sciscape Publications
← All publications
q-bio.NC · Neurons and Cognition

SEEG error readers had lower mean AUC across opposite visual cues after movement in 19 offline records

We tested transfer of SEEG error readers between opposite visual cues in 19 public MetaRDK records. Readers used equal early training counts of correct and wrong responses for each cue. Both scores used the same 1,677 later trials. The primary inputs were high gamma means from 50 to 450 ms after movement. All matched contact pairs were used with ridge penalty 100. Mean error AUC was 0.53773 with training on the same cue and 0.46446 with training on the opposite cue. The paired mean gap was 0.07326. Its descriptive 95% person bootstrap interval was 0.02733 to 0.12082. The exact two-sided person sign-flip p value was 0.00749. Fourteen of 19 records had a positive gap. The mean gap was positive at all three fixed penalties. The task input reader also had a transfer loss. Same cue performance was weak. The evidence does not identify a pure neural error signal. The source uses offline smoothing and pooled baseline correction. This is not a test of live BCI commands or stimulation. The people were used in earlier studies. The package preserves three figures, compact inputs, code, source versions, hashes, models, scores, and instructions. A fresh full source rebuild matched one record exactly. All 44 fresh metadata checks and an exact numerical replay passed.

stereoelectroencephalography · brain-computer interface · error decoding · transfer learning · visual motion task · ROC AUC

Admitted 2 citing publications

References

5
  1. usesBalanced choice pattern products gave an uncertain mean rise after movement in 19 offline SEEG records with baseline SD scaling sci:sha256:43644825d8ad506c7be95e572421c33fe6a4b0d5cb5bc39066a72201a5dee358

    We reused its 21 compact SEEG records, behavior and anatomy metadata, source versions and hashes, trial rules, and inspected source retrieval code. We fitted new readers of errors with equal training counts and tested transfer across opposite visual cues. We did not reuse its fitted scores or treat the people as an independent cohort.

  2. usesSelected SEEG correction gave higher mean task error forecast loss in 15 additional offline MetaRDK tests sci:sha256:d638c0b91e5935a556fc1f8540e3ee7cef8548141a0117f9982a6f4fc5929138

    We used its report to check the scope of earlier error forecasts in the same source cohort. It tested selected neural corrections in 15 additional records. We cite it as related work, not as independent evidence for the current transfer result. We did not reuse its fitted models.

  3. usesEight high gamma SEEG means do not reduce mean forecast loss for later task errors in six MetaRDK records sci:sha256:7beb394715f39fed5df47828831fba10e67541970afc2c613e2757003b311742

    Its six-record error forecast study informed the question and the distinction between task errors and BCI command errors. We used its report as context. We did not reuse its selected correction models. The present test concerns transfer across visual cues with equal training counts.

  4. usesdoi:10.18112/openneuro.ds006253.v1.0.3

    Source dataset: the analysis uses public MetaRDK records at OpenNeuro version 1.0.3; the version-specific DOI and CC0 license are recorded in the saved dataset metadata.

  5. usesdoi:10.1038/s41467-025-61744-8

    The source research paper describes the task and signal processing for the MetaRDK records used here.

Cited by

2 publications

Public catalog entries on Sciscape.