TRANSLATIONAL EVIDENCE FOR DOPAMINERGIC ALTERATION OF BASAL GANGLIA FUNCTIONAL CONNECTIVITY IN PERSONS WITH SCHIZOPHRENIA

In prior work, a transgenic mouse model of the striatal dopamine dysfunction observed in persons with schizophrenia (PSZ), the D2-receptor-overexpressing mouse (D2-OE), exhibited dopamine-related neuroplasticity in the basal ganglia. These mice show a plastic increase in the density of axonal collaterals from direct pathway D1R-expressing medium spiny neurons of the associative striatum to the globus pallidus externus (GPe; indirect pathway). This phenotype of increased “bridging collaterals” has not yet been demonstrated in PSZ. We aimed to test whether unmedicated PSZ show analogous alterations in connectivity between dorsal caudate (DCa) and GPe using functional magnetic resonance imaging (fMRI) acquired in the resting state (RS-fMRI) and in a self-ordered working memory (WM) task state (TS-fMRI). We predicted that functional connectivity (FC) alterations would be most pronounced during task engagement. Further, we examined whether these alterations are associated with deficits in WM task performance and dopamine function assessed by neuromelanin-sensitive MRI (NM-MRI) and [11C]PHNO positron emission tomography (PET). We conducted a case-control study of unmedicated PSZ and demographically matched healthy controls (HC). 37 unmedicated PSZ and 30 HCs completed RS-fMRI; 29 PSZ and 29 HC also completed TS-fMRI (~30 minutes per paradigm; TR=850 ms; 2 mm voxels). 22 PSZ and 20 HC additionally completed NM-MRI; 7 PSZ and 4 HC also completed PHNO PET with amphetamine challenge. Outcome measures included: 1) group differences in resting-state FC (RS-FC) and task-state FC (TS-FC) between DCa and GPe, 2) NM-MRI contrast ratio in substantia nigra-ventral tegmental area voxels previous shown to be associated with psychotic symptom severity (SN/VTA), and 3) baseline and amphetamine-induced change in PHNO binding potential in DCa. TS-fMRI data were residualized to task-evoked responses and only “task on” TS-fMRI volumes were analyzed. RSFC and TS-FC between DCa and GPe were calculated as partial correlations within a basal ganglia-thalamo-cortical network, controlling for signal in non-pair nodes (dorsolateral prefrontal cortex, globus pallidus internus, substantia nigra, mediodorsal thalamus). Dunn-Šidák correction was performed for the two primary comparisons (RS-FC and TS-FC). We performed robust multivariable regressions to relate DCa-GPe TS-FC to WM capacity, NM-MRI contrast in SN/VTA, baseline PHNO BPND, and amphetamine-induced PHNO ΔBPND in DCa; models included age, sex assigned at birth, and diagnosis. PSZ exhibited greater DCa-GPe TS-FC than HC (PSZ mean[SD]=0.11[0.10], HC=0.05[0.09]; P=0.0252). DCa-GPe RS-FC did not differ between groups (PSZ median[IQR]=0.08[0.12], HC=0.05[0.16]; P=0.233). Greater DCa-GPe TS-FC was associated with lower WM capacity (β*[SE] = -0.31[0.13]; P=0.020) and greater NM-MRI contrast in SN/VTA (β*[SE] = 0.40[0.17]; P=0.023). Greater DCa-GPe TS-FC was associated with reduced DCa baseline D2R availability (PHNO BPND; β*[SE] = -0.45[0.17]; P=0.039) and increased DCa dopamine release (more negative ΔBPND; β*[SE] = -0.82[0.27]; P=0.021). Unmedicated PSZ demonstrate task-dependent DCa-GPe hyperconnectivity that covaries with WM impairment and markers of striatal dopamine dysfunction. These results provide translational evidence for a dopamine-associated abnormality in associative striato‑pallidal circuitry in PSZ, consistent with compensatory “bridging collateral” plasticity observed in the D2-OE mouse, and suggest a potential developmental mechanism for WM deficits in PSZ.