jangwonjin / pines Goto Github PK

The goal of Pattern Recognition Grand Challenge (PRGC) 2018 is to predict a subject's emotional state from the fMRI image of his or her brain. In the PINES dataset below, 183 subjects' fMRI images and emotional states are provided. Build a model to predict a subject's emotional state (Y) from an fMRI image (X).

• See here for a breif introduction about this study.
• For more detail, read the Method section of the original article

https://neurovault.org/collections/503/

This dataset contains single trial responses to 30 images from the International Affective Picture Set from 182 subjects. These data were used to train the Picture Induced Negative Emotion Signature described in Chang et al., 2015.

The details of the experiment can be seen in Gianaros et al., 2014 (doi:10.1016/j.biopsych.2013.10.012).

• Negative photographs (Pictures: 2053, 3051, 3102, 3120, 3350, 3500, 3550, 6831, 9040, 9050, 9252, 9300, 9400, 9810, 9921) depicted bodily illness and injury (10 photographs), acts of aggression (2 photographs), members of hate groups (1 photograph), transportation accidents (1 photograph) and human waste (1 photograph).
• Neutral photographs (Pictures: 5720, 5800, 7000, 7006, 7010, 7040, 7060, 7090, 7100,7130, 7150, 7217, 7490, 7500, 9210) depicted inanimate objects (10 photographs) or neutral scenes (5 photographs).

• S1_Data.csv from https://s3-eu-west-1.amazonaws.com/pstorage-plos-3567654/2129473/S1_Data.csv
• Subject's reponses are stored in column Rating.
• Use Holdout column to divide the data into training (Trainig) and test (Test) datasets.

• loadData.ipynb - Shows you how to load the dataset

You can download all the data using downloadData.py.

In order to read the provided fMRI images, you need nibabel package. Here is how you can install it.

See http://nipy.org/nibabel/gettingstarted.html for more detals about nibabel.

1. Create a virtual environment

   python3 -m venv ~/ni
   

2. Activate the virtual envrinonment

   source ~/ni/bin/activate
   

3. Install Jupyter

   pip3 install jupyter
   

4. Install nibabel in it

   pip3 install nibabel
   

Once you're done, you can activate the environment (as shown in Step 2) and use it to load *.nii.gz files.