One of the fundamental challenges that exist in the ?eld of psychiatry and cognitive science is to better under- stand how multiple brain areas work together to produce emergent behavior and cognition. The brain's ability to precisely coordinate these areas in the face of a constantly changing environment is important for maintaining normal psychosocial and cognitive functioning. Complex mental disorders, such as schizophrenia, affect multiple interacting dynamical systems (including sensory processing and attention) subserving cognition. The degree to which each system is affected may be related to the varying degrees of cognitive impairment observed in these disorders. Characterizing the dynamical states of these systems non-invasively via electroencephalography (EEG) can provide an avenue for identifying clinically useful biomarkers. This research proposal seeks to develop a novel computational method based on dynamical systems theory to detect nonlinear systems architectures hidden in EEG signals and to relate these dynamical signatures to psychosocial/cognitive functioning and the underlying brain network dynamics. Speci?cally, the three aims of the proposal are: (1) to develop a new clustering method to group brain signals based on nonlinear dynamical states, (2) to apply the method to a large EEG dataset to identify nonlinear dynamical features associated with psychosocial functioning, and (3) to relate nonlinear dynamical states of EEG signals to the underlying network interactions. A large EEG recording dataset from patients diagnosed with schizophrenia will be used to probe dynamical states corresponding to neurocognitive de?cits, and a simultaneous EEG-fMRI dataset from patients with absence epilepsy will be used to establish a relationship between EEG dynamical features and functional networks critical for cognition and attention. Together, these aims have potential to identify unique EEG dynam- ical states that are indicative of cognitive de?cits and network dysfunction present in neuropsychiatric disorders. In summary, investigating EEG nonlinear system features can reveal disorder-speci?c biomarkers that can help identify individuals at risk, make early diagnosis, and monitor intervention/treatment responses.
Identifying individuals in the early disease course and delivering appropriate and prophylactic interventions can immensely bene?t the ?eld of clinical psychiatry. However, there are currently not many reliable diagnostic and prognostic biomarkers available for mental illnesses. The proposed work, which aims to develop a novel com- putational method to characterize unique brain signal features associated with cognitive de?cits, could provide a next-generation class of noninvasive biomarkers that can help identify individuals at increased risk for a mental illness and predict the course of the disease.
