Deep Learning Approaches in the Early Diagnosis of Parkinson's Disease

1. Introduction

Parkinson's disease (PD) is the commonest movement disorder worldwide, of no proven neuroprotective or neurorestorative therapies (Smimih et al., 2023). PD is a multisystem disorder characterized by both motor symptoms; rigidity, akinesia/bradykinesia, and tremors (El Khiat et al., 2023), and non-motor symptoms (NMSs) (El-Mansoury et al., 2023), including neuropsychiatric abnormalities, sensory and autonomic dysfunction. (Masato et al., 2019; Poewe, 2008; Zesiewicz, 2019; El-Mansoury et al., 2023). These NMSs have been underrecognized, but now are of increasing interest because of their negative impact on the patients’ quality of life (QOL) (Aarsland et al., 2021; Chaudhuriet al., 2006 ; Chaudhuri et Schapira, 2009). Despite the lack of systematic longitudinal follow-up studies, there are currently insufficient data on the natural course of PD (W. H. Poewe & Wenning, 1998). The understanding of PD has been better understood and treated in the past 200 years, which has led to numerous developments in the history of the disease (Draoui et al., 2020). It is well known that the motor symptoms of PD are caused by the degeneration of dopaminergic neurons in the midbrain's substantia nigra (SN) pars compacta (SNpc), which leads to DA deficiency, dysregulation of the nigrostriatal pathway, and ultimately abnormal subcortico-cortical interactions (El-Mansoury et al., 2023). Several risk factors are believed to contribute to the development of PD including exposure to pesticides (Saad et al., 2023), hydrocarbons (El Baz et al., 2023), genetic causes (Aimrane et al., 2023), among others, along with epigenetic mechanisms (Tsalenchuk et al., 2023).