Hi - we use connected-components clustering:
1) threshold the Z-stat image at the specified Z threshold level (ie zero
anything under this level).
2) Find connected-component clusters (where connection is allowed in 3D
for any of a voxel's 26 neighbours).
3) Test each cluster's size against the size threshold derived from
standard Gaussian-random-field theory using the P-threshold set by the
user.
The code can be seen in fsl/stable-src/cluster/cluster.cc
Hope this answers the question! Thanks, Steve.
On Mon, 5 Aug 2002, TetsuoKoyama wrote:
> Dear
>
> I would like to ask you a simple question.
>
> I would like to know what the clustering algorithm is used for FEAR
> analysis?
>
> Looking forward to hearing from you.
>
> Regards,
>
> Tetsuo
> ----
> Tetsuo Koyama, MD, PhD
> Department of Neurobiology and Anatomy
> Wake Forest University School of Medicine
> Winston-Salem, 27157-1010, NC
> TEL: 336-716-9302 or 4284
> FAX: 336-716-4534
> E-mail: [log in to unmask]
>
Stephen M. Smith
Head of Image Analysis, FMRIB
Oxford University Centre for Functional MRI of the Brain
John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
|