Multicondition training for PD diagnosis with voice features from diadochokinesis test
Description
The dataset includes acoustic features extracted from voice recordings using a diadochokinesis (DDK) test protocol. Realistic environmental noise was introduced into the recordings prior to feature extraction to simulate real-world conditions. The dataset comprises voice samples from 30 Parkinson's disease (PD) patients and 30 healthy controls. All samples were captured via smartphone without audio compression to preserve data quality.
Files
Steps to reproduce
This dataset has been derived and used in the paper: Mario Madruga Escalona, Yolanda Campos-Roca, Carlos Javier Pérez Sánchez, Enhancing noise robustness of automatic Parkinson’s disease detection in diadochokinesis tests using multicondition training, Expert Systems with Applications, Volume 260, 125401, 2025, https://doi.org/10.1016/j.eswa.2024.125401