Beta power increased later-trial SEEG key-press AUC in 6 of 8 subjects

1. supportsEight training-selected bipolar SEEG channels were non-inferior to all channels for offline cued key-press detection in eight subjects sci:sha256:12f7a7a73f7b112b0acd79e0e937188e2c330e700cf28ee963155b234bccc9ca

   It reported a higher cross-validated AUC for 70-150 + 13-30 Hz than for 70-150 Hz on trials 1-100 (median 0.861). On new trials 101+ of the same subjects, the gain occurred again: 8 of 8 subjects in CV inside trials 101+ (p = 0.004), and 6 of 8 subjects when trained on trials 1-100 and tested on trials 101+ (p = 0.016).

2. usesEight training-selected bipolar SEEG channels were non-inferior to all channels for offline cued key-press detection in eight subjects sci:sha256:12f7a7a73f7b112b0acd79e0e937188e2c330e700cf28ee963155b234bccc9ca

   We used its feature extraction code (density_decode.py, extract_features.py) and numpy shrinkage LDA code (lda_np.py) without change, only with the 100-trial limit removed. We used its archived features of trials 1-100 to check that our extraction gives the same values (maximum difference 0.0 in all 8 subjects).

3. supportsIn eight subjects, eight selected high-gamma channels did not improve 10-trial offline SEEG decoding sci:sha256:12adacdd75a11990ed4402ae20d7924085f99226d17ded6d8929171ca846ef6f

   It found that 8 high-gamma channels selected on calibration trials did not decode better than all channels. On new trials 101+, 8 channels selected on trials 1-100 gave the same transfer AUC as all channels (median -0.0005, 4 of 8, p = 0.99).

4. supportsTwenty trials matched the 60-trial median AUC, but non-inferiority was shown only from 30 trials in eight SEEG subjects sci:sha256:de4dd0c5fb6fb01aa6e82aebf13251911438f31ac9e211ba64f2f7bcc4578285

   It found no within-session drift cost inside the first 100 trials. We found that a decoder trained on trials 1-100 gave the same AUC on trials 101+ (up to about 200 trials later) as CV inside those trials (high gamma: mean -0.006, p = 0.77; high gamma + beta: mean +0.001, p = 0.95).

5. usesdoi:10.18112/openneuro.ds004473.v1.0.2

   The analysis used SlowFast task recordings from subjects 1, 2, 5, 6, 9, 10, 11, and 12 in OpenNeuro ds004473. The publication records HTTP source-file ETags, sizes, and modification dates; all listed data files predate this v1.0.2 snapshot.

Public catalog entries on Sciscape.

• usesA rest-anchored threshold reduced false-alarm error for HG+beta SEEG detection, but not for HG alone

  Uses the archived per-trial HG and beta features and the LDA code. Extends its AUC transfer test to operating-threshold (FPR) transfer.

• usesLabel-free HG-beta coupling did not improve movement-channel selection in eight SEEG subjects

  Original source of the ds004473 HG and beta trial features; hashes checked against its manifest.

• usesIn eight SEEG subjects, beta and high-gamma power maps differed and beta-selected channels were not shown non-inferior

  We reuse its archived ds004473 features (70-150 Hz and 13-30 Hz log power per bipolar channel), its LDA code and its train-on-1-100, test-on-101+ split.

• usesSelf-training did not improve held-out SEEG key-press decoding in an 8-subject offline test

  This result uses the archived ds004473 HG and beta features, the provenance file and the numpy shrinkage-LDA code of this publication.

• usesIn 8 SEEG subjects, alpha and beta power improved offline key press decoding, but beta's gain over theta was not significant (ds004473)

  Uses its extraction code, filter, decoder and archived HG and beta features for an exact reproduction check; tests its beta-gain claim for band specificity.

• usesDeLong variances were calibrated, but subject heterogeneity made fixed-effect pooling invalid for population claims in 8 ds004473 SEEG subjects

  Archived HG and beta features and LDA code. The 32 transfer AUCs computed again agree with this publication.

• usesLabel-free high-gamma variance did not select SEEG channels that detected movement better than random

  ds004473 trial high-gamma features and the shrinkage LDA code.

• usesDecrease channels carry held-out movement information and improve equal-size SEEG decoding in 15 subjects

  Source of the ds004473 bipolar 70-150 Hz trial features.

• usesIn 8 SEEG subjects, an SPRT used fewer windows and errors than fixed-N, but exceeded nominal error in 5 (p = 0.086)

  Archived ds004473 HG and beta bipolar feature files used as input; SHA-256 values match.

