load_haxby_example: option to share the block order across runs
- Dominant language
- Python
- Stars
- 130
- Forks
- 48
- Avg merge
- 6h 42m
- Merged PRs (30d)
- 34
Description
`load_haxby_example` draws a fresh random block order for every run (`rng.permutation(_HAXBY_CONDITIONS)` per run), so intersubject correlation across runs treated as subjects is zero by construction until the TRs are re-sorted into one shared condition sequence. The quickstart does that with a small helper on the page. A keyword such as `block_order="shared"` (default unchanged) that draws the order once and reuses it for every run would let ISC, ISRSA and alignment examples use the runs directly. Small; keep the return contract.
Contributor guide
Research direction
Start at the `load_haxby_example` entry point and read the quickstart helper that currently re-sorts TRs into a shared condition sequence. Trace how the block order is generated for each run and preserve the existing return contract. Done means an opt-in shared order is reused across runs while the default behavior remains unchanged, supporting direct ISC, ISRSA, and alignment examples.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- data
- Issue type
- Feature
- Difficulty
- 2/5
- Estimated time
- 1-3 hours
- Activity status
- Active
- Clarity
- Mostly clear
- Newbie friendliness
- 74/100