Tools Development and Open Source Neuroscience
Author: Hernan Bocaccio | Email: hbocaccio@gmail.com
Hernan Bocaccio 1°2°, Marisol Domínguez 4°, Bettina Mahler 3°, Juan C. Reboreda 3°, Gabriel Mindlin 1°2°
1° Universidad de Buenos Aires, Facultad de Ciencias Exactas y Naturales, Departamento de Física, Ciudad Universitaria, 1428 Buenos Aires, Argentina.
2° CONICET – Universidad de Buenos Aires, Instituto de Física Interdisciplinaria y Aplicada (INFINA), Ciudad Universitaria, 1428 Buenos Aires, Argentina.
3° Universidad de Buenos Aires, Facultad de Ciencias Exactas y Naturales, Departamento de Ecología, Genética y Evolución, Ciudad Universitaria, 1428 Buenos Aires, Argentina.
4° Unit of Evolutionary Biology/Systematic Zoology, Institute of Biochemistry and Biology, University of Potsdam, Germany
Birdsong patterns are complex emergent behaviors that serve roles in communication, mate attraction, and territory defense among avian species. In recent years, machine learning techniques applied to audio field recordings of birdsongs have yielded successful results in studying population distributions and identification of individuals for their monitoring in a variety of bird species, offering promising possibilities in the study of biodiversity and conservation strategies for birds. In this work, we employed deep learning models on sonograms of audio field recordings to explore vocalization statistics in the endangered Yellow Cardinal, a novel application for this species. Our results indicate the presence of vocal signatures that reflect similarities in songs of individuals that inhabit the same region, determining dialects, but which also show differences between individuals that can be exploited by a deep learning classifier to discriminate the bird identities through their songs. Our approach reinforces existing research while automating the characterization of cultural units within the species. When combined with genetic data, this method could enhance management unit delineation, supporting reintroduction initiatives for the Yellow Cardinal. The innovation of neural network-based individual classification, despite limited data availability, holds promise for non-invasive acoustic monitoring, with substantial conservation implications.