The startle response and prepulse inhibition in psychosis and violence: A combined electromyography and electroencephalography study

There is a pressing need for biomarkers of violent behavior risk in psychosis. Previous research indicates that electrophysiological measures of automatic defensive reactions may have potential. The purpose of this study was to investigate associations between violent behavior in individuals with and without psychosis and electromyography (EMG) and electroencephalography (EEG) responses to startling auditory stimuli. Electromyography and EEG were recorded during an auditory startle paradigm from healthy controls (HC, <i>n</i> = 211), individuals with psychosis and a history of violent behavior (violent-PSY, <i>n</i> = 18), individuals with psychosis without a history of violence (nonviolent-PSY, <i>n</i> = 32), and individuals with a history of violence without psychosis (violent non-PSY, <i>n</i> = 22). We estimated the auditory startle response (ASR) and prepulse inhibition (PPI) using EMG (ie, EMG_ASR and EMG_PPI) and the auditory-evoked potential (ie, AEP_ASR and AEP_PPI) of the EEG. There were no significant effects of group on the EMG_ASR (<i>P</i> = .10) or the 30-, 60-, and 120-ms prepulse + pulse EMG_PPI amplitudes (<i>P</i> = .11, <i>P</i> = .19, and <i>P</i> = .50, respectively). The N1 amplitude of the AEP_ASR was reduced in the violent-PSY group (<i>P</i> < .001) and the nonviolent-PSY group (<i>P</i> = .015) compared with HC. The P2 amplitude of the AEP_ASR was reduced in violent-PSY relative to nonviolent-PSY (<i>P</i> = .003), violent non-PSY (<i>P</i> = .016), and HC (<i>P</i> < .001). Together, these results show that EEG-based neural responses to startling auditory stimuli are promising biomarkers of violence risk in psychosis.