GCaMP6 variants are on addgene. Three flavors, fast kinetics or big signals. Bigger responses than OGB-1, some are MUCH bigger. The responses to drifting gratings in visual cortex are spectacular. Sorry no pics for now. Hopefully the reviewers will be nice so we can all read about it soon. Still work to be done getting true 1AP resolution when simultaneously imaging large populations of neurons, but for single neuron imaging in vivo, this has 1AP resolution. If you have been waiting for the GCaMPs that will blow your expectations away, these are them.
From the SfN abstract :
Using structure-guided mutagenesis and high-throughput screening, we increased the fluorescence change in response to single action potentials (APs) by >10-fold compared to GCaMP3. We also accelerated the kinetics by ~2-fold. These new GECIs reliably report single APs in single trials in vivo with near 100% accuracy. In the mouse visual cortex, we detected ~5-fold more visually responsive neurons. The sensitