Although several advances have been obtained in the field of neurological diseases, the diagnosis of Parkinson’s Disease (PD) is still challenging. Non motor symptoms, such as autonomic dysfunction, could be present before the onset of motor symptoms, when diagnosis can be performed using the current recommended criteria. As a consequence, patients with PD may be for a long time with no proper treatment. Therefore, this study aims at assessing the RR-time series and the cardiac autonomic control, analyzed through artificial intelligence (AI), as biomarkers of PD. We will assess electrocardiogram (ECG) from participants with Diagnosis of PD and age-matched controls. The main outcomes will be linear and non-linear indexes of the cardiac autonomic control, obtained by the analysis of heart rate variability (HRV). When the database is complete, convolutional neural networks will be applied to the signal of RR time series. Moreover, the Receiver Operating Characteristic (ROC) curve will be applied in order to establish a cutoff point for HRV indexes and then, sensitivity and specificity for HRV indices in discriminating healthy from PD participants will be calculated. The expected results of the current project are: 1) AI applied to RR time series will be able to discriminate the signal patterns between healthy participants and patients with PD; 2) artificial intelligence applied to RR time series will provide a better marker than linear and nonlinear HRV indexes and chronotropic response to a cardiac stress testing; and 3) artificial intelligence applied to RR time series will discriminate mild and moderate stages of PD. If the hypotheses of the current study are confirmed, this study has the potential, as future perspectives, to generate a new low-cost, quick, and accessible method to improve the diagnosis of PD. Finally, we state this study will be conducted in accordance with the “use of Artificial Intelligence (AI) applications and models” as requested by UKBiobank.
