Changing the preceding trial changed fixation theta power under a centered FIR filter in 14 SEEG patients

We tested a filter effect in OpenNeuro ds004752 version 1.0.1. The test used 14 patients, 96 contacts with source hippocampal labels, and 593 eligible current trials. Each current trial stayed fixed while its preceding trial changed. A centered FIR filter had 257 coefficients and cutoffs at 4 and 8 Hz. The mean absolute fixation power change with equal patient weight was 0.738 dB. Its descriptive 95% patient bootstrap interval was 0.694 to 0.784 dB. The interval assumes independent, exchangeable patients and fixes the sampled records and donor draws. After a native endpoint screen, the mean was 0.715 dB across 545 targets. Three other filters also changed fixation power. A finite filter proof protected samples from 1 to 7 seconds for the main filter. Separate reflection padding gave identical results after trial order changed. All 14 full source hashes passed. Every compact signal sample rebuilt exactly. The complete analysis repeated exactly. The package preserves signals, source pins, code, powers, donor indices, figures, and reproduction steps. This result describes a computation on released trial arrays. It does not establish a neural history effect.

stereoelectroencephalography · finite impulse response filter · trial boundaries · theta power · signal processing · OpenNeuro ds004752