 |
 |
Hi there, My colleague emailed back in December with some problems he was having with FEAT (email thread pasted below). Matthew concluded that the problem originated with a difference in how input data with low variance is masked between 4.0 and 4.1, and sent us a corrected film_gls, which worked perfectly on our Mac OS X system. We have recently moved our data to a server running Red Hat Linux Ver 4.5 (64bit). We installed the most recent version of FSL (4.1.2), and the same problem has popped up again (see attached image), so perhaps this fix was omitted from the most recent version? We would really appreciate it if you could send/make available the same fix for our new system. Thanks a million, Clare On 18 Dec 2008, at 13:33, Mennes, Maarten wrote: > Hi Matthew, > > thanks for getting back at me. 'Good' for me to know its only a bug > and not me doing something wrong. > I'm using Mac OSX 10.5 (leopard) running on an intel core. > > Thanks again! > Maarten > > > > -----Original Message----- > From: Matthew Webster [mailto:[log in to unmask]] > Sent: Thu 12/18/2008 7:58 AM > To: Mennes, Maarten > Subject: Re: Analysis inconsitencies between FSL4.0 and FSL4.1.2 > > Hi, > I've had a look at your data - the bug comes from a difference in > how input data with low variance is masked between 4.0 and 4.1. We'll > be issuing a patch in January, but I can send you a corrected film_gls > binary before then if you let me know what platform you are using... > > Many Regards > > Matthew > > > >> From: "Mennes, Maarten" <[log in to unmask]> >> Date: 15 December 2008 15:29:24 GMT >> To: "Steve Smith" <[log in to unmask]> >> Subject: RE: Analysis inconsitencies between FSL4.0 and FSL4.1.2 >> >> Dear Professor, >> >> the file is uploaded, the upload ID is 334053 and the filename is >> CSC-FSL412.tar.gz >> >> Thanks for the quick response! >> Maarten >> >> >> >> >> -----Original Message----- >> From: Steve Smith [mailto:[log in to unmask]] >> Sent: Mon 12/15/2008 10:09 AM >> To: Mennes, Maarten >> Subject: Re: Analysis inconsitencies between FSL4.0 and FSL4.1.2 >> >> Hi - I thought maybe this was down to a chance in the masking - but >> no, LOADS of voxels are zero in the latest version - clearly a >> problem >> here. Could you upload the full input data and output FEAT directory >> (for just 4.1.2 is enough) and we'll take a look. >> >> Please upload the files in a single compressed tarfile to >> http://www.fmrib.ox.ac.uk/cgi-bin/upload.cgi >> And then email me the upload ID. >> >> Cheers. >> >> >> On 15 Dec 2008, at 15:00, Mennes, Maarten wrote: >> >>> >>> Dear Professor Smith, >>> >>> I recently started working with FSL when moving to the NYU Child >>> Study Center, having worked with SPM before. In getting to know the >>> FSL way on doing fMRI analysis I found a strange difference between >>> results produced with FSL4.0 and FSL4.1.2. I had installed the >>> latest version on my computer but could not produce the same results >>> on some sample data from the lab as others had produced before. It >>> turned out that doing the same analysis with FSL4.0 did return the >>> same results, leaving us puzled about a possible cause for this. >>> >>> What seems to happen in the FSL 4.1.2 results is that in the zstat >>> maps some voxels are set to 0, while others just keep their z-value. >>> The setting to 0 seems to happen at random. Is this something that >>> you are aware of? >>> >>> I have attached the design files and zstat maps of the FSL 4.0 and >>> FSL 4.1.2 analysis. If you need additional information please let me >>> know. >>> >>> Sincerely, >>> Maarten Mennes >>> >>> ------------------------------------------------------------ >>> This email message, including any attachments, is for the sole use >>> of the intended recipient(s) and may contain information that is >>> proprietary, confidential, and exempt from disclosure under >>> applicable law. Any unauthorized review, use, disclosure, or >>> distribution is prohibited. If you have received this email in error >>> please notify the sender by return email and delete the original >>> message. Please note, the recipient should check this email and any >>> attachments for the presence of viruses. The organization accepts no >>> liability for any damage caused by any virus transmitted by this >>> email. >>> ================================= >>> < >>> design_FSL40 >>> .fsf><design_FSL412.fsf><zstat1_FSL40.nii.gz><zstat1_FSL412.nii.gz> >> -- A.M. Clare Kelly, Ph.D., Postdoctoral Research Fellow, NYU Child Study Center, 215 Lexington Ave, 14th Floor, New York, NY 10016