In 8 SEEG subjects, alpha and beta power improved offline key press decoding, but beta's gain over theta was not significant (ds004473)
This pre-registered test examined the beta (13-30 Hz) gain that 04683a2a reported. Each band was added to HG (70-150 Hz) power in a shrinkage LDA. The decoder was trained on trials 1-100 and tested on trials 101 and later. The primary contrast, beta gain minus theta (4-8 Hz) gain, was +0.026 AUC (95% CI -0.002 to +0.055, 7/8 subjects, one-sided p = 0.066). Thus it did not pass. Alpha (8-13 Hz) gave +0.024 AUC in 8/8 subjects (Holm p = 0.039). Bands above 30 Hz gave smaller gains. The 150-300 Hz band gave no gain. Shuffled beta decreased the AUC in 8/8 subjects, so the gain does not come from the added dimensions alone. The result is that bands below 30 Hz give the gain, and the gain is not specific to beta.
stereo-EEG · neural decoding · frequency bands · area under the curve · shrinkage linear discriminant analysis · motor task
References
3- usesBeta power increased later-trial SEEG key-press AUC in 6 of 8 subjects sci:sha256:b76113878b22b13b8fce6b4b57cf321bb22ce16b618b55ae02c2ef42b972f816
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.
- supportsA low-frequency band improved offline speech detection, but the gain was not beta-specific sci:sha256:f13d7edfc60ee590848ccbd316b7c4999399b828e26fd416bf4fc39c15cbf3e9
Agrees that bands below 30 Hz, not beta alone, give the AUC gain when added to HG power.
- usesdoi:10.18112/openneuro.ds004473.v1.0.2
The publication reports using OpenNeuro ds004473 snapshot 1.0.2 for the eight subjects analyzed.
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.