In this project, we used data from the Shanghai At Risk for Psychosis (SHARP) program. As part of this research program, a large sample of adolescents and young adults (mostly between 15 and 25 years of age) with early signs of impending psychosis was recruited at the Shanghai Mental Health Center. Clinical symptoms and cognitive functioning, as well as brain MRI data were collected and processed by labs at Harvard and MIT. For the current project, we used MRI data to reconstruct the brain network for each individual participant and analyzed these networks using graph theoretical analysis.
We mapped functional brain networks based on functional MRI data. We were specifically interested in the way that the functional connectome is organized into distinct communities. These communities or modules consist of sets of brain areas that are highly interconnected but have relatively poor connectivity with regions outside of their community. We found that youth with warning signs of impending psychosis, particularly those that go on to develop a psychotic episode, show abnormalities in the organization of these functional communities. Specifically, a number of brain regions that have been linked to (early-course) schizophrenia were found to change community affiliation. For example, superior temporal gyrus, which comprises the auditory cortex and has been implicated in auditory verbal hallucinations, changed affiliation from the sensorimotor to the limbic module. These findings are consistent with hypotheses that abnormal interactions between sensory and limbic systems may be involved in the development of psychotic symptoms. These findings were published in Molecular Psychiatry, a top-ranking journal in biological psychiatry and were picked up by several news outlets including SciTech Daily (
https://scitechdaily.com/neuroscientists-identify-brain-activity-pattern-linked-to-schizophrenia/(öffnet in neuem Fenster)). In addition, our findings were presented at international scientific meetings including the Schizophrenia International Research Society (SIRS) 2018 conference, International Early Psychosis Association (IEPA) 2018 conference, and American College of Neuropsychopharmacology (ACNP) 2018 annual meeting.
In a follow-up study, we performed a prediction analysis to assess if functional brain connectivity and modular connectome organization are predictive of psychosis development. Using machine learning and leave-one-out cross-validation, we found that a prediction model that combined connectivity metrics with clinical and cognitive predictors yielded superior outcome prediction. This model outperformed connectivity-only and clinical-only prediction models. These findings were published in Neuroimage: Clinical and presented at the Society for Biological Psychiatry (SOBP) 2019 conference.
In addition, we had access to an independent dataset of children with a familial risk for serious mental illness because they had a parent or sibling diagnosed with schizophrenia or an affective psychotic disorder (i.e. bipolar disorder or psychotic depression). Using this dataset, we replicated our earlier findings that brain network disturbances predate illness onset. Moreover, we found that disturbances in functional brain networks that process self-relevant information were associated with behavioral precursors of future psychopathology, including internalizing problems, withdrawn behavior, and thought problems. These findings suggest that early interventions that target these specific networks may not only reduce early internalizing problems, but may also help avert the progression from early behavioral problems to mental illness in at-risk youth. This insight is thus of key interest to the development of neurobiologically-informed early interventions that seek to slow or prevent the development of serious mental illness. These findings were published in two separate papers in Schizophrenia Research (2020) and Frontiers in Psychiatry (2021), and will be presented at the upcoming SIRS 2021 conference.
