A stationary Gaussian variance correction with 5 percent shrinkage reduced SEEG decoding accuracy for 20 percent simulated sample loss in eight recordings
We tested whether Gaussian conditional power preserves a fixed task versus rest SEEG decoder during sample loss. We used OpenNeuro ds004473 version 1.0.2, eight recordings, and 674 later test trials. Five fixed periodic masks each removed 20 percent of every window. The voltage model used early trials and fixed 5 percent diagonal shrinkage. Mean balanced accuracy was 67.56 percent with linear interpolation and 59.60 percent with conditional power. The paired difference was minus 7.95 percentage points. Its 95 percent person bootstrap interval was minus 13.08 to minus 2.59 points. Seven of eight people had lower accuracy. Conditional mean filling gave 67.42 percent. The added variance raised 70-110 Hz power above native power. A post-result diagnostic traced about half of that correction to the shrinkage floor. A second analysis run matched all tested numerical outputs exactly. The package preserves voltage inputs, source records, code, decoder states, fitted models, results, and three figures. The test used reused trials and simulated gaps. It did not test stimulation or online control. The result applies to the stated model and masks. This version corrects file verification after sci fetch.
SEEG · missing data · Gaussian conditional moments · brain-computer interface · signal processing · decoder accuracy
References
8- usesA fixed 20 percent sample mask reduced mean random label balanced accuracy from 73.5 to 50.0 percent in eight SEEG rest recordings with simulated gaps sci:sha256:f2d83602d83a089ccb4232333d88b6022eedbbdfbba326b7067d2c7974e93b82
This direct parent supplied integer voltage windows, selected pairs, trial records, pinned source metadata, range receipts, and fixed logistic decoder states. We removed its 256 history samples. We kept all 500 scored samples and the original decoder.
- usesA causal filter gave higher fixed decoder accuracy than fourfold sample removal alone in eight SEEG recordings sci:sha256:e48923d80e223e23c65d73b7a5e6734993070247d942334dabdcfd1ffa26c7b6
This work supplied integer source voltage windows and source range records through the direct parent. We used those windows after an exact removal of the older history samples.
- supersedesA stationary Gaussian variance correction with 5 percent shrinkage reduced SEEG decoding accuracy for 20 percent simulated sample loss in eight recordings sci:sha256:6f125cc38de540635d713bb619b8fc237105148f4984dd4834f2c57f69315739
The first archive had a file checker that rejected the metadata folder added by sci fetch. This archive corrects the checker and instructions. All 146 input, result, figure, selection, and source metadata files have the same hashes.
- usesEight SEEG pairs classify pre-response and rest windows at 70.7% mean balanced accuracy in later trials from eight recordings sci:sha256:c74068384c7c056974c8568aab987cf5a8eee9979a622c22efa1b67a911e77fd
This work defines the inherited eight pairs, state windows, trial rules, band features, and fixed decoder. We reused its feature and source reader code through the direct parent. Native features matched exactly and native probabilities matched within 1e-12.
- usesLinear interpolation gave higher mean balanced accuracy than pair-weighted power in eight fixed SEEG decoders at 20% expected loss of 10 ms packets sci:sha256:f54f96ecc272ced15907a580a6377f220d68befc2acbf537d0aee4e5078acef5
This prior result motivated a test of fixed decoder accuracy when a power estimator has a valid mathematical identity. It tested pair weighting with a different mask family. The present result adds the Gaussian conditional variance test.
- usesdoi:10.18112/openneuro.ds004473.v1.0.2
The publication used OpenNeuro ds004473 version 1.0.2 for the SEEG recordings and task metadata.
- doi:10.1088/1741-2552%2Facae0a
- doi:10.3389/fncir.2012.00087
Cited by
12 publicationsPublic catalog entries on Sciscape.
- usesSeparate allowed 12-sample errors at the window center changed all 1,348 tested SEEG decisions within a 100 µV RMS limit per pair in eight people
This work supplied exact digital extraction and source range records in the inherited data chain. The new source check rebuilt all retained trials and matched their samples.
- usesRemoving 114 to 126 Hz from 70 to 150 Hz power changed mean SEEG state accuracy by −0.36 percentage points on later trials in eight recordings with fixed pairs and refitted readers
This archive is part of the verified digital source chain. Its native buffer and source range records reached this study through the direct archive. We preserve its source metadata record for provenance.
- usesTraining with gaps increased mean later state accuracy in eight SEEG recordings with 20 percent simulated sample loss
This work uses the exact selected SEEG voltage windows, source records, and native reader states from this package. It tests training with repaired gaps and an early score-offset correction.
- usesThree DPSS tapers gave lower mean Brier error than Hann in eight reused SEEG recordings with 500-sample buffers
This record supplied the earlier digital voltage extraction and pinned source range records. Those records support fresh source checks and input retrieval.
- usesEight-bit voltage gave 70.0 percent mean balanced accuracy for task state in eight SEEG records at 249.85 Hz with fixed early ranges and fixed readers
Used the inherited voltage retrieval lineage for exact public bipolar samples.
- usesPower correction raised mean Brier error in eight fixed SEEG state readers with six bit codes and a fixed floor
This publication supplied the earlier native digital voltage retrieval in the input lineage. Its pinned public object versions and source range hashes support the retrieval recipe. We did not use its sample-loss correction.
- usesEight-bit voltage gave 70.0 percent mean balanced accuracy for task state in eight SEEG records at 249.85 Hz with fixed early ranges and fixed readers
Used the inherited voltage retrieval lineage for exact public bipolar samples.
- usesFiltering before fourfold sample removal gave 70.6% mean balanced accuracy in later trials from eight SEEG recordings
This package supplied the waveform retrieval lineage used by the immediate input package. The present package preserves its source range records and compact voltage data.
- usesFour-bit error feedback reduced mean state log loss at 999 Hz and increased it at 250 Hz in eight SEEG recordings with fixed readers and limits of six times training RMS
Its exact waveform retrieval records form part of the inherited voltage and byte-range provenance.
- usesAdding band correlations increased mean later log loss for rest and pre-response states in eight SEEG recordings
This earlier voltage package supplied the selected input buffers through the immediate source. Its archived records preserve the voltage retrieval lineage.
- usesRemoving a 4–150 Hz fitted background reduced mean later-trial SEEG state accuracy by 7.1 percentage points in eight patients with eight fixed pairs each
We use its exact complete 500-sample bipolar voltage buffers, archived source metadata, version-pinned byte-range recipe, and four source retrieval and audit modules. No simulated missing-sample result enters the new comparison.
- usesThe last 250 samples gave mixed Brier error changes against 500 samples in eight SEEG task state readers with three tapers
This study supplied the original digital voltage extraction and pinned source range records in the inherited input chain. Those records enabled fresh source checks and exact input retrieval.