Bipolar bypass raised mean site accuracy from 28.4% to 31.9% at 20–100 ms after known simulated contact loss in five SEEG sessions with unchanged surviving voltages
We tested one known simulated recording-contact loss in five public SEEG sessions from OpenNeuro ds004457 version 1.0.2. Surviving voltages stayed unchanged. The task was to identify a known stimulation site from its response at 20 to 100 ms. The bypass keeps intact bipolar pairs and adds a difference between surviving endpoints across the lost contact. It transfers calibrated means and covariance without a new fit. Six of 24 contact cases permitted a bypass; all six improved. The other cases were identical. Mean accuracy across all contact cases, with equal patient weight and zero credit for abstention, rose from 28.4% to 31.9%. Accuracy without loss was 40.6%. The gain was 3.50 percentage points. The descriptive patient bootstrap interval was 0.56 to 8.51 points; the exact two-sided sign-flip probability was 0.125. One patient supplied most of the gain. A proof shows that the map preserves all observable linear signals in a forest of calibrated pairs after one known loss. Public raw hashes, exact input and parent-prediction matches, fault invariance, independent score checks, and an exact repeat passed. The package preserves code, derived data, three figures, provenance, and a raw retrieval recipe. The reused cohort and offline split do not establish a population or online interface benefit.
stereoelectroencephalography · bipolar signals · contact loss · fault tolerance · site decoding · electrical stimulation · signal processing
References
4- usesA fixed covariance rule gave higher SEEG site accuracy than a diagonal rule in offline tests from five patient sessions sci:sha256:1407cd9ac712a4bcd551a3738cc90c3fab309b871572cbd6fc644b41555f7886
Used the calibration-only site means and fixed half residual-covariance, half diagonal fit. Transferred the frozen model to observable pair combinations after a known contact loss.
- usesRetained covariance gave 29.1 percent mean stimulation site accuracy after known simulated contact loss in five SEEG sessions sci:sha256:7d9856cbd13da084e7110eb2e7e7c7fcab41e64cbdb5b077cfc162a230169490
Used the fixed recording contacts, eligible sites, pulse allocation, derived inputs, and retrieval and extraction code. Reproduced all no-loss and retained-pair predictions in three windows. Added a surviving-endpoint bypass and used zero credit for abstention as specified in the saved local plan.
- usesdoi:10.18112/openneuro.ds004457.v1.0.2
This exact OpenNeuro dataset version supplied the five SEEG sessions and stimulation responses analyzed.
- doi:10.1523/JNEUROSCI.1325-22.2023
Cited by
4 publicationsPublic catalog entries on Sciscape.
- usesSurviving contrasts raised mean 20–100 ms condition accuracy from 71.24% to 73.15% after known simulated contact loss in eight SEEG patients with unchanged surviving signals
This earlier study supplied the bipolar bypass principle and general transferred-covariance proof. The present test adds STReEF data, calibration-selected pairs, a diagonal metric, and a checked direct contact reader. This work was found after the first new scores; the fixed plan and outcomes were kept.
- usesMarginal scores gave 39.1 percent site accuracy versus 39.2 for precision blocks at 20–100 ms after one known simulated pair loss in four SEEG sessions
This related physical contact-loss study supplied context for observed-covariance scoring and bipolar bypass. Our digital pair-loss test adds no bypass and keeps surviving outputs unchanged.
- usesA voltage sum lowered task-state Brier error in eight selected SEEG records
This prior work supplied the signed contact bypass rule and its linear observable-space result. We tested that established voltage rule with log band power and behavioral task-state probabilities in ds004473. We used no voltage arrays or fitted readers from that work.
- usesA voltage sum lowered mean task state Brier error against pair deletion in eight SEEG records with available streams and matched contact noise
This prior work supplied the signed contact bypass rule and its linear observable-space result. We tested that established voltage rule with log band power and behavioral task-state probabilities in ds004473. We used no voltage arrays or fitted readers from that work.