy. Epilepsia 2013;54:2048–59. [40] Van Paesschen W, Goffin K, Van Laere K. PET and SPECT in presurgical evaluation of epilepsy. In: Shorvon SD, Perucca E, Engel JJ, editors. The treatment of epilepsy. 3rd Edition. Oxford: Wiley-Blackwell; 2009. p. 821–8.
[41] Hashiguchi K, Morioka T, Yoshida F, Kawamura T, Miyagi Y, Kuwabara Y, et al. Thalamic hypometabolism on 18FDG-positron emission tomography in medial temporal lobe epilepsy. Neurol Res 2007;29:215–22. [42] Rubin E, Dhawan V, Moeller JR, Takikawa S, Labar DR, Schaul N, et al. Cerebral metabolic topography in unilateral temporal lobe epilepsy. Neurology 1995;45:2212–23.
Contents lists available at ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
The relationship between neuropsychological functioning and FDG-PET hypometabolism in intractable mesial temporal lobe epilepsy Alex A. Knopman a,b, Chong H. Wong c,d, Richard J. Stevenson a, Judi Homewood a, Armin Mohamed c,e, Ernest Somerville f,g, Stefan Eberl c,e, Lingfeng Wen c,e, Michael Fulham c,e, Andrew F. Bleasel c,d,⁎ a
Department of Psychology, Macquarie University, NSW, Australia Department of Medical Psychology, Westmead Hospital, NSW, Australia Sydney Medical School, University of Sydney, NSW, Australia d Departments of Neurology, Westmead Hospital and The Children's Hospital at Westmead, NSW, Australia e Department of Molecular Imaging, Royal Prince Alfred Hospital, NSW, Australia f Institute of Neurological Sciences, Prince of Wales Hospital, NSW, Australia g Faculty of Medicine, University of New South Wales, NSW, Australia b c
a r t i c l e
i n f o
Article history: Received 25 September 2014 Revised 20 January 2015 Accepted 21 January 2015 Available online xxxx Keywords: Focal epilepsy Episodic memory Epilepsy surgery Positron emission tomography Cognitive functioning Frontal lobes
a b s t r a c t We examined the relationship between baseline neuropsychological functioning and 18-fluorodeoxyglucose positron emission tomography (FDG-PET) in intractable mesial temporal lobe epilepsy (MTLE). We hypothesized relationships between dominant temporal lobe hypometabolism and verbal memory and between nondominant temporal lobe hypometabolism and nonverbal memory in line with the lateralized material-specific model of memory deficits in MTLE. We also hypothesized an association between performance on frontal lobe neuropsychological tests and prefrontal hypometabolism. Thirty-two patients who had undergone temporal lobectomy for treatment of MTLE and who completed both presurgical FDG-PET and comprehensive neuropsychological investigations with widely used standardized measures were included. Age-adjusted composite measures were calculated for verbal memory, nonverbal memory, relative material-specific memory, IQ, executive function, attention/working memory, and psychomotor speed. Fluorodeoxyglucose positron emission tomography was analyzed with statistical parametric mapping (SPM) to identify hypometabolism relative to healthy controls. Pearson's correlation was used to determine the relationship between regions of hypometabolism and neuropsychological functioning. Dominant temporal lobe hypometabolism was associated with relatively inferior verbal memory, while nondominant temporal lobe hypometabolism was associated with inferior nonverbal memory. No relationship was found between performance on any frontal lobe measures and prefrontal hypometabolism. Statistical parametric mapping-quantified lateralized temporal lobe hypometabolism correlates with materialspecific episodic memory impairment in MTLE. In contrast, prefrontal hypometabolism is not associated with performance on frontal lobe measures. We suggest that this is because frontal lobe neuropsychology tests may not be good measures of isolated frontal lobe functioning. © 2015 Elsevier Inc. All rights reserved.
1. Introduction Temporal lobe epilepsy (TLE) is associated not only with episodic memory and language deficits characteristic of temporal lobe dysfunction [1,2] but also with cognitive deficits that extend beyond those thought to be mediated by the temporal lobe. These include IQ, attention, visuospatial deficits, speed of processing, and higher-level executive deficits [1]. Similarly, abnormalities extend well beyond the temporal lobe in TLE on fluorodeoxyglucose positron emission tomography (FDG-PET). Specifically, in addition to ipsilateral temporal lobe hypometabolism, ⁎ Corresponding author at: Department of Neurology, Westmead Hospital, PO Box 533, Wentworthville, NSW 2145, Australia. Tel.: +61 2 9845 6753; fax: +61 2 9635 6684. E-mail address: [email protected] (A.F. Bleasel).
http://dx.doi.org/10.1016/j.yebeh.2015.01.023 1525-5050/© 2015 Elsevier Inc. All rights reserved.
