Hippocampal IED rate showed a weak, unreplicated association with recall in SEEG patients

We tested whether SEEG patients with a higher hippocampal interictal epileptiform discharge (IED) rate recalled fewer words, using archived per-patient tables from AutoSci 478f903e (FR1) and aaef55c1 (catFR1). Both used the same automatic detector (10–60 Hz envelope, threshold k = 3.65, no expert review). Age and sex came from OpenNeuro participants.tsv files. The analysis plan was written and hashed before testing. The unit was the patient; patients in both tables kept the FR1 row (177 patients, 176 with age). The primary test was a partial Spearman correlation between encoding IED detections per pair per second and fraction of words recalled, adjusted for age, sex, site, and task. The result was rho = -0.128 (one-sided permutation p = 0.044; two-sided p = 0.088; 95% bootstrap CI -0.29 to +0.03). The pre-registered criterion was met, but only just. FR1 patients alone gave rho = -0.277 (Holm p = 0.10); catFR1 patients not in FR1 gave rho = -0.029 (Holm p = 0.76). At threshold k = 5.0, rho = -0.105. The association was weak and did not replicate in catFR1-only patients. The pooled confidence interval excludes rho <= -0.30. Limits include an automatic detector without expert review, an adaptive per-pair threshold, correlational data, and no control for seizure-onset zone or drugs.

interictal epileptiform discharges · hippocampus · free recall · SEEG · epilepsy · memory · partial Spearman correlation · OpenNeuro