Running recon-all on data collected at 7T

See also: SubmillimeterRecon

Bias correction prior to recon-all

Compared to 3T MPRAGE and MEMPRAGE scans collected at 7T have stronger intensity inhomogeneity, which won't be fully corrected in the default recon-all. Therefore, to process 7T data we recommend the following steps:

1) Bias-correct your input images using SPM12's unified segmentation algorithm (Ashburner and Friston, 2005)

2) Run recon-all as usual with your bias-corrected image as input. (If your 7T voxel size is less than 1mm3 and you would like to process data at native resolution, see SubmillimeterRecon page on how to run recon-all)

Recommended bias-correction parameters:

For Martinos Center MEMPRAGE with FOCI pulses

If the gray matter of the temporal lobes appears particularly dark, your pial surfaces may be misplaced in those regions as illustrated in the example below.

pial_surface_misplacement.jpg

To fix this, examine the image intensities in problematic regions by clicking around and add a line to your expert options file specifying maximum intensities of the gray matter and CSF as follows:

mris_make_surfaces -max_gray_at_csf_border 60 -max_csf 35

putting the values that you found by examining your image instead of 60 and 35. For more explanations see this email thread.