Training energy normalization gave a positive mean five-pulse error gap for 20 to 500 ms responses under uniform partitions of eight SEEG recordings
We tested a five-pulse estimate of later waveform error in public SEEG data from eight people. The study used 317 stimulation conditions and 1,183 condition and recording pair cases. For 20 to 500 ms responses, the mean gap divided by training energy was 0.265 for the first five pulses and 0.272 for uniform pulse partitions. Their difference was -0.0066, with a descriptive 95 percent person bootstrap interval from -0.0296 to 0.0137. A finite-data proof gives zero mean raw gap across uniform partitions. The random denominator can give a positive mean ratio gap. The 20 to 100 ms window had an additional ordering gap. This secondary result does not establish a neural cause. The package provides all-pulse matrices, code, source hashes, proofs, and figures. This is a new analysis of a reused cohort, not a clinical BCI test.
stereoelectroencephalography · electrical stimulation · calibration · waveform error · energy normalization · finite-data identity · uniform partitions · brain-computer interface
References
4- usesPair selection within each fold reduced observed mean normalized error underestimation at 20–500 ms in eight SEEG sessions sci:sha256:cc8dd031a38bd1409ae47b5eded349970f98f6e00038879f4af3c5e34f5c80dc
Its calibration study reported a normalized fixed-pair gap and stated that energy normalization has no general unbiasedness guarantee. These results motivated our explicit partition and reporting-scale diagnostic. The estimators and selected cases differ.
- usesA fixed noise correction reduced later 20 to 500 ms waveform error at SEEG pairs selected from metadata in eight recordings sci:sha256:3ab281a187b6b380f718cf4fbf3224ffdfad27c27435b3231e9fda7c80a5d3be
We used the metadata selection, exclusions, raw-file manifest, and retrieval code. We used its five-pulse matrices and later scalar products to check our new all-pulse extraction.
- usesdoi:10.18112/openneuro.ds004696.v1.0.1
The analysis used OpenNeuro ds004696 version 1.0.1; the source text identifies this exact dataset version.
- doi:10.1523/JNEUROSCI.2201-22.2023
Cited by
5 publicationsPublic catalog entries on Sciscape.
- usesCalibration range limits reduced 20 to 500 ms donor SEEG error at penalty one after five fit pulses in eight people, but a fixed template had lower error
This supplied all-pulse signal matrices, native examples, source event maps, metadata selection code, and complete MEF retrieval records. We checked every reused matrix against a fresh raw reconstruction and used it to test calibration range limits.
- usesCalibration range limits reduced 20 to 500 ms donor SEEG error at penalty one after five fit pulses in eight people, but a fixed template had lower error
This supplied all-pulse signal matrices, native examples, source event maps, metadata selection code, and complete MEF retrieval records. We checked every reused matrix against a fresh raw reconstruction and used it to test calibration range limits.
- usesA noise rule gave a mean error estimate below the mean test error under random pulse orders in eight SEEG recordings
Uses its 4,732 all-pulse Gram matrices, source event records, and metadata readout rules from OpenNeuro ds004696 version 1.0.1. Uses a fixed full-recording energy scale to test the error estimate at an adaptive stop. Adds new partial raw checks.
- usesFixed shared pulse weights preserved linear readout changes and gave no clear 20–500 ms SEEG forecast gain in eight sessions
We used its all-pulse matrices, conditions, example voltages, source metadata, hashes, and reconstruction code. We checked inherited hashes. We rebuilt one raw bipolar example pair. The new test measures shared-weight forecast error and consistency under fixed linear readout changes.
- usesAdding pulse interval increased mean SEEG size forecast error by 0.13% at 20 to 500 ms in eight tests that excluded one person from shared slope fitting
This study uses the all-pulse Gram matrices, condition records, native metadata, selection code, and pinned raw file manifests from this publication. It adds actual preceding-pulse interval features, forecasts with slopes fit in seven other people, and a bootstrap that refits those slopes. Fresh downloads and decoding check all inherited matrices.