Characterizing Transdiagnostic Processes Underlying the Drive to Socially Engage: A Multimeasurement Factor Analytic...

Altered drive to socially engage is a transdiagnostic feature across multiple psychopathologies. Yet, lack of clarity regarding specific processes that constitute social drive, along with insufficient measurement methods, has hindered understanding in this area. This study ascertained the feasibility of approximating difficulties within specific fine-grained social drive processes as proposed by 2 theoretical frameworks: "orienting," "wanting," "pursuing," "liking," "learning," and "reticence" within a reward processing framework and "orienting," "seeking and maintaining," and "liking" within a social motivation framework.

Secondary data from the Healthy Brain Network were analyzed, comprising 2,663 participants (age range 5-21 years, mean [SD] age = 10.22 [3.38] year; 67% male) with diagnosed neurodevelopmental and/or neuropsychiatric conditions. Factor analyses used items from 7 caregiver-report questionnaires. Replicability of the factor solution was examined across different sex (male, female), age (≤12, >12), and diagnosis (autism, internalizing, other neurodevelopmental disorders) subsamples. Generalized additive modeling explored the association of identified social drive factors with IQ, age, sex, and diagnosis.

A 3-factor solution comprising social reticence, seeking, and maintaining factors provided the best fit to the data. The factor structure and item-factor loadings remained consistent across tested subsamples, with good-to-excellent internal consistency reliability. Factors showed differential associations with IQ, age, and sex. The autism subsample exhibited the most pronounced challenges, followed by internalizing and other neurodevelopmental disorders.

A multi-instrument approach offers potential to capture difficulties within specific social drive processes beyond the capabilities of standard single-instrument scoring. However, the limited coverage of social orienting, wanting, liking, and learning processes within existing questionnaires underscores a need for future measurement development.