An exact check gave more 7 Hz phase certificates after SEEG gaps in 250 records with bounded missing voltage

We derived an exact phase check for missing real voltage in a fixed linear reader. We tested a causal 7 Hz reader in 250 public hippocampal SEEG records from seven people. Each missing centered sample was assumed to stay within five times early voltage RMS. The center used zero completion. The exact check certified 22.88% of tested outputs. A circular bound certified 19.34%. Each person had equal weight. The gain was 3.54 percentage points. Its descriptive 95% person bootstrap interval was 2.72 to 4.52 points. Both checks used a strict 45-degree phase limit. All 39 channel records certified at the gap end lost their certificate at a later query. Each output thus needs a new check. In 2,000 artificial gaps, 1,947 hidden segments met the assumed voltage limit. None of their 13,825 certified queries reached 45-degree error. The limit failed in 53 controls. Available voltage in the first second after real gaps exceeded that limit in 376 of 500 channel records. These adjacent samples cannot test hidden gap voltage. The real-gap claim remains conditional. Fresh downloads matched all 250 EDF hashes. A full repeat matched twelve numerical output files. The package supplies source data, provenance, proof, reusable reader code, scores, four figures, and repeat instructions. This method limits the effect of missing filter inputs. It does not establish neural phase, artifact recovery, decoder accuracy, or stimulation benefit.

SEEG · intracranial EEG · phase certification · missing samples · brain-computer interface · signal processing · causal filters · uncertainty bounds · stimulation