Dear Michael,

Unfortunately, I am now confused about the slice timing altogether. My basic understanding is that an interleaved acquisition is, by default, ascending (from foot to head, F->H) with odd slices first. The Siemens scanner have a different convention, in that if the number of slices is odd, it starts by odd slices, and if the number of slices is even, it starts by even slices. Anyway, I don't think slice timing correction has much of an impact on slow temporal fluctuations, but with the current infos found in the release table (which is missing the scanner type) it may not be possible to get the slice timing right.

As a side comment, one also has to take into consideration how the volume is stored. Depending on the software you are using, slice timing correction is usually applied in the voxel space. So if the acquisition was F->H (ascending), but the first slice in the volume is on the H side, this needs to be coded as a descending scheme ... I wish there were a reliable way to derive slice timing information ...

Best regards,

Pierre[/userquote]