Hi Vasudev,The quality of the alignment certainly will influence the parcellation. You would align the images within subject using FLIRT, then the subjects with each other using FNIRT (that is non-linear). Consult the FSL documentation.All the best,AndersonOn 11 January 2017 at 15:10, Dev vasu <vasudevamurthy.devulapally@gmail.com > wrote:VasudevThanksDear Sir,Can i align rfMRI +DTI to each subjects Structural data using flirt, my only concern is whether such alignment will influence the parcellation that i would like to perform on group level.On 11 January 2017 at 13:55, Anderson M. Winkler <[log in to unmask]> wrote:AndersonAll the best,Hi Vasudev,Yes, you can. FLIRT works with images that have different types of contrast. Consider using FA when aligning the DTI, though, as opposed to the raw diffusion-weighted images.On 10 January 2017 at 15:10, Dev vasu <vasudevamurthy.devulapally@gmail.com > wrote:VasudevThanksDear Sir,Can i use FLIRT for aligning DTI+fMRI and sMRI ? , I am not sure if such multimodal alignment will work properly with FLIRT .On 7 January 2017 at 13:04, Anderson M. Winkler <[log in to unmask]> wrote:Hi Vasudev,For the alignment, have you tried FLIRT?Regarding the overall task, I came across this paper yesterday (via Twitter): https://arxiv.org/abs/1701.01315 . Perhaps it's something along these lines you need?All the best,AndersonOn 6 January 2017 at 14:00, Dev vasu <vasudevamurthy.devulapally@gmail.com > wrote:VasudevThanksDear Sir,Is there any approach that i can use for "multi modal data alignment " , i would like accomplish within the subject alignment of multiple modalities like Structural, functional and DTI scans ?.On 6 January 2017 at 12:59, Anderson M. Winkler <[log in to unmask]> wrote:Hi Vasudev,I'm unaware of a tool that would parcellate in the way as you'd like, although you should be able to do it yourself, e.g., by building a similarity matrix using some measure or feature of interest that would encompass all the modalities you are want, then applying a clustering algorithm.All the best,AndersonOn 5 January 2017 at 11:58, Dev vasu <vasudevamurthy.devulapally@gmail.com > wrote:VasudevThanksI have even read the work of Craddock and his Software tool PyclusterROI, but the software is only applicable for rfMRI not DTI scans, i would like to know if there is any other journal or software tool which can perform group wise multimodal parcellation of T1w, rfMRI,and DTI scans of 2 groups of subjects.Dear Sir,I have read the publication from Matthew Glasser, but the scripts from Glasser are valid if we have full brain scans of T1w and T2w images but there is a problem with my T2w data acquisition, I did no acquire full brain T2w images so i cannot use HCP scripts for my study.On 5 January 2017 at 11:52, Anderson M. Winkler <[log in to unmask]> wrote:Hi Vasudev,There are many different ways in which the brain can be parcellated, based on anatomy, function, and combinations thereof. The paper by Brett et al. (2002, http://www.nature.com/nrn/journal/v3/n3/full/nrn756.html ) had a different original purpose but it's useful here too in reviewing what different features can mean. And major multimodal parcellation effort using data from the HCP has just been published (Glasser et al, 2016, http://www.nature.com/nature/j ournal/v536/n7615/full/nature1 ).8933.html Given what you seem interested in clustering, perhaps consider as a starting point the work by Craddock et al (Human Brain Mapping, 2012, https://www.ncbi.nlm.nih.gov/pubmed/21769991 ). Some additional details and the scripts are available freely here and here.All the best,AndersonOn 4 January 2017 at 19:50, Dev vasu <vasudevamurthy.devulapally@gmail.com > wrote:VasudevThanksDear Sir,Is there any approach where i can perform Group wise multimodal parcellation, I want to include T1w,T2w,rfMRI, and DTI or 2 groups of subjects (25 healthy and 25 diseased controls with BLVP) and perform multimodal clustering on whole brain. If you have some Journals or some approaches which can useful in performing the analysis i would be really grateful for your help.On 4 January 2017 at 15:38, Dev vasu <vasudevamurthy.devulapally@gmail.com > wrote:---------- Forwarded message ----------
From: Dev vasu <vasudevamurthy.devulapally@gmail.com >
Date: 4 January 2017 at 13:58
Subject: Re: [FSL] Group Average parcellation
To: FSL - FMRIB's Software Library <[log in to unmask]>VasudevThanksI know that we can easily define a seed region on an Atlast ( MNI 152 or Harvard oxford atlas ) but i would like use clustering approach for whole brain parcellation.Dear Sir,I would like to measure the neural activity of visual vestibular interaction between healthy controls and patients with bilateral vestibulopathy ,for this i have thought of using seed based functional connectivity to examine visual vestibular changes in BLVP patients in my study,
I would like to perform whole brain clustering on 2 groups of subjects that i have ( 25 healthy and 25 BLVP ) and extract 600 - 700 ROIs using temporal correlation and i feel this is better way to define a seed region to investigate functional connectivity changes , kindly let me know if there is any thing wrong in my approach, your suggestions are greatly appreciated.On 4 January 2017 at 12:32, Anderson M. Winkler <[log in to unmask]> wrote:Hi Vasudev,What feature would you like to drive the parcellation? The ICA itself is a form of soft (fuzzy) parcellation. I'm unsure what you'd like to do.All the best,AndersonOn 2 January 2017 at 12:58, Dev vasu <vasudevamurthy.devulapally@gmail.com > wrote:VasudevThanksI would like to know the better approach for Group wise parcellation ( any previous Publications concerning the same topic would be greatly appreciated ).Dear FSL community,I have 50 Subjects ( 25 healthy controls and 25 patients with BLVP ), I have performed Group mean ICA , I would like to perform whole brain parcellation on Group Mean ICA of Healthy controls and Group mean ICA of patients with BLVP.