Fast Whole-Brain, Geometry-Aware Functional Alignment for Cross-Subject Decoding
2026-07-12 • Machine Learning
Machine Learning
AI summaryⓘ
The authors address the problem of differences in brain activity patterns between people, which makes it hard to create brain decoders that work well for everyone. They propose a new method called SpectralOT that aligns brain data from different individuals by combining brain function with the brain's shape. This helps to compare brain activity more accurately while keeping computations efficient. Their approach improves how brain data is matched across people before analyzing it.
fMRIfunctional alignmentbrain decodinginter-individual variabilityLaplace-Beltrami eigenmodescortical geometrypopulation-level decodersSpectralOT
Authors
Pierre-Louis Barbarant, Florent Meyniel, Bertrand Thirion
Abstract
Decoding brain activity is useful for characterizing brain processes and understanding the functional architecture underlying cognition. However, the inter-individual variability in brain response patterns limits the development of decoders that generalize across individuals. A solution to this challenge is functional alignment: aligning functional data across individuals before training population-level decoders. The core issue is to strike the balance between aligning functional features and preserving the anatomical structure, while maintaining computational efficiency. We introduce a new functional alignment method for fMRI, SpectralOT, that embeds cortical geometry into Laplace-Beltrami eigenmodes along functional data to regularize the alignment.