Predicting cortical signal mixing in coarse-resolution fMRI: a forward-operator analysis of surface mapping and the conditional benefit of correction

A forward-operator model for coarse-resolution surface fMRI predicts net far-cortex leakage exceeding 5% in 4%, 10%, and 22% of HCP 7T cortex at 1.6 mm and emulated 2.4 and 3.2 mm.

• Why it matters: Coarse fMRI projection can mix signals from distant cortical locations along the folded sheet, risking biased surface maps; correction is needed but must be quantified against references.
• What they did: They built an anatomy- and acquisition-informed linear operator and inverted it with regularised surface reconstruction, then evaluated residual error using finer-resolution cross-run references in HCP development (n = 15) and a confirmatory family-independent cohort (158 participants, 86 families).
• The result: Regularised reconstruction reduced residual error and improved coarse-map predictions, with gains increasing with a response-weighted contamination measure in HCP and an independent 3T test (35 participants), but not for multi-run averages or robustly across separately acquired sessions.

AI summary of the abstract · may contain errors · not medical advice · read the paper.