Low-density EEG-based Functional Connectivity Discriminates Minimally Conscious State plus from minus

Area of Science:

• Neuroscience
• Clinical Neurology
• Computational Psychiatry

Background:

• Patients in a Minimally Conscious State (MCS) present a spectrum of consciousness, with some exhibiting high-level behavioral responses (MCS+) and others not (MCS-).
• Accurate differential diagnosis between MCS+ and MCS- is critical for tailoring rehabilitation and pharmacological treatments, as these groups differ in prognosis and underlying physiology.
• Current diagnostic methods rely on behavioral assessments, which carry a significant risk of misdiagnosis, highlighting the need for objective physiological markers.

Purpose of the Study:

• To investigate if low-density electroencephalography (EEG)-based graph metrics can distinguish between MCS+ and MCS- patients.
• To determine if brain network characteristics derived from EEG correlate with behavioral responsiveness in post-comatose patients.
• To assess the physiological significance of standard behavioral assessments for quantifying responsiveness.

Main Methods:

• A prospective observational study involving 57 MCS patients (30 MCS-, 28 males) was conducted.
• Resting-state, closed-eyes EEG recordings (30 minutes) were obtained at intensive rehabilitation admission, alongside consciousness diagnosis.
• EEG data underwent preprocessing, and graph metrics were calculated using various connectivity measures across multiple connection densities and frequency bands (α, θ, δ). Machine learning models were employed for cross-validated classification of MCS+/- outcomes.

Main Results:

• The MCS- group exhibited lower brain activity integration in the alpha (α) band compared to the MCS+ group.
• In the delta (δ) band, the MCS- group showed higher clustering (weighted clustering coefficient) than the MCS+ group.
• An Elastic-Net regularized logistic regression model achieved the highest discrimination accuracy of 79% (sensitivity 74%, specificity 85%) for differentiating MCS+/- patients.

Conclusions:

• Low-density EEG, a routine clinical tool, can differentiate between MCS+ and MCS- patients by analyzing brain network properties.
• Graph-theoretical features derived from EEG successfully discriminate between these neurophysiologically distinct conditions.
• These findings suggest that EEG-based graph metrics can serve as a valuable adjunct to clinical diagnosis, supporting more accurate patient stratification.