Dear Shary
Please note that KW has supplied extensive instructions within his
stat_threshold.m function.
To provide an example of the function usage you might wish to try
stat_threshold(690,[],1,50,[],3.7,[1.423, 0.633, 0.654, 0.478])
This will return a listing of P_val_extent for the clusters given as
adj-resel.
Bill Moorhead
Quoting "[log in to unmask]" <[log in to unmask]>:
> Dear all,
>
> I wonder whether any of you have used Dr. Moorehead's team's
> empirical method. I am trying to understand how the corrected
> cluster level p-values been corrected for underlying smoothness in
> their paper. I am sure I am doing something wrong and cannot get the
> correct results when I use the data from the paper (Table 2, the firs
> row where Adjusted cluster P value is 0.015).
>
> Based on what I read in the web site
> (http://www.math.mcgill.ca/keith/fmristat/). Here is how the
> variables should be entered for stat_threshold:
>
> stat_threshold(whole brain resels, voxels whole brain, smoothness,
> df, Peak p-value, cluster threshold, p_val_extent)
>
>
>
> I calculated whole brain voxels bu summing up the unadjusted voxels
> in each cluster (28489 voxels). The whole image RPV is 0.232x10e-3
> according to the table. So, I first entered the whole brain resels,
> number of voxels in whole brain, 1 for smoothness because I am using
> resels, and 50 for df and got the following output:
>
>
> -------------------------------------------------------------------------------------------------
>>> stat_threshold(0.323*10^-3, 28489, 1, 50)
> peak_threshold =
> 1.7941
>
> Cluster_threshold =
> 3.2618
>
> peak_threshold_1 =
> 4.5315
>
> extent_threshold =
> NaN
>
> extent_threshold_1 =
> 0.4014
>
> ans =
> 1.7941
>
> ---------------------------------------------------------------------------------------------------------------------
>
> In the following I re-entered the same data plus peak_threshold_1,
> Cluster_threshold and extent_threshold_1 taken from above to
> p-values as shown below
> --------------------------------------------------------------------------------------------------------------
>
>>> stat_threshold(0.323*10^-3, 28489, 1, 50, 4.5315, 3.2618, 0.4014)
> P_val_peak =
> 3.0331e-005
>
> P_val_peak_1 =
> 0.0500
>
> extent_threshold =
> NaN
>
> extent_threshold_1 =
> 0.0660
>
> ans =
> 3.0331e-005
>
> ------------------------------------------------------------------------------------------------------------------------------
>
>
> I also tried entering the adjusted cluster resel (0.633), unadjusted
> voxels in the same cluster (1229), 1 for FWHM, 50 for df. Peak
> -value, cluster threshold and p_val_extent were taken from the first
> output shown above (whole brain) and got the following output:
> ----------------------------------------------------------------------------------------------------------------------------
>
>
>
>>> stat_threshold(0.633, 1229, 1, 50, 4.5315, 3.2618, 0.4014)
> P_val_peak =
> 7.5990e-004
>
> P_val_peak_1 =
> 0.0500
>
> extent_threshold =
> NaN
>
> extent_threshold_1 =
> 0.0660
>
> ans =
> 7.5990e-004
>>>
>
>
> ----------------------------------------------------------
>
> I just cannot get the Adjusted cluster P value (0.015) shown in the
> table 2. I am sure I am missing something. But, what am I doing
> wrong?
>
>
> Has anyone used this method for any publication? If you, could let
> me know what I am doing wrong. Your help is highly appreciated.
>
>
> regards,
> Shary
>
|