Replicate "A nonoscillatory, millisecond-scale embedding of brain state provides insight into behavior" in other dandisets

[rly]

Parks et al. (2024) found that sleep and wake states could be classified using 1 to 10 ms of high-resolution neural activity recorded in ten mouse brain regions over 24 h, and switches in the local embedding of brain state correlate with changes in behavior. Abstract below.

This finding is nice for data reanalysis because the claim is independent of brain area and specific behavioral/task states besides wake/sleep. DANDI has several datasets that include extracellular recordings during wake and sleep. It would be interesting to replicate the findings in datasets during different behaviors, from other recording setups, using different spike sorting / data processing methods, in other brain areas, and in other species. Replicating the finding would evaluate its robustness to these variables and perhaps broaden or limit its claims. Some of these datasets involve experimental manipulations that might also be interesting to explore in combination with the finding. This could be a nice NeuroDataReHack project.

The data from the original paper are not publicly available, though the authors state "We are excited to share data upon reasonable request and as technical limitations make possible." They use a White Matter eCube Server for neural recordings and video recording. "Recordings were made continuously for between 2 weeks and 3 months (Servernode software, White Matter LLC)." Code is available at https://github.com/hengenlab.

It would be nice to:

1. Adapt their code to run on an arbitrary NWB dataset (or write a simplified, independent version that should produce the same results)
2. Contact the authors, convert their data to NWB, and upload it to DANDI. Such long-term recordings from many brain areas would be a great dataset for others to use in data reanalysis projects.
3. Reproduce their findings using the adapted code on the NWB form of their data (this validates items 1 and 2)
4. Replicate their findings on one of more DANDI datasets

Items 2 and 3 are not necessary for replication, but nice to do.

I do not have the bandwidth to work on this; I'm just proposing the idea. We could contact the authors and see if they would be interested in doing this. In my opinion, this is a well-defined project that might be nice for a rotation student or grad student.

Abstract:

The most robust and reliable signatures of brain states are enriched in rhythms between 0.1 and 20 Hz. Here we address the possibility that the fundamental unit of brain state could be at the scale of milliseconds and micrometers. By analyzing high-resolution neural activity recorded in ten mouse brain regions over 24 h, we reveal that brain states are reliably identifiable (embedded) in fast, nonoscillatory activity. Sleep and wake states could be classified from 10^0 to 10^1 ms of neuronal activity sampled from 100 µm of brain tissue. In contrast to canonical rhythms, this embedding persists above 1,000 Hz. This high-frequency embedding is robust to substates, sharp-wave ripples and cortical on/off states. Individual regions intermittently switched states independently of the rest of the brain, and such brief state discontinuities coincided with brief behavioral discontinuities. Our results suggest that the fundamental unit of state in the brain is consistent with the spatial and temporal scale of neuronal computation.