Replicate “Minute-scale oscillatory sequences in medial entorhinal cortex” in other dandisets

[stephprince]

Motivation

Gonzalo Cogno et al. (2024) found that neural activity in the mouse medial entorhinal cortex (MEC) can be organized into ultraslow oscillations, with periods ranging from tens of seconds to minutes. During these slow oscillations, neural activity was also organized into periodic sequences. These sequences were not related to location, running direction, or scheduled rewards, and they were specific to MEC.

This finding is a good candidate for a data reanalysis study because the claim is independent of specific behavioral/task variables and was observed in both calcium imaging and electrophysiological data. The authors also shared their data and their code!

There are a few dandisets with recordings from entorhinal cortex, and it would be interesting to replicate these findings in these datasets; incorporating additional electrophysiological recordings, applying different data preprocessing methods, using data from other species, and using data recorded under different experimental conditions.

Work performed so far

As part of NeuroDataReHack 2024, @rly, @bendichter, and I made some progress on an initial data reuse demonstration. Code can be found here.

1. We replicated part of their findings (Figure 1) using their shared data and our own Python code: (figure1_autocorr.ipynb, and figure2_2pautocorr.ipynb)
2. We replicated part of their findings using one of the shared datasets on DANDI (dandiset 000053) by Mallory et al., 2021: (figure1_newdata.ipynb)

Next steps

1. Continue to replicate the other findings in the paper. We developed some code to replicate their finding of minute-scale oscillations at a single cell level, but it would be great to replicate their findings of oscillatory sequences at a population level.
2. Contact the authors, convert their data to NWB, and upload it to DANDI. The shared neuropixels recordings were saved in NWB format, but the calcium imaging data and some additional metadata was shared in other formats. It would be great to convert their full dataset to NWB to improve the interoperability of any analysis code across datasets and for future reuse.
3. Leverage the author’s data combined with other dandisets to expand upon the original findings. The authors mention in the discussion that their work brings up further questions about how these sequences present across a variety of behavioral conditions and how they interface with the dynamics of grid cells and head direction cells. A data reanalysis project could be a great opportunity to pursue questions such as:
   • Is this type of oscillatory activity also present during navigation tasks?
   • How do these slow oscillatory cells relate to grid cells? Are they grid cells or are they not?
   • Is this phenomenon present in other species?
   • Does this oscillatory phenomenon relate to behavior or task accuracy?

If you’re interested

We would like to propose this as a potential project for collaboration and/or hand it off to a new researcher to continue to develop. We could also contact the authors of the initial paper and dandiset 000053 to see if they would like to be involved.