Data-Driven Subtyping of Parkinson’s Using Acoustic Analysis of Sustained Vowels and Cluster Analysis: Findings in the Parkinson’s Voice Initiative Study

Abstract

People diagnosed with Parkinson’s (PwP) exhibit a diverse manifestation of heterogeneous symptoms which likely reflect different subtypes. However, there is no widely accepted consensus on the criteria for subtype membership assignment. We explored clustering PwP using a data-driven approach mining speech signals. We used data from the three English-speaking cohorts (Boston, Oxford, Toronto) in the Parkinson’s Voice Initiative (PVI), where speech and basic demographic information were collected over the standard telephone network. We acoustically characterized 2097 sustained vowel /a/ recordings from 1138 PwP (Boston cohort) using 307 dysphonia measures. We applied unsupervised feature selection to select a concise subset of the dysphonia measures and hierarchical clustering combined with 2D-data projections using t-distributed stochastic neighbor embedding (t-SNE) to facilitate visual exploration of PwP groups. We assessed cluster validity and consistency using silhouette plots and the cophenetic correlation coefficient. We externally validated cluster findings on the Oxford and Toronto PVI cohorts (n = 285 and 107 participants, respectively). We selected 21 dysphonia measures and found four main clusters which provide tentative insights into different dominating speech-associated characteristics (cophenetic coefficient = 0.72, silhouette score = 0.67). The cluster findings were consistent across the three PVI cohorts, strongly supporting the generalization of the presented methodology towards PwP subtype assignment, and were independently visually verified in 2D projections with t-SNE. The presented methodology with mining sustained vowels and clustering may provide an objective and streamlined approach towards informing PwP subtype assignment. This may have important implications towards developing more personalized clinical management of symptoms for PwP.

Guest Editorial Pervasive Computing in Healthcare

“…Overall, digital technologies (e.g., smartphones and smartwatches) are becoming increasingly affordable and have already been embraced by many people including elders [2], [3], facilitating large scale investigations and clinical trials. Furthermore, ubiquitous devices such as standard telephones, for example, have been used to collect speech signals for healthcare assessments, enabling large studies (∼10000 people) within months, across multiple countries, with minimal cost [5]. Nation-wide studies reaching 100000+ people who contribute their data have become possible, such as the U.K. BioBank (https://www.ukbiobank.ac.uk/), which has enabled novel data explorations into incident cardiovascular disease at scale [6].…”

Using the recurrence plots as indicators for the recognition of Parkinson’s disease through phonemes assessment

“…People diagnosed with PD exhibit a diverse manifestation of heterogeneous symptoms which likely reflect different subtypes [5]. Speech tasks are among the most robust ways of evaluating PD and further, assessing the stage of PD to which a patient is.…”

Comparing Lab-based and Telephone-based Speech Recordings Towards Parkinson's Assessment: Insights from Acoustic Analysis

“…Typically, sustained vowel /a/ phonations are used in speech clinical assessments because they overcome linguistic effects [3]. Indicative speech-PD applications include: (a) differentiating PwP from age-and gender-matched controls with almost 99% accuracy [5]; (b) accurately replicating the standard clinical scale denoting PD symptom severity [6]- [10]; (c) assessing voice rehabilitation in PD home monitoring systems [11]; (d) providing early PD precursors which may lead to early accurate diagnosis [12], [13]; (e) clustering PwP towards identifying PD subtypes and developing more targeted approaches for individuals [14]; and (f) speech articulation kinematic models to characterize PD dysarthria, thus providing tentative insights into the underlying physiology [15].…”