Most people should be using Jacobian scaled data, which is written as floating
point. There is also the option to use a different data type for smoothed
images, which was intended to get around this problem.
Best regards,
-John
On Tuesday 04 November 2008 15:43, Thomas Nichols wrote:
> Dear List,
>
> Off-list we solved Adam's problem with the VBM-non-isotropic/NS-toolbox
> error "Number of clusters does not much in spm_max and spm_clusters!" For
> the record, the problem was with discreteness in the statistic image
> (induced by char datatype of VBM processed data) combined with small
> clusters.
>
> spm_clusters returned a list of 20 clusters, but spm_max only reported 19
> maxima. This mismatch was due to a single 2-voxel cluster where both
> voxels had the same statistic value, and hence no unique maximum existed
> for that cluster. While you might say "What are the chances of that
> happening?", it's not so unexpected, given that SPM's segment writes out
> 8-bit (c1/c2/c3) images, so with a small dataset the number of unique
> statistic values can be far from infinite. Once Adam changed the datatype
> to float the problem went away.
>
> -Tom
>
> On Wed, Oct 22, 2008 at 11:55 PM, Satoru Hayasaka
<[log in to unmask]>wrote:
> > I took a look at spm_max_nS and spm_list_nS. The only reason I can think
> > of for your error message below is that spm_max and spm_clusters really
> > are producing different number of clusters. Make sure you have the
> > updated version of following functions:
> >
> > spm_bwlabel.m
> > spm_get_lm.m
> > spm_clusters.m
> > spm_max.m
> >
> > They all should have the date 03/12/16. If they are up to date and you
> > are still getting the same error, then probably the clustering function
> > (a compiled routine, I believe) may have something to do with the error.
> > But I am afraid that is beyond my expertise though...
> >
> > -Satoru
> >
> >
> > -----Original Message-----
> > From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] On
> > Behalf Of Thomas, Adam (NIH/NIMH) [E]
> > Sent: Wednesday, October 22, 2008 7:26 PM
> > To: [log in to unmask]
> > Subject: Re: [SPM] VBM2 toolbox: non-isotropic smoothness correction
> > error
> >
> > Satoru,
> >
> > The previous output I sent was still using the spm_getSPM.m from the VBM2
> > toolbox. My apologies. Below is the output of a completely redone
> > analysis with the VBM2 toolbox not installed. The same error seems to
> > persists.
> >
> > Thanks for any advice you can provide as to what might be causing this
> > error.
> >
> > -Adam
> >
> >
> >
> > SPM2: spm_results_ui (v2.42) 18:59:05 - 22/10/2008
> > ========================================================================
> >
> >
> > SPM2: spm_getSPM (v2.51) 18:59:06 - 22/10/2008
> > ----------------------------------------------------------------
> > SPM computation : ...initialising
> > SPM computation : ...done
> > SPM computation : ...done
> > ??? Error using ==> spm_max_nS at 60
> > Number of clusters does not much in spm_max and spm_clusters!
> >
> > Error in ==> spm_list_nS at 349
> > [N Z XYZ A Rs] = spm_max_nS(zscores,varargin{3}.XYZ,VRPV);
> >
> > Error in ==> spm_results_ui_nS at 33
> > spm_list_nS('list',SPM.xVol.VRpv,xSPM,hReg);
> >
> > Error in ==> spm_ns at 8
> > [hReg,xSPM,SPM]=spm_results_ui_nS;
> >
> > Error in ==> spm at 1189
> > evalin('base',varargin{2}(i).prog)
> >
> > ??? Error using ==>
> > spm('TBlaunch',get(gcbo,'UserData'),get(gcbo,'Value')-1), se
> > t(gcbo,'Value',1)
> > Error using ==> spm_max_nS at 60
> > Number of clusters does not much in spm_max and spm_clusters!
> >
> > ??? Error while evaluating uicontrol Callback
> >
> > > -----Original Message-----
> > > From: Thomas, Adam (NIH/NIMH) [E]
> > > Sent: Tuesday, October 21, 2008 7:36 AM
> > > To: [log in to unmask]; Satoru Hayasaka
> > > Cc: Thomas, Adam (NIH/NIMH) [E]
> > > Subject: Re: VBM2 toolbox: non-isotropic smoothness correction error
> > >
> > > Hi Satoru,
> > >
> > > Thanks for your reply. I get the same error using the nS toolbox (see
> > > below).
> > >
> > > Thanks,
> > > -Adam
> > >
> > > ------------------------------
> > > Use local RPV values to correct for non-stationarity of smoothness.
> > > ??? Error using ==> spm_max_nS at 60
> > > Number of clusters does not much in spm_max and spm_clusters!
