Project work 1 (1st-dl 19th January at 10:00)
The main objectives of the first project work are that you get familiar with how fMRI data looks like and will be able to perform the standard fMRI data analysis using general linear modeling. The results from this work will also form the basis for the next weeks' project works, so it's important that you get this done. That said, no worries if you have more questions than answers after the first week.
I suggest you do the following:
- Get familiar with the fMRI dataset. Describe the spatial & temporal properties of the dataset and the stimuli/task that the subjects were performing. Don't copy-paste this from the article where the data was originally published but think yourself what are the relevant information needed to understand the data.
- Load the data to matlab. You might want to uncompress/extract the packed nifti files first. For reading the nifti images to matlab, use for example, readnifti.m (see attached files; originally from http://www.neuro.mcw.edu/~chumphri/matlab/)
- Visualize some part of the data (e.g., a few slices of the image data using 'imagesc.m' or plot a time-course of a few voxels).
- Make a design matrix based on the information about the stimuli. See below for an example of a hemodynamic response function (see attached 'hrf.mat', which includes also a version downsampled to the resolution of the fMRI data). Visualize the design matrix. You might find useful the matlab function: 'conv.m'
- Fit the general linear model to the data using the design matrix. Remember that standard fMRI data analysis is done voxel-by-voxel. You might find useful the matlab function: 'glmfit.m' (or '\' operator).
- Visualize results (e.g., an example T-map).
- Think whether you could improve the result by making some modifications to the design matrix. Could you include extra regressors?
Return the first version of your project work by Tuesday, January 19, at 10:00. Even if you have had trouble with the analysis, do return something that we can then discuss during the Tuesday's Q&A session. You will get 5 points for returning your preliminary results and actively contributing during the Q&A session.