Cross-trial estimates of 50–250 ms response energy were negative in 9 of 18 bipolar SEEG files from ds006065

We tested finite trial variation in late hippocampal pulse responses. We recovered 1,212 pulse blocks from 18 public SEEG files and seven patients in ds006065 version 1.0.0. For bipolar responses from 50 to 250 ms with no added filter, cross-trial energy was negative in 9 files, although average-waveform energy was positive. The mean correction fraction was 175.8% across patients; its patient bootstrap interval was 84.2% to 265.1%. Under independent trials with a common mean, this correction removes finite trial variance bias. A 500 microvolt rule left 7 negative estimates. The result depends on the reference and quality rule. Negative estimates do not prove absent neural responses. The package preserves selected raw data, full source hash checks, source versions, code, all results, and offline reproduction instructions.

stereoelectroencephalography · hippocampus · evoked response · trial variability · cross-trial energy · intracranial EEG