Objective: Our goal is to develop a data-driven approach to identify optimal Deep brain stimulation (DBS) parameters for gait enhancement in patients with Parkinson’s disease (PD).

 ​​​​​​​

Background: DBS has shown promise in alleviating various PD symptoms like tremors and bradykinesia, but challenges persist in addressing advanced gait-related issues such as hypokinetic gait patterns and increased gait variability. DBS programming, primarily based on appendicular symptoms does not fully capture gait disorders in PD. Additionally, the effects of different stimulation parameters on gait have been variable.

Method: Patients’ DBS settings are altered within safety ranges to investigate their impacts on gait. Local field potentials (LFPs) were recorded from the globus pallidus (GP) and a cortical paddle overlying the motor cortex of three PD patients using a bidirectional, investigational device. Gait kinematics were captured by inertial measurement unit sensors. To analyze how gait is influenced by DBS setting changes, we developed a Walking Performance Index (WPI) capturing gait kinematics including stride velocity, step length and time variabilities, and arm swing amplitude. We employed a Gaussian Process Regressor (GPR), a data-driven methodology, for modeling the relationship between DBS settings and the WPI across individual subjects.

Results: We identified DBS settings that resulted in an overall 18% improvement in the WPI, including increases in stride velocity (16%) and arm swing amplitudes (5%), and decreases in step length variability (37%) and step time variability (31%). Further investigation revealed a neurophysiological basis for the observed improvements in WPI. In the corticomotor and somatosensory areas, we noted a significant negative correlation (p < 0.01) between the alpha frequency band of LFP power during the double support period and enhancement in WPI. Moreover, we detected a significant positive correlation between the coherence of the GP with the corticomotor and somatosensory network and the improvement in WPI.

Conclusion: Our findings highlight the potential of integrating the WPI and GPR to optimize DBS parameters, underscoring the importance of personalized, data-informed interventions. The revealed neurophysiological basis offers a foundation for designing adaptive DBS systems to dynamically enhance the gait functions in PD.

« Back to 2024 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/personalized-optimization-of-deep-brain-stimulation-parameters-for-gait-enhancement-in-parkinsons-disease-a-data-driven-approach/