Abstracts

Previous studies have reported that measures of delayed recall and learning of paired word associates are particularly sensitive to mesial temporal sclerosis in presurgical epilepsy patients. Studies examining the role of mesial temporal structures in verbal recognition memory have been mixed. The purpose of this study was to examine how accurately measures of delayed recall of stories, word pairs, and word lists discriminate surgical candidates with and without left mesial temporal sclerosis. A similar analysis was conducted with measures of verbal recognition memory., Subjects were 27 candidates for left temporal lobectomy who underwent comprehensive neuropsychological assessment as part of their presurgical evaluation. Mean age was 35.4 [plusmn] 10.5 years, mean education was 13.0 [plusmn] 2.1 years, and mean seizure duration was 24.0 [plusmn] 12.6 years. Subjects were classified as having mesial temporal sclerosis (MTS; n=21) or not (NMTS; n=6) according to MRI findings read by a consensus of experienced epileptologists and neuroradiologists. The two groups did not differ in age or education, but the MTS group did have longer seizure duration than the NMTS group. Wada testing revealed that 21 subjects had left-hemisphere dominant language, while 6 showed bilateral language functions (5 showed L[gt]R). Subjects were administered the California Verbal Learning Test-II and the Logical Memory and Verbal Paired Associates subtests of the Wechsler Memory Scale-III. All scores were corrected for age., Correlational analyses revealed strong associations between the measures of delayed recall (all r[apos]s [gt].62, p[lt].001). Strong associations between the two measures of recognition memory were also revealed (r= .67, p[lt] .001). Discriminant function analyses, with MTS vs. NMTS as the dependent variable, indicated that all measures of delayed recall were good individual predictors of MTS group (73-77% of cases correctly classified), but classification accuracy was modestly improved to 80% when the measures were combined. An additional discriminant function analysis indicated that the recognition memory measures were very strong predictors of MTS, correctly classifying 96% of cases., The results suggest that commonly-used clinical measures of verbal delayed recall can discriminate between patients with and without MTS, and that classification accuracy is improved when these measures are used in combination. Additionally, very high classification accuracy was observed when measures of delayed recognition memory were employed. Our results may differ from those of previous studies reporting weak associations between MTS and verbal recognition memory because of differences in task complexity and length of delay.,