Post by Stephanie Williams
What's the science?
Computations in the brain often occur at the network level. How individual neurons participate in these computations, and how they are coupled to local (neurons nearby) and distal (neurons located further away) dynamics is still under investigation. This week in Nature Neuroscience, Clancy and colleagues investigated how the coupling of neurons to local and distant networks can dynamically change across different behavioral states.
How did they do it?
The authors measured activity patterns in different brain regions of 25 mice while the mice exhibited different behaviors (eg. staying still vs. running on a wheel). The authors used different techniques to measure the activity of neurons, including electrophysiological recordings, two-photon imaging and a technique called wide-field calcium imaging. Calcium imaging relies on recording fluorescent signals emitted from neurons. When neurons spike, calcium flows into cells, activating a calcium-sensitive fluorescent protein and causing it to glow brighter. The authors recorded spiking in several regions, including: (1) visual cortex (specifically, a region called V1) and (2) retrosplenial cortex, a region known to be involved in spatial navigation. To record spiking activity from individual units, the authors inserted silicon probes into visual and retrosplenial brain areas of mice. Then, they simultaneously imaged activity across dorsal cortex using wide-field calcium imaging. They analyzed how the activity in faraway brain regions were related to the single units they were recording. The authors used the relationship between the single units and distal regions to create correlation maps of individual units with different brain areas. A major mystery of cortical activity is how variable cortical neurons are — some neurons seem to do things very differently from their neighbours. The authors investigated whether the activity of neurons that didn't seem to follow the spiking of their neighbours was more likely to be correlated with distant brain areas, which would suggest they might be directly driven by long-range projections. The authors examined how behavioral state (e.g. locomotion) impacted the activity of the neurons they recorded, and how it changed the way that individual neurons were coupled to activity in local and distal regions.
What did they find?
The authors found that many neurons showed activity similar to other neurons in the same area. However, some neurons went against this pattern and were correlated with activity in distal regions. When mice switched from quietly sitting to running on a wheel, the authors found that the coupling patterns of neurons to local and distal regions dynamically changed. They found that the firing of neurons in the visual brain area called V1 became more correlated with local activity, and more similar to one another. In contrast, the neurons in the retrosplenial cortex that were correlated with local activity when the mice were not moving became different from one another, and more correlated with activity in distant areas. This suggests that behavioral state of an animal determines how individual neurons are coupled to activity in distant regions. The authors suggest their findings support the idea that locomotion induces a major network reorganization in which the strongly locally correlated neurons become silenced, and the distally correlated neurons become “unmasked”. These changes may gate how sensory information is processed in the retrosplenial cortex.
What's the impact?
The author’s work expands upon previous findings, which had suggested neurons are primarily coupled locally, and instead shows that many cortical neurons are correlated with activity in diverse distant regions. Their work shows that behavioral state can shift how neurons are coupled both locally and distally, and that this impact of running on different brain regions is distinct, perhaps reflecting how these different areas contribute to processing information relevant to navigation.
Clancy, K et al. Locomotion-dependent remapping of distributed cortical networks. Nature Neuroscience (2019). Access the original scientific publication here.