abnormalities have been reported in the contralateral temporal lobe and frontal, parietal, and thalamic regions mostly ipsilateral to the seizure focus [3–5]. To date, little research on the relationship between these functional abnormalities identified by FDG-PET and neuropsychological impairment in TLE exists. This stands in contrast with a sizable body of literature addressing the link between cognitive impairment and structural brain abnormalities in TLE both within and beyond the temporal lobes analyzed with quantitative MRI. This research has suggested that more widespread cognitive dysfunction, including lower IQ and executive dysfunction, is related to anatomical abnormalities outside of the temporal lobe [6]. The aim of the current study was to explore the relationship between FDG-PET hypometabolism and neuropsychological test scores in the most common form of focal epilepsy — unilateral mesial temporal
A.A. Knopman et al. / Epilepsy & Behavior 44 (2015) 136–142
lobe epilepsy (MTLE). Left TLE is reliably associated with verbal learning and memory deficits [1], while right TLE is generally thought to be associated with nonverbal memory deficits [2] (but see [7]). Indeed, several authors have reported a relationship between verbal memory and left temporal lobe hypometabolism [8–10], while nonverbal memory has been linked with right temporal lobe hypometabolism [8]. However, these studies have used either nonstandardized experimental memory stimuli [8] or memory measures which have been since replaced by more valid and reliable tests [9,10] — limitations that restrict clinical utility and practical reproducibility. Two studies have not found a correlation between temporal lobe hypometabolism and episodic memory measures in TLE [11,12]. Similarly, superseded memory measures were used in one of these studies [11], while the other study restricted their analysis to hypometabolism of the left hippocampal and parahippocampal gyri only, thereby not considering the potential role of the whole temporal lobe [12]. Thus, further investigation is warranted to clarify the relationship between episodic memory measures and lateralized temporal lobe hypometabolism in MTLE. Several lines of evidence suggest a relationship between prefrontal integrity and performance on higher-level neuropsychological tests measuring IQ, attention, working memory, psychomotor speed, and executive functions like problem-solving, abstract reasoning, response inhibition, and phonemic fluency (“frontal lobe abilities”) [13–19]. Only two studies have investigated whether extratemporal hypometabolism in TLE is correlated with neuropsychological functioning [3,4]. The first study reported metabolic asymmetry of the prefrontal cortex to be strongly related to IQ [3]. The second study employed quantitative statistical parametric mapping (SPM) methodology to report an association between prefrontal hypometabolism and executive dysfunction (set-shifting) in patients with MTLE, although these authors used a nonstandardized and experimental measure of executive functioning [4]. The current study examined for the first time the relationship between lateralized temporal lobe hypometabolism and separate material-specific episodic memory abilities in one sample of patients using standardized neuropsychological instruments. Additionally, the relationship between prefrontal hypometabolism and performance on frontal lobe neuropsychological tests has not previously been examined in the same sample of patients and not at all with standardized neuropsychological measures. The following hypotheses were made: first, dominant MTLE is associated with verbal memory impairment, and nondominant MTLE is associated with nonverbal memory impairment [8], with an analogous relationship between lateralized temporal lobe hypometabolism and memory using current memory measures. That is, it is expected that dominant temporal lobe hypometabolism is associated with poor verbal memory and that nondominant temporal lobe hypometabolism is associated with poor nonverbal memory; second, a relationship between poor performance on frontal lobe neuropsychological tests and hypometabolism of the prefrontal cortex is predicted. 2. Method
137
was obtained from most patients. No patient with dual pathology or psychiatric comorbidity (major depression, major anxiety, or psychosis) was included. For the majority of patients, hemispheric language dominance was investigated with the intracarotid amytal procedure (IAP) or functional magnetic resonance imaging (fMRI). Where these data were not available, left hemispheric language dominance was assumed for right-handed patients only. Exclusions were hippocampal lesions other than HS (n = 5), IQ b 70 (n = 4) as set in similar studies [20], left-handedness but no fMRI or IAP investigation (n = 3), previous neurosurgery (n = 3), bilateral language representation on fMRI or IAP investigations (n = 2), comorbid neurological disease (Parkinson's disease, n = 1; cerebral palsy, n = 1), history of brain injury (n = 1), normal pathology (n = 1), and English as a second language (n = 1). This left a final sample of 32 patients, four of whom demonstrated reversed hemispheric language dominance. Seventeen patients had a history of febrile convulsion, while 8 patients had a family history of epilepsy. All patients were on antiepileptic polypharmacotherapy, with the majority on two types of medication, but more specific information was not available given that extant data were used. Following surgery, 17 patients were classified as Engel Class 1 [21], 13 as Engel Class 2, and two as Engel Class 3, with a mean follow-up duration of 9.08 years (SD = 3.09). Other demographic data and clinical attributes are recorded in Table 1. The group with nondominant MTLE was older than the group with dominant MTLE (p = 0.03), but this was considered to be of low importance given that neuropsychological and FDG-PET measures were calculated relative to age-matched controls. There were no significant differences between groups on IQ or other demographic and clinical characteristics (t-tests or chi-square). 2.2. Neuropsychological measures Patients underwent comprehensive presurgical neuropsychological assessment with standard measures widely used in clinical practice [22] (Table 2). Raw scores on all psychometric tests were converted to age-adjusted z-scores based on published normative tables (mean = 0, SD = 1), except for performance on the Wisconsin Card Sorting Test (WCST), for which age and education-adjusted normative data were used. Cognitive domains were created based on concurrent and divergent validity studies showing that the different component tests measure the same cognitive construct [23]. This approach has been used elsewhere [20]. Domains comprised the mean z-score of contributing tests where available and included Verbal Memory [24,25], Nonverbal Memory [24,25], IQ [26,27], Executive Functioning [26–29], Attention/Working Memory [26,27], and Psychomotor Speed [26,27, 30] (Table 2). Performance on the WCST was also included as a separate measure of executive functioning given its reported sensitivity to frontal lobe integrity [13,15]. Note that different editions of the Wechsler scales are highly correlated with each other [31]. A more sensitive measure of relative material-specific memory deficit (Relative Memory) was also calculated for each patient by subtracting the scores for Nonverbal
2.1. Patients The study was approved by the ethics committees of Western and Central Sydney Area Health Services. Patients were retrospectively selected from the epilepsy surgery databases at the Westmead and Royal Price Alfred Hospitals, Sydney, Australia. All patients at least 17 years of age with MTLE, who had undergone standard anterior temporal lobectomy, who had completed both presurgical neuropsychological and FDG-PET investigations between 1994 and 2004, who had been followed up for at least two years (median = 9.08 years; range = 3.01– 15.20 years), and who had hippocampal sclerosis (HS) proven pathologically, were included (n = 54). Mesial temporal lobe epilepsy was diagnosed by clinical history, seizure semiology, interictal and ictal scalp electroencephalography (EEG), and MRI changes consistent with HS. Ictal single-photon emission computerized tomography (SPECT)
Table 1 Demographic and clinical attributes.
