The Human Voice and Whole-Body Kinetics Open Dataset

This contains all the supporting data for the paper "The Human Voice Aligns With Whole-body Kinetics" (to be) published in the Proceedings B (Authors: Wim Pouw, Lara Burchardt, Raphael Werner, Luc Selen). Variables and concepts are also explained in detail in the Rmarkdown provided on GitHub (https://github.com/wimPouw/kineticsvoice).

This dataset contains biomechanical and acoustics measurements from an experiment investigating how arm movements affect vocalization through biomechanical connections. The study examined 17 participants (7 female, 10 male) performing different arm movements while vocalizing to determine whether and how these movements create perturbations that affect voice amplitude. The experiment used a within-subject design with three manipulated factors: movement condition (no movement, extension, flexion, external rotation, internal rotation), weight condition (no weight, 1kg wrist weight), and vocalization condition (vocalize, expire). Each participant completed 80 trials across these conditions. The dataset includes synchronized recordings of multiple physiological systems captured at high temporal resolution (2500 Hz). 1. Surface EMG activity from four muscle groups: pectoralis major (internal rotator), infraspinatus (external rotator), rectus abdominis (postural muscle), and erector spinae (postural muscle). These measurements track both the focal muscles directly responsible for arm movements and the postural muscles that stabilize the body during movement. 2. Postural measurements from a balance board, including center of pressure changes in anteroposterior and mediolateral directions, and ground reaction forces. These measurements capture how the body maintains stability during arm movements. 3. Movement kinematics derived from video tracking, including wrist position coordinates, movement speed, and acceleration. These data quantify the movements performed. 4. Voice measurements, including the amplitude envelope of vocalization. These measurements track modulations in voice intensity during movements.
All data streams were synchronized using LabStreamingLayer (LSL) with sub-millisecond precision, allowing for detailed analysis of the temporal relationships between muscle activity, postural adjustments, movement, and vocalization. The data is provided in multiple formats: raw time series for each trial, z-scaled (normalized) time series, and aggregated peak values for statistical analysis. This dataset allows researchers to investigate the biomechanical chain linking arm movements to the voice.