Cognitive function and the longitudinal hippocampal axis in mesial temporal sclerosis.

Objective: The relationship of preoperative memory deficits in patients with mesial temporal lobe epilepsy (mTLE) and hippocampal sclerosis (HS) to the distribution of neuronal loss is uncertain. Building on the material specificity theory, we tested the hypothesis that visual memory deficits are associated with posterior hippocampal atrophy, whereas verbal memory deficits are associated with anterior hippocampal atrophy.

Methods: We studied 22 adults with mTLE and HS, calculating hippocampal head, body, and tail volumes, correcting for estimated total intracranial volume, using automated segmentation. Preoperative memory ability was evaluated with the Wechsler Memory Scale (WMS-II: logical memory, verbal paired associates, family pictures, and faces subtests). We correlated memory ability with hippocampal division volumes using SPSS 26.1 (repeated measures ANOVAs, one-way ANOVAs, Pearson r correlations) for statistical analysis.

Results: We found a significant main effect of hippocampal subdivision, reporting volumetric differences between the head, body, and tail. Pairwise comparisons reported that the hippocampal head had significantly greater volume than both the body and tail (p < 0.001). For both left and right focus groups, the ipsilateral hippocampi were significantly smaller than the contralateral. Linear regression reported a left hippocampal model (head, body, and tail volumes) predicted performance on logical memory with the left hippocampal tail volume being the strongest predictor. A right hippocampal model (head, body, and tail volumes) predicted memory ability for family pictures and verbal paired associates at a trend level.

Conclusions: Ipsilateral hippocampal head and tail seem more vulnerable to injury than the body in both the left and right mTLE. Our study suggests there may be functional differences along the hippocampal longitudinal axis, particularly for the left hippocampal tail with verbal memory. Our findings are consistent with material-specific right-left differences in memory processing.