Gender (M/F) Agea (mean, SD) Educationa IQ Age at seizure onset a Duration of epilepsy a SGTCSs (present/absent)
Dominant MTLE (n = 14)
Nondominant MTLE (n = 18)
Total (n = 32)
2/12 29.6 (8.8) 9.9 (0.7) 85.4 (10.3) 8.0 (6.8) 21.9 (10.3) 6/8
7/11 37.7 (11.2)b 10.8 (1.7) 90.7 (15.6) 13.3 (10.2) 24.4 (11.2) 13/5
9/23 34.1 (10.9) 10.4 (1.4) 88.3 (13.6) 11.0 (9.1) 23.3 (10.8) 19/13
MTLE, mesial temporal lobe epilepsy; M/F, male/female; IQ, intelligence quotient; SGTCSs, secondary generalized tonic–clonic seizures. a Years. b p b 0.05.
A.A. Knopman et al. / Epilepsy & Behavior 44 (2015) 136–142
Table 2 Neuropsychological test battery. Domain Verbal Memory Nonverbal Memory
IQa Executive Functioninga
Attention/Working Memorya
Psychomotor Speeda
–
Test Logical Memory II from WMS-R/III Verbal Paired Associates II 9 Visual Paired Associates II > > = Visual Reproduction II from WMS-R/III > Faces II > ; Family Pictures II Full Scale IQ from WAIS-R/III Controlled Oral Word Association Test 9 Similarities = Block Design from WAIS-R/III ; Matrix Reasoning 9 Digit Span > > = LNS from WAIS-R/III > Spatial Span > ; Visual Memory Span 9 TMT-A = DSymb or DSymbC ; from WAIS-R/III Symbol Search b WCST
WMS-R/III, Wechsler Memory Scale — Revised or Third Edition; WAIS-R/III, Wechsler Adult Intelligence Scale — Revised or Third Edition; LNS, Letter Number Sequencing; TMT-A, Trail Making Test — Part A; DSymb, Digit Symbol; DSymbC, Digit Symbol Coding; WCST, Wisconsin Card Sorting Test. a The tests comprising these domains are sensitive to frontal lobe dysfunction [15,19,23]. b Number of perseverative errors.
Memory from the scores for Verbal Memory, where positive values represent relatively stronger verbal memory skills and negative values represent relatively stronger nonverbal memory skills. 2.3. FDG-PET acquisition and image processing Detailed procedures for presurgical FDG-PET acquisition and image processing for patients and healthy volunteer controls have been reported elsewhere [5]. Analysis of FDG-PET images was performed using Statistical Parametric Mapping (SPM 2, Wellcome Department of Cognitive Neurology, London, UK). Both patient and control images were spatially normalized onto the standard Montreal Neurological Institute (MNI, McGill University, Montreal, Quebec, Canada) brain PET template. Regional cerebral metabolic rate of glucose utilization (rCMRGlu) for each voxel was normalized to the whole-brain average of rCMRGlu using a proportional scaling method. Normalized images were then smoothed by convolution with FWHM Gaussian kernel (10 mm full width at half maximum). The 16 age-matched healthy volunteers who served as control subjects had no history of medical or neurological illness and did not take any medications known to affect FDG-PET studies. Fluorodeoxyglucose positron emission tomography was analyzed voxel-by-voxel with SPM 2 to identify significant regions of hypometabolism relative to the control group using a two-sample t-test at the individual voxel-level with a threshold of p ≤ 0.01 (z-score of 2.34). Only clusters with N250 contiguous voxels (volume of 2 cm3) and cluster significance of p ≤ 0.05 (uncorrected) were included. Significant regions of hypometabolism were superimposed onto a normalized, spatiallyregistered MNI brain template [32] with anatomic areas in each hemisphere including the anteromesial temporal cortex (temporal pole, hippocampus, parahippocampal gyrus, amygdala, and fusiform gyrus), whole temporal cortex (anteromesial cortex and lateral temporal cortex), prefrontal cortex (orbitofrontal, dorsolateral, and mesial frontal), central cortex, parietal cortex, occipital cortex, cingulate gyrus, insula, basal ganglia, and thalamus. The extent of hypometabolism comprised the number of significant voxels in each anatomical area.
