Connections that characterize callousness: Affective features of psychopathy are associated with personalized patterns of resting-state network connectivity

Abstract

Background: Psychopathic traits are hypothesized to be associated with dysfunction across three resting-state networks: the default mode (DMN), salience (SN), and central executive (CEN). Past work has not considered heterogeneity in the neural networks of individuals who display psychopathic traits, which is likely critical in understanding the etiology of psychopathy and could underlie different symptom presentations. Thus, this study maps person-specific resting state networks and links connectivity patterns to features of psychopathy.

Methods: We examined resting-state functional connectivity among eight regions of interest in the DMN, SN, and CEN using a person-specific, sparse network mapping approach (Group Iterative Multiple Model Estimation) in a community sample of 22-year-old men from low-income, urban families (N = 123). Associations were examined between a dimensional measure of psychopathic traits and network density (i.e., number of connections within and between networks).

Results: There was significant heterogeneity in neural networks of participants, which were characterized by person-specific connections and no common connections across the sample. Psychopathic traits, particularly affective traits, were associated with connection density between the DMN and CEN, such that greater density was associated with elevated psychopathic traits.

Discussion: Findings emphasize that neural networks underlying psychopathy are highly individualized. However, individuals with high levels of psychopathic traits had increased density in connections between the DMN and CEN, networks that have been linked with self-referential thinking and executive functioning. Taken together, the results highlight the utility of person-specific approaches in modeling neural networks underlying psychopathic traits, which could ultimately inform personalized prevention and intervention strategies.

Keywords: Community sample; Default mode network; GIMME; Person-specific; Psychopathic traits; fMRI.

Fig. 1

Final GIMME networks for six illustrative participants. Solid lines depict contemporaneous connections and dashed lines depict lagged connections. Regions in red are within the salience network. Regions in blue are within the default mode network. Regions in green are within the central executive network. There was no group-level structure; thus, all lines depict individual-level connections (uniquely estimated for the participant) that also have associated β weights. Red lines depict connections with positive β weights and blue lines depict connections with negative β weights. 1A.) Participant with high levels of affective features (score = 23; Range in the sample 7 – 25); x² (73) = 106.06, RMSEA = 0.05, SRMR = 0.05, CFI = 0.96, NNFI = 0.95. 1B.) Participant with high levels of lifestyle features (score = 23; Range in the sample 7 – 28); x²(75) = 105.77, RMSEA = 0.05, SRMR = 0.05, CFI = 0.96, NNFI = 0.94. 1C.) Participant with first order model; x²(81) = 111.75, RMSEA = 0.05, SRMR = 0.06, CFI = 0.95, NNFI = 0.93. 1D.) Participant with first order model; x²(80) = 102.50, RMSEA = 0.04, SRMR = 0.06, CFI = 0.96, NNFI = 0.95. 1E.) Participant with second order model; x²(13) = 186.84, RMSEA = 0.04, SRMR = 0.05, CFI = 0.96, NNFI = 0.91. 1F.) Participant with third order model; x²(19) = 259.25, RMSEA = 0.04, SRMR = 0.04, CFI = 0.97, NNFI = 0.92. dlPFC = dorsolateral prefrontal cortex; vmPFC = ventromedial prefrontal cortex; PPC-R = right posterior parietal cortex; PPC-L = left posterior parietal cortex; INS-R = right insula; INS-L = left insula; PCC = posterior cingulate cortex; ACC = anterior cingulate cortex. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2

Psychopathic traits were associated with increased positive density between the default mode network and central executive network across all participants. 2A.) Schematic depiction of possible connections between the default mode network and central executive network regions of interest. Regions in red are within the salience network. Regions in blue are within the default mode network. Regions in green are within the central executive network. 2B.) Scatter plot represents bivariate correlations between total psychopathic traits and positive density between the DMN and CEN for each participant. 2C. Scatter plot represents bivariate correlations between affective traits of psychopathy and positive density between the DMN and CEN for each participant. DMN = default mode network. CEN = central executive network. dlPFC = dorsolateral prefrontal cortex; vmPFC = ventromedial prefrontal cortex; PPC-R = right posterior parietal cortex; PPC-L = left posterior parietal cortex; INS-R = right insula; INS-L = left insula; PCC = posterior cingulate cortex; ACC = anterior cingulate cortex. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)