Default mode network integrity across neuropsychiatric disorders and its relation to social dysfunction: A normative modelling approach

Structural and functional default mode network (DMN) alterations are common in neuropsychiatric disorders and may contribute transdiagnostically to social dysfunction. Normative modelling enables assessment of DMN alterations at the individual level. This study investigates whether individual deviations in cortical thickness, surface area, and between-network functional connectivity of the DMN differ between schizophrenia (SZ), major depressive disorder (MDD), Alzheimer's disease (AD), and healthy controls (HC), and whether these deviations transdiagnostically relate to social dysfunction. Social dysfunction was assessed using a composite score from the Social Functioning Scale and De Jong-Gierveld Loneliness scale. Structural MRI data was collected for 329 participants (SZ=86, MDD=44, AD=82, HC=117) and resting-state fMRI data for 317 participants. Individual deviation scores of DMN integrity were computed by adapting existing normative models of cortical thickness (N = 58,836), surface area (N = 43,524), and between-network functional connectivity (N = 21,515). Extreme deviations were quantified using a z-threshold of ±1.96. DMN deviation scores were not transdiagnostically associated with social dysfunction across the sample (ps>0.05). AD patients had more extreme negative deviations in DMN cortical thickness than all other groups (ps<0.0001; z = -4.14 to -6.34) and fewer extreme positive deviations in DMN surface area relative to SZ and HC (ps<0.05; z = 2.10 to 2.71). For between-network functional connectivity of the DMN, AD and SZ patients had more extreme negative deviations than MDD and HC (ps<0.05; z = -2.09 to -3.54). To conclude, normative modelling reveals differences in individual deviations of DMN integrity between neuropsychiatric groups, but these deviations do not transdiagnostically relate to social dysfunction.