Tweetstorm on behalf of twitterless Cameron Higgins and @markwoolrich.

Our brains at rest spontaneously replay recently acquired information. How is this process orchestrated to avoid interference with ongoing cognition?

We can see replay in humans with magnetoencephalography (MEG) (eg Liu 2019). MEG similarly lets us measure whole brain resting state network activity on very rapid timescales (Vidaurre2018). We investigated whether these phenomena were linked.

We found a strong relationship between replay and activation of the default mode and parietal alpha networks (DMN and PAN). Both these networks are associated with inhibiting sensory input and an inward orientation of attention.

Furthermore, we found replay activity to display a very organised temporal structure. Replay did not occur randomly in time, but was packaged into transient bursts. These burst events coincided with concentrated activity in the DMN and PAN.

We then asked what specific electrophysiological patterns characterised these networks. We found they were associated with very large increases in network coherence; that is, the level of synchronisation of network oscillations.

More specifically, these synchronised oscillations separated into two independent frequency bands: fronto-temporal DMN areas oscillated in synchrony in a low delta/theta band, while parietal PAN regions oscillated at the alpha band.

Finally, we investigated how resting state networks correlated with activity in the sharp wave ripple (SWR) band. We found that the DMN exclusively among networks correlated with increases in SWR band power, that source localised to temporal cortex.

Synchronised oscillations and sharp wave ripples are both known electrophysiological mechanisms for information transfer between disparate cortical areas. These together suggest a mechanism by which replay may be coordinated across the cortex.

This suggests a tight correspondence between two widely studied phenomena of neural physiology. Does the DMN coordinate replay bursts to minimize interference with ongoing cognition?

Oh no - I forgot the link to the paper. Never let someone else post your tweet-storm: https://twitter.com/biorxiv_neursci/status/1275830135254630400?s=20