Abstracts

Rationale: Among partial-onset epilepsies in adults temporal lobe epilepsy (TLE) is the most common and most refractory to medical management. Anterior temporal lobectomy (aTL) is often therapeutic; however, the potential for adverse cognitive outcomes, notably in language and memory, is significant. Presently identification of surgical candidates relies on the Wada procedure, an invasive angiographic approach for anatomically lateralizing these cognitive domains. Functional MRI (fMRI) is a potentially non-invasive alternative. In particular, alterations in the resting- state fMRI signal, such as the default mode network (DMN), have been hypothesized to predict the cognitive impact of aTL in TLE patients. Here, we 1) confirm the vaule of resting state fMRI in lateralizing seizure onset, and 2) attempt to identify a resting-state fMRI measure that outperforms Wada testing in predicting the impact of aTL on memory. Methods: Twenty patients with TLE undergoing surgical evaluation were included. Those with a unilateral temporal seizure focus underwent 1) Wada testing and 2) preoperative neuropsychological testing as routine preoperative workup, as well as 3) 3T fMRI as part of a prior research study. For analysis of the fMRI, resting signal was extracted, then averaged within each of 90 anatomical regions of interest. Pearson correlation coefficient was calculated between each hippocampus and the left and right halves of the posterior cingulate cortex (PCC). Most patients additionally underwent postoperative neuropsychological testing. Those who had completed testing were analyzed. The postoperative change in two neuropsychological parameters (measures of visuospatial and of verbal memory) were linearly regressed against possible predictors: Wada memory asymmetry index and functional connectivity between regions of interest (ROIs) in the DMN. Results: Functional connectivity between the left hippocampus and the posterior cingulate cortex (PCC) was significantly decreased in left compared to right TLE ( p < .051 for left PCC, p < .06 for right PCC). No corresponding difference was seen in the right hippocampus. The functional connectivity between the hippocampi and the PCC, or between the anterior cingulate cortex and PCC, did not correlate with memory change postoperatively. Conclusions: We confirmed that elements of the DMN are useful in lateralizing TLE. Our results suggest that resting state fMRI may not contribute to the prediction of postoperative memory change.