 |
 |
Dear List, We have data from a long FMRI-session (896 volumes at 3.66 sec TR). In general, motion has not been a big problem (<1mm), but we have a handful of subjects with nasty point-like movement artefacts. They are higly likely due to coughing during a volume acquisition: the symptoms are a "grated" appearance, where, due to interleaved imaging, consecutive slices have different spin excitation histories. Timecourses of voxels show sharp intensity gradients. We would like not to have to throw these subjects away. Expectedly, mcflirt cannot cope with such radical events. My first instinct was simply to go through the data (in AFNI), mark the one or max two (consequtive) volumes with the "grated" appearance and splice them out with avwroi (naturally adjusting the design matrix accordingly). We would pretend those volumes were never acquired. However, a colleague pointed out that an option would be to interpolate the contaminated volumes, i.e. replace them with the average of the preceeding and succeeding volumes. This would then possibly not affect the autocorrelations in the data. This could, however, be considered more severe tampering of the data. Any comments appreciated, Christopher Bailey -- Research Assistant (MSc) Center for Functionally Integrative Neuroscience (CFIN) Aarhus University Hospital, Denmark [log in to unmask] phn: +45 89 49 43 78 fax: +45 89 49 44 00