FDG-PET variables (whole temporal cortex and prefrontal cortex) was necessary for normalization. Neuropsychological and FDG-PET variables were compared between the group with dominant MTLE and the group with nondominant MTLE with t-tests. Pearson's correlation was used to examine the relationship between regions of hypometabolism and neuropsychological functioning for the sample as a whole. Alpha here was set at 0.025 (i.e., 0.05/2) after the Bonferroni correction for the two interrelated temporal lobe regions (anteromesial temporal lobe and whole temporal lobe) for each hemisphere. Onetailed tests of significance were employed given the expected relationship between the material-specific memory and the corresponding temporal lobe based on similar research with FDG-PET [8–10]. Following a Bonferroni adjustment for the five frontal lobe abilities, alpha was set at 0.01 (i.e., 0.05/5) for investigation into whether each of these abilities was correlated with hypometabolism of the prefrontal cortex. Here, two-tailed tests of significance were employed given the uncertainty of the relationship between these particular brain regions and frontal lobe abilities. The only missing data were for six patients on the WCST. 3. Results 3.1. Neuropsychological functioning Fig. 1 shows the scores for Verbal Memory and Nonverbal Memory by group. There was a main effect for nonverbal memory, where nondominant MTLE was associated with poorer nonverbal memory compared with dominant MTLE (t(30) = 2.41, p = 0.02). In contrast, no main effect was found for verbal memory by group (t(30) = 1.31, p = 0.20). However, patients with dominant MTLE demonstrated weaker verbal memory than nonverbal memory (t(30) = 2.85, p = 0.01), which was not the case for patients with nondominant MTLE (t(30) = 1.04, p = 0.31). Relative Memory was also significantly different between groups (t(30) = 2.81, p = 0.009), whereby dominant MTLE was associated with worse relative verbal memory deficit compared with nondominant MTLE. Together, these results support the prediction that patients with unilateral MTLE demonstrate material-specific episodic memory deficits according to laterality. For frontal lobe abilities, there was no significant difference between groups on executive functioning (t(30) = 1.66, p = 0.11), attention/ 0 -0.2
Test performance (z-score)
138
-0.4 -0.6 -0.8 -1 -1.2 -1.4 -1.6 -1.8 -2 -2.2 Nondominant MTE
Dominant MTE
Verbal Memory Nonverbal Memory
2.4. Statistical analysis Data were analyzed with PASW Statistics (GradPack; version 18; SPSS Inc., Chicago, IL, U.S.A.). Square root transformation of some
Fig. 1. Mean (and standard error) for memory by group. Patients with nondominant MTLE demonstrated significantly poorer nonverbal memory compared with patients with dominant MTLE. Patients with dominant MTLE demonstrated significantly weaker verbal memory than nonverbal memory.
A.A. Knopman et al. / Epilepsy & Behavior 44 (2015) 136–142
working memory (t(30) = 0.37, p = 0.71), psychomotor speed (t(30) = 0.41, p = 0.69), or performance on the WCST (t(24) = 0.73, p = 0.47). Fig. 2 shows performance on frontal lobe ability measures for the whole sample. The greatest impairment was found for IQ (88.3 or z = −0.78), although a greater range was evident on WCST performance, with some patients demonstrating severe impairment. Overall, these results suggest that MTLE is associated with mildly impaired performance on frontal lobe neuropsychology tests. Education was significantly correlated with verbal memory (r = 0.41, p = 0.02, d = 0.90), IQ (r = 0.67, p b 0.001, d = 1.81), executive functioning (r = 0.54, p = 0.001, d = 1.28), attention/working memory (r = 0.48, p = 0.005, d = 1.09), and psychomotor speed (r = 0.56, p = 0.001, d = 1.35) but not with nonverbal memory (r = 0.20, p = 0.28, d = 0.41). Consequently, education was treated as a control variable for correlational analyses between regions of hypometabolism and neuropsychological functioning. Education was not held constant for analyses with WCST performance given that scores on this task were already education-adjusted. No other demographic or clinical attributes were related to neuropsychological functioning.
3.2. FDG-PET hypometabolism Ipsilateral temporal lobe hypometabolism was demonstrated in 91% of the patients (100% of the patients with dominant MTLE and 83% of the patients with nondominant MTLE). Bilateral temporal lobe hypometabolism was demonstrated in 19% of the patients (14% of the patients with dominant MTLE and 27% of the patients with nondominant MTLE). Hypometabolism beyond the ipsilateral temporal lobe was demonstrated in 69% of the patients. The most common pattern of extratemporal hypometabolism involved the ipsilateral frontal lobe and insula. Prefrontal hypometabolism was found in 47% of the patients. Patients with dominant MTLE showed greater hypometabolism of both the dominant anteromesial temporal cortex (t(30) = 8.95, p b 0.001) and the dominant whole temporal lobe (t(30) = 12.10, p b 0.001) relative to patients with nondominant MTLE. Patients with nondominant MTLE, in turn, showed greater hypometabolism of both the nondominant anteromesial temporal cortex (t(30) = 3.73, p = 0.001) and the nondominant whole temporal lobe (t(30) = 4.36, p = b0.001) relative to patients with dominant MTLE. Hypometabolism of neither the whole brain (t(30) = 0.48, p = 0.64) nor the prefrontal cortex (t(30) = 1.57, p = 0.13) differed by group.
