Forgotten FR1 words that later became prior-list intrusions showed no significant encoding-power difference in 26 SEEG patients

We compared encoding power for forgotten free-recall words that later returned as prior-list intrusions (P) and forgotten words never recalled again (N) in 26 pure-SEEG patients performing FR1 in session 0. The analysis used archived 0–1.6 s band power from bipolar pairs and events.tsv sample/fs timing, with list and serial-position means removed and z values averaged over pairs. The 138 P words and 4,475 N words yielded a primary tilt (HFA minus theta) difference of +0.059 z (95% CI −0.029 to +0.145; sign-flip p = 0.21; Holm p = 0.43; 17/26 subjects positive). HFA was +0.042 z (Holm p = 0.26) and theta was −0.017 z (Holm p = 0.62). The primary test was not significant and does not show that forgotten words that later intrude have stronger encoding signals. It also does not reject a graded-strength account: the same pipeline did not detect the recalled-versus-forgotten effect in these subjects (tilt +0.038 z, p = 0.13), and the P-minus-N confidence interval includes both zero and that reference effect. All point estimates had the direction of a strength account. Prior-list intrusions were rare (2–15 per subject), and a more informative test needs more words from more subjects or sessions and more sensitive pair selection. The analysis includes 32 archived subjects, of whom 26 met inclusion criteria (at least 2 P words and 10 N words). The last list was excluded because its words cannot become prior-list intrusions. We assume events.tsv sample/fs corresponds to EDF time, as in the library standard and archived feature source. A P word is one said again in a later recall period; the analysis does not use intrusion time or list lag. The archived event files came from the OpenNeuro ds004789 S3 mirror described as serving the latest version; exact dataset release applicability is uncertain. The package preserves source URLs, input hashes, the pre-analysis plan and hash, code, event files, and per-subject results.

free recall · prior-list intrusion · encoding · memory · SEEG · high-frequency activity · theta · FR1