BVQXtools code samples
From BVwiki
Contents |
Requirements
All samples require the installation of the BVQXtools toolbox in the MATLAB path. The version with which the example has been created (which can be seen as the minimum requirement) is denoted in the example respectively.
Design matrices
Concatenation
Available as of version 0.5a
In case you have two (or several) runs and want to create a combined design matrix (that can be used for PSC/z-transformed concatenated time courses), there is a very simple way to come to this:
>> r1 = BVQXfile('run_1.rtc');
>> r2 = BVQXfile('run_2.rtc');
>> r1and2 = r1.Concatenate(r2);
>> r1and2.SaveAs('run_1and2.rtc');
As part of this process, ambiguous regressor names will be appended with the study (run) number. The result then looks like this
| + | 
| |
| = | 
| ||
GLM's
Concatenating FFX GLMs
Available as of version 0.5c In case you have many, many subjects and yet want to use full (per-study) predictor separation, the calculation may take a while. BrainVoyager QX actually creates a design matrix that allows to code between-subject interactions, which normally (unless you explicitely and manually do so) are not part of the design matrix. Hence the calculus can be sped up by splitting the MDM in some partitions, calculate the FFX models separately and then joining them together again:
>> mdm = BVQXfile('study.mdm'); % read MDM
>> mdm.NrOfStudies = 10; % take only first ten
>> mdm.XTC_RTC = mdm.XTC_RTC(1:10,:);
>> mdm.SaveAs('first10.mdm');
% (...) do so for 11-20, etc.
>> glm = BVQXfile('first10.glm');
>> glm = glm.JoinFFX(BVQXfile('second10.glm'));
>> glm = glm.JoinFFX(BVQXfile('remaining.glm'));
>> glm.SaveAs('fullgroup.glm');
VMRs
Creation of the Talairach VMR
Available as of version 0.5b
The TalairachBrain.vmr file can be, principally, self generated from the Talairach Daemon database with the following commands
>> vmr = BVQXfile('vmr');
>> vd = vmr.VMRData;
>> ba = [1:13, 17:25, 27:47];
>> for bc = 1:length(ba), bvc = tal2bv(tdlocal(6, sprintf('Brodmann area %d', ba(bc))))';
for cc = 1:size(bvc), vd(bvc(cc,1)+1, bvc(cc,2)+1, bvc(cc,3)+1) = uint8(226 + mod(bc-1, 20)); end
end
>> vmr.VMRData = vd;
>> vmr.SaveAs('TalBrainAlternative.vmr');
On the first call to tdlocal, a cache file will be created that, for future sessions, will speed up this a lot!
Caution: the resulting VMR file will not be 100 per cent identical!
Cleaning, and brain peeling of VMR
Available as of version 0.7a, Snapshot 37 (Aug/26/07)
The code is relatively simple (for a better understanding, check the BVQXtools SBS VMR cleaning guide).
Part 1: cleaning the V16
>> v16 = BVQXfile('*.v16');
>> v16_clean = v16.CleanVMR;
>> v16_clean.LimitVMR;
>> v16_clean.SaveAs;
Part 2: Peeling the VMR
>>vmr = BVQXfile('*.vmr');
>> vmr_peeled = vmr.PeelBrain;
>> vmr_peeled.SaveAs;
>> vmr = BVQXfile('*.vmr');
Caution: these functions are still experimental!
