Welcome to Food Regulation fMRI’s documentation!¶
Links¶
- Source code: https://github.com/danieljwilson/cogReg_fMRI
- Contact: daniel.j.wilson@gmail.com
Getting Started¶
This documentation will cover some of the basics of the project, and hopefully put you in a position to run the experiment and analyze data yourself.
Running the Experiment¶
The experiment protocol can be found on Dropbox at
DJW_Projects/02_FOOD_REG/PAPERWORK/fMRI_Food_Regulation_Experiment_Protocol.gdoc
The protocol includes information regarding:
- Booking Scanner Time
- Recruiting Participants
- Running Study
The study questionnaire is a Google Form, and lives on the lab’s Google Drive account.
Data Storage¶
All data (raw and processed) are stored on external hard drive CH_ext_001.
All files are in the folcer 2019_FoodReg_fMRI/.
Note that this drive is password protected.
Updating Documentation¶
Note
Keep in mind that if you are using the Read the Docs documentation there is always the option to add/edit.
Just look for this image on the top right of the page:
Click on it (which automatically forks it), make your edits and then create a pull request.
Folder Structure¶
The folder structure for the project follows the format illustrated below:
1_project_management¶
This is not uploaded to git.
2_ethics_governence¶
This is not uploaded to git.
3_experiment¶
This is where most of the project lives.
3_1_inputs¶
Refers to the tools used to capture information, including:
- Experiment code
- including all assets (e.g. photos)
- Questionnaires
3_2_data¶
Raw data lives here.
3_3_data_analysis¶
This includes:
- Scripts for preprocessing and cleaning data
- Processed data
- Scripts for analyzing processed data
4_dissemination¶
Presentations, publications and publicity live here.
docs¶
There is also a folder that has been added by Sphinx, and is where all the documentation lives.
Note
Keep in mind that if you are using the Read the Docs documentation there is always the option to add/edit.
Just look for this image on the top right of the page:
Click on it (which automatically forks it), make your edits and then create a pull request.
Experiment Details¶
Protocols¶
The experiment protocol is located on Dropbox at:
DJW_Projects/02_FOOD_REG/PAPERWORK/fMRI_Food_Regulation_Experiment_Protocol.gdoc
The fMRI scan protocol is here.
Main Task¶
The main task involved three discreet phases.
- Pre-Scan
- Food liking ratings
- Main task training
- Scan
- In-scanner trials
- 9 runs
- In-scanner trials
- Post-Scan
- Food liking ratings (repeated)
- Food taste ratings
- Food health ratings
The code for the Main Task (including the instructions)
was written in MATLAB and uses Psychtoolbox.
The Main Task Code includes many files, but the key scripts are:
runPreMRI.m- Launches the experiment instructions and the initial Food Liking Rating
runSession.m- Launches a run of the experiment in-scanner
runPostMRI.m- Launches the post task ratings of Liking, Taste, and Health
Localizers¶
The localizer task invovled two discreet phases.
- Pre-Scan
- Localizer task training
- Scan
- In-scanner trials
- go-nogo: 2 runs
- switching task: 1 run
- In-scanner trials
The code for the Localizers was written in Python,
using the Psychopy toolbox.
Go-NoGo¶
The Go-NoGo task scripts include both a practice and fMRI version.
The main difference is that the fMRI version waits for the scanner
to send a 5 to progress.
The Go-NoGo task was based on Wager et al. 2005. We used the letters ‘m’ and ‘w’ as the ‘go’ and ‘no-go’ stimuli, requiring the execution or withholding, respectively, of a keypress response (counterbalanced).
After presentation of a 500ms fixation cross, participants had 450ms to respond to the stimulus.
There were two types of blocks presented: low-go blocks, in which 20% of the trials required a response, and high-go blocks, in which 50% of trials required a response. The beginning and ends of these blocks were not indicated to participants.
A total of 24 blocks (12 of each condition), containing 12 trials each–a total of 288 trials–were presented. The rapid event-related design with clustered events is expected to maximize power (Liu 2004). The task was broken into two equal (144 trial) sessions for scanning to reduce fatigue.
Switching Task¶
The switching task scripts also include both a practice and
fMRI version. The main difference is that the fMRI version waits
for the scanner to send a 5 to progress.
The attention switching task showed subjects a pair of images, one face and one house, on each trial. The images were overlaid directly on top of each other with each image’s opacity reduced so that both images could be clearly deciphered. On each trial subjects were directed to focus their attention either on the Face or the House image, which was indicated both by text on screen indicating “Face” or “House” and the background color of the image (i.e. a different background color for Faces andHouses). On Face trials, subjects had to determine the face’s gender, using a keypress to indicate their response. On House trials, subjects had to indicate whether it was an old or modern house, using a keypress to indicate their response.
There were four total response possibilities, and four corresponding buttons to press. Participants had up to 1 second to respond. The inter trial interval was between 1s and 6s, uniformly distributed. A total of 80 trials were presented in a single session.
Questionnaires¶
Upon completion of all tasks we asked subjects to complete the following questionnaires:
There is also a questionnaire key, that will be helpful for data analysis.
Data Analysis¶
fMRI Preprocessing¶
fmriprep¶
fmriprep is a pipeline developed by the Poldrack lab at Stanford University for use at the Center for Reproducible
Neuroscience (CRN), as well as for
open-source software distribution.
fmriprep is designed to provide an easily accessible,
state-of-the-art interface that is robust to variations in scan acquisition
protocols and that requires minimal user input, while providing easily
interpretable and comprehensive error and output reporting.
It performs basic processing steps (coregistration, normalization, unwarping, noise component extraction, segmentation, skullstripping etc.) providing outputs that can be easily submitted to a variety of group level analyses, including task-based or resting-state fMRI, graph theory measures, surface or volume-based statistics, etc.
The fmriprep workflow takes as principal input the path of the dataset
that is to be processed.
The input dataset is required to be in valid BIDS (Brain Imaging Data
Structure) format, and it must include at least one T1w structural image and
(unless disabled with a flag) a BOLD series.
We highly recommend that you validate your dataset with the free, online
BIDS Validator.
The exact command to run fmriprep depends on the Installation method.
The common parts of the command follow the BIDS-Apps definition.
Example:
fmriprep data/bids_root/ out/ participant -w work/
GLMs¶
General linear model scripts were run using MATLAB and SPM 8.
The model regressors are specificed by the files ending in analyze2.
The contrasts are calculated in the files that start with contrast2.
The second level/group analyses are performed by the rfx_par script.
While the analyze and contrast scripts can be run just with the
function, you need to use the following syntax to run the rfx_par
script. Note that you need to provide access to a contrast file.
Example:
f = fullfile('8_pre_liking', preproc_version, 'm8_pre_liking_cons.mat');
load(f);
for con = 1:length(cname)
rfx_par('8_pre_liking',cname(con),good_subjects,preproc_version)
end
Behavioral¶
A Jupyter notebook (using an R kernel) for the behavioral results.
DDM¶
We fit both a base model and constant model (with an additional constant parameter that is added to the drift).
The DDM model scripts.
A Jupyter notebook (using an R kernel) for the ddm results.
Neural¶
A Jupyter notebook (using an R kernel) for the neural results.
Correlations¶
TO BE ADDED…