Bioinspired closed-loop deep brain stimulation for disorders of decision-making

Historically, Parkinson's disease (PD) was conceptualised as a movement disorder. However it is increasingly clear that neuropsychiatric symptoms occur both as a result of the underlying disease and its treatment. For example, subthalamic nucleus deep brain stimulation (STN-DBS) has been implicated in the dysregulation of impulse control, leading to impulsive and compulsive behaviours (ICBs).
Experiment 1
The STN is pivotal in response inhibition during high-conflict decision-making and this is thought to occur through theta-band (4-8Hz) coupling of the medial prefrontal cortex (mPFC) and the STN. At therapeutic frequencies (60-150Hz) STN-DBS disrupts this mPFC-STN coupling, reducing reaction time and increasing error rates in high-conflict decision-making. To test the hypothesis that electrophysiological correlates of behaviour can inform stimulation paradigms, and that these electrophysiological signatures encode bidirectional communication between brain regions that modulate behaviour, I will conduct a study of theta-frequency STN-DBS. Previous studies have demonstrated the safety of this but whole-brain network effects have not been investigated. PD patients with and without ICBs who have internalised STN-DBS implants will be recruited into a trial of theta-frequency STN-DBS during a high-conflict decision-making task. Behavioural outcomes will be compared against therapeutic-frequency STN-DBS and STN-DBS off. Simultaneous electrophysiological recordings will be taken (MEG/EEG) to identify mPFC theta-coherence during task and a whole-brain modelling pipeline will identify the wider decision-network effects.
Experiment 2
In humans, a decision-making action initiation network has been identified from fMRI data, with correlates in non-human primates. Notably, these include DBS targets such as the pedunculopontine nucleus. Transcranial Ultrasound Stimulation (TUS) has demonstrated reproducible effects on this decision-making network in non-human primates but this has not yet been tested in humans. To test the hypothesis that TUS of network nodes in this decision-making network can lead to reproducible and reversible behavioural change in humans, I will conduct a study investigating TUS of targets within this network in healthy volunteers and PD patients. Participants will complete high-conflict decision-making tasks during TUS using safe parameters, and will undergo fMRI imaging to detect network modulation.
Experiment 3
Based on the observed findings, a subset of these targets will be used to identify electrophysiological correlates of decision network modulation. To test the hypothesis that these network effects are bidirectional and correlate with observed task performance, PD patients with implanted DBS devices will be recruited into a TUS/DBS study. This will utilise the capacity of modern DBS devices to record and telemetrically readout local field potentials (LFPs) after internalisation:
Phase 1: LFP signals of decision-making will be recorded during TUS of key network nodes. These LFP recordings will be correlated with behavioural outcomes.
Phase 2: Correlated LFP signatures will be used to define stimulation settings. Implanted DBS devices will be programmed to these settings and participants will repeat the decision-making tasks without TUS. Behavioural outcomes will be compared to those during TUS.
Phase 3: These LFP signatures of decision-making will be used in a proof-of-concept closed-loop stimulation paradigm to demonstrate reproducibility within individuals.