Back to PLS Helptest data information
Posted on 05/18/09 18:13:14
Number of posts: 2
Hello, I am a new user of PLS software. I just want to play with the test data, specifically fMRI data, which is download from this website. Is there any information or readme file for this data, i.e. what are the stimuli, and further descriptions of this experiment? Thank you very much.
BTW: I read the paper "Spatiotemporal analysis of event-related fMRI data using partial least squares" published in 2004. But the experimental setup described there doesn't seem to match the test data.
Jianing
BTW: I read the paper "Spatiotemporal analysis of event-related fMRI data using partial least squares" published in 2004. But the experimental setup described there doesn't seem to match the test data.
Jianing
Replies:
Untitled Post
I'm Online
Posted on 05/18/09 19:54:24
Number of posts: 291
As far as I know, some of the test data were generated manually just for test purpose. The activation map in the brain were not obtained by stimuli from any experiment. I do not have many information about those test data, since they are just served to demonstrate the procedures that I described in the User's Guide.
Untitled Post
Posted on 05/19/09 07:49:52
Number of posts: 394
The test data are from a simulation that is described in: Della-Maggiore et al, Neuroimage,17, 19-28 (2002). The empirical data used for the 2004 paper are not part of the test data set. It would be too large to store on the FTP server, so we opted for the smaller simulated data.
Untitled Post
Posted on 05/20/09 10:29:22
Number of posts: 2
Thank you very much for your reply. Is there anyway to know where are those truly active pixels (since this is simulated data, we should have ground truth)?
Untitled Post
I'm Online
Posted on 05/20/09 12:29:23
Number of posts: 291
Active voxels can be defined as those having salient Brain LV values, including both positive values and negative values.
Here's a simple example from our Testdata:
By the way, you can also adjust the positive/negative threshold if you think it is necessary. This will of cause change the activation map.
Here's a simple example from our Testdata:
- Using dataset_001 and onsets.txt to create s1_fMRIsession.mat/s1_fMRIdatamat.mat for subject1; Using dataset_002 and onsets.txt to create subject2; Do same thing to create subject3, subject4, and subject5;
- Put all five subjects into one group; Run Mean-Centering PLS analysis with 500 permutation tests and 81 bootstrap tests;
- You can find active voxels from the result file by yourself, here are some of the voxels: [41 45 2] on Lag2, [24 20 5] on Lag1, etc. I don't want to enumerate all of them.
By the way, you can also adjust the positive/negative threshold if you think it is necessary. This will of cause change the activation map.
Login to reply to this topic.