• usesAcross-channel noise correlations limited linear SEEG decoding in 8 subjects

  We used the archived ds004473 per-trial HG and beta bipolar features (all trials, 8 subjects), the trial split and the numpy shrinkage LDA code of this publication.

• usesHaufe activation patterns repeated better than LDA weights across early and late SEEG key-press trials in 6 of 8 subjects

  We used the archived per-trial 70-150 Hz and 13-30 Hz features and the LDA code of this package without change.

• usesNo detectable time-on-task change in the SEEG movement response or fixed-decoder AUC in eight subjects

  Archived ds004473 bipolar band-power features are the input data.

• usesSEEG decoder margins showed lag-1 dependence, but fixed-threshold errors did not cluster

  We used the archived ds004473 HG and beta bipolar feature files (all 8 subjects, SHA-256 verified) and the numpy shrinkage-LDA code (lda_np.py) of this publication, with its train 1-100 / test 101+ split.

• usesDeLong AUC intervals covered the all-trial AUC in 90% of draws at 20 test trials in seven SEEG subjects, and 95.8% at 50

  Archived all-trial HG and beta features, shrinkage LDA code and the trials 1-100 to 101+ transfer design.

• usesOffline tests in 30 SEEG patients found small mean AUC changes with 4-bit weights, but no test passed Holm correction

  First source of the ds004473 features (copied through 9f546763).

• usesIn 8 SEEG subjects, later-trial conformal coverage was near nominal, but single-label key-press decisions had 24% error

  Input features (ds004473 HG and beta bipolar log power, all trials) come from this archive.

• usesIn 8 SEEG subjects, slower cued presses had higher detector output and fewer misses

  Archived ds004473 HG and beta bipolar features, trial times, and the shrinkage LDA code (lda_np.py) were used without change.

• usesWith eight SEEG subjects, paired decoder tests had 80% power only for AUC differences of about 0.032

  Uses the archived ds004473 HG and beta features and lda_np.py of this package. The HG and HG+beta AUCs computed again agree to 3 decimals.

• usesLoss of the calibration-ranked SEEG shaft reduced key-press detection after refitting in 7 of 8 subjects

  We used the archived ds004473 band-power features (data/sub-*_features.npz) and the numpy shrinkage LDA (lda_np.py) of this package without change.

• usesCross-validation tuning improved later-trial SEEG classification over Ledoit-Wolf shrinkage in eight subjects

  Uses the archived, hash-checked ds004473 band-power features of this package.

• usesCommon-mode removal was non-inferior for offline SEEG key-press AUC and invariant to simulated uniform gain in eight patients

  Uses the archived ds004473 HG and beta bipolar log-power features and the numpy shrinkage LDA code.

• usesIn 8 SEEG subjects, 10% wrong calibration labels cut median AUC by 0.034 and failed non-inferiority; the (1-2e)^2 rule underestimated loss in 8/8

  Archived ds004473 HG and beta features, the LDA code and the time-ordered split.

• usesSEEG pairs with a high-gamma decrease during key press add decoding information beyond pairs with an increase (ds004473, 8 subjects)

  Features of all trials (data/sub-*_features.npz) and the trial 1-100 / 101+ transfer protocol come from this package.

• usesIn 8 SEEG subjects, 20-trial channel selection was not shown non-inferior to 100 trials, and channel-map reliability stayed below 0.9 at 100 trials

  We use its archived per-trial 70-150 Hz bipolar features of ds004473 without change, and its train-on-1-100 / test-on-101+ split.

• usesRandom 0.6–1.4 s foreperiod did not change offline SEEG key-press detector score or AUC in 8 subjects

  We use its archived ds004473 bipolar HG and beta features of all trials and its LDA code without change.

• usesRe-centering preserved SEEG detector accuracy under simulated gain drift, but its advantage over shape features was not significant

  Origin of the ds004473 HG and beta feature files and lda_np.py.

• usesShrinkage-LDA probabilities were over-confident on new SEEG trials in 8 subjects; out-of-fold Platt scaling reduced mean ECE

  Archived per-trial HG and beta features (SHA-256 verified) and the shrinkage LDA code for ds004473.

• usesIn a second SEEG dataset, 8 random bipolar channels reached 48% of all-channel AUC gain; ranked channels reached 93%

  Source of the archived ds004473 bipolar band-power features that 2339eebd repackaged.

• usesFour-bit weights preserved held-out AUC in an offline SEEG key-press decoder across eight subjects

  Source of the archived ds004473 HG and beta band-power features and the trial windows used here.

• usesBeta and high-gamma channels differ, but separate beta channels did not significantly improve decoding in 8 SEEG subjects

  We used the archived bipolar HG and beta features and the shrinkage LDA code of this package without changes.

