My main suggestion would be to use SPM8 - rather than SPM5. The manual
says rather more about setting up a VBM analysis. After this, the main
thing to consider would be how you plan to do global normalisation to
model out the variability of brain (or TIV) sizes.
The processing (with the new segmentation and DARTEL) is rather simpler
in SPM8.
Best regards,
-John
On Fri, 2009-07-31 at 19:43 +0200, Enrico D'Ambrosio wrote:
> Dear John and SPM users,
> I would like to run a VBM analysis in SPM5 using the "unified
> segmentation" and to normalize the segments by using a DARTEL
> template.
>
> Could you please telling me if the following parameters are correct?
>
> 1) Segmentation
>
> TASKS --> Spatial --> Segment
>
> ---Output files---
> Grey matter: Native Space
> White matter: Native Space
> Cerebro-Spinal fluid: None
> Bias corrected: Don't Save Corrected
> Clean up any partitions: Don't do cleanup
>
> ---Custom---
> Tissue probability maps:
> /usr/local/spm5/tpm/grey.nii
> /usr/local/spm5/tpm/white.nii
> /usr/local/spm5/tpm/csf.nii
> Gaussians per class: 2 2 2 4
> Affine Regularisation: ICBM space template
> Warping Regularisation: 1
> Warp frequency cut-off: 25
> Bias regularisation: very light
> Bias FWHM: 60 mm cut off
> Sampling distance: 3
> Masking image: none
>
> 2) Creation of DARTEL template
>
> TASKS --> Tools --> DARTEL Tools --> Initial Import
>
> Bounding box: NaN NaN NaN; NaN NaN NaN;
> Voxel size: 1.5
> Image option: none
> Grey Matter: Yes
> White Matter: Yes
> CSF: No
>
> TASKS --> Tools --> DARTEL Tools --> Run DARTEL (create Templates)
>
> Regularisation Form: Linear Elastic Energy
> Outer Iterations: default settings
> Optimisation Settings
> LM Regularisation: 0.01
> Cycles: 3
> Iterations: 3
>
> TASKS --> Tools --> DARTEL Tools --> Create Warped
>
> Modulation: Modulation
> Time Steps: 64
> Interpolation: Trilinear
>
> 3) Normalize using DARTEL
> TASKS --> Spatial --> Normalise --> Normalise: Estimate
>
> Template Image: /usr/local/spm5/tpm/grey.nii
> Template Weighting Image: None
> Source Image Smoothing: 8 mm
> Template Image Smoothing: 0
> Affine Regularisation: ICBM space template
> Nonlinear Frequency Cutoff: Inf
> Nonlinear Interations: 0
> Nonlinear Regularisation: Inf
>
>
> Have you got any suggestion?
>
> Thank you, Enrico D'Ambrosio
>
--
John Ashburner <[log in to unmask]>
|