Tuning Deep Brain Stimulation Parameters: An Adaptive and Individualized Approach

Masoumeh Heidari Kapourchali1, Bonny Banerjee

  • 1The University of Memphis

Details

19:30 - 20:30 | Tue 6 Mar | Caribbean ABC | TuPO.2

Session: Poster Session # 2 and BSN Innovative Health Technology Demonstrations

Abstract

Deep brain stimulation (DBS) is an established treatment for Parkinson's disease (PD) based on chronic high-frequency stimulation of the basal ganglia nuclei. Currently, DBS parameters are chosen empirically using a trial and error approach. Recently, a set of objective methods have shown promising results for automatic selection of DBS parameters. These approaches provide more efficient therapy and use less battery power. However, there are significant limitations and drawbacks to these proposals. An adaptive DBS system should be able to handle the dynamic nature of the brain, take into account the progressive nature of the disorder, alleviate symptoms while minimizing the side effects, consider personal characteristics, and have low computational cost to be installed in the implanted DBS device. In this work, we propose a data-driven technique for DBS programming which is efficient and adaptive without any changes in the surgical procedures. Our model is accurate and efficient in terms of the computational cost. It does not presume the shape of oscillatory patterns. The stimulation is done only when needed. This makes the battery life longer and leads to less post therapeutic visits and surgeries. The proposed algorithm is extensively evaluated using a computational model of the symptoms. We generated signals of various patterns with injection of different levels of noise to evaluate the proposed model since in reality, the oscillations are modulated with different amounts of noise. Our predicted parameters were compared with other adaptive DBS techniques and also clinical measurements reported in the literature for individual patients. Our results indicate that the proposed model is highly reliable and accurate.