Deep Learning for DBS Parameter Recognition in Parkinson’s Disease

Deep Learning for DBS Parameter Recognition in Parkinson’s Disease

 Dr. Shashikala S V M.Tech ,Ph.D

 Department of Computer Science , BGS Institute of Technology,

Mandya, India hodcse@bgsit.ac.in

Samanvitha R Adagur 20CSE070

Department of Computer Science, BGS Institute of Technology,

Mandya, India samanvitharadagur@gmail.com

Abstract— Deep brain stimulation (DBS) presents a patient-specific treatment approach for Parkinson’s disease (PD), crucially dependent on quantifying therapeutic response. Inertial sensor systems have significantly improved the precision of assessing movement disorder characteristics, offering wearable and wireless solutions mounted on the hand dorsum to objectively measure tremor severity,Utilizing AI calculations, remarkable order precision has been accomplished. As of late evolved conformal wearable and remote inertial sensor frameworks, looking like wraps, permit advantageous mounting on the hand dorsum utilizing glue mediums and empower remote information transmission to get distributed computing conditions. This study aimed to quantify and differentiate PD tremor responses across various DBS amplitude settings, including 'Off' as a baseline and utilizing deep learning convolutional neural networks via TensorFlow and quantifying tremor with the mentioned sensor system. Remarkable classification accuracy was achieved, indicating potential for optimizing DBS treatment for PD.

Keywords: Deep brain stimulation, Parkinson’s disease, inertial sensor systems, convolutional neural networks, TensorFlow, wearable sensors, classification accuracy.