ADAPTING METHODS TO REFINE THE EMPIRICAL STRUCTURE OF PERSONALITY AND PSYCHOPATHOLOGY

Figure 2. Diagram of the full Group Iterative Multiple Model Estimation (GIMME) solution for 7 variables assessed by daily diary over 100-days in a sample of individuals (N=35) diagnosed with borderline personality disorder. Note, this model is preliminary, and is intended to demonstrate proof of concept, not provide definitive statements about the group, sub-group, and individual dynamic structures of daily behavior in borderline personality disorder.

An enduring challenge to the empirical study and effective treatment of personality disorders is the lack of a scientifically supported “structure” of personality pathology.  That is to say, that although the Diagnostic and Statistical Manual of Mental Disorders (DSM) provides a summary of 10 personality disorder diagnoses, in fact these do a poor job of mapping the actual phenotypic diversity of this form of psychopathology.  An ongoing aim of the lab is to enlist quantitative methods to further clarify the empirical structure of personality pathology (and psychopathology more generally).  Past efforts in the lab have used a number of latent variable modeling techniques applied to psychiatric interview data and dispositional self-report measures (e.g., Wright et al., 2012; Wright, Hallquist, et al., 2013; Wright, Krueger, et al., 2013; Wright & Simms, 2014, 2015; Sharp et al., 2015).  Currently, we are working with colleagues to adapt Group Iterative Multiple Model Estimation (GIMME), developed for the study of connectivity networks in fMRI data, for the study of the structure of psychopathology.  GIMME fits individual structural equation models to multivariate time-series, and uncovers full-group, sub-group, and individual-specific paths among variables.  Developed for the study of individual differences in connectivity among regions of interest in the brain, we are studying the applicability of these models to behaviors sampled intensively and repeatedly in naturalistic settings (e.g., daily diaries).  Initial results (see Figure 2) are encouraging, and suggest the method holds unique promise for the simultaneous uncovering of shared (i.e., nomothetic) and person-specific (i.e., idiographic) psychological processes in samples of interest.