To investigate whether electrophysiology, combined with machine learning techniques, can support the selection of optimal electrode contacts for deep brain stimulation (DBS) in patients with Parkinson's disease.
Key Findings:
The model predicted therapeutic windows with a correlation of r = 0.45 (p < 0.001) in the original cohort and r = 0.30 (p < 0.001) in an independent cohort.
The model primarily relied on fast (>35 Hz) subthalamic activity and STN-cortex coherence across various frequency bands.
The approach allows for faster identification of optimal electrode contacts.
Interpretation:
The study demonstrates the feasibility of using electrophysiological features to predict therapeutic windows, potentially aiding in automated contact selection for DBS.
Limitations:
The study's sample size was limited, with only 45 patients in the original cohort and 8 in the independent cohort, which may affect the robustness and generalizability of the findings.
The model's performance may vary across different patient populations and clinical settings.
Conclusion:
This research highlights the potential of integrating electrophysiological markers and machine learning to enhance DBS programming for Parkinson's disease.
by Fayed Rassoulou, Abhinav Sharma, Alexandra Steina, Markus Butz, Christian J. Hartmann, Bahne H. Bahners, Jan Vesper, Alfons Schnitzler, Jan Hirschmann
