Hello! I'm trying to extract TSV information from my fsl preprocessed data using eventstats and am finding that the TSV curves that are associated with my active voxels do not match the template curve I specified, but instead just look like noise. I believe that the active voxels are correct as they match what would be expected (e.g. ACG activation to a target condition), so the breakdown seems to be in the extraction of the TSV. Has anyone else ever had this problem that could maybe give me some troubleshooting tips?
The _avg files are extracted first, and are then used in the correlations to determine the active voxels (_avg_percent are rescaled to percent signal change) so it would be surprising if they didn't match the shape of the template, especially since the active voxels match what you expect. Is there a TR specified in the input .bxh files and is it in milliseconds? Does the "timepoints:" list in the prefix_LOG_EPOCHAVG.txt file match what you're expecting? (these units are seconds)
Thanks Syam! I think that may have fixed the problem! My TR is 1.5 but the TR in the eventstats files was still at 2. I just reran one subject and the TSVs look great for that person now! Thanks again!
Eventstats on FSL data
Posted - Apr 07 2008 : 11:58:16 AM
