Detecting Parkinson’s disease with sustained phonation and speech signals using machine learning techniques

Abstract

This study investigates the processing of voice signals for detecting Parkinson's disease. This disease is one of the neurological disorders that affect people in the world most. The approach evaluates the use of eighteen feature extraction techniques and four machine learning methods to classify data obtained from sustained phonation and speech tasks. Phonation relates to the vowel /a/ voicing task and speech to the pronunciation of a short sentence in Lithuanian language. The audio tasks were recorded using two microphone channels from acoustic cardioid (AC) and a smartphone (SP), thus allowing to evaluate the performance for different types of microphones. Five metrics were employed to analyze the classification performance: Equal Error Rate (EER) and Area Under Curve (AUC) measures from Detection Error Tradeoff (DET) and Receiver Operating Characteristic curves, Accuracy, Specificity, and Sensitivity. We compare this approach with other approaches that use the same data set. We show that the task of phonation was more efficient than speech tasks in the detection of disease. The best performance for the AC channel achieved an accuracy of 94.55%, AUC 0.87, and EER 19.01%. When using the SP channel, we have achieved an accuracy of 92.94%, AUC 0.92, and EER 14.15%.

Identifying Parkinson’s Disease Through the Classification of Audio Recording Data

“…The results are better than those from [26], but not as good as those achieved using the LRNN. The LRNN results show a clear improvement in particular over the smartphone results from [27], which are probably more comparable to ours, then the professional microphone ones.…”

“…Uniform data Table I provides the results for the two neural networks designed here, and for comparison those from [26]. [27], and those produced using Rapid Miner. The clinical diagnosis of Parkinson's, as mentioned previously, has a misdiagnosis rate of 19%-24%.…”

“…We have also included results from [27] in the table. As explained, these cannot be compared directly, since they use different but similar data.…”

“…Another recent study shows good results for detecting Parkinson's disease from audio recording [27]. Here, new recordings are done, using either a smartphone or professional recording equipment.…”

Intelligent Sensory Pen for Aiding in the Diagnosis of Parkinson’s Disease from Dynamic Handwriting Analysis

“…Another case of successful use of ML and signal analysis for the detection of PD was done by Jefferson S. Almeida using sustained phonation obtained from speech tasks. The audio recorded from two microphones was processed by multiple classifiers using eighteen different extracted features and achieved between 92.94% and 94.55% accuracy for assessing parkinsonian patients [ 21 ]. In the same topic, Laureano Moro-Velázquez obtained 87% accuracy via Kinect change analyses using speech recognition techniques in different application domains [ 22 ].…”