Common HATSA Pitfalls and Solutions

Guide to avoiding common issues when using HATSA.

Pitfall 1 - Too Many Components

Using too many components relative to your data size can lead to overfitting.

Solution: Use hatsa_suggest() or keep components < 10

Pitfall 2 - Mismatched Dimensions

All subjects must have the same number of voxels (spatial alignment).

Solution: Ensure all subjects are in the same space/parcellation.

Pitfall 3 - Including Bad Subjects

Subjects with excessive motion or artifacts can degrade alignment.

Solution: Pre-screen subjects and exclude outliers.

Pitfall 4 - Wrong Preprocessing

HATSA expects centered data (zero mean per voxel).

Solution: Always scale/center your data before HATSA.

Pitfall 5 - Ignoring Convergence Warnings

Warnings about convergence indicate potential issues.

Solution: Check your data quality and try different parameters.