MUSIC TRACK · PROPOSED LISTENER-PROFILE LAYER
What someone has heard can help explain what feels familiar. We will test whether MI adds useful musical structure to that history.
Study-design illustration · No account connection, audio download or MI extraction occurs here.
Participants may provide track records from Spotify, Apple Music or another source. Keep title, artist and available recording identifiers. Counts, dates and listening durations are used only when supplied; a list of names cannot establish frequency or recency.
Locate candidate recordings through catalogues or public pages, including YouTube. Confirm artist, version and duration rather than accepting the first title match. Record live/cover/remaster differences and unresolved matches. Track-level listening duration does not identify which moments were heard.
Analyse the verified recording using a frozen MI configuration. R³ describes acoustic structure; T³ describes its temporal organisation. A separate comparison adds the six prespecified C³ outputs already shown in the architecture. Their names remain computational labels.
Combine those features with the available history to estimate exposure to musical structures. Prespecify aggregation, recency weighting and missing-data rules in development data. Report matched-track and listening-time coverage where possible. A profile is an estimate to test, not a diagnosis or a reconstruction of autobiographical memory.
Use short recognition and continuation tasks before judging the profile's value. Separate calibration from evaluation. Newly created, acoustically matched foils allow false-alarm estimation; absence from a platform export alone does not establish novelty. Pleasure is measured in a separate block.
Compare self-report, self-report plus listening-history and artist/genre baselines, and the same model with MI features. Compare R³ → R³ + T³ → R³ + T³ + C³ with identical data splits and fitting budgets. Hold out listeners and recording families. History donation does not replace the controlled learning experiment.
A track-level observation, with incomplete coverage across devices, services and everyday listening.
An MI-informed prediction that must agree with independent behavioral responses to be useful.
Candidate discovery is separate from permission to acquire and analyse audio. Extraction will use authorised sources and documented applicable-use review. Spotify platform restrictions and YouTube access terms are not removed by academic purpose, non-distribution or later deletion.
Temporary analysis audio will not be redistributed. Delete it after extraction and quality control, within seven days of acquisition, including unresolved or failed items. Keep only necessary pseudonymous features, identifiers, checksums and audit records under the consented retention policy. Audio used for participant playback requires a separate authorised source and retention plan.
Parsers, recording matching, the profile model and research data collection remain to be implemented and validated. The existing MI map depicts audio routes; it does not depict an already implemented personal-memory system.
Source requirements: Spotify export fields · Spotify Developer Policy · Apple Music API scope · YouTube terms. Reviewed 7 October 2026.