Data-Driven Models for Objective Grading Improvement of Parkinson’s Disease

Abstract

Parkinson’s disease (PD) is a progressive disorder of the central nervous system that causes motor dysfunctions in affected patients. Objective assessment of symptoms can support neurologists in fine evaluations, improving patients’ quality of care. Herein, this study aimed to develop data-driven models based on regression algorithms to investigate the potential of kinematic features to predict PD severity levels. Sixty-four patients with PD (PwPD) and 50 healthy subjects of control (HC) were asked to perform 13 motor tasks from the MDS-UPDRS III while wearing wearable inertial sensors. Simultaneously, the clinician provided the evaluation of the tasks based on the MDS-UPDRS scores. One hundred-ninety kinematic features were extracted from the inertial motor data. Data processing and statistical analysis identified a set of parameters able to distinguish between HC and PwPD. Then, multiple feature selection methods allowed selecting the best subset of parameters for obtaining the greatest accuracy when used as input for several predicting regression algorithms. The maximum correlation coefficient, equal to 0.814, was obtained with the adaptive neuro-fuzzy inference system (ANFIS). Therefore, this predictive model could be useful as a decision support system for a reliable objective assessment of PD severity levels based on motion performance, improving patients monitoring over time.

Ensemble deep model for continuous estimation of Unified Parkinson’s Disease Rating Scale III

“…[ 15 ] with -0.56, equal or slightly lower than Refs. [ 17 , 25 ] and lower than Ref. [ 18 ] with 0.88.…”

“…Our analysis indicated such a possibility with a high correlation of and low MAE = 5.95 when using an ensemble of three deep-learning models. Most of the existing work for UPDRS III estimation requires PwP’s active engagement to perform the specific tasks used in the UPDRS-III procedure [ 15 , 17 , 18 , 25 ]. Unlike these approaches, our algorithm could estimate UPDRS III as the patients performed a variety of ADL without the need for performing constrained tasks.…”

“…A review of the methods proposed for estimating the severity of PD is shown in Table 2 . Comparing our algorithm to task-dependent approaches (i.e., obtrusive methods) [ 15 , 17 , 18 , 25 ] indicates that our method provides comparable performance with the advantage of not constraining PwP’s activities. For example, it has a better correlation than Ref.…”

Internet of Things Technologies and Machine Learning Methods for Parkinson’s Disease Diagnosis, Monitoring and Management: A Systematic Review

“…Nevertheless, in another approach [64], SVM models achieve 87.75-94.5% accuracy, which corresponds to the highest performance obtained for H&Y scale. Additionally, in [76,77] wrist-and foot-worn IMU signals are combined to quantify PD severity. In the first case, CNNs and long short-term memory (LSTM) networks are combined to estimate UPDRS scores, while in the second, the adaptive neuro-fuzzy inference system (ANFIS) is proposed.…”

An Evaluation of KELVIN, an Artificial Intelligence Platform, as an Objective Assessment of the MDS UPDRS Part III

“…To help address this variability, several Artificial Intelligence (AI) tools are in development. Studies inputting data from wearable sensors and voice recordings into machine-learning techniques have demonstrated high levels of accuracy of automated systems to score items from the MDS-UPDRS such as bradykinesia [11][12][13][14] and tremor [12] [15] -18.…”