NeuroVoz: a Castillian Spanish corpus of parkinsonian speech

TL;DR

NeuroVoz provides the first comprehensive publicly available Castilian Spanish Parkinsonian speech corpus, consisting of 112 native speakers (54 PD, 58 HC) recorded in ON state and annotated with GRBAS along with detailed phonation, intensity, speed, resonance, intelligibility, and prosody metadata. The dataset encompasses diverse tasks—sustained vowels, a DDK test, 16 LR utterances, and spontaneous monologues—plus manual transcriptions and a structured data organization across audio, features, GRBAS, transcriptions, and metadata, totaling ~1.01 GB with 2,977 audio files. Prior studies using NeuroVoz achieved up to 89% PD-screening accuracy, and the work discusses language- and dialect-specific considerations and cross-corpora challenges to be addressed for language-agnostic PD analysis. Access is governed by a Data Usage Agreement and FAIR principles, with accompanying Python scripts and open tools to enable baseline analyses and feature extraction, facilitating reproducible research in Parkinsonian speech for Spanish speakers and beyond.

Abstract

The screening of Parkinson's Disease (PD) through speech is hindered by a notable lack of publicly available datasets in different languages. This fact limits the reproducibility and further exploration of existing research. To address this gap, this manuscript presents the NeuroVoz corpus consisting of 112 native Castilian-Spanish speakers, including 58 healthy controls and 54 individuals with PD, all recorded in ON state. The corpus showcases a diverse array of speech tasks: sustained vowels; diadochokinetic tests; 16 Listen-and-Repeat utterances; and spontaneous monologues. The dataset is also complemented with subjective assessments of voice quality performed by an expert according to the GRBAS scale (Grade/Roughness/Breathiness/Asthenia/Strain), as well as annotations with a thorough examination of phonation quality, intensity, speed, resonance, intelligibility, and prosody. The corpus offers a substantial resource for the exploration of the impact of PD on speech. This data set has already supported several studies, achieving a benchmark accuracy of 89% for the screening of PD. Despite these advances, the broader challenge of conducting a language-agnostic, cross-corpora analysis of Parkinsonian speech patterns remains open.

Figures (21)

• Figure 1: Sex distribution of participants.
• Figure 2: Age distribution of participants.
• Figure 3: Distribution of years since the diagnosis of PD.
• Figure 4: GRBAS total sum distribution of participants.
• Figure 5: The free monologue is guided by asking the patient to describe the scenes presented in this illustration.