Amelie Rauland (all star via IRTG student w/ @sbe.bsky.social ) will be presenting "Benchmarking Reconstruction Methods for Bundle Segmentation in Single-Shell dMRI" on Friday and Saturday, Poster #1283

Analysis Methods: 11:00 – 11:30
Harmonize: Leakage in data harmonization
JuSpace: cross-modal brain map correlation

Applied Concepts: 11:30 – 12:00
Causal modeling for Confounds in ML
Subtypes of major depression

Fully agreed. This adds another layer of complexity

We likely need to differentiate three aspects here
- "True" changes of the trait
- State-changes due to, e.g., sleep
- Measurement error

All three can and will superimpose

In my view, the take-home message is slightly different from the BWAS paper. The latter argued you need large N.

This paper shows that if your target is unreliable, even very large numbers can’t save you

Also, association vs prediction