 |
 |
Hi,
I suggest running single-subject ICAs first, just to see if the data quality per se is OK or if you happen to have strange artefacts in your data. If you cannot make this work then feel free to upload a single data set so we can see what's going on...
hth
Christian
On 31 May 2013, at 11:53, Ahmad Omar <[log in to unmask]> wrote:
> Hi Steve,
>
> Thank you for your reply. Unfortunately, we are only getting what in some places are randomly scattered activations in the midbrain or concentrated activations in the frontal and orbito-frontal regions where there is already some epi-related distortions so not be relied upon . I was meaning to ask this follow-up question. Shouldn't I turn off temporal high pass filtering as I am interested in the ultra-low frequency BOLD
> signals? I tried a run doing that and the activation areas changed a bit to being around the ventricles and in the putamen which also looked strange to us.
>
> We very much appreciate your help on this.
>
> Ahmad
>
> Ahmad Omar
>
>
>
> On Thu, May 30, 2013 at 1:00 AM, Stephen Smith <[log in to unmask]> wrote:
> Hi - there's nothing in this description or the log that jumps out at me - did you find *some* RSNs that you recognised?
> Steve.
>
> On 29 May 2013, at 16:31, Ahmad Omar <[log in to unmask]> wrote:
>
>>
>> Re-send. Thanks.
>>
>>
>> Ahmad Omar
>>
>>
>>
>> ---------- Forwarded message ----------
>> From: Ahmad Omar <[log in to unmask]>
>> Date: Mon, May 27, 2013 at 4:53 PM
>> Subject: Resting state fMRI using MELODIC
>> To: FSL Mailing List <[log in to unmask]>
>>
>>
>> Hello All,
>>
>> I am trying to apply MELODIC (ver. 3.13, FSL ver 5.0.4 on a virtual machine) to some resting-state fMRI data we are acquiring here at the radiology center. The scanner is a 3T Siemens Skyra. The sequence is the standard ep2d sequence acquiring 150 time points with TR = 2.57 sec, TE = 36 msec, flip angle = 75 deg, distance factor = 20, PAT = 2, Fat Sat. on, phase partial Fourier off, bandwidth = 1698 Hz/Px and 1.7mm x 1.7 mm x 5 mm (in the z-axis direction) resolution with interleaved slice acquisition. Motion correction and temporal high pass filtering were enabled in the scanner sequence. When I view and loop through the raw data in FSL View, it seems fine. No excessive translations or head rotation and no severe intensity variations. In MELODIC I set the following settings: Delete Volumes = 3, High pass filter cutoff (s) = 100, Motion Correction : MCFLIRT, Slice timing correction: Interleaved, spatial smoothing FWHM (mm) = 5, Temporal filtering: Highpass, Threshold IC maps left at 0.5 and output the full stats folder was enabled. We were trying to reproduce the work done by Tang, L et al. in the paper "Thalamic Resting-State Functional Networks: Disruption in Patients with Mild Traumatic Brain Injury". MELODIC finishes processing fine but we don't know why we are not seeing any meaningful activations in say the Default Mode Network areas let alone activations in the thalamus. I have included below a part of the log.txt file in the filtered_func_data.ica folder.
>>
>> Thanks and your help is much appreciated.
>>
>> Melodic Version 3.13
>>
>> /usr/local/fsl/bin/melodic -i filtered_func_data -o filtered_func_data.ica -v --nobet --bgthreshold=3 --tr=2.5699999332 --report --guireport=../../report.html -d 0 --mmthresh=0.5 --Ostats
>> ---------------------------------------------
>>
>> Melodic results will be in filtered_func_data.ica
>>
>> Create mask ... done
>> Reading data file filtered_func_data ... done
>> Estimating data smoothness ... done
>> Removing mean image ... done
>> Normalising by voxel-wise variance ... done
>> Excluding voxels with constant value ... done
>>
>> Data size : 147 x 106223
>>
>> Starting PCA ... done
>> Start whitening using 49 dimensions ...
>> retaining 88.3302 percent of the variability
>> ... done
>>
>> Starting ICA estimation using symm
>>
>> Step no. 1 change : 0.422912
>> Step no. 2 change : 0.509312
>> Step no. 3 change : 0.458818
>> Step no. 4 change : 0.490924
>> .
>> .
>> .
>> Step no. 63 change : 5.50092e-05
>> Step no. 64 change : 5.21687e-05
>> Step no. 65 change : 5.06082e-05
>> Step no. 66 change : 4.94706e-05
>> Convergence after 66 steps
>>
>> Sorting IC maps
>>
>> Writing results to :
>> filtered_func_data.ica/melodic_IC
>> filtered_func_data.ica/melodic_Tmodes
>> filtered_func_data.ica/melodic_mix
>> filtered_func_data.ica/melodic_FTmix
>> filtered_func_data.ica/melodic_PPCA
>> filtered_func_data.ica/melodic_ICstats
>> filtered_func_data.ica/mask
>> ...done
>> Creating report index page ...done
>>
>>
>> Running Mixture Modelling on Z-transformed IC maps ...
>> IC map 1 ...
>> calculating mixture-model fit
>> saving probability map: filtered_func_data.ica/stats/probmap_1
>> saving mixture model fit: filtered_func_data.ica/stats/MMstats_1
>> re-scaling spatial maps ...
>> thresholding ...
>> alternative hypothesis test at p > 0.5
>> saving thresholded Z-stats image: filtered_func_data.ica/stats/thresh_zstat1
>> creating report page ... done
>> .
>> .
>> .
>> IC map 49 ...
>> calculating mixture-model fit
>> saving probability map: filtered_func_data.ica/stats/probmap_49
>> saving mixture model fit: filtered_func_data.ica/stats/MMstats_49
>> re-scaling spatial maps ...
>> thresholding ...
>> alternative hypothesis test at p > 0.5
>> saving thresholded Z-stats image: filtered_func_data.ica/stats/thresh_zstat49
>> creating report page ... done
>>
>> Ahmad
>>
>
>
> ---------------------------------------------------------------------------
> Stephen M. Smith, Professor of Biomedical Engineering
> Associate Director, Oxford University FMRIB Centre
>
> FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
> +44 (0) 1865 222726 (fax 222717)
> [log in to unmask] http://www.fmrib.ox.ac.uk/~steve
> ---------------------------------------------------------------------------
>
> Stop the cultural destruction of Tibet
>
>
>
>
>