EEG spectral power and negative symptoms in at-risk individuals predict transition to psychosis.

INTRODUCTION: EEG power in the delta, theta and beta1 bands has been shown to be positively correlated with negative symptoms in first episode psychotic patients. The present study investigates this correlation in an "at risk mental state for psychosis" (ARMS) with the aim to improve prediction of transition to psychosis.
METHODS: Thirteen ARMS patients with later transition to psychosis (ARMS-T) and fifteen without (follow-up period of at least 4 years) (ARMS-NT) were investigated using spectral resting EEG data (of 8 electrodes over the fronto-central scalp area placed according to the 10-20 system) and summary score of the Scale for the Assessment of Negative Symptoms (SANS). Linear regressions were used to evaluate the correlation of SANS and EEG power in seven bands (delta, theta, alpha1, alpha2, beta1, beta2, beta3) in both ARMS groups and logistic regressions were used to predict transition to psychosis. Potentially confounding factors were controlled.
RESULTS: ARMS-T and ARMS-NT showed differential correlations of EEG power and SANS in delta, theta, and beta1 bands (p<.05): ARMS-T showed positive and ARMS-NT negative correlations. Logistic regressions showed that neither SANS score nor EEG spectral power alone predicted transition to psychosis. However, SANS score in combination with power in the delta, theta, beta1, and beta2 bands, respectively, predicted transition significantly (p<.03).
CONCLUSIONS: ARMS-T and ARMS-NT show differential correlations of SANS summary score and EEG power in delta, theta, and beta bands. Prediction of transition to psychosis is possible using combined information from a negative symptom scale and EEG spectral data.