Yong Liu

Background: Consciousness assessment in the ICU is frequently confounded by endotracheal intubation, sedation, and neuromuscular blockade, rendering conventional behavioral scales (GCS, FOUR, CRS-R) unreliable. Advanced imaging (fMRI, PET) is impractical for immobilized patients, while standard 19-channel EEG is cumbersome, time-intensive, and unsuited for routine bedside deployment. Commercially available tools such as the Bispectral Index (BIS) reflect anesthetic depth rather than the quality of consciousness. We propose a simplified 4-channel EEG protocol combined with Integrated Information Theory (IIT)-based Φ quantification as a practical, theory-driven, and real-time consciousness monitor for ICU settings. The protocol was validated using single-session repetitive median nerve stimulation (RMNS) as an arousal challenge. Methods: Patients with disorders of consciousness admitted to the ICU were enrolled in a randomized, sham-controlled trial. Each patient underwent one session (RMNS or sham). The EEG protocol comprised a 10-minute baseline recording, 20 minutes of RMNS (20 mA, 40 Hz, 300 μs, left median nerve) or Sham stimulation, and a 10-minute post-stimulus recovery recording. Primary analyses focused on 4-channel EEG microstructural features: spectral power (δ/θ/α/β/γ bands), weighted phase lag index (wPLI) functional connectivity, and EEG microstate parameters (duration, coverage, occurrence, transition probability). Φ values (PyPhi 3.0) were computed from the 4-channel subset and compared against the full 19-channel montage (Bland-Altman analysis, intraclass correlation coefficient [ICC]). Secondary comparisons benchmarked Φ against BIS, spectral edge frequency, and alpha-delta ratio. Results: Data collection is ongoing. Preliminary analyses will characterize RMNS-induced EEG microstructural changes, compare the agreement between 4-channel and 19-channel Φ values, and evaluate whether Φ outperforms conventional EEG indices in detecting post-stimulus arousal. Full results including spectral, connectivity, microstate, and Φ analyses will be presented at the conference. Conclusions: This study provides the first systematic evaluation of a 4-channel IIT-based consciousness monitoring framework in ICU patients. The findings will clarify whether simplified EEG montages can capture meaningful Φ changes and whether single-session RMNS alters EEG microstructure in DoC. If validated, the protocol could serve as a practical, theory-driven bedside consciousness meter for immobilized ICU patients.

刘永是南方医科大学深圳医院重症监护医学部主任兼主任。他获得同济医科大学医学博士学位和博士学位。拥有超过25年的ICU经验，他的临床和研究兴趣集中在神经重症监护，特别是意识评估和意识障碍中的神经调节。他目前领导一项基于脑电图的意识监测随机对照试验，利用整合信息理论对重症监护患者进行意识监测。他在《重症护理》和《转化医学年鉴》等期刊上发表了10多篇论文。