In patients with temporal lobe epilepsy undergoing temporal lobe resections (TLR), hippocampus (hc) sparing surgical procedures are typically considered as effective in preventing postoperative memory deterioration. Despite this function-preserving approach, some patients may experience postoperative memory decline, possibly due to a postoperative shrinkage of the ipsilateral hc. The objective of the present study was to investigate whether this shrinkage and hippocampal volume loss is associated with postoperative verbal memory performance and to identify possible other risk factors for memory decline after surgery.

Analyses included data of 32 patients who underwent hc sparing TLR (resection of the temporal pole and corpus amygdaloideum) between 2010 and 2020. We included data on (1) pre- to postoperative verbal memory decline in a German version of the Rey Auditory Verbal Learning Test (learning capacity, long-term recall, long-term retention), (2) lateralization of surgery, (3) presence of postoperative hippocampal shrinkage as determined by visual inspection of the postoperative MRI by an expert clinical radiologist, and (4) volume loss of the hc as determined by a semi-automated software that analyzed pre- and postoperative T1-weighted structural 3 Tesla MRI. Postoperative MRI scans were obtained six months after surgery and postsurgical resection extent was manually traced and coregistered to presurgical scans for precise quantification of lesion extent and volume loss.

The results showed (1) no significant difference in overall learning capacity and long-term recall or retention change between patients with and without postoperative hippocampal shrinkage (all p’s ≥ .31); (2) no significant association between verbal memory change and hippocampal volume loss (all p’s ≥ .54); (3) that patients with language dominant TLR exhibited a significantly increased memory decline compared to patients with non language-dominant TLR (all p’s ≤ .04); (4) that the factors lateralization of surgery and preoperative memory performance significantly predicted postoperative memory change (p < .001; R² = .37).

Our data suggest that there seems to be no strong relationship between verbal memory change and hippocampal shrinkage as well as hippocampal volume loss after hc sparing TLR. Rather, as with surgical procedures including the hc, high preoperative memory performance and language dominant side of the resection seem to be crucial risk factors in the prediction of postsurgical memory change. Our findings indicate that even hc sparing surgical procedures in TLR may not necessarily prevent verbal memory decline. Disruptions to the broader (medial) temporal lobe memory network may be sufficient to cause memory change, particularly if resections are performed in the language-dominant hemisphere and if memory is unaffected prior to surgery.