Abnormal neural oscillations in clinical high risk for psychosis: a magnetoencephalography method study.

Background: Neural oscillations directly reflect the rhythmic changes of brain activities during the resting state or while performing specific tasks. Abnormal neural oscillations have been discovered in patients with schizophrenia. However, there is limited evidence available on abnormal spontaneous neural oscillations in clinical high risk for psychosis (CHR-P). The brain signals recorded by the magnetoencephalography (MEG) technique are not to be disrupted by the skull and scalp. Methods: In this study, we applied the MEG technique to record the resting-state neural activities in CHR-P. This was followed by a detailed MEG analysis method including three steps: (1) preprocessing, which was band-pass filtering based on the 0.5-60 Hz frequency range, removal of 50 Hz power frequency interference, and removal of electrocardiography (ECG) and electrooculography (EOG) artefacts by independent component analysis; (2) time-frequency analysis, a multitaper time-frequency transformation based on the Hanning window, and (3) source localisation, an exact low-resolution brain electromagnetic tomography. The method was verified by comparing a participant with CHR-P with a healthy control during the MEG recordings with an eyes-closed resting state. Results: Experimental results show that the neural oscillations in CHR-P were significantly abnormal in the theta frequency band (4-7 Hz) and the delta frequency band (1-3 Hz). Also, relevant brain regions were located in the left occipital lobe and left temporo-occipital junction for the theta band and in the right dorsolateral prefrontal lobe and near orbitofrontal gyrus for the delta band. Conclusions: Abnormal neural oscillations based on specific frequency bands and corresponding brain sources may become biomarkers for high-risk groups. Further work will validate these characteristics in CHR-P cohorts.