Detecting EEG index of arousal in sleeping patients with unresponsive wakefulness syndrome

First published online September 24, 2026. EEG spectral slope tracked sleep-related arousal changes, while complexity measures were more useful for separating unresponsive wakefulness syndrome from minimally conscious state in two task datasets. These research measures are not a stand-alone clinical diagnosis.

Li M, Pan J, Han J, Bao X, Wu H, Wang F, Yu Y, Li A, Liu Y, Chen D, Yang X, He Y, Qin P, Li Y.
iScience · Publication date: 2026-09-24
Added to site: 2026-10-17
DOI: 10.1016/j.isci.2026.117448

First published online September 24, 2026; collected in the October 16, 2026 issue feed. Historical addition.

Modalities

EEG, Sleep EEG, BCI, ERP

Clinical focus

Disorders of consciousness, Arousal, Covert awareness

Population

Patients diagnosed with unresponsive wakefulness syndrome or minimally conscious state, plus healthy controls

Sample

Discovery sleep dataset: 52 participants, including 16 UWS, 18 MCS and 18 healthy controls; independent healthy sleep dataset: 37; separate BCI datasets: 78 and 10 patients with disorders of consciousness

Intervention / method

The study analyzed overnight sleep EEG and repurposed P300, SSVEP and auditory-response BCI recordings. Spectral slope was evaluated as an arousal feature, while spectral entropy and Lempel-Ziv complexity were evaluated as awareness-related features.

Comparator

Sleep-stage changes and diagnostic-group comparisons across UWS, MCS and healthy participants, with two independent BCI task datasets

Clinician summary

Spectral slope tracked arousal-related sleep-stage variation but did not distinguish UWS from MCS. Awareness-related spectral entropy and Lempel-Ziv complexity produced 82.05% accuracy with AUC 0.84 in one BCI dataset and 90% accuracy with AUC 0.92 in a second small dataset; the arousal feature was near chance for diagnostic separation.

How to interpret the findings

Small discovery cohort, heterogeneous and partly repurposed datasets, incompletely controlled medication effects, limited channels in the discovery sleep recording, no separation of oscillatory and aperiodic components in the broadband slope method, and very small second BCI dataset. Clinical UWS diagnosis does not exclude covert awareness, and the classifiers are not validated stand-alone diagnostic tests.

Equipment and disclosures

Discovery sleep recordings used a Compumedics Grael Profusion PSG system; independent recordings used SynAmps2/Neuroscan systems where specified. These are source-reported devices, with no BrainMaster attribution.

Supported by Chinese national, Guangdong provincial and National Natural Science Foundation of China grants, including 2022ZD0208900 and 32371098. The authors declared no competing interests. No indexed correction, retraction or concern notice appeared in the retained PubMed or Crossref records; this is a bounded check. AI-assisted review.

Related evidence

Compare findings from the same clinical area, including mixed results.

• Targeting central thalamus with transcranial ultrasound reduces arousal in healthy volunteers · Positive with limitations
• Comparative Efficacy of Noninvasive Neuromodulation in Patients with Disorders of Consciousness: A Systematic Review and Network Meta-Analysis of Randomized Controlled Trials · Positive with limitations
• Clinically unresponsive patients with acute brain injury activate a brain-computer interface · Positive with limitations
• HD-tACS at 40 Hz Improves Consciousness and Brain Connectivity in Prolonged Disorders of Consciousness: A Randomized Sham-Controlled Clinical Trial. · Positive with limitations