The abstract is read at the publisher; the summary is JClub's.
Gamma power, theta power, and spectral entropy are the only EEG features that simultaneously discriminate ALS from controls, show within-subject reliability in both ALS cohorts, and keep consistent effects across three cohorts.
- Why it matters: ALS EEG biomarkers have not reached clinical deployment, partly due to limited cross-cohort validation; a pre-registered framework is needed to screen candidates that can translate across sites and protocols.
- What they did: The team pre-registered a three-criterion screening approach and tested 18 EEG features across EEGET-ALS (6 ALS, 170 controls; home recordings over 17 days), UCL/Penn State (8 ALS; clinic recordings over 1–2 months), and a 60-participant test-retest reference, using Cohen’s d > 0.5, ICC(3,1) > 0.6, and cross-cohort effect direction checks.
- The result: Cross-cohort permutation testing found no feature meeting the full reliability intersection in any of 1000 permutations (Monte Carlo p = 0.001); bootstrap resampling showed the selected set’s composition is sample-sensitive, with inclusion probabilities at ICC > 0.6 of 57% (gamma), 27% (spectral entropy), and 19% (theta).
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.