Back

A Network Target for Memory Dysfunction Derived from Brain Lesions and Stimulations

Howard, C. W.; Madan, S.; Garimella, A.; Schaper, F.; Kletenik, I.; Ng, M. C.; Mosley, P.; Grafman, J.; Bakshi, R.; Glanz, B.; Fosdick, L.; Johnson, A.; Colyer, R.; Lyketsos, C. G.; Morton-Dutton, M.; Giftakis, J.; Temel, Y.; Rouhl, R. P. W.; Ko, J. H.; Onur, O.; Schmahl, R.; Baldermann, J. C.; Andrade-Montemayor, P.; Visser-Vandewalle, V.; Kuhn, J.; Corbetta, M.; Fisher, R. S.; Picht, T.; Faust, K.; Hermiller, M.; Voss, J.; Chitnis, T.; Kahana, M. K.; Smith, G. S.; Lozano, A.; Siddiqi, S. H.; Horn, A.; Fox, M. D.

2026-03-12 neurology
10.64898/2026.03.10.26348082 medRxiv
Show abstract

Therapeutic brain stimulation holds promise in treating memory dysfunction, but recent clinical trials have produced heterogenous results. Uncertainty in the neuroanatomical target may contribute to this heterogeneity, with over 13 different brain regions targeted to date. To derive a neuroanatomical target, we studied verbal episodic memory changes across 1247 patients in 12 independent datasets, including patients with focal lesions (n = 985), DBS sites (n = 207), and TMS sites (n = 72). We found lesion and stimulation sites causing verbal memory changes converged on a common brain network. In 3 held-out datasets, connectivity to this convergent memory network explained more variance than connectivity to modality-specific maps or other neuroanatomy previously implicated in memory. In a meta-analysis of 21 prior brain stimulation trials for Alzheimer Disease, overlap between stimulation site and our convergent memory network correlated with trial effect size. In conclusion, we find Lesions, DBS, and TMS sites influencing verbal memory converge upon a single memory network, and this network may inform targeting in memory neuromodulation trials.

Matching journals

The top 3 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.