Parkinson’s disease (PD) is the most common neuro-degenerative movement disorder . Characterized by the degradation of dopaminergic neurons, PD patients develop severe motor symptoms and cognitive impairment. Unfortunately, treatment options for PD symptoms do not reverse the underlying neuronal degradation and disease progression . Currently, the best method of defense against PD is early detection. However, patients are often diagnosed with PD by severe motor dysfunction, occurring around 80% degeneration of dopamine neurons . Therefore, it is important to diagnose PD before its advancement to preserve neuron integrity and slow progression.
One strategy for early detection of PD is speech pattern recognition. PD vocal dysfunction may be identified 5 years before traditional diagnoses by changes including, reduced volume and tongue flexibility, narrow pitch range, and long pauses . In this project, we used various speech-related data sets to classify and predict PD disease severity.