Abstract

Music perception, cognition, and production research have progressed significantly from examining neural correlates of musical components to a better understanding of the interplay of multiple neural pathways that are both unique and shared among other higher neurocognitive processes. The interactions between the neural connections to perceive an abstract entity like music and how musicians make music are an area to be explored in greater depth. With the abstract nature of music and cultural differences, carrying out research studies using ecologically valid stimuli is becoming imperative. Artificial intelligence (AI) and machine learning (ML) models are data-driven approaches that can investigate whether our current understanding of the neural substrates of musical behaviour can be translated to teach machines to perceive, decode, and produce music akin to humans. AI algorithms can extract features from human-music interaction. Training ML models on such features can help in information retrieval to look at the brain's natural music processing, recognizing the patterns concealed within it, deciphering its deeper meaning, and, most significantly, mimicking human musical engagements. The question remains how these models can be generalized for knowledge representation of human musical behaviour and what would be applications in a more ecologically valid manner.