ARIbrain: Fast and flexible statistical analysis of fMRI data using True Discovery Proportions.
Presented During: Poster Session 3
Friday, June 27, 2025: 01:45 PM - 03:45 PM
Friday, June 27, 2025: 01:45 PM - 03:45 PM
Presented During: Poster Session 4
Saturday, June 28, 2025: 01:45 PM - 03:45 PM
Saturday, June 28, 2025: 01:45 PM - 03:45 PM
Presented During:
Saturday, June 28, 2025: 11:30 AM - 12:45 PM
Brisbane Convention & Exhibition Centre
Room:
M4 (Mezzanine Level)
Poster No:
1558
Submission Type:
Abstract Submission
Authors:
Wouter Weeda1, Lucas Peek1, Xu Chen2, Jelle Goeman3
Institutions:
1Leiden University, Leiden, ZH, 2University of Essex, Colchester, Essex, 3Leiden University Medical Center, Leiden, South Holland
First Author:
Co-Author(s):
Introduction:
Statistical analysis of functional MRI (fMRI) data requires correct control for multiple comparisons. Most often this is done using cluster-extent based thresholding. One of the drawbacks of this method is that it requires the setting of an arbitrary cluster-forming threshold (usually Z > 3.1) that determines the size/shape of the clusters used for further analysis. Once this threshold is set, researchers are not allowed, statistically, to redo the analysis with a different threshold. Together with the knowledge that the correct interpretation of a significant cluster is: 'there is at least one active voxel in this cluster' and not 'all voxels in this cluster are active', this leads to a (statistically) suboptimal way of analysing fMRI data.
Recent advances in statistics have led to a new range of methods based on the True Discovery Proportion (TDP) [1-3]. These methods estimate the lower bound of the number of truly active voxels, i.e.TDP, within a cluster, for any cluster in the data, as many times a researcher wants, with full control of the family-wise error rate. They do not require the setting of an arbitrary threshold as the TDP provides a simultaneous bound on the number of active voxels over all possible clusters.
In practice, these methods thus give the researcher almost unlimited freedom in selecting and analysing clusters. One can use different thresholds for different clusters, calculate the TDP for an a priori defined cluster, or search for clusters with at least a certain TDP level. This flexibility requires software that goes beyond what is available in current statistical analysis packages like FSL or SPM.
Recent advances in statistics have led to a new range of methods based on the True Discovery Proportion (TDP) [1-3]. These methods estimate the lower bound of the number of truly active voxels, i.e.TDP, within a cluster, for any cluster in the data, as many times a researcher wants, with full control of the family-wise error rate. They do not require the setting of an arbitrary threshold as the TDP provides a simultaneous bound on the number of active voxels over all possible clusters.
In practice, these methods thus give the researcher almost unlimited freedom in selecting and analysing clusters. One can use different thresholds for different clusters, calculate the TDP for an a priori defined cluster, or search for clusters with at least a certain TDP level. This flexibility requires software that goes beyond what is available in current statistical analysis packages like FSL or SPM.
Methods:
Here we introduce the improved ARIbrain analysis framework [4]. The ARIbrain Explorer is a newly developed Python application that allows for flexible statistical analysis of fMRI data based on TDPs. It is a graphical user interface that allows researchers to explore and adjust all clusters of activity in the brain with full family-wise error control (see Figures 1 and 2). The ARIbrain Explorer also incorporates recent improvements of the original Simes based method [1] based on permutations [2] for better TDP estimates. The ARIbrain Explorer can read in images from all popular analysis packages like FSL, SPM, or NiLearn.
Results:
Figures 1 and 2 show screenshots of the ARIbrain Explorer. Researchers can upload z- or t-statistics images derived from any analysis package or directly load data from a previous ARIBrain analysis.
After the images are loaded a gradient map with the maximum TDP for each voxel is shown. Once a minimal TDP threshold is chosen, a list of clusters with at least this TDP is shown in a table as well as visuallly on a functional or anatomical template. All clusters can then be further explored: they can be 'drilled-down', made smaller to increase TDP, or increased in size if the TDP is deemed too large. In addition, they can be intersected with an anatomical mask or functional atlas, in order to test hypotheses like: 'how much signal is there in the amygdala?'. All these analyses can be performed as many times the researcher wants without compromising statistical validity, in an interactive and highly visual way.
Given this flexibility, the ARIbrain Explorer contains many functions to aid researchers in visualizing and reporting results of their analysis. It has an intuitive user-interface that allows users to visually adjust each cluster. In addition, it allows researchers to easily incorporate anatomical and functional templates. Next to these visual exploration features it also outputs more traditional analyses augmented with TDP statistics.
After the images are loaded a gradient map with the maximum TDP for each voxel is shown. Once a minimal TDP threshold is chosen, a list of clusters with at least this TDP is shown in a table as well as visuallly on a functional or anatomical template. All clusters can then be further explored: they can be 'drilled-down', made smaller to increase TDP, or increased in size if the TDP is deemed too large. In addition, they can be intersected with an anatomical mask or functional atlas, in order to test hypotheses like: 'how much signal is there in the amygdala?'. All these analyses can be performed as many times the researcher wants without compromising statistical validity, in an interactive and highly visual way.
Given this flexibility, the ARIbrain Explorer contains many functions to aid researchers in visualizing and reporting results of their analysis. It has an intuitive user-interface that allows users to visually adjust each cluster. In addition, it allows researchers to easily incorporate anatomical and functional templates. Next to these visual exploration features it also outputs more traditional analyses augmented with TDP statistics.
Conclusions:
The ARIbrain Explorer allows researchers to perform statistical analyses based on TDPs in a flexible manner. The software integrates seamlessly with standard analysis packages and incorporates the latest advances in statistical methods. It is a useful and statistically sound way to improve cluster-extent based analyses of fMRI data.
Modeling and Analysis Methods:
Activation (eg. BOLD task-fMRI) 2
Methods Development 1
Univariate Modeling
Keywords:
Data analysis
FUNCTIONAL MRI
Open-Source Software
Statistical Methods
Other - multiple comparisons; closed testing; cluster-extent
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] Andreella, A., et al. (2023). Permutation-based true discovery proportions for fMRI cluster analysis. Statistics in Medicine, 42(14): 2311-2340.
[3] Goeman, J.J., et al. (2023). Cluster extent inference revisited: quantification and localization of brain activity. J Roy Stat Soc B, 85(4), 1128–1153.
[4] https://wdweeda.github.io/ARIbrain-project/