Generation of open-source fMRI noise templates
Poster No:
1590
Submission Type:
Abstract Submission
Authors:
Kristian Galea1, Nina Attard Montalto2, Aitor Alberdi Escudero2, Doneka Lonaiz Aranguren3,2, Paola Galdi4, Kenneth Scerri5, Liam Butler2, Claude Bajada2,1
Institutions:
1University of Malta Magnetic Resonance Imaging Platform, University of Malta, Msida, Malta, 2Department of Physiology and Biochemistry, Faculty of Medicine and Surgery, University of Malta, Msida, Malta, 3Mondragon University, Mondragon, Spain, 4School of Informatics, University of Edinburgh, Edinburgh, United Kingdom, 5Department of Systems & Control Engineering, Faculty of Engineering, University of Malta, Msida, Malta
First Author:
Kristian Galea
University of Malta Magnetic Resonance Imaging Platform, University of Malta
Msida, Malta
University of Malta Magnetic Resonance Imaging Platform, University of Malta
Msida, Malta
Co-Author(s):
Nina Attard Montalto
Department of Physiology and Biochemistry, Faculty of Medicine and Surgery, University of Malta
Msida, Malta
Department of Physiology and Biochemistry, Faculty of Medicine and Surgery, University of Malta
Msida, Malta
Aitor Alberdi Escudero
Department of Physiology and Biochemistry, Faculty of Medicine and Surgery, University of Malta
Msida, Malta
Department of Physiology and Biochemistry, Faculty of Medicine and Surgery, University of Malta
Msida, Malta
Doneka Lonaiz Aranguren
Mondragon University|Department of Physiology and Biochemistry, Faculty of Medicine and Surgery, University of Malta
Mondragon, Spain|Msida, Malta
Mondragon University|Department of Physiology and Biochemistry, Faculty of Medicine and Surgery, University of Malta
Mondragon, Spain|Msida, Malta
Kenneth Scerri
Department of Systems & Control Engineering, Faculty of Engineering, University of Malta
Msida, Malta
Department of Systems & Control Engineering, Faculty of Engineering, University of Malta
Msida, Malta
Liam Butler
Department of Physiology and Biochemistry, Faculty of Medicine and Surgery, University of Malta
Msida, Malta
Department of Physiology and Biochemistry, Faculty of Medicine and Surgery, University of Malta
Msida, Malta
Claude Bajada
Department of Physiology and Biochemistry, Faculty of Medicine and Surgery, University of Malta|University of Malta Magnetic Resonance Imaging Platform, University of Malta
Msida, Malta|Msida, Malta
Department of Physiology and Biochemistry, Faculty of Medicine and Surgery, University of Malta|University of Malta Magnetic Resonance Imaging Platform, University of Malta
Msida, Malta|Msida, Malta
Introduction:
fMRI analysis is commonly preceded by multiple preprocessing steps such as head motion correction, rigid-body transformation and slice-timing correction. For analyses involving multiple participants,each subject's data is spatially normalised to a standard space, such as the MNI template (1–3). Although essential, preprocessing may propagate artifacts in the data (4). For example, preprocessing steps such as spatial normalisation make use of interpolation which has been shown to introduce artifacts, which we collectively call geometric effects (5,6). The lack of standard datasets to explore these effects across different preprocessing pipelines limits the reproducibility of artifact analyses. This study proposes the use of noise templates generated from fMRI data as standard data sets which may be used to investigate the magnitude of the geometric effects caused by fMRI data preprocessing. The templates consist of uncorrelated (voxel-wise and volume-wise) Rician noise, to mimic scanner-specific noise measurements generated from fMRI data of a participant (7,8).
