Abstracts

Rationale: This study explored the concept of the "functional deficit zone" (the brain area showing abnormal functioning in the interictal period) by correlating neuropsychology measures and PET hypometabolism. It was hypothesised that (a) verbal memory is associated with dominant temporal lobe (TL) hypometabolism, (b) nonverbal memory is associated with nondominant TL hypometabolism, (c) FSIQ, executive, attention/working memory and/or speed of thinking is associated with prefrontal hypometabolism. Additionally, we examined whether any of these neuropsychology variables could predict poor outcome following temporal lobectomy. Methods: Thirty-four patients (23 females, mean age 34.7±10.8) with FSIQ>69 were included in the study, all of whom had medically refractory mesial temporal lobe epilepsy (TLE), had undergone temporal lobectomy (14 dominant) and were followed up for at least one year. Presurgical FDG-PET was analysed voxel-by-voxel using SPM to identify significant regions of glucose hypometabolism relative to normal controls. Presurgical neuropsychology variables included FSIQ (WAIS-R/III), delayed verbal and nonverbal memory indices (WMS-R/III), executive functioning (Wisconsin Card Sorting Task perseverative errors (WCSTpe) or a composite of WAIS-R/III Similarities, Block Design, Matrix Reasoning and COWAT), attention/working memory (composite of WAIS-R/III Digit Span, LNS and Spatial Span) and speed (composite of WAIS-R/III Coding, Symbol Search and TMT-A). Results: Nondominant TLE patients demonstrated poorer visual memory than dominant patients (t=2.29, p<0.05). Dominant TLE patients demonstrated weaker verbal memory relative to visual memory (t=2.85, p<0.05). Poor verbal memory was associated with dominant anteromesial TL hypometabolism (r=-0.36, p<0.05), whereas poor nonverbal memory was associated with nondominant TL hypometabolism (r=-0.39, p<0.05). FSIQ was not associated with prefrontal hypometabolism (r=0.23, p=0.19). There was no association between prefrontal hypometabolism and executive (WCSTpe: r=0.2, p=0.32; composite: r=0.26, p=0.14), attention/working memory (r=0.21, p=0.23) or speed (r=0.10, p=0.58) measures. No neuropsychology variables were able to predict surgical outcome (Engel class 1). Conclusions: Standard neuropsychological memory indices correlate with unilateral temporal lobe hypometabolism in the expected material-specific way, a relationship not previously identified for both verbal and nonverbal memory. FSIQ is not correlated with prefrontal hypometabolism, in contrast with the findings reported by Jokeit et al. (Brain 120:2283,1997). Furthermore, other specific neuropsychological indices do not correlate with prefrontal hypometabolism, an unexpected finding given the widely held view that the prefrontal region broadly subserves cognition. This suggests remote PET hypometabolism (Wong et al. Epilepsia 51:1365,2010) is not concordant with the functional deficit zone. Neuropsychological measures do not predict seizure outcome following temporal lobectomy.