Multi-View Information Fusion for EEG Signal in Health Care Monitoring: A Systematic Review for Modeling New Strategies

Abstract

Electroencephalography (EEG) is a critical health monitoring tool that captures real-time brain activity. However, single-channel data usually used by traditional analysis methods severely limits the depth of insight. This systematic review addresses multi-view information fusion techniques used in EEG signal processing to enhance diagnostic accuracy and monitoring functionality. Searches were conducted in EEG analysis multi-view fusion strategies across multiple databases. It reviewed evaluations of various methodologies, applications, and challenges with these techniques. Results show that applications that use multi-view information fusion have increased accuracy in seizure detection, cognitive state assessment, and neurological disorder diagnosis. These techniques combine data from different views: spatial, temporal, and frequency to provide a 'holistic' picture of brain activity. Nevertheless, data heterogeneity, high computational complexity, and real-time implementation still exist. This is a promising advance in neuro-physiological monitoring. The development of models that will be able to adapt to individuals within a fixed workload capacity is at its initial stages. Personalized and adaptable models that emerge, coupled with existing emerging technologies like artificial intelligence and IoT, have countless opportunities to reinforce health monitoring practices. Further research is necessary to overcome current challenges towards the full realization of the benefit of multi-view information fusion in clinical applications.