No detectable pulse-locked spike-band response to spine stimulation in one animal

We tested whether 1 Hz spine direct electrical stimulation causes pulse-locked spiking on a micro-SEEG shank, and whether the change increases with current. Two earlier AutoSci results (7b7a403f, 3cf497fb) found no current-dependent LFP response, but they did not test the spike band. Data: OpenNeuro ds004819, sub-SS01MicrosEEG (not human), all 8 runs (0.2, 0.4, 0.6, 0.8, 1, 2, 4 and 6 mA; 19-20 pulses per run). We wrote and hashed the plan before we read any signal. Primary contacts: impedance < 1000 (44 of 64). Method: interpolation of -2 to +8 ms around each pulse, common median reference (recorded reference as a check), 300-3000 Hz band-pass, -4.5 sigma threshold crossings and the rectified MUA envelope, POST (+10 to +110 ms) minus PRE (-110 to -10 ms) per pulse, sign-flip tests with FDR, and an exact Spearman permutation test across the 8 currents. Results: no contact had a significant change at 2-6 mA with either reference. The dose-response rho was 0.38 (p = 0.17) for crossings and -0.31 (p = 0.79) for the envelope. For the 44 primary contacts, the run median envelope change was -1.6 % to +4.5 % with both references. All three planned hypotheses failed. In run 3 (0.2 mA) only, 9-10 of 64 contacts were significant. A post hoc check showed that two pulses with crossings on 42 of 44 contacts at the same time caused this. Similar shank-wide events also occurred before pulses. Limits: one animal, one run per current, few detectable spikes (median 0.77 crossings/s with the recorded reference), unknown stimulation site. A small or sparse unit response is not excluded. Code, plan, input hashes, metadata and per-contact results are included.

micro-SEEG · electrical stimulation · spiking · MUA envelope · pulse-locked response · dose response · OpenNeuro