Methods:
A volunteer was scanned with a 3T Siemens MAGNETOM Vida MRI scanner using an echo planar imaging (EPI) pulse sequence (repetition time=1.84s, echo time=34ms, flip angle=65°, 23∧mm∧3 isotropic voxels, 393 volumes). Using a morphological mask, only voxels residing in the background outside of the brain and skull were considered in the noise generation. Then, an empty array of voxels with the same dimensions as the EPI image was created as the noise template. For each voxel in the template, two separate samples were acquired from a Gaussian distribution. The mean and standard deviation of the Gaussian distribution were equivalent to those of the noise voxels. Then, to model Rician noise, the square root of the sum of squares of the random samples was computed. To simulate the time-series component of the fMRI data, a total of 393 noise volumes were generated to construct a 4-dimensional noise template composed of spatially and temporally uncorrelated Rician noise realisations. The procedure was repeated 100 times to simulate noise from 100 fMRI acquisitions. Furthermore, to confirm that the templates were generated with no correlations between voxels or volumes, Pearson correlation was computed. The procedure was carried out using Python, and the resulting noise files were saved as Python-compressed pickle files.
Results:
A total of 100 4-dimensional Rician noise templates (e.g. Figure 1a, overlaid on the EPI image) that simulate realistic noise distributions from 100 fMRI acquisitions were generated. Voxel intensities extracted from the background field of view outside the brain and skull had a mean(SD) of 22.59(69.58). The voxel-voxel Pearson's correlation using 500,000 random sample pairs of voxels (neighbouring) resulted in a mean correlation of 0 with a standard deviation of 0.051 while the temporal Pearson's correlation of volumes using the 10,000 random pairs of voxels resulted in a Pearson's correlation of 0 with a standard deviation of 0.001. Finally, one of the noise templates was used to populate the fMRI data of the volunteer (Figure 1b). The noise templates will be accessible on the OSF link https://doi.org/10.17605/OSF.IO/CQFZX.
Conclusions:
This work produced a publicly available dataset of 100 Rician noise templates generated from background noise from volunteer fMRI data. This open dataset could be used to consistently assess geometric effects resulting from preprocessing and analysis pipelines. Specifically, these datasets will (i) allow for consistent assessment of novel toolboxes and pipelines, (ii) facilitate reproducibility of results across studies and research groups, and (iii) allow novel pipelines to be consistently compared to already validated methods.
Modeling and Analysis Methods:
Motion Correction and Preprocessing 1
Neuroinformatics and Data Sharing:
Databasing and Data Sharing
Workflows 2
Keywords:
FUNCTIONAL MRI
Open Data
Spatial Normalization
Spatial Warping
Statistical Methods
STRUCTURAL 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.
1. Fonov V et al. Unbiased nonlinear average age-appropriate brain templates from birth to adulthood. NeuroImage. 2009 Jul 1;47:S102.
2. Evans AC et al. Brain templates and atlases. Neuroimage. 2012;62(2):911–22.
3. Brett M. The problem of functional localization in the human brain. Nat Rev Neurosci. 2002 Mar;3(3):243–9.
4. Farrugia C et al. Effects of preprocessing on local homogeneity of fMRI data. 2023 Jul
5. Farrugia C et al. Local gradient analysis of human brain function using the Vogt-Bailey Index. Brain Struct Funct. 2024 Mar 1;229(2):497–512.
6. Ciantar KG et al. Geometric effects of volume-to-surface mapping of fMRI data. Brain Struct Funct. 2022 Sep 1;227(7):2457–64.
7. Agrawal U et al. Model-based physiological noise removal in fast fMRI. NeuroImage. 2020 Jan 15;205:116231.
8. Wink AM et al. BOLD Noise Assumptions in fMRI. Int J Biomed Imaging. 2006;2006(1):012014.
Funding Sources
The study is financed by Xjenza Malta, for and on behalf of the Foundation for Science and Technology, through FUSION: Space Upstream Programme (Project: Operation TOM, Grant ID: SUP-2023-01).
Acknowledgements
The authors gratefully acknowledge the provision of scanning services by the University of Malta’s MRI Platform (UMRI).