Huge thanks to the many people who made this project possible by generously making their data available! @urihasson.bsky.social‬‬ ‪@chrishoney.bsky.social‬ @sparsity.bsky.social ‪@lukas-rier.bsky.social‬ Matt Brookes and Elena Boto!

These findings highlight OPM-MEG’s potential as a flexible, noninvasive tool capable of capturing complex neural activity with high reliability. 🧠✅

Finally, we compared z-scores across conditions. Permutation testing identified areas with significant reliability differences.

4. Surprisingly–in some regions–SNR estimates between OPM & fMRI (red) exceeded OPM alone. In others OPM alone exceeded the OPM & fMRI comparison (green).

📊 Fig 7

3. OPM signals correlated more strongly with fMRI than iEEG, especially in visual areas. OPM to fMRI correlations followed the expected inverse relationship with BOLD signals: negative correlations at low frequencies, positive correlations at higher frequencies.

📊 Fig 5

2. Between-subject correlation maps were more spatially focused in sensory regions

📊 Fig 4

1. OPMs showed high test-retest reliability, especially at lower frequencies, often on par with reliability in fMRI and iEEG.

Z-scores reflect signal correlation relative to correlations of phase-shuffled data.

📊 Fig 3

Three separate groups watched a short movie twice while neural activity was recorded via OPMs, fMRI, or iEEG.

We quantified:
1. Within-subject reliability
2. Shared signal across subjects (ISC)
3. Shared signal across methods
4. SNR within & across OPM subjects, and between OPM & fMRI

📊 Fig 1