Earlier repeats did not predict fixed-position MTL HFA, but a 26-patient test was insensitive

We tested whether the within-list decrease in medial temporal lobe (MTL) high-frequency activity (HFA, 70–150 Hz) during repeated-word free recall followed the number of earlier new words or serial position. The analysis used archived per-word features from 26 SEEG patients in OpenNeuro ds005411 v1.0.1 (session 0): 6,146 first presentations, with HFA measured 0.10–0.75 s after word onset. The preregistered model controlled for serial position with separate position indicators. Replacing one earlier new word with one earlier repeat at a fixed position had an estimated effect of c = −0.010 z (95% CI −0.043 to +0.021; 12/26 positive; sign-flip p = 0.54), so the predicted novelty-load effect was not supported. The full novelty account predicted c = +0.0043 z from the linear position slope (−0.0043 z per position, p = 0.028); the confidence interval was about seven times wider than that prediction, leaving the test unable to distinguish the accounts. A secondary additive linear model without position indicators suggested a novelty effect (b_new = −0.015, p = 0.001; b_old = +0.007, p = 0.037; difference p = 0.003). Post hoc checks found that this difference was not significant after adding log(1 + position) (difference +0.013, p = 0.12) or when restricting to positions ≥4 (−0.006, p = 0.36), consistent with a curved serial-position effect explaining the additive result. A post hoc power calculation estimated that about 3,069 subjects would be needed for 80% power because N_old varied little at a fixed position (median SD 0.55 words). Exploratory analyses found no effect of a repeat at the previous position on HFA (p = 0.22) and no effect of N_old on recall of once-presented words (p = 0.84). The results are limited to one session, mostly entorhinal and parahippocampal pairs, one time window, one feature pipeline, and first presentations.

free recall · serial position · high-frequency activity · medial temporal lobe · intracranial EEG · novelty · preregistration