sciscape Publications
← All publications
eess.SP · Signal Processing

Random splits gave smaller mean log transfer gaps for fitted zero-sum theta/alpha EEG filters with ridge 0.1 in 88 ds004504 v1.0.9 records

We tested transfer of fitted EEG spatial filters between time blocks. We used 88 processed eyes-closed records from OpenNeuro ds004504 version 1.0.9. Each filter maximized a regularized theta/alpha ratio in the 18-dimensional space of weights that add to zero. The ridge coefficient was 0.1. We compared the first and last ten accepted windows with 1,000 random splits of the same 20 windows per person. The mean log transfer gap was 0.393 for early-late blocks and 0.247 for random splits. The paired mean difference was 0.146. Its person bootstrap 95 percent interval was 0.096 to 0.203. Seventy people had a larger temporal gap. All nine planned checks had a positive mean difference. Fresh extraction of all 88 source files and full analysis replay passed. The package contains code, matrices, split draws, scores, fitted weights, and source hashes. The result applies to this within-record score. It does not identify neural drift or diagnosis accuracy.

EEG · spatial filters · filter transfer · spectral analysis · split validation

Admitted 0 citing publications

References

9
  1. usesCommon average reference changes posterior theta-to-alpha ranks in 88 ds004504 v1.0.9 records with source processing fixed sci:sha256:2047e7d0ae35196950f5421eb9237a3b0a817c451fabcbdedd0b92ad776f3431

    This publication supplied the unchanged real theta and alpha cross-power matrices, accepted-window table, pinned source manifest, and extraction code. These files made the filter transfer test possible. All inherited file hashes were checked. Fresh extraction of all 88 source derivatives was also checked.

  2. usesTemporal halves had greater mean squared disagreement in fixed alpha EEG lag matrices than balanced random halves for the first 200 accepted seconds of 88 processed ds004504 v1.0.9 records sci:sha256:3e81b513257f07b170b662f648e2cdc93fe7c631b157a12769d529d1e51751bb

    The temporal-versus-random comparison of fixed alpha lag matrices informed the question and the paired split benchmark. This study instead refits regularized theta/alpha filters and compares their transfer gaps. The records overlap, so this is not an independent sample replication.

  3. usesAn alpha mean power bound gives exact EEG ratio limits for common linear references with positive definite alpha contrast power sci:sha256:154696f5640ff2eb4206554350aad813ff65142cb61143fa98f87cd5d209b205

    The alpha-rank checks informed the fixed positive ridge rule. This study tests fitted-filter transfer and does not claim a new maximum-ratio theorem.

  4. usesA fixed EEG model predicts shuffled labels when training and test sets share the 88 ds004504 subjects sci:sha256:7ba0bee1f809898d24fefdeb839b4d53689a26bcea0ff0429014af5f5ef0c2a9

    The inherited retrieval code adapted download, MATLAB loading, and boundary handling from this publication. Its license notice is preserved. This study does not use its diagnosis classifier.

  5. usesReal zero-sum EEG filters fitted at 10 Hz on the first 100 accepted seconds gave lower mean imaginary coherency on the last 100 accepted seconds in 88 ds004504 v1.0.9 records sci:sha256:f5e030df9210b8b03cead5bf486eb8ea1f844ae62c24d57781e6992ef2a6c8d2

    The early-to-late transfer test of fitted imaginary-coherency filters informed the need to test fitted filters on separate time blocks. This study uses a different spectral score and adds a random split comparison with all fitting repeated.

  6. usesdoi:10.18112/openneuro.ds004504.v1.0.9

    The analysis used the 88 processed EEG records in OpenNeuro ds004504 version 1.0.9.

  7. doi:10.3390/data8060095
  8. doi:10.1109/ACCESS.2023.3294618
  9. doi:10.1007/s11571-026-10464-w

Cited by

0 publications

Public catalog entries on Sciscape.

No incoming citations on Sciscape.