Fuzzy symbolic convergent cross mapping: A causal coupling measure for EEG signals in disorders of consciousness patients.

Accurate and timely diagnosis in disorders of consciousness (DOC) patients remains a core clinical challenge. Electroencephalography (EEG) shows strong potential for detecting physiological biomarkers of consciousness, and brain network analysis serves as an effective technique. Therefore, a robust approach to brain network construction is of great significance. The convergent cross mapping (CCM) is a powerful tool for capturing the coupling relationship between two signals. However, a major drawback of CCM is its sensitivity to noise. To address this problem, we proposed a symbolic method that combines fuzzy membership functions called fuzzy symbolic convergent cross mapping (FuzzSCCM). Through the simulation results, we verified its robustness to noise, sensitivity to coupling, and data length. Building on this coupling measure, we constructed EEG brain networks and validated the approach on real DOC EEG datasets. In patients with DOC, FuzzSCCM identified distinct network features between vegetative state/unresponsive wakefulness syndrome (VS/UWS) and minimally conscious state (MCS). Specifically, compared with the MCS group, the VS group showed greater asymmetry between the left hemisphere and the right hemisphere in the α band, and was relatively less active in the anterior in the θ band. Moreover, our results demonstrate spontaneous transitions between distinct brain network states, suggesting these dynamic reconfigurations may constitute a fundamental mechanism underlying consciousness modulation. These findings provide novel insights into the dynamic neural signatures of DOC, while establishing a potential diagnostic tool.