 |
 |
Susceptibility issue is also a major problem of anterior temporal lobe (due to ear channels). The signal drops there dramatically (it might be a drop of 80-90%). People invent different protocols and scan sequences to overcome those issues. Here, for example, a paper which explains nicely the options: http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0008160 Other than that, there is a defaults.mask.thresh parameter in spm_defaults.m which defines a threshold for the SPM mask. In SPM 5 this value is set to 0.8. As far as I understand, this means that any voxel with signal value bellow 80% of global brain average signal would not be included in the 1-st level analyze mask. My experience is that lowering this mask to let's say 0.5 does not helps too much though a little more voxels indeed are included in the analyze. On Sat, Mar 12, 2011 at 7:10 PM, John Fredy <[log in to unmask]> wrote: > Hello SPMers, in my experience the dark spots in the basal area in the > temporal lobe means loss of signal because the suceptibility effects. The > BOLD effect is around the 5% of change of signal in 3T, it cannot explain > the total loss of signal, please correct me if I am wrong. > > John Ochoa > > > On Sat, Mar 12, 2011 at 9:26 AM, Vladimir Bogdanov <[log in to unmask]>wrote: > >> Dear Roberto Viviani, >> >> Thank you for your reply. >> >> Though, I did not understand the last sentence. "It adds to eny >> > exeisting inhomegenity; if this latter has led to signal >> > loss, you'll have little from which get the BOLD signal." >> >> I have no deep knowledge of the physics of the processes. I am interested >> if the "dark spots" mean the loss or decrease of BOLD responses? Or, >> alternatively, the BOLD response is independent of the baseline level of the >> signal. It can be just additive. In this case and the only difference >> between "dark" and "normal" areas is in the "baseline" of the signal. >> >> Please correct me if my logic is wrong. >> >> Sincerely yours, >> Vladimir >> >> --- On Sat, 3/12/11, [log in to unmask] < >> [log in to unmask]> wrote: >> >> > From: [log in to unmask] <[log in to unmask]> >> > Subject: Re: [SPM] Explicit masking to improve signal in the OFC? >> > To: "Vladimir Bogdanov" <[log in to unmask]> >> > Date: Saturday, March 12, 2011, 2:18 PM >> > Dear Vladimir, >> > >> > the loss of signal has to do with the physics of the >> > generation of the MR signal. The effect of magnetic >> > inhomogeneity on the signal is the following: >> > >> > -- there is loss of signal (less with shorter TE's) >> > -- there is displacement of signal (irrespective of TE's) >> > >> > In typical T1-weighted structural images (typical in the >> > sense of being obtained with gradient recall, like EPIs), >> > these problems are less marked because the voxels are >> > smaller, and hence less homogeneous. >> > >> > The BOLD signal is itself the effect of voxel >> > inhomogeneity, modulated by oxygen content in blood, through >> > its effects on the Fe in the heme group. It adds to eny >> > exeisting inhomegenity; if this latter has led to signal >> > loss, you'll have little from which get the BOLD signal. >> > >> > Best >> > Roberto Viviani >> > >> > >> > Quoting Vladimir Bogdanov <[log in to unmask]>: >> > >> > > Dear Roberto Viviani, >> > > >> > > This is often the case in our studies. We indeed use >> > 3T scanner. There are a lot of visible "black holes" >> > in EPI images (see the picture). In some subjects it >> > is more pronounced, in others - less. >> > > >> > > What does it mean? There should be no task-related >> > BOLD signal responses from those areas? How to treat >> > those inhomogeneities? >> > > >> > > Thank you in advance for your opinion! >> > > >> > > Sincerely yours, >> > > Vladimir >> > > >> > > >> > > --- On Sat, 3/12/11, Roberto Viviani <[log in to unmask]> >> > wrote: >> > > >> > >> From: Roberto Viviani <[log in to unmask]> >> > >> Subject: Re: [SPM] Explicit masking to improve >> > signal in the OFC? >> > >> To: [log in to unmask] >> > >> Date: Saturday, March 12, 2011, 1:12 PM >> > >> Dear Deborah, >> > >> >> > >> before concluding that signal in the OFC is being >> > cut off, >> > >> I'd check that this is indeed the case by >> > overlaying the spm >> > >> image over the mean EPI image as a template. >> > Magnetic field >> > >> inhomogenity in that region results in the brain >> > shape in >> > >> the EPI images to differ from that of the T1 >> > template brains >> > >> used in displays, sometimes substantially. You >> > might >> > >> discover that you have no brain where you think >> > the OFC is. >> > >> This may be especially the case if you are using a >> > 3T >> > >> scanner. >> > >> >> > >> To mask your analysis, choose 'explicit masking' >> > from the >> > >> menu, after having prepared a 1/0 mask image for >> > selection. >> > >> >> > >> Best, >> > >> Roberto Viviani >> > >> >> > >> Quoting Deborah Tang <[log in to unmask]>: >> > >> >> > >> > Dear SPM Community, >> > >> > >> > >> > While analyzing results from a recent fMRI >> > study, I >> > >> have discovered >> > >> > that I have very little activation in the >> > lower >> > >> regions of the brain. I >> > >> > am concerned that SPM may have truncated my >> > data in >> > >> the lower regions >> > >> > of the brain because they were below >> > threshold. Since >> > >> I am particularly >> > >> > interested in activation in the OFC, this is >> > a >> > >> problem. Upon reviewing >> > >> > my data (ie. mask.img), it looks like SPM cut >> > out >> > >> brain activity in the >> > >> > OFC after first-level analysis. >> > >> > >> > >> > To resolve this issue, I am thinking of >> > trying to >> > >> force SPM to use an >> > >> > explicit mask of the entire brain. Does >> > anyone know >> > >> how to implement >> > >> > this in script format? >> > >> > >> > >> > Alternatively, if anyone has better ideas on >> > how to >> > >> improve signal in >> > >> > the OFC, please let me know. >> > >> > >> > >> > Many thanks in advance for all your help! >> > >> > Deborah Tang, Doctoral Candidate >> > >> > McGill University >> > >> > Montreal Neurological Institute >> > >> > 3801 University St, Room WB 214D >> > >> > Montreal, QC H3A 2B4 >> > >> > Canada >> > >> > >> > >> > Phone: (514) 398-6644 Ext. 089469 >> > >> > Fax: (514) 398-8948 >> > >> > [log in to unmask] >> > >> >> > > >> > > >> > > >> > >> > >> > >> >> >> >> >