Evaluation of Deep Brain Stimulation Effect on Motor Signs of Individuals with Movement Disorders Through Gaussian Models

In late stages of Parkinson's Disease (PD) or Essential Tremor (ET), surgical treatments, such as Deep Brain Stimulation (DBS), are promising. However, conventional DBS devices are open-loop systems in which the stimulation intensity is constant regardless of the dynamic behavior of motor manifestations. In this sense the seek to objective strategies that can assess the efficacy of DBS is relevant for the development of precise medicine, promoting the implementation of closed-loop DBS. This study proposes the use of individualized Gaussian models to evaluate the motor signs of people with PD and ET considering DBS in ON and OFF states. Inertial data were collected from the left and right hands and forearms of the participants. After data preprocessing, amplitude and variability-based features were extracted to estimate specific Gaussian models for each individual. An objective metric that is sensitive to the distance of models from DBS in ON and OFF states was proposed. The specific Gaussian models allowed for the estimate of tremor probability in the DBS states. The estimated values of the proposed metric discriminated between the DBS ON and OFF states, and the verification of the DBS effect on the motor sign of each individual. The specific Gaussian models can be used as part of a closed-loop DBS system to estimate the tremor severity of patients.