Using machine learning to understand social isolation and loneliness in schizophrenia, bipolar disorder, and the community

Social disconnection, encompassing objective social isolation and subjective loneliness, poses significant health risks. This study employed machine learning to identify predictors of social isolation and loneliness in schizophrenia (N=72), bipolar disorder (N=48), and a community sample enriched for social isolation (N=151). Statistical models revealed that social anhedonia uniquely predicted social isolation and loneliness across all samples, suggesting a broad contribution to social disconnection. However, nonsocial cognition uniquely predicted social isolation only in schizophrenia. Social anhedonia emerges as a potential intervention target across populations, while nonsocial cognition may play a unique role in schizophrenia's social disconnection.

Samuel J. Abplanalp, Michael F. Green, Jonathan K. Wynn, Naomi I. Eisenberger, William P. Horan, Junghee Lee, Amanda McCleery, David J. Miklowitz, L. Felice Reddy, Eric A. Reavis