> > >
> > > Error in ==> spm_getSPM at 581
> > > [N Z2 XYZ2 A2 V2R] = spm_max_nS(Z,XYZ,SPM.xVol.VRpv);
> > >
> > > Error in ==> spm_results_ui at 260
> > > [SPM,xSPM] = spm_getSPM;
> > >
> > > Error in ==> spm_results_ui_nS at 30
> > > [hReg,xSPM,SPM] = spm_results_ui;
> > >
> > > Error in ==> spm_ns at 8
> > > [hReg,xSPM,SPM]=spm_results_ui_nS;
> > >
> > > Error in ==> spm at 1189
> > > evalin('base',varargin{2}(i).prog)
> > >
> > > ??? Error using ==>
> > > spm('TBlaunch',get(gcbo,'UserData'),get(gcbo,'Value')-1),
> > > set(gcbo,'Value',1)
> > > Error using ==> spm_max_nS at 60
> > > Number of clusters does not much in spm_max and spm_clusters!
> > >
> > > ??? Error while evaluating uicontrol Callback
> > >
> > >
> > >
> > >
> > >
> > > On Mon, 20 Oct 2008 19:26:17 -0400, Satoru Hayasaka
> > >
> > > <[log in to unmask]> wrote:
> > > >Adam,
> > > >
> > > >The error was returned from spm_getSPM.m function, so this might be a
> > >
> > > change unique to VBM2 Toolbox. Try our nS toolbox
> > >
> > > >http://www.fmri.wfubmc.edu/cms/NS-General
> > > >and see if you get the same error.
> > > >
> > > >-Satoru
> > > >
> > > >-----Original Message-----
> > > >From: Thomas, Adam (NIH/NIMH) [E] [mailto:[log in to unmask]]
> > > >Sent: Saturday, October 18, 2008 3:33 PM
> > > >To: Satoru Hayasaka
> > > >Subject: FW: VBM2 toolbox: non-isotropic smoothness correction error
> > > >
> > > >Dr. Hayasaka,
> > > >
> > > >I'm having some trouble with the non-isotropic smoothness correction
> > >
> > > implemented in the VBM2 Toolbox. I posted a question to the SPM mailing
> > > list
> > > but didn't get a response. Do you have any idea what might be causing
> > > the
> > > error I describe below?
> > >
> > > >Thank you for your time.
> > > >-Adam Thomas
> > > >
> > > >--
> > > >Adam Thomas [log in to unmask]
> > > >Functional MRI Facility, NIMH/NIH/DHHS
> > > >10 Center Dr, Room 1D80
> > > >Bethesda MD. 20892-1148
> > > >Phone:301-402-6351
> > > >
> > > >
> > > >-----Original Message-----
> > > >From: Thomas, Adam (NIH/NIMH) [E]
> > > >Sent: Tuesday, October 14, 2008 5:22 PM
> > > >To: [log in to unmask]
> > > >Subject: VBM2 toolbox: non-isotropic smoothness correction error
> > > >
> > > >
> > > >I'm using the VBM2 Toolbox (v1.07) to perform a simple paired T-test,
> > >
> > > however when I select the non-isotropic smoothness correction I get the
> > > error
> > > below regarding a mismatch between the number of clusters from spm_max
> > > and
> > > spm_clusters. Can anyone explain what causes this error?
> > >
> > > >-------------------------------------
> > > >
> > > >Use local RPV values to correct for non-stationarity of smoothness.
> > > >
> > > >numCl =
> > > >
> > > > 35
> > > >
> > > >
> > > >numCl2 =
> > > >
> > > > 36
> > > >
> > > >??? Error using ==> spm_max_nS at 60
> > > >Number of clusters does not much in spm_max and spm_clusters!
> > > >
> > > >Error in ==> spm_getSPM at 581
> > > > [N Z2 XYZ2 A2 V2R] = spm_max_nS(Z,XYZ,SPM.xVol.VRpv);
> > > >
> > > >Error in ==> spm_results_ui at 260
> > > >[SPM,xSPM] = spm_getSPM;
> > > >
> > > >??? Error using ==> [hReg,xSPM,SPM] = spm_results_ui;
> > > >Error using ==> spm_max_nS at 60
> > > >Number of clusters does not much in spm_max and spm_clusters!
> > > >
> > > >--------------------------
> > > >
> > > >Thanks,
> > > >-Adam
> > > >
> > > >
> > > >--
> > > >Adam Thomas [log in to unmask]
> > > >Functional MRI Facility, NIMH/NIH/DHHS
> > > >10 Center Dr, Room 1D80
> > > >Bethesda MD. 20892-1148
> > > >Phone:301-402-6351
|