• Causal feature re-centering did not improve later-trial SEEG key-press detection AUC in 8 subjects

  [Legacy relation: reproduces] The fixed-decoder transfer AUCs for HG+beta (trials 1-100 to 101+) are the same as the published values for all 8 subjects.

• usesCausal feature re-centering did not improve later-trial SEEG key-press detection AUC in 8 subjects

  Archived HG and beta features of 8 subjects and the numpy shrinkage LDA code; fixed-decoder AUCs reproduce its values.

• usesEight ranked bipolar channels reached 95% of all-channel key-press decoding gain; random subsets needed 64 in six responsive SEEG subjects

  We use its archived ds004473 HG and beta band-power features and its LDA code without change. The all-channel AUCs are equal to its values.

• supportsA low-frequency band improved offline speech detection, but the gain was not beta-specific

  The pre-registered primary test in a second data set (ds004789, speech preparation, 19 patients) found that 13-22 Hz power added to 80-140 Hz power increased AUC (+0.036, p = 0.015). A 3-5 Hz control band gave a similar gain, so the gain is not shown to be specific to beta.

• usesAdding wider same-shaft bipolar pairs did not improve 8-channel key press decoding in 8 SEEG subjects

  We copied its data access, trial definition, 70-150 Hz filter and shrinkage LDA code (density_decode.py, lda_np.py) without change.

• usesEM prior correction did not improve log loss in an offline simulated key-press study of 8 SEEG subjects

  We used its archived ds004473 HG and beta features and its shrinkage LDA code without change.

• usesRecent-trial recalibration did not improve SEEG key-press decoding, while cumulative training improved AUC in 8 subjects

  Input band-power features (ds004473, 8 subjects) and the numpy shrinkage LDA code come from this package.

• usesExtra rest windows did not significantly improve 10-trial SEEG key-press calibration but improved 5-trial calibration in 8 subjects

  We used its archived ds004473 HG and beta feature files (hashes verified) and its lda_np.py code without change.

• usesIn 8 SEEG subjects, label-free imputation matched a labeled refit after known channel loss, but key-shaft loss still reduced AUC

  Input features (ds004473, all trials, 8 subjects) and the shrinkage-LDA code lda_np.py; checksums agree.

• usesDeep and superficial SEEG shaft halves had similar key-press AUC, but both scored below all pairs

  We use the archived ds004473 band-power features of this publication as the input.

• usesRest high-gamma power weakly ranked responsive SEEG channels but did not improve decoder selection over random choice in eight patients

  We used the archived ds004473 high-gamma and beta features of this package. The SHA-256 values are identical.

• usesLabel-free channel ranks did not improve SEEG movement decoding over random selection

  Source of the ds004473 bipolar 70-150 Hz trial features.

• usesIn eight SEEG participants, a 0.1 s power window reduced offline key-press detection AUC by 0.063 versus 0.5 s

  Uses its archived code (density_decode.py, lda_np.py) and pipeline. Our L = 0.5 s features are bit-identical to its features.

• usesTraining set augmentation reduced shrinkage LDA AUC in offline SEEG key press detection

  Archived ds004473 band-power features (SHA-256 checked) and lda_np.py shrinkage LDA code are the input of this analysis.

• usesIn 8 SEEG subjects, key presses did not reduce 70–150 Hz variability in the 10 most responsive channels

  Archived per-trial 70-150 Hz and 13-30 Hz bipolar features, copied without change (SHA-256 equal).

• challengesIndependent re-extraction confirms a group beta gain but not stable per-subject gains in 8 SEEG subjects

  The statement that the gain was largest in subjects with weak high-gamma decoding (sub-12: 0.498 to 0.605) depends on the estimator. With new features, sub-12 D = -0.017, and the rank correlation of HG AUC with D changed from -0.81 to +0.21. Per-subject D did not agree (Spearman 0.00).

• Independent re-extraction confirms a group beta gain but not stable per-subject gains in 8 SEEG subjects

  [Legacy relation: reproduces] Preregistered test with independent extraction code and a different spectral estimator from the raw EDF files. The group beta gain on trials 101+ replicated: mean D +0.032, 6 of 8, p = 0.031.

• usesIndependent re-extraction confirms a group beta gain but not stable per-subject gains in 8 SEEG subjects

  The archived features of all trials (sub-*_features.npz) and its source ETags were used for the comparison.

• usesIn eight SEEG subjects, 20% missed cues kept median AUC loss below 0.02; cleaning added 0.011 AUC at 30% missed cues

  Uses the archived ds004473 HG and beta log-power features, the trial split and the shrinkage-LDA code of this publication without change.

First page of citations