Hippocampal IEDs Were Not Linked to More Errors in 14 SEEG Patients

We tested whether hippocampal interictal epileptiform discharges (IEDs) during a verbal Sternberg working-memory task were associated with errors or slower responses. A prespecified automated detector was applied to sessions 1–2 from 14 patients in OpenNeuro ds004752 (1,358 complete trials; 121 errors). IEDs occurred in 55% of trials. During maintenance, trials with an IED had fewer errors than trials without one (7.07% vs 10.17%; Mantel–Haenszel odds ratio 0.579; permutation p = 0.0285), opposite to the hypothesized direction; after Holm correction, p = 0.057. Retrieval IEDs were not associated with a significant accuracy difference (odds ratio 0.779; p = 0.468). Fixation and encoding IEDs were not associated with significant accuracy differences. Excluding trials marked by the dataset Artifact flag gave a similar maintenance result (odds ratio 0.530; p = 0.0165). Correct-trial response time did not differ significantly with maintenance or retrieval IEDs; encoding IEDs were associated with slightly shorter response times in an uncorrected secondary test (dz = −0.118; p = 0.0259). These results do not support the claim that hippocampal IEDs in the tested windows impaired task performance in this dataset. Interpretation is limited by few errors, an automatically detected event set without expert visual validation, high detection rates in four patients, a fixed 1-second retrieval window, atlas contact labels, and use of sessions 1–2 only.

hippocampus · interictal epileptiform discharges · working memory · Sternberg task · intracranial EEG · SEEG · epilepsy · permutation test