Uncovering Key Predictive Channels and Clinical Variables in the Gamma Band Auditory Steady-State
Holton, Kristina M, MS, University of DelawareHiggins, Amy, BS, McLean HospitalBrockmeier, Austin J., PhD, University of DelawareHall, Mei-Hua, PhD, McLean Hospital, Harvard Medical School
Poster not on display
Psychotic disorders are characterized by abnormalities in synchronization of neuronal responses. The 40 Hz gamma band deficit measured by auditory steady-state response (ASSR) electroencephalogram (EEG) is a robust observation in psychosis, and is associated with symptoms and functional deficits. However, most scalp EEG studies report using only one or a few electrodes to classify cases and controls, nor do they investigate if clinical variables are correlated with 40Hz ASSR signals in patients with early stage psychosis (ESP). There is also a lack of longitudinal 40Hz ASSR ESP studies.In this study, we establish ESP 40Hz ASSR deficits using all channels, classify ESP status using an ensemble of machine learning techniques with all channels, correlate EEG channels with modules of clinical/demographic/functioning variables to find associations, and assess whether baseline 40Hz ASSR deficits predict short-term functional outcome. Consistent with the literature, deficits in phase locking and power are already detectable in the ESP patients. Our machine learning models indicate that phase locking has a more predictive and parsimonious signature than power. Phase locking is also correlated with cognitive processes. Longitudinal functional outcome can be predicted from the baseline 40Hz ASSR signals from the FCz channel and other channels in both phase locking and evoked power.