The effect of spherical projection on spin tests for brain maps
Bazinet, V.; Liu, Z.-Q.; Misic, B.
Show abstract
Statistical comparison between brain maps is a standard procedure in neuroimaging. Numerous inferential methods have been developed to account for the effect of spatial autocorrelation when evaluating map-to-map similarity. A popular method to generate surrogate maps with preserved spatial autocorrelation is the spin test. Here we show that a key component of the procedure -- projecting brain maps to a spherical surface -- distorts distance relationships between vertices. These distortions result in surrogate maps that imperfectly preserve spatial autocorrelation, yielding inflated false positive rates. We then confirm that targeted removal of individual spins with high distortion reduces false positive rates. Collectively, this work highlights the importance of accurately representing and manipulating cortical geometry when generating surrogate maps for use in map-to-map comparisons.
Matching journals
The top 3 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- Enhancement of brain atlases with laminar coordinate systems: Flatmaps and barrel column annotations 95%
- The Individualized Neural Tuning Model: Precise and generalizable cartography of functional architecture in individual brains 94%
- Spatial-extent inference for testing variance components in reliability and heritability studies 94%
Similar papers in this journal
- Functional gradients in the human lateral prefrontal cortex revealed by a comprehensive coordinate-based meta-analysis 93%
- Different computations over the same inputs produce selective behavior in algorithmic brain networks 93%
- A statistical framework to assess cross-frequency coupling while accounting for modeled confounding effects 93%
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.