An empirical evaluation of functional alignment using inter-subject decoding

• Published in: NeuroImage (Orlando, Fla.) Vol. 245; p. 118683
• Main Authors: Bazeille, Thomas, DuPre, Elizabeth, Richard, Hugo, Poline, Jean-Baptiste, Thirion, Bertrand
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 15.12.2021; Elsevier Limited; Elsevier
• Subjects: Accuracy; Algorithms; Artificial Intelligence; Bioinformatics; Brain - diagnostic imaging; Brain architecture; Brain Mapping - methods; Brain research; Computational neuroscience; Computer Science; Datasets; fMRI; Functional alignment; Functional morphology; Humans; Inter-subject variability; Magnetic Resonance Imaging - methods; Mathematics; Neural coding; Performance evaluation; Predictive modeling; Statistics; Structure-function relationships; Usability; Functional alignment; fMRI; Predictive modeling; Inter-subject variability; inter-subject variability; predictive modeling; functional alignment
• DOI: 10.1016/j.neuroimage.2021.118683
• ISSN: 1053-8119; 1095-9572

•Methods that improve inter-subject decoding accuracy reduce inter-individual variability without losing signal specificity.•Functional alignment methods consistently improve inter-subject decoding on several datasets, with the best methods recovering half of the signal lost in anatomical-only alignment.•For whole-brain alignment, piecewise alignment (performed in non-overlapping regions) is more accurate and much more efficient than searchlight alignment.•Shared Response Model and Optimal Transport yield highest decoding accuracy gains. Inter-individual variability in the functional organization of the brain presents a major obstacle to identifying generalizable neural coding principles. Functional alignment—a class of methods that matches subjects’ neural signals based on their functional similarity—is a promising strategy for addressing this variability. To date, however, a range of functional alignment methods have been proposed and their relative performance is still unclear. In this work, we benchmark five functional alignment methods for inter-subject decoding on four publicly available datasets. Specifically, we consider three existing methods: piecewise Procrustes, searchlight Procrustes, and piecewise Optimal Transport. We also introduce and benchmark two new extensions of functional alignment methods: piecewise Shared Response Modelling (SRM), and intra-subject alignment. We find that functional alignment generally improves inter-subject decoding accuracy though the best performing method depends on the research context. Specifically, SRM and Optimal Transport perform well at both the region-of-interest level of analysis as well as at the whole-brain scale when aggregated through a piecewise scheme. We also benchmark the computational efficiency of each of the surveyed methods, providing insight into their usability and scalability. Taking inter-subject decoding accuracy as a quantification of inter-subject similarity, our results support the use of functional alignment to improve inter-subject comparisons in the face of variable structure-function organization. We provide open implementations of all methods used.