Read on Twitter
We just put up a preprint describing experiments we've done to try to determine what the fiber photometry signal represents in the striatum of awake mice. Work was led by @AlexLegariaM and @jlicholai. https://www.biorxiv.org/content/10.1101/2021.01.20.427525v1
We've been wondering about this since we first recorded photometry signals and noted that they didn't look like the patterns of spiking activity we saw in the striatum. For example, here's avg spiking and photometry around movement onset in the striatum https://www.jneurosci.org/content/38/14/3547)
Spiking activity increases ~1-2 seconds before the movement onset, while the photometry signal ramps up starting ~10 seconds before movement onset (work led by @tanisha_london).
This made us wonder if photometry signals reflected subthreshold dendritic calcium, driven by increasingly coherent excitatory input as upstream structures prepared the animal to move. Once these inputs broke a threshold they drove striatal spiking, which preceded movement.
In support of this idea: 1) Genetically encoded calcium indicators like GCaMP6 express well in neuropil
2) Striatal medium spiny neurons have extensive dendritic arbors. Beautiful reconstruction from Preston, Bishop, and Kitai, 1980 https://pubmed.ncbi.nlm.nih.gov/7353139/
3) Neuropil calcium is known to "contaminate" somatic calcium in imaging experiments. Many people have tried to mitigate this contamination with analytical approaches or recently biological approaches ie: soma-targeted GCaMP @zaknight @eboyden3
So we recorded GCaMP6s in striatum through a GRIN lens with an endoscopic epifluorescence microscope @mightexsystems. We extracted somatic signals with the amazing CaImAn toolbox https://elifesciences.org/articles/38173 that implements CNMF-E in an online notebook https://elifesciences.org/articles/28728
We also extracted the whole field photometry signal, as well as non-somatic neuropil signals from multiple places in the images with @FijiSc #ImageJ, to test how the photometry signal correlated with both somatic and neuropil calcium signals.
While we expected to see some correlation between the photometry signal and neuropil calcium, we were surprised to see how strong this correlation was, and how weak the correlation was with somatic calcium signals.
To test the relationship between striatal photometry and spiking in a more direct way, we also simultaneously recorded extracellular spiking and fiber photometry.
Here, we asked whether larger changes in photometry signals (ie: bigger calcium transients) were associated with more spiking, larger changes in spiking, or larger numbers of responsive units, than small transients. In all cases the answer was no. Photometry =/= spiking.
We conclude that fiber photometry should not be interpreted to reflect spiking activity of striatal neurons, but rather the total changes in calcium of these cells, of which dendritic calcium appears to be the largest contributor.
There are several limitations to this work. (Does this only apply to striatum? Would different GECIs produce different results?). More to do but the story so far is online and we'd be happy to receive feedback. Data and analysis code are also online at https://osf.io/8j7g2/