Leveraging Simultaneous EEG-fMRI for Functional Connectivity Biomarker Estimation in Schizophrenia: Insights from EEG Neurofeedback Training in Healthy Individuals

Current medications for schizophrenia (SCZ) remain ineffective, highlighting the growing need for targeted treatments addressing abnormal brain states. Functional connectivities (FCs) in resting-state functional magnetic resonance imaging (rs-fMRI) have successfully identified brain states associated with both diagnosis and symptoms. These FC-based biomarkers have been developed for several neuropsychiatric disorders, including SCZ. Furthermore, FC-based neurofeedback training (FCNef) utilizing these biomarkers has shown promise in normalizing specific brain states, leading to improvements in corresponding symptoms. EEG is a more cost-effective and less physically demanding method compared to fMRI, and EEG-based neurofeedback (EEG-NF) is gaining popularity due to its ease of use. Developing methods to predict brain states in FC-based biomarkers from EEG data is crucial for EEG-NF. In this study, aiming to perform EEG-NF for normalizing brain states in SCZ patients, we proposed a prediction method for fMRI biomarkers (fMRI-BM), named biomarker-based brain state prediction (BioBSP). Through three-day EEG-NF training in a single-blind manner (SCZ-NF: N = 11; sham-NF: N = 10), the SCZ-NF group successfully demonstrated the change in SCZ-BM than the sham-NF group with a significant improvement in working memory performance without any adverse effects. Our findings suggest that BioBSP may be a possible alternative tool and a novel approach for treating SCZ symptoms.