EEG biomarkers for assessment, prognosis, and monitoring of natural upper limb recovery after stroke: a systematic review.

The review did not combine results into a meta-analysis because the studies used different EEG measures, tasks, frequency bands, brain regions and statistical methods. Many included studies were small, with sample sizes as low as 10 participants, which leaves their correlations open to chance and overfitting. Most studies were rated fair rather than good, and common problems included no sample size justification, no repeated exposure assessment, no blinding of outcome assessors and no adjustment for confounders. The evidence is observational and correlational, so it cannot show that EEG changes cause, drive or reliably predict upper limb recovery. The literature focuses heavily on impairment measures, especially the Fugl-Meyer Assessment Upper Extremity scale, and much less on activity and participation outcomes such as daily arm use. Interventional studies were excluded, so the review describes natural recovery patterns and cannot say how EEG biomarkers behave during specific therapies. The review included only English-language, peer-reviewed adult studies with at least 10 participants, so smaller, non-English or paediatric evidence was not assessed.

The supplied text lists funding from several Chinese provincial and national science foundations and innovation programs. It does not include an author conflict-of-interest declaration beyond those funding sources.

Do not read these EEG associations as evidence that a biomarker can direct upper limb therapy. The review combined only observational studies, most of fair quality, and it did not test whether EEG changes cause or reliably predict recovery.