Hi Joann,

to assert that single time point results are comparable to results from subjects with more than 1 time point, the single time point needs to be send through a fake/artificial longitudinal stream. It will thus undergo the same processing steps (including interpolation to the template). The method first normalizes the pose (make the head upright and straight) and uses that as the template space. It then processes this image as base and constructs a longitudinal run by initializing from the base. These steps are identical to how you'd process other subjects with more than 1 time point.

About editing: after editing the cross, you always need to check and edit the base. You can see the base as another (very important) time point. In some cases, it may even be sufficient, to only edit the base (e.g. brainmask), which can save some time. Surfaces need to be accurate in the base, as all longitudinal time points start fine tuning based on that initial surface. See https://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalEdits for details. Also some edits are automatically transferred from cross to long, e.g. control points. Since the long images are in a different location (the base space), it may be tricky to find where the points went with respect to slice coordinates, but they are at the same anatomical location.

Best, Martin

On 04/27/2016 04:44 AM, Joann Poh wrote:

Dear All,

When I was inspecting the longitudinal data for single time-point subjects, it differed quite a bit from the cross section. Can someone please explain why this is/could be so? How is the template generated for a single time-point subject?

I'm using FS 5.3. I did my edits at the cross sectional, generated the template and ran the longitudinal processing. The edits I made in the cross was not carried over to the template, and it also did not show in the long.

Thank you very much.

Regards, Joann

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer