Movie Stimuli (NAT Paradigm)
Short films (~4 minutes each) shown during naturalistic encoding sessions
(ses-19 through ses-28). The movie schedule is defined in
movies/MMM movies - Sheet1.csv.
- Count: ~40 unique titles across 10 sessions
- Duration: ~4 minutes each
- Styles: Animated, live-action, and stop-motion; with or without speech
- Source: Mix of short films and clips from longer films (sourced via a
“memory search project” with collaborators Jordan & Kahlyn). Movie files are
stored on Google Drive (links in
short films 4 minutes.rtf). - Movie files: 60
.movfiles inmovies/movie_files/, named{Title}_trimmed_normalized_filtered.mov - Movie cues: 60
.jpgrecall cue images inmovies/movie_cues/, named{Title}_cue.jpg(see Movie Cue Images below)
Movie Cue Images
During free recall, participants are prompted with a movie title. If they cannot recall which movie the title refers to, they are shown a cue image — a single representative frame from the film. The cue is designed to orient the participant without biasing their memory of the movie’s content.
- Count: 60 images (one per movie)
- Format: JPG
- Location:
movies/movie_cues/{Title}_cue.jpg
Recall Conditions
Each movie in a session is assigned one of three recall conditions:
| Condition | Label | Description |
|---|---|---|
| 1 | Same-session recall | Viewed and recalled within the same session |
| 2 | Next-session recall | Viewed this session, recalled in the following session |
| 3 | Multiple-repeat | Recurring film shown across many/all sessions |
Two films recur across all 10 sessions as condition 3: “From Dad To Son” (animated, no speech) and “The Bench” (live-action, speech).
Session Structure
Each session contains 6–8 movies:
- ~4 condition 1 movies (same-session recall)
- ~2 condition 2 movies (recalled next session)
- 2 condition 3 movies (recurring)
- Sessions 2–10 also include 2 carry-over movies from the prior session (positions 7–8 at recall) that are recalled but not re-viewed
Movie Schedule Metadata
MMM movies - Sheet1.csv columns:
| Column | Description |
|---|---|
Session |
NAT session number (1–10, mapping to ses-19 through ses-28) |
Condition |
Recall condition: 1 = same-session, 2 = next-session, 3 = multi-repeat |
Movie name |
Title of the film |
Movie style |
Style description (e.g., “Animated, no speech”, “Live-action, speech”) |
Movie duration |
Duration (currently empty — all ~4 minutes) |
Cue position at Recall |
Order position (1–8) of this movie’s recall cue; N/A for condition-2 movies |
Blank rows in the CSV separate sessions. Rows without a condition value are carry-over movies from the prior session.
Computational Features (viz2psy)
Movie frames and cue images are processed with
viz2psy, producing per-movie temporal
feature timeseries in movies/viz2psy_scores/.
For each movie, the output includes:
{Title}_scores.csv— one row per sampled frame, indexed bytime(seconds), with ~2,900 feature columns (memorability, emotion, CLIP embeddings, etc.){Title}_scores.meta.json— feature definitions and provenance{Title}_scores_dashboard.html— interactive visualization of features over time{Title}_scores_frames/— extracted frame images (JPG)
Cue images are scored separately in a consolidated cue output file.
See the viz2psy documentation for full feature definitions.