0
Test performance (z-score)
-0.2
139
3.3. Association of nondominant temporal lobe hypometabolism and memory scores Fig. 3 shows the relationship between memory scores and temporal lobe regions of interest. In line with our prediction, hypometabolism of the whole nondominant temporal lobe was associated with nonverbal memory (r01.2(29) = −0.48, p = 0.003, d = 1.09). That is, poorer nonverbal memory was associated with increasing metabolic abnormality of the nondominant temporal lobe. Hypometabolism of the whole nondominant temporal lobe was also associated with Relative Memory, whereby relative nonverbal memory deficit was related to increasing metabolic abnormality of the nondominant temporal lobe (r01.2(29) = 0.47, p = 0.004, d = 1.06). Hypometabolism of the whole nondominant temporal lobe was not significantly correlated with verbal memory (r01.2(29) = 0.20, p = 0.14, d = 0.41). Similarly, hypometabolism of the more circumscribed nondominant anteromesial temporal lobe was associated with poor nonverbal memory (r01.2(29) = −0.41, p = 0.01, d = 0.90) but not with verbal memory (r01.2(29) = 0.11, p = 0.28, d = 0.22). Relative Memory was associated with nondominant anteromesial temporal lobe hypometabolism, but this relationship was not significant following the Bonferroni adjustment (r01.2(29) = 0.34, p = 0.03, d = 0.72). 3.4. Association of dominant temporal lobe hypometabolism and memory scores Hypometabolism of both the whole dominant temporal lobe and the dominant anteromesial temporal lobe was significantly associated with Relative Memory (r01.2(29) = − 0.46, p = 0.004, d = 1.04 and r01.2(29) = − 0.53, p = 0.00, d = 1.25, respectively). Specifically, relative verbal memory deficit was associated with increasing metabolic abnormality in the dominant temporal lobe. Poor verbal memory was associated with hypometabolism of the dominant anteromesial temporal lobe (r(30) = −.36, p = 0.02, d = 0.77), but this became a nonsignificant trend when controlling for education (r01.2(29) = −0.29, p = 0.06, d = 0.61). There was also a nonsignificant trend for an association between poor verbal memory and hypometabolism of the whole dominant temporal lobe (r01.2(29) = −0.17, p = 0.18, d = 0.35). Nonverbal memory was also significantly associated with hypometabolism of both these regions but in a positive direction (r01.2(29) = 0.52, p = 0.001, d = 1.22 and r01.2(29) = 0.46, p = 0.004, d = 1.04, respectively). That is, increasing hypometabolism of the whole dominant temporal lobe or the dominant anteromesial temporal lobe was associated with better nonverbal memory. We interpreted these results as being related to the fact that metabolic abnormality in the dominant temporal lobe was more likely in patients with dominant MTLE, who demonstrated intact nonverbal memory. 3.5. Association of prefrontal hypometabolism and frontal lobe neuropsychological measures
-0.4 -0.6
Hypometabolism of the prefrontal cortex was not significantly correlated with IQ (r(30) = 0.28, p = 0.12, d = 0.58), executive functioning (r(30) = 0.28, p = 0.12, d = 0.58), attention/working memory (r(30) = 0.23, p = 0.21, d = 0.47), psychomotor speed (r(30) = 0.14, p = 0.47, d = 0.28), or performance on the WCST (r(24) = 0.29, p = 0.15, d = 0.61). Thus, we could not support our second prediction that performance on frontal lobe neuropsychological tests is associated with hypometabolism of the prefrontal cortex.
-0.8 -1 -1.2 -1.4 -1.6 -1.8
4. Discussion
-2
Frontal lobe abilies Fig. 2. Mean (and standard error) for frontal lobe abilities for the group with dominant MTLE and for the group with nondominant MTLE combined. Attn/WM, attention/working memory; speed, psychomotor speed; WCST, Wisconsin Card Sorting Test.
We aimed to investigate the relationship between FDG-PET hypometabolism and widely used standardized neuropsychological measures in a well-defined, homogeneous sample of unilateral MTLE. Our first set of predictions was supported in that a relationship between
140
A.A. Knopman et al. / Epilepsy & Behavior 44 (2015) 136–142
A
B
Nondominant TL Whole
Whole
Anteromesial Pearson Correlaon
Pearson Correlaon
**
0.2 0.0 -0.2 -0.4 -0.6
Anteromesial
0.6
0.6 0.4
Dominant TL
** Relave Memory
*
0.4
**
**
0.2 0.0 -0.2 -0.4 -0.6
Verbal Memory
**
**
Nonverbal Memory
Fig. 3. A. Hypometabolism of both the whole nondominant temporal lobe and the more circumscribed anteromesial nondominant temporal lobe was correlated with Nonverbal Memory but not with Verbal Memory. The whole nondominant temporal lobe was correlated with Relative Memory. B. Hypometabolism of both the whole dominant temporal lobe and the more circumscribed anteromesial dominant temporal lobe was correlated with Relative Memory and Nonverbal Memory. Note that higher scores on Relative Memory (Nonverbal Memory subtracted from Verbal Memory) correspond to relatively better verbal memory and that lower scores correspond to relatively better nonverbal memory. *p b 0.025, **p b 0.005; TL, temporal lobe.
