Abstract
Traumatic brain injury (TBI) affects millions of people worldwide, many of whom are affected with post-TBI mood disorders or behavioral changes, including aggression or social withdrawal. Diminished functionality can persist for decades after TBI and delay rehabilitation and resumption of employment. It has been established that there is a relationship between these mental disorders and brain injury. However, the etiology and causal relationships behind these conditions are poorly understood. Rodent models provide a helpful tool for researching mood disorders and social impairment due to their natural tendencies to form social hierarchies. Here, we present a rat model of mental complications after TBI using a suite of behavioral tests to examine the causal relationships between changes in social behavior, including aggressive, hierarchical, depressive, and anxious behavior. For this purpose, we used multivariate analysis to identify causal relationships between the above post-TBI psychiatric sequelae. We performed statistical analysis using principal component analysis, discriminant analysis, and correlation analysis, and built a model to predict dominant-submissive behavior based on the behavioral tests. This model displayed a predictive accuracy of 93.3% for determining dominant-submissive behavior in experimental groups. Machine learning algorithms determined that in rats, aggression is not a principal prognostic factor for dominant-submissive behavior. Alternatively, dominant-submissive behavior is determined solely by the rats' depressive-anxious state and exploratory activity. We expect the causal approach used in this study will guide future studies into mood conditions and behavioral changes following TBI.