Mean bits per source word were higher inside detected voltage runs for fixed lossless 100 ms packets in 21 ds006519 SEEG records

We measured exact SEEG packet size before, inside, and after detected periodic voltage runs. We used OpenNeuro ds006519 version 1.0.0. The main set had 21 records from 14 people with agreement in the final detector landmark. Each independent 100 ms packet used a fixed four-mode lossless code and a 32-byte header. The equal-person mean rose from 18.053 to 18.868 bits per source word inside the runs. The paired increase was 0.815 bits, with a 95% person bootstrap interval of 0.447 to 1.211. The ratio of cohort means increased by 4.51%. An equal-person mean of 64.6% of run packets exceeded each record's before-period 95th percentile size. The increase remained after exclusion of channels on the stimulation shaft. All 45,683 scored packets decoded to exact source bits. The package has code, source hashes, metadata, packet examples, tables, and three figures. The result applies to the selected voltage runs and stored float32 words. Hardware pulse times, neural effects, and device power remain untested.

lossless compression · stereo-EEG · intracranial EEG · brain interfaces · packet size · electrical stimulation · data links