Early-Life Chick Vocalisations: VPA Neurodevelopmental Model Dataset for Bioacoustic and Computational Analysis

This dataset contains audio recordings and annotations of domestic chick (Gallus gallus) vocalisations from the study "Embryonic Exposure to VPA Influences Chick Vocalisations: A Computational Study" (Torrisi et al., 2026). The dataset includes 46 audio recordings (WAV, 44.1 kHz) collected from visually-naïve chicks tested within 12 hours after hatching during a social choice experiment. Recordings comprise 27 VPA-exposed chicks (16 males, 11 females) and 19 control chicks (8 males, 11 females), totalling 4 hours and 18 minutes of audio. Both manual annotations (22,430 calls) and automatic detections (23,549 calls) are provided, along with complete experimental metadata. Automatic detections were validated against manual annotations, achieving F1-score = 0.933 for onset detection and F1-score = 0.885 for offset detection. This dataset enables bioacoustic research, computational analysis of animal vocalisations, and studies investigating neurodevelopmental disorders in animal models.