Authors :
Presenting Author: Sherry Liu, MBBS, MRCS – Institute of Neurology, UCL
Jane de Tisi, BSc – NHNN, UCL
Lorenzo Bianchi, PhD – UCL, Department of Clinical and Experimental Epilepsy; Epilepsy Society, MRI Unit
Davide Giampiccolo, MD – NHNN, UCL
Anna Miserocchi, MD, FRCS – UCL, Department of Clinical and Experimental Epilepsy; Epilepsy Society, MRI Unit
Andrew McEvoy, FRCS – UCL, Department of Clinical and Experimental Epilepsy; Epilepsy Society, MRI Unit
Aidan O'Keeffe, PhD – University of Nottingham
Peter Taylor, PhD – Newcastle University
Sjoerd Vos, PhD – UCL, Department of Clinical and Experimental Epilepsy; Epilepsy Society, MRI Unit
Sallie Baxendale, PhD – UCL, Department of Clinical and Experimental Epilepsy; Epilepsy Society, MRI Unit
Meneka Sidhu, MRCP, PhD – UCL, Department of Clinical and Experimental Epilepsy; Epilepsy Society, MRI Unit
Gavin Winston, MD/PhD – Queen's University
John duncan, DM FRCP FMedSci – NHNN, UCL
Lawrence Binding, PhD – Centre for Medical Imaging and Computing, UCL
Rationale:
Anterior temporal lobe resection (ATLR) for drug-resistant temporal lobe epilepsy (TLE) leads to significant postoperative visual memory decline in 30% of right- and < 10% of left-sided resections, which can be be debilitating. To minimize these postoperative deficits, alternative approaches with smaller lateral temporal neocortical resections, such as selective amygdalohippocampectomy and temporal pole resections, have been suggested. These approaches, however, result in similar memory deficits. We tested the hypothesis that damage to subcortical white matter tracts might be the underlying cause for memory decline in such patients. This study examines the relationship between surgically disconnected structural circuits and post-ATLR visual memory decline.
Methods:
Visual memory encoding (immediate recall) and retrieval (delayed recall) outcomes in 138 left TLE (LTLE) and 150 right TLE (RTLE) patients who underwent ATLR were z-scored against healthy controls at three timepoints: pre-operative, and 3- and 12-months postoperatively. White matter tracts in temporal lobe implicated in visual memory—ventral cingulum (vCing), fornix, medial occipital longitudinal tract (MOLT), and middle longitudinal fasciculus— were reconstructed using an atlas-based template generated from a subset of these patients with preoperative diffusion MRI (dMRI, N=59). The reconstructed tracts, overlaid on the resection cavity masks, were considered disconnected if ≥50% of its volume overlapped with resected voxels. Visual memory outcomes measured by changes in Z-scores from preoperative to 3- and 12-months postoperatively were correlated with tract disconnection to identify white matter tracts associated with visual memory. Additionally, preoperative dMRI tractography was mapped on to resection masks to estimate the percentage of fibre bundle transected and postoperative dMRI tractography was used to calculate the remaining percentage of fibre bundle. These were similarly correlated with memory decline outcomes.
Results:
Individuals with RTLE showed significantly greater visual memory decline at both 3- and 12-month post-operatively than did those with LTLE. In RTLE patients, disconnection of the vCing was associated with decline in visual encoding, but not retrieval. In LTLE patients, MOLT disconnection was associated with visual encoding decline. A greater proportion of vCing fibre transection and reduced remaining tract volume on the right significantly correlated with visual encoding decline in RTLE, and similar findings for MOLT were observed in LTLE.
Conclusions:
Our results suggest that visual memory may be supported by white matter connections between the cortical-subcortical and cortical-cortical regions with memory deficits arising from their transection. These findings may inform a revised surgical approach to minimise damage to the implicated tracts to optimise postoperative memory outcome.
Funding: The presenter of the abstract is funded by Singapore National Clinician-Scientist Residency Program funding