Abstracts

RATIONALE: This study reports the association of MRI defined hippocampal abnormality (volume loss and increased T2 signal) with memory performance. Although correlations between volume loss and memory impairment have been previously reported, they are modest in magnitude. Furthermore, the MRI defined syndrome of MTS also includes hippocampal signal abnormality, but there is no information about the relation of signal abnormality to memory performance.
METHODS: A consecutive sample of 236 patients taking part in a multicenter quality of life outcome study underwent MRI and neuropsychological examination. MRI scans obtained at several different centers were all read by the same neuroradiologist (RB). Neuropsychological examination included assessment of intellectual functioning (WAIS-R), verbal memory (CVLT) and nonverbal memory (Rey CFT). All Rey figures were scored by the same rater (MW) for reliability.
RESULTS: Approximately one-half of the sample (50.4%) was male, ranging in age from 16-65 years (average FSIQ=90.1). Sixty-seven patients had a history of febrile seizures. Two separate sets of analyses, univariate and multivariate, were performed. Univariate analysis was performed to assess the association of volume loss with material specific (verbal or nonverbal) memory impairment. Separate univariate analysis examined the assocation of signal change with memory impairment. Multivariate analyses (Regression and MANOVA) were then performed to investigate the combined power of volume loss and signal change for predicting degree of baseline memory impairment, controlling for confounding variables such as IQ, age, gender, and seizure history variables. Results confirm previously described relationships between left hippocampal atrophy and verbal memory impairment (p=.0188). We also observed a relationship between right hippocampal atrophy and nonverbal memory impairment (p=.04). Signal abnormality was not, by itself, significantly related to verbal or nonverbal memory performance. However, in a regression model that included both volume loss and signal abnormality, left hippocampal signal abnormality contributed significantly to the overall prediction of verbal memory impairment. MANOVA (using all 10 memory variables) revealed both right and left signal abnormality were significantly associated with overall differences in memory performance.
CONCLUSIONS: The present results indicate that the combination of atrophy and signal change may be a better predictor of memory function than either measure alone when considered in a multivariate model that also accounts for confounding variables such as gender, history of febrile seizures, and lateralization of EEG defined seizure onset. This may provide a better model for predicting memory outcome in temporal lobectomy patients.
Support: Supported in part by NIH Grant RO1NS32375