Analysis for individual differences in mental information decoded from predicted voxel response

This program show individual-difference reflection (IDR) of predictted-response decoding for measured-response decoding.

.
├── data
│   ├── TVAdMovieSet
│   │   └── {subject}
│   │       ├── SceneDescription
│   │       ├── ImpressionRatings
│   │       └── AdEffectivenessIndices
│   ├── WebAdMovieSet
│   │   └── {subject}
│   │       ├── SceneDescription
│   │       ├── ImpressionRatings
│   │       ├── AdPreferenceVotes
│   │       └── PreferenceRatings
│   └── subject_data.scv
├── (result)
│   ├── TVAdMovieSet
│   └── WebAdMovieSet
├── config.py
├── labels.py
├── pair_dist.py
├── plot.ipynb
├── plot.py
├── subject_selector.py
└── util.py

Data

There are two different sets of movies, with multiple categories.

We provide predicted decodings by measured and predicted voxel response.

1. Web ad movie set
   • Scene descriptions (data/WebAdMovieSet/{subject}/SceneDescription/, 40 subjects)
   • Impression ratings (data/WebAdMovieSet/{subject}/ImpressionRatings/, 40 subjects)
   • Ad effectiveness indices (data/WebAdMovieSet/{subject}/AdEffectivenessIndices/, 40 subjects)
2. TV ad movie set
   • Scene descriptions (data/TVAdMovieSet/{subject}/SceneDescription/, 28 subjects)
   • Impression ratings (data/TVAdMovieSet/{subject}/ImpressionRatings/, 28 subjects)
   • Ad preference votes (data/TVAdMovieSet/{subject}/AdPreferenceVotes/, 28 subjects)
   • Subjective preference ratings (data/TVAdMovieSet/{subject}/PreferenceRatings/, 14 subjects)

※In subjective preference ratings, IDR is also calculated for manual ratings, which were annotated from fMRI subjects.

Usage

Execute plot.py. The result pictures will be stored in result/ directory.

Option	Choices	Description
-m	web tv	Movie set as stimlus
-t	sd ir ae ap pr	Task decoded by predicted- and measured-response decoding

Example

python pair_dist.py -m web tv -t sd ir ae ap pr

You can execute the same program as above with plot.ipynb.

Code is formatted by Black Formatter in Visual Studio Code.