Regarding understanding different fMRI dataset

[NidhiSultan]

HI @akcarsten, Thanks for explaining everything so easily, I tried the code you have provided to understand the nature of different fMRI dataset., mentioned below.

1. (http://www.cs.cmu.edu/afs/cs.cmu.edu/project/theo-81/www/)
2. (https://openfmri.org/dataset/ds000204/)

A. How do I analyze a mat format data like the one mentioned in 1 to understand which brain part is active? B. Also, the '2' dataset is in .nii format which I could utilize like the hdr dataset you explained. But got correlation as negative for this and a similar dataset which is of a normal human.

Mainly I am trying to utilize the methods explained to understand the active parts of the brain.

Thanks for the help, your guidance would be really helpful to me.

Regards Nidhi

[akcarsten]

Hi @NidhiSultan,

1. the .mat indicates that the data is saved as a Matlab file. You can try importing it to Python with SciPy.
2. Its hard to tell you anything without knowing what kind of analysis you have done. If you upload the code somewhere I can have a look and give you some feedback.

Best, Carsten

[NidhiSultan]

Hi @akcarsten,

1. I was able to use the SciPy for loading and using .mat format. Though, my concern is that the data in it is not a similar format as mentioned in the data link provided by you.
2. I had used the same code as provided by you, to analyze the fMRI. It gave me a negative correlation and hence no region with any activity. Thanks for the prompt reply.
3. If you think for the different datasets we need to use different strategies to do the analysis, please explain that as well. I am keen on understanding fMRI more, any link for the same would be helpful as well.

Regards Nidhi

[akcarsten]

Hi @NidhiSultan,

As I said it is hard to give you concrete advice without having seen your code. The data in the .mat file is from a different experiment, so you will have to adapt the design matrix to match the experimental conditions. This means adapting the timings etc. Also in the linked data two stimuli were shown with the second one being of variable length ("...on the screen for four seconds, or until the subject pressed the mouse button, ..."). This has to be taken into account as well for the analysis. As a general remark: The notebooks in this repository are meant to outline the general process of fMRI data analysis. There are many things that are not covered in the notebooks but have to be taken into account if you want to perform a serious analysis. However there are toolboxes available for Python as well as for Matlab that have lots of useful features already implemented. I would advice you to switch to one of those libraries. Finally on the page that provides the Matlab dataset, code and documentation is linked which should be a good starting point. Further it provides references to peer-reviewed papers that implemented various ways of analyzing the dataset.

[NidhiSultan]

Hi @akcarsten, Thanks for the support. I'll research a bit on this and if required will come back with further doubts. Regards Nidhi