Deep Beats, Deep Thoughts? Predicting General Cognitive Ability from Natural Music-Listening Behavior

Area of Science:

• Cognitive Psychology
• Computational Social Science
• Digital Phenotyping

Background:

• Music listening is a complex behavior engaging cognitive systems.
• Individual differences in music listening patterns may reflect general cognitive ability (GCA).
• Digital behavioral data offers novel avenues for cognitive assessment.

Purpose of the Study:

• To investigate if everyday music listening patterns predict general cognitive ability (GCA).
• To explore the utility of smartphone-based behavioral data for cognitive insights.
• To identify which aspects of music listening behavior are most predictive of GCA.

Main Methods:

• Collected five months of smartphone-based music listening records from 185 participants.
• Quantified listening habits (duration) and music preferences (audio/lyrical characteristics).
• Employed linear LASSO regression and random forest models to predict GCA scores from behavioral features.

Main Results:

• The random forest model showed modest, reliable associations between music listening behavior and GCA.
• Lyrics-based preferences were the most predictive feature group, followed by listening habits.
• Audio characteristics of music contributed minimally to predicting GCA.

Conclusions:

• Cognitive ability may be subtly reflected in micro-patterns of everyday, non-achievement-related music listening behavior.
• Digital behavioral data can complement traditional cognitive assessments, though challenges remain.
• Future research should refine methods for using digital footprints to understand cognition.