Dissociation of putative open loop circuit from ventral putamen to motor cortical areas in humans I: high-resolution connectomics

Human movement is partly organized and executed by cortico-basal ganglia-thalamic closed-loop circuits (CLCs), wherein motor cortical areas both send inputs to and receive feedback from the basal ganglia, particularly the dorsal putamen (PUTd). These networks are compromised in Parkinson’s disease (PD) due to neurodegeneration of dopaminergic inputs primarily to PUTd. Yet, fluid movement in PD can sporadically occur, especially when induced by emotionally arousing events. Rabies virus tracing in non-human primates has identified a potential alternative motor pathway, wherein the ventral putamen (PUTv) receives inputs from subcortical limbic areas (such as amygdala nuclei) and sends outputs to motor cortical areas putatively via the nucleus basalis of Meynert (NBM). We hypothesize that this separable open loop circuit (OLC) may exist in humans and explain the preservation of movement after CLC degradation. Here, we provide evidence for the normal human OLC with ultra-high field (7T), multi-echo functional magnetic resonance imaging. We acquired resting-state functional connectivity (FC) scans from 21 healthy adults (avg. age=29, 12M/9F, all right-handed) and mapped left-hemisphere seed-to-voxel connectivity to assess PUTv FC with putative subcortical nodes and motor cortical areas. We found that putative OLC node (basolateral amygdala, NBM) FC was greater with PUTv (p<0.05), while CLC subcortical seed (ventrolateral nucleus of thalamus) FC was greater with PUTd (p<0.01). Striatal FC patterns varied across cortical motor areas, with cingulate (p<0.0001) and supplementary (p<0.0001) motor areas showing greater FC with PUTv vs. nucleus accumbens. SMA had greater FC with PUTd vs. PUTv (p<0.0001), while cingulate and primary motor areas showed no significant differences in FC between PUTd and PUTv (p>0.1). Collectively, these results suggest that PUTv is functionally connected to motor cortical areas and may be integrated into a separable motor OLC with subcortical limbic inputs.