GLM-Express
This is a package for modeling functional neuroimaging tasks. As the name implies, it's optimized to be simple and straightforward! The task_information.json
file stores all of the regressors and modeling specifications for each task; modifying this file allows you to test a range of analytical outcomes.
Included
This package comes equipped with the following modeling objects:
-
Subject
is a first-level modeler for subject-specific functional neuroimaging data -
GroupLevel
is a second-level modeler that is optimized to aggregate contrast maps derived by theSubject
object -
Aggregator
applies first-level models to all subjects in your BIDS project (not efficient for larger datasets) -
RestingState
for analyses of subject-level resting state functional connectivity
Assumptions
We assume the following about your data:
- Your data is in valid
BIDS
format
- Your data has been preprocessed via
fmriprep
- Your preprocessed data are stored in a
derivatives
folder nested in yourBIDS
project
- You have adequate events TSV files for all of your functional tasks
- Any parametric modulators are stored within each event file
- Otherwise, you can build custom design matrices and feed them into the modeling function
task_information.json
About Glossary of keys in the task_information
file; manipulating these allow you to quickly and effectively customize your modeling parameters without editing any source code
-
tr
: Repetition time (defined here, but can be overriden for any one subject) -
excludes
: Subjects in your project you need to exclude for a given task -
condition_identifier
: Column in your events file that denotes trial type; NOTE this will be changed totrial_type
in the script -
confound_regressors
: Regressors to include fromfmriprep
output -
modulators
: Parametric modulators to weight trial type (these should be in your events file) -
block_identifier
: Column in your events file that denotes block type; defaults tonull
-
design_contrasts
: Your defined contrasts! Include as few or as many as you see fit