Uncovering the common factors of chemical exposure and behavior: Evaluating behavioral effects across a testing battery using factor analysis.

Although specific environmental chemical exposures, including flame retardants, are known risk factors for neurodevelopmental disorders (NDDs), direct experimental evidence linking specific chemicals to NDDs is limited. Studies focusing on the mechanisms by which the social processing systems are vulnerable to chemical exposure are underrepresented in the literature, even though social impairments are defining characteristics of many NDDs. We have repeatedly demonstrated that exposure to Firemaster 550 (FM 550), a prevalent flame retardant mixture used in foam-based furniture and infant products, can adversely impact a variety of behavioral endpoints. Our recent work in prairie voles (Microtus ochrogaster), a prosocial animal model, demonstrated that perinatal exposure to FM 550 sex specifically impacts socioemotional behavior. Here, we utilized a factor analysis approach on a battery of behavioral data from our prior study to extract underlying factors that potentially explain patterns within the FM 550 behavior data. This approach identified which aspects of the behavioral battery are most robust and informative, an outcome critical for future study designs. Pearson's correlation identified behavioral endpoints associated with distance and stranger interactions that were highly correlated across 5 behavioral tests. Using these behavioral endpoints, exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) extracted 2 factors that could explain the data: Activity (distance traveled endpoints) and Sociability (time spent with a novel conspecific). Exposure to FM 550 significantly decreased Activity and decreased Sociability. This factor analysis approach to behavioral data offers the advantages of modeling numerous measured variables and simplifying the data set by presenting the data in terms of common, overarching factors in terms of behavioral function.