Does FLAME support repeated-measures crossover designs? (visual task, pre/post)

1

Hi everyone,

I’m analyzing a visual task (block design) in FMRIB Software Library (FSL) and I’m unsure about the correct way to use FLAME for my design.

Design

  • 22 subjects
  • Within-subject crossover
  • 3 conditions
  • Two runs per condition: pre-drug and post-drug
    → Total: repeated measures within subjects

Current setup

  • First level: GLM per run, one COPE per run
  • I want to test:
    • pre vs post within each condition
    • condition differences accounting for repeated measures

Questions

  1. Does FLAME (mixed effects) properly support this kind of repeated-measures / crossover design?
  2. Should this be modeled using:
  • a single higher-level design matrix with subject regressors, or
  • multiple paired contrasts (e.g., pre vs post per condition)?
  1. Is there a recommended way to construct the design matrix and contrasts for this case in FSL?

Any guidance or examples would be very helpful. Thanks!