Doubling current increased early CCEP amplitudes about twofold while preserving most of the map in two SEEG patients

Question: at the same SEEG contact pair, what does a change from 1 mA to 2 mA single pulses do to the early CCEP (10-50 ms)? Data: OpenNeuro ds005448 v1.0.0, patients STREEF06 and STREEF07, 21 directional sites (11 contact pairs) that received both currents, 1164 site-channel combinations at 10 mm or more. We used the archived single-trial table of 4b7a0539 (copy in 5ea568d6) and did not read raw signals. The single-trial amplitude is a projection on one template made from the pulses of both currents. A plan was written before the analysis. Results: the responder amplitude sum increased at all 11 contact pairs (median gain 2.0x, mean log2 gain 1.55, exact sign-flip p = 0.001). It was not different from linear doubling (p = 0.073). The gain at far responders (25 mm or more) was not larger than at near responders (-0.28 log2, 3 of 7 pairs, p = 0.64). The 1 mA and 2 mA maps had a median correlation of 0.67, and 0.85 after noise correction. At a fixed sham-based threshold, 2 mA gave 307 detected site-channels and 1 mA gave 196. The channels found only at 2 mA were a little farther away (median 24 vs 18 mm). This agrees with a proportional gain, not with a new spatial pattern. Limits: 2 patients, one center, 5 pulses per current, the 1 mA block always came first (order and current are confounded), and waveform shape changes are not measured. The package has the code (numpy only), the subset input with SHA256, the plan, per-site and per-channel tables, and a figure.

CCEP · SEEG · brain stimulation · current intensity · epilepsy · brain-computer interface