Age-associated alterations in thalamocortical structural connectivity in youths with a psychosis-spectrum disorder

Abstract

Psychotic symptoms typically emerge in adolescence. Age-associated thalamocortical connectivity differences in psychosis remain unclear. We analyzed diffusion-weighted imaging data from 1254 participants 8-23 years old (typically developing (TD):N = 626, psychosis-spectrum (PS): N = 329, other psychopathology (OP): N = 299) from the Philadelphia Neurodevelopmental Cohort. We modeled thalamocortical tracts using deterministic fiber tractography, extracted Q-Space Diffeomorphic Reconstruction (QSDR) and diffusion tensor imaging (DTI) measures, and then used generalized additive models to determine group and age-associated thalamocortical connectivity differences. Compared to other groups, PS exhibited thalamocortical reductions in QSDR global fractional anisotropy (GFA, p-values range = 3.0 × 10^-6-0.05) and DTI fractional anisotropy (FA, p-values range = 4.2 × 10^-4-0.03). Compared to TD, PS exhibited shallower thalamus-prefrontal age-associated increases in GFA and FA during mid-childhood, but steeper age-associated increases during adolescence. TD and OP exhibited decreases in thalamus-frontal mean and radial diffusivities during adolescence; PS did not. Altered developmental trajectories of thalamocortical connectivity may contribute to the disruptions observed in adults with psychosis.

Fig. 1. Group differences in two measures of thalamocortical connectivity.

Group differences in thalamocortical fractional anisotropy (A) and global fractional anisotropy (B) in typically developing youth (blue) vs. youth with other psychopathologies (yellow) vs. psychosis-spectrum youth (red). The y-axis represents residualized data (age and sex regressed out) from the thalamocortical tracts, with the overall mean added to the residuals to facilitate one’s ability to interpret the findings.

Fig. 2. Neurodevelopmental trajectories of global fractional anisotropy and fractional anisotropy for psychosis-spectrum youth, typically developing youth, and youth with other psychopathologies.

Partial residual plots of (A–H) global fractional anisotropy trajectories, and (I–P) fractional anisotropy trajectories, in tracts between the thalamus and the lateral prefrontal cortex (A, I), medial prefrontal cortex (B, J), orbitofrontal cortex (C, K), sensorimotor cortex (D, L), lateral temporal cortex (E, M), medial temporal cortex (F, N), parietal cortex (G, O), and occipital cortex (H, P) for typically developing youth (blue) youth with other psychopathologies (yellow) and psychosis-spectrum youth (red). The partial residual plots reflect the relationship between age (x-axis) and the respective neuroimaging measures (y-axis), given the covariates in the model. For each group, the thick line reflects the line of best fit. The bars underneath the age plots reflect the derivative of the slope, i.e., the rate of change taking place at a particular age. Darker blue indicates that there is a stronger decrease in the respective QSDR or DTI measure taking place at that age, while brighter red indicates a stronger increase in the respective QSDR or DTI measure. Dotted lines and areas of lightly shaded blue indicate times when there was a significant “difference in smooths” between typically developing and psychosis spectrum youth.

Fig. 3. Diffusion weighted imaging analysis pipeline.

A Thalamic region of interest (light blue). B Cortical regions of interest: the lateral prefrontal cortex (red), medial prefrontal cortex (yellow), orbitofrontal cortex (orange), sensorimotor cortex (pink), lateral temporal cortex (green), medial temporal cortex (dark blue), parietal cortex (purple), occipital cortex (light blue). C Tracts between the thalamus and the lateral prefrontal cortex (pink), medial prefrontal cortex (dark blue), orbitofrontal cortex (purple), sensorimotor cortex (red), lateral temporal cortex (white), medial temporal cortex (beige), parietal cortex (green), occipital cortex (yellow).