lateralized temporal lobe hypometabolism and material-specific episodic memory impairment was found. Specifically, hypometabolism within the dominant temporal lobe was related to relatively poorer verbal memory, and nondominant temporal lobe hypometabolism was related to absolute nonverbal memory impairment. We were unable to show a relationship between performance on frontal lobe neuropsychological measures and prefrontal hypometabolism. Consequently, we could not support our second hypothesis. Our findings support the limited literature showing a relationship between lateralized temporal lobe hypometabolism and materialspecific episodic memory impairment in TLE [8–10]. However, to our knowledge, the current study is the first to show this relationship separated by verbal memory and nonverbal memory in the same sample of patients using standardized neuropsychological instruments. We also showed that patients with nondominant MTLE demonstrated poorer nonverbal memory compared with patients with dominant MTLE and that patients with dominant MTLE demonstrated poorer verbal memory than their own nonverbal memory. Together, these sets of results are in line with the lateralized material-specific memory deficit model of TLE [33]. Although its validity has been questioned in recent years [7], our data provide unique support for this model by evidencing materialspecific episodic memory correlates of SPM-quantified lateralized temporal lobe hypometabolism. In the context of the observed relationships between episodic memory and lateralized temporal lobe hypometabolism, we were unable to support a relationship between performance on frontal lobe neuropsychological tests and hypometabolism of the prefrontal cortex. This is surprising because of the widely held view that the frontal lobes are sensitive to performance on neuropsychological measures of IQ, attention, working memory, psychomotor speed, and executive abilities [13–17]. However, two other studies cast doubt on the putative ‘expected’ relationship between frontal lobe or executive cognitive abilities and frontal lobe imaging abnormalities in TLE. The first study failed to find an association between frontal lobe volume on quantitative MRI and performance on several measures of executive functioning in TLE [34]. The second study showed that frontal lobe thinning on quantitative MRI could not distinguish between varying degrees of cognitive impairment. Instead, a systematic and stepwise relationship was reported between a combination of generalized cortical thinning and abnormalities of subcortical structures and white matter with increasing levels of cognitive impairment across all cognitive domains [35]. The current results suggest that hypometabolism on FDG-PET is partially discordant with neuropsychological functioning. That is, while lateralized temporal lobe hypometabolism is associated with material-
specific episodic memory impairment both in temporal lobe epilepsy and in extratemporal epilepsy [36], the same metabolic–cognitive relationship may not apply for hypometabolism of the prefrontal cortex and frontal lobe neuropsychological test performance. One possibility is that these tests may not be good measures of frontal lobe functioning in isolation. Indeed, subcortical structures seem also to play a role in frontal lobe cognitive abilities: atrophy of the thalamus and basal ganglia has been reported to be correlated with executive functioning in patients with TLE [34,37]. Such findings are in line with our understanding of the role these structures play in relaying information processing along cortico-subcortical circuits which connect the prefrontal cortex, basal ganglia, and cerebellum via the thalamus [38]. In our sample, too few patients demonstrated metabolic abnormalities of subcortical regions to allow meaningful analysis of whether these regions correlate with frontal lobe abilities. Another viewpoint is that frontal lobe neuropsychological measures are instead sensitive to a distributed neural network of frontal and nonfrontal brain regions [13]. Indeed, emerging research on resting state connectivity using fMRI analysis points to the abnormality of various resting state networks in TLE including the default mode and executive control networks — known to be involved during higher-level cognitive tasks [39]. Our findings are in contrast with previous reports of hypometabolism of the prefrontal cortex being correlated with low IQ [3] and executive dysfunction [4] in unilateral TLE. However, there are important differences between the current methodology and that used by these authors. The first study used hand-drawn regions of interest with a liberal threshold of N10% asymmetry value to determine prefrontal metabolic abnormality in patients with TLE, only a subset of whom had HS [3]. Visual analysis of FDG-PET is affected by known biases and multiple sources of variation [40], and asymmetry measures may hide bilateral hypometabolism [41]. In contrast, we employed whole brain SPM in a welldefined sample with homogeneous MTLE, with a high threshold of 250 contiguous voxels required to detect metabolic abnormality. This precise quantitative methodology offers greater objectivity and sensitivity, particularly when compared with a healthy control group. In the second study, despite SPM methodology being used, the FDG-PET scans of the 21 patients with MTLE were not compared with those from a healthy control group, and the authors employed an experimental and unstandardized measure of executive functioning [4]. In contrast, our study utilized FDG-PET scans from a healthy control group and a range of commonly used standardized neuropsychology measures. Another technical difference between our study and that by Takaya and colleagues [4] is that they measured the mean intensity of hypometabolism in selected regions of interest, which is subject to volume-averaging effects [42].
A.A. Knopman et al. / Epilepsy & Behavior 44 (2015) 136–142
An important consideration for future research is that the precise mechanism of FDG-PET hypometabolism distant to the seizure focus in MTLE remains incompletely understood. Diaschisis seems to play a significant role with epilepsy disturbing neuronal activity along anatomical frontotemporal pathways [4]. Repeated seizure propagation along the same pathways may result in additional remote hypometabolism. The current results and others [8–10,36] suggest that regional temporal lobe hypometabolism is associated with regional cognitive dysfunction. However, the hypothesis that distant hypometabolism of the prefrontal cortex is associated with regional cognitive dysfunction is supported by some research [3,4,18] but not ours. Limitations of the current study include its retrospective nature. This did not allow control over the latency between neuropsychological assessment and FDG-PET scanning, although this was typically within a few months, and data on the latency between the last seizure and the FDG-PET scan were not collected. The fact that our sample consisted of highly selected surgical patients with MTLE for whom FDG-PET was part of the presurgical decision for surgery can be seen as another limitation. However, it is unlikely that patients were excluded from surgery because of the FDG-PET result with otherwise concordant investigations in the context of HS. We note that the heterogeneous collection of measures included as frontal lobe abilities are not perfect measures of frontal lobe functioning and that they tap into other brain areas too. However, very few neuropsychological tests measure frontal lobe functioning in isolation, and all of the measures employed in our study involve the frontal lobes to some degree. Antiepileptic medication is another factor we were unable to control, and we cannot rule out medications contributing to reductions in cognitive functioning. 5. Conclusions We employed voxel-wise quantitative analysis of FDG-PET and widely used standardized neuropsychological measures to show a correlation between lateralized temporal lobe hypometabolism and material-specific episodic memory impairment in a well-defined homogeneous sample of pathologically confirmed MTLE. In contrast, performance on frontal lobe neuropsychological measures was not related to hypometabolism of the prefrontal cortex. The results suggest that frontal lobe neuropsychology tests are not good measures of frontal lobe functioning in isolation and that clinical interpretation of FDG-PET should be made with caution. Acknowledgments This work was supported, in part, by the University of Sydney Postgraduate Award, Millennium Institute Stipend, and Pfizer Neuroscience Research Grant (03-01-07) to Dr. Chong H. Wong. Disclosures None of the authors has any conflicts of interest to disclose. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. References [1] Hermann BP, Seidenberg M, Schoenfeld J, Davies K. Neuropsychological characteristics of the syndrome of mesial temporal lobe epilepsy. Arch Neurol 1997;54:369–76. [2] Baxendale S, Paesschen W, Thompson PJ, Connelly A, Duncan JS, Harkness WF, et al. The relationship between quantitative MRI and neuropsychological functioning in temporal lobe epilepsy. Epilepsia 1998;39:158–66. [3] Jokeit H, Seitz RJ, Markowitsch HJ, Neumann N, Witte OW, Ebner A. Prefrontal asymmetric interictal glucose hypometabolism and cognitive impairment in patients with temporal lobe epilepsy. Brain 1997;120:2283–94. [4] Takaya S, Hanakawa T, Hashikawa K, Ikeda A, Sawamoto N, Nagamine T, et al. Prefrontal hypofunction in patients with intractable mesial temporal lobe epilepsy. Neurology 2006;67:1674–6.
141
[5] Wong CH, Bleasel A, Wen L, Eberl S, Byth K, Fulham M, et al. The topography and significance of extratemporal hypometabolism in refractory mesial temporal lobe epilepsy examined by FDG-PET. Epilepsia 2010;51:1365–73. [6] Bell B, Lin JJ, Seidenberg M, Hermann B. The neurobiology of cognitive disorders in temporal lobe epilepsy. Nat Rev Neurol 2011;7:154–64. [7] Glikmann-Johnston Y, Saling MM, Chen J, Cooper KA, Beare RJ, Reutens DC. Structural and functional correlates of unilateral mesial temporal lobe spatial memory impairment. Brain 2008;131:3006–18. [8] Harris IM, Fulham MJ, Miller LA. The effects of mesial temporal and cerebellar hypometabolism on learning and memory. J Int Neuropsychol Soc 2001;7: 353–62. [9] Rausch R, Henry TR, Ary CM, Engel Jr J, Mazziotta J. Asymmetric interictal glucose hypometabolism and cognitive performance in epileptic patients. Arch Neurol 1994;51:139–44. [10] Weintrob DL, Saling MM, Berkovic SF, Berlangieri SU, Reutens DC. Verbal memory in left temporal lobe epilepsy: evidence for task-related localization. Ann Neurol 2002; 51:442–7. [11] Arnold S, Schlaug G, Niemann H, Ebner A, Luders H, Witte OW, et al. Topography of interictal glucose hypometabolism in unilateral mesiotemporal epilepsy. Neurology 1996;46:1422–30. [12] Griffith HR, Pyzalski RW, Seidenberg M, Hermann BP. Memory relationships between MRI volumes and resting PET metabolism of medial temporal lobe structures. Epilepsy Behav 2004;5:669–76. [13] Alvarez JA, Emory E. Executive function and the frontal lobes: a meta-analytic review. Neuropsychol Rev 2006;16:17–42. [14] Baldo JV, Shimamura AP, Delis DC, Kramer J, Kaplan E. Verbal and design fluency in patients with frontal lobe lesions. J Int Neuropsych Soc 2001;7:586–96. [15] Demakis GJ. A meta-analytic review of the sensitivity of the Wisconsin Card Sorting Test to frontal and lateralized frontal brain damage. Neuropsychology 2003;17: 255–64. [16] Keller SS, Baker GA, Downes JJ, Roberts N. Quantitative MRI of the prefrontal cortex and executive function in patients with temporal lobe epilepsy. Epilepsy Behav 2009;15:186–95. [17] Stuss DT, Alexander MP, Shallice T, Picton TW, Binns MA, Macdonald R, et al. Multiple frontal systems controlling response speed. Neuropsychologia 2005;43:396–417. [18] Takaya S, Mikuni N, Mitsueda T, Satow T, Taki J, Kinoshita M, et al. Improved cerebral function in mesial temporal lobe epilepsy after subtemporal amygdalohippocampectomy. Brain 2009;132:185–94. [19] Glascher J, Tranel D, Paul LK, Rudrauf D, Rorden C, Hornaday A, et al. Lesion mapping of cognitive abilities linked to intelligence. Neuron 2009;61:681–91. [20] Hermann BP, Seidenberg M, Lee EJ, Chan F, Rutecki P. Cognitive phenotypes in temporal lobe epilepsy. J Int Neuropsychol Soc 2007;13:12–20. [21] Engel Jr J, Van Ness PC, Rasmussen TB, Ojemann LM. Outcome with respect to epileptic seizures. In: Engel JJ, editor. Surgical treatment of the epilepsies. Second ed. New York: Raven Press; 1993. p. 609–21. [22] Djordjevic J, Bellerose J, Rudy L, Jones-Gotman M. Neuropsychological assessment of patients with epilepsy: an international survey. American Epilepsy Society 63rd Annual Meeting, December 4–8 Boston, MA; 2009. [23] Strauss E, Sherman E, Spreen O. A compendium of neuropsychological tests: administration, norms, and commentary. Third ed. New York: Oxford University Press; 2006. [24] Wechsler D. Wechsler Memory Scale — Revised. San Antonio, TX: The Psychological Corporation; 1987. [25] Wechsler D. Wechsler Memory Scale — Third Edition. San Antonio, TX: The Psychological Corporation; 1997. [26] Wechsler D. Wechsler Adult Intelligence Scale — Revised. San Antonio, TX: The Psychological Corporation; 1981. [27] Wechsler D. Wechsler Adult Intelligence Scale — Third Edition. San Antonio, TX: The Psychological Corporation; 1997. [28] Heaton RK, Chelune GJ, Talley JL, Kay GG, Curtis G. Wisconsin Card Sorting Test (WCST) manual, revised and expanded. Odessa, FL: Psychological Assessment Resources; 1993. [29] Tombaugh TN, Kozak J, Rees L. Normative data stratified by age and education for two measures of verbal fluency: FAS and animal naming. Arch Clin Neuropsychol 1999;14:167–77. [30] Tombaugh TN. Trail Making Test A and B: normative data stratified by age and education. 2004;19:203–14. [31] Tulsky D, Zhu J, Ledbetter M. WAIS-III WMS-III Technical Manual. San Antonio, TX: The Psychological Corporation; 1997. [32] Tzourio-Mazoyer N, Landeau B, Papathanassiou D, Crivello F, Etard O, Delcroix N, et al. Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage 2002;15:273–89. [33] Milner B. Interhemispheric differences in the localization of psychological processes in man. Brit Med Bull 1971;27:272–7. [34] Tuchscherer V, Seidenberg M, Pulsipher D, Lancaster M, Guidotti L, Hermann BP. Extrahippocampal integrity in temporal lobe epilepsy and cognition: thalamus and executive functioning. Epilepsy Behav 2010;17:478–82. [35] Dabbs K, Jones J, Seidenberg M, Hermann BP. Neuroanatomical correlates of cognitive phenotypes in temporal lobe epilepsy. Epilepsy Behav 2009;15:445–51. [36] Knopman AA, Wong CH, Stevenson RJ, Homewood J, Mohamed A, Somerville E, et al. The cognitive profile of occipital lobe epilepsy and the selective association of left temporal lobe hypometabolism with verbal memory impairment. Epilepsia 2014; 55:e80–4. [37] Riley JD, Moore S, Cramer SC, Lin JJ. Caudate atrophy and impaired frontostriatal connections are linked to executive dysfunction in temporal lobe epilepsy. Epilepsy Behav 2011;21:80–7.
142
A.A. Knopman et al. / Epilepsy & Behavior 44 (2015) 136–142
[38] Alexander GE, Crutcher MD, DeLong MR. Basal ganglia-thalamocortical circuits: parallel substrates for motor, oculomotor, “prefrontal” and “limbic” functions. Prog Brain Res 1990;85:119–46. [39] Cataldi M, Avoli M, de Villers-Sidani E. Resting state networks in temporal lobe epilepsy. Epilepsia 2013;54:2048–59. [40] Van Paesschen W, Goffin K, Van Laere K. PET and SPECT in presurgical evaluation of epilepsy. In: Shorvon SD, Perucca E, Engel JJ, editors. The treatment of epilepsy. 3rd Edition. Oxford: Wiley-Blackwell; 2009. p. 821–8.
[41] Hashiguchi K, Morioka T, Yoshida F, Kawamura T, Miyagi Y, Kuwabara Y, et al. Thalamic hypometabolism on 18FDG-positron emission tomography in medial temporal lobe epilepsy. Neurol Res 2007;29:215–22. [42] Rubin E, Dhawan V, Moeller JR, Takikawa S, Labar DR, Schaul N, et al. Cerebral metabolic topography in unilateral temporal lobe epilepsy. Neurology 1995;45:2212–23.
