Investigation of the effects of dopaminergic medication and deep brain stimulation on neural dynamics of self-initiated action in Parkinson’s disease

Akinesia, the inability to initiate movement, is a key clinical sign of Parkinson’s disease. Both dopaminergic medication and deep brain stimulation can be effective for the treat-ment of akinesia. However, the neural dynamics of akinesia in cortex and basal ganglia and the influence of dopamine and deep brain stimulation on these dynamics remain elu-sive. Uncovering the neural mechanisms underlying akinesia and its treatment may lead to new therapeutic paradigms for Parkinson’s disease, where deep brain stimulation is triggered in a temporally targeted manner to support phasic dopamine release in rein-forcing healthy movement patterns. We recorded invasive brain signals from 25 patients (12 female) with Parkinson’s dis-ease undergoing deep brain stimulation surgery. Local field potentials from the subtha-lamic nucleus and subdural electrocorticography from the sensorimotor cortex were ob-tained while patients performed self-initiated upper limb movements. Machine learning-based brain signal decoders were then trained to detect movement intention. The latency between movement intention and movement execution, measured via machine learning classification and readiness potentials, was compared for recordings OFF therapy (N=22), ON dopaminergic medication (N=18), and ON subthalamic deep brain stimula-tion (N=8). We found that decoders trained on subthalamic local field potentials and sensorimotor electrocorticography were able to separate resting from movement intention periods at above-chance accuracy. However, decoders derived from electrocorticography were sig-nificantly more accurate than from subthalamic local field potentials. We also found that latencies between movement intention and movement execution in sensorimotor cortex, measured via decoding output and readiness potentials, were significantly longer OFF therapy than ON dopaminergic medication and ON subthalamic deep brain stimulation. To examine the motor preparation period in more detail, we measured directed cortico-subthalamic connectivity via Granger causality. We found that cortico-subthalamic drive in the theta band (4 – 8 Hz) correlated with motor execution latencies ON dopaminergic medication. Importantly, administration of both dopaminergic medication and deep brain stimulation shifted cortico-subthalamic connectivity from antikinetic beta (13 – 35 Hz) to prokinetic theta activity during motor preparation. This study demonstrates the modulation of cortical activity in Parkinson’s disease pa-tients during movement preparation through both dopaminergic medication and subtha-lamic deep brain stimulation. Our findings suggest a therapeutic mechanism for the alle-viation of akinesia via modulation of cortico-subthalamic oscillatory network activity. This work may pave the way toward a new paradigm for clinical brain-computer interfaces in movement disorders, where decoding of the patient’s intention to move is translated into temporally specific stimulation that accelerates action initiation.