Validation of freely-available pitch detection algorithms across various noise levels in assessing speech captured by smartphone in Parkinson's disease. (April 2020)

• Record Type: Journal Article
• Authors: Illner, Vojtech; Sovka, Pavel; Rusz, Jan
• Is Part Of: Biomedical signal processing and control. Volume 58(2020)
• Publication Date: 2020-04
• Subjects: Pitch -- Fundamental frequency -- Speech -- Voice -- Dysarthria -- Smartphones -- Parkinson's disease; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2019.101831
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: Pitch estimation is pivotal for connected speech assessment in Parkinson's disease. Performance of pitch trackers against noise in evaluation by smartphone was tested. SWIPE was the most robust algorithm for correct estimation of pitch sequences. Monopitch distinguished Parkinson's disease from controls at a low 6 dB SNR level. Abstract: Measuring the fundamental frequency of the vocal folds F 0 is recognized as an important parameter in the assessment of speech impairments in Parkinson`s disease (PD). Although a number of F 0 trackers currently exist, their performance in smartphone-based evaluation and robustness against background noise have never been tested. Monologues from 30 newly-diagnosed, untreated PD patients and 30 matched healthy control participants were collected. Additive non-stationary urban and household noise at different SNR levels was added to the recordings, which were subsequently assessed by 10 freely-available and widely-used pitch-tracking algorithms. According to the comparison of all investigated pitch detectors, sawtooth inspired pitch estimator (SWIPE) was the most robust and accurate method in estimating mean F 0 and its standard deviation. However, at a low 6 dB SNR level, a combination of more algorithms may be needed to achieve the desired precision. Monopitch, calculated as F 0 standard deviation and estimated by SWIPE, proved to be robust in distinguishing between the PD and healthy control groups ( p < 0.001). We anticipate that monopitch may serve as a quick and inexpensive biomarker of disease progression based on longitudinal data collected via smartphone, without any logistical or time constraints for patients and physicians.