Correlation between Neurophysiological Measures of Consciousness and BCI Performance in a Locked-in Patient

This exploratory study searched for systematic correlations between measures of consciousness in resting-state EEG and the performances reached in subsequently performed P300-based BCI tasks. We opted for measures of consciousness (LZC & power-law exponent) which are reflective for two different properties of brain dynamics, i.e., the complexity and criticality of neural signals. To unravel systematic relationships, we performed correlation analyses on a small, clinical dataset from a locked-in ALS-patient and on the publicly accessible dataset from Won et al. [1] containing data from 55 healthy participants. We detected opposing correlation patterns for the two samples. Whereas increased brain criticality and complexity seemed to be related to higher performance in the healthy participants, strong correlations between both our measures of consciousness and BCI performance in the ALSpatient indicate that decreased brain criticality and decreased brain complexity seemed to be advantageous to reach higher BCI performance in ALS-patients. We interpret this pattern regarding the known increases in functional connectivity in ALS-patients and put up for discussion if the increased functional connectivity with simultaneously decreasing structural connectivity in ALS-patients induces a brain dynamic of higher or even super-criticality, which is disadvantageous for information processing. In this case the brain criticality must be reduced to reenter a range of criticality in brain dynamics to facilitate the performance of tasks in need of a high amount of information processing - a pattern which we observed here.