Simulated pulse schedules matched with and without 25-sample gaps at a 301-sample interval in seven SEEG rest records with fixed readers and delays

We tested whether sensing gaps change later phase-based output times. We used four SEEG pairs from each of seven people in OpenNeuro ds006065 version 1.0.0. Each fixed reader used 250 past samples. Output delay and the main gap were each 25 samples. Each simulated output caused its own later input gap. It caused no signal change outside that gap. A proof gave the sufficient interval H+D+G+1, where H is history length, D is delay, and G is gap length. A crossing uses two consecutive forecasts. At 301 samples, about 602 ms, all 28 gap schedules matched no-gap schedules at that same interval. All 840 tested gap pair-conditions covered by the bound also matched. At the earlier 250-sample interval, hold gaps changed 44.07% of exact no-gap event times. The descriptive person interval was 39.09–49.12%. Total event count was unchanged, and mean phase error changed by 0.16 degrees. The main bound reduced pulse rate by 14.93%. The proof assumes fixed finite input support, exact delays, identical initial states, and no effect outside each gap. This exploratory study reuses rest data. It does not test a neural response or live stimulation benefit. The package preserves selected voltage, fixed models, every event schedule, proof, code, provenance, four figures, and reproduction instructions.

sensing gaps · finite input support · SEEG · phase forecasting · pulse scheduling · feedback replay