Individual functional segregation maps gained from resting state fMRI group ICA
Poster No:
1462
Submission Type:
Abstract Submission
Authors:
Philipp Sämann1, Michael Czisch1
Institutions:
1Max Planck Institute of Psychiatry, Munich, Germany
First Author:
Co-Author:
Introduction:
Resting state fMRI (rsfMRI) is broadly used in clinical settings, and while being informative on brain activity at the macro-network level, its challenge lies in its high dimensionality and a multitude of analysis concepts (despite marked progress regarding the standardization of preprocessing [1] for multi-site samples). In brief, this high dimensionality of rsfMRI data at the analytical level leads to many 'user degrees of freedom' (general analysis concept [independent component analysis [ICA], seed analysis, temporal lag analysis], various atlas systems) that hampers the development of rsfMRI biomarkers. Several methods exist to re-condensate the voxel-by-voxel functional connectivity (FC) information, for example FC density mapping that averages FC values of a voxel to all other voxels [2], or amplitudes of low frequency fluctuations [3]. However, such strong re-collapsing of information may be under-informative regarding functional integration vs. segregation aspects. Here we suggest a new type of map (functional segregation map [FSM]) that informs on the certainty of the 'network membership': low values indicate unclear membership status, whereas high values include clear assignment to a functional network.
Methods:
Sample, Preprocessing and group ICA: Using a subsample of N=100 healthy subjects of the BeCOME Study, a deep phenotyping study focusing in anxiety and depression [4], we first performed state-of-the-art preprocessing (motion and slice timing correction, spatial normalization using DARTEL after segmentation of a T2-EPI, motion correction and first residualisation, followed by aCompCor based phyiosological denoising and spatial smoothing 6 x 6 x 6 mm3) and group ICA [infomax algorithm, Matlab based GIFT) to extract 42 networks. ICASSO was used to gain network stability metrics. The latter number (lower than e. g. suggestions of 75 networks [4]) was chosen empirically as the preprocessing already corrected for artefacts and physiological noise, and as the dimensionality question is not key to demonstrate the principle. Calculation of functional segregation map (FSM): Resulting components were analyzed at the group level and for two individuals (after reconstruction of individual components). Fast Fourier transformation spectra were analyzed to differentiate neural vs. noise components using a low/high frequency power ratio. Eventually, per voxel the component with the highest Z-value was searched and the difference between this value and the second highest Z-value calculated. We refer to this value as the difference measure (DM).
Results:
At the group level, the DM maps revealed a large set of mainly cortical regions that showed markedly high DM values ( Fig. 1A). Depending on the thresholding and displaying, border regions were more attached to the detected hubs ( Fig. 1B). As successful proof of concept, DM values in the WM and CSF spaces (despite contained in the mask) were very low. Defining the frequency ratio for component selection more lenient (low/high ratio 2 compared with 1.5 [standard]) led to false high values in the ventricular space ( Fig. 1C). Two examples of individual FSMs generated with the same principle show marked differences between participation of the frontomesial/frontopolar region (yellow arrow) and the precuneus area (light blue arrow) ( Fig. 2A-B).
·Difference measure as calculated from the group ICA result. (A) and (B) represent differently scaled color systems. (C) demonstrates the effect of not neglecting the noise components.
·Two individual DM maps exhibiting marked hub expression differences
Conclusions:
Using group ICA we demonstrate a bi-modal distribution of regions regarding the certainty of network membership: About 40% of the cortex seem to represent areas with high certainty (high DM) of the voxel/network assignment, but sharply delineated remaining areas with low DM are interposed. While the general topology was detectable in individual subjects, the DM hubs were differently expressed in individuals, which is promising for fMRI phenotyping, biotyping and automated classification. Further data are needed to clarify the influence of the ICA dimensionality and the noise component criteria.
Modeling and Analysis Methods:
Connectivity (eg. functional, effective, structural) 2
fMRI Connectivity and Network Modeling 1
Methods Development
Keywords:
FUNCTIONAL MRI
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
[2] Tomasi D, Volkow ND. Functional connectivity density mapping. Proc Natl Acad Sci U S A. 2010;107):9885-90
[3] Zou QH., Zhu CZ., Yang Y, Zuo XN., Long XY, Cao QJ, Wang YF, Zang YF. An improved approach to detection of amplitude of low-frequency fluctuation (ALFF) for resting-state fMRI: fractional ALFF. J. Neurosci. Methods. 2008;172:137–41
[4] Allen EA et al. A Baseline for the Multivariate Comparison of Resting-State Networks. Front Syst Neurosci. 2011;5