Mind Melodies: An NLP platform to examine music cognition

Music is a fundamental semiotic system that shapes our daily experience by evoking thoughts and emotions. Across genres, a key driver of these effects is musical mode (the scale variation employed in a given piece or passage). To provide quantitative, ecologically valid insights on these phenomena, we developed Mind Melodies, an interactive platform that combines art, science, and technology to reveal individual and collective effects triggered by each Greek mode (Ionian, Dorian, Phrygian, Lydian, Mixolydian, Aeolian, and Locrian). Here we aim to (i) describe the project's rationale; (ii) outline the platform's interfaces for data collection, feature extraction, and visualization; and (iii) report preliminary findings from a pilot study. In Mind Melodies, participants listen to highly controlled musical compositions in different modes and to type out the mental scenes thus elicited. Natural language processing and deep learning algorithms are used to identify and visualize the main concepts, emotions, and imagery elicited by each mode. The resulting metrics are subjected to statistical analysis to capture individual and collective patterns. In a pilot study with 142 participants, we found that each mode was distinctly associated with a different semantic field and partly dissimilar emotions (e.g., Ionian was linked to 'life' and 'joy'; Phrygian to 'solitude', 'sadness', and 'fear'; Lydian to 'nature' and 'joy'; and Aeolian to 'mystery' and 'sadness'). Also, specific modes (e.g., Dorian and Mixolydian, Phrygian and Locrian) presented greater similarity than others. Overall, Mind Melodies represents a scalable transdisciplinary framework to inform science-first initiatives in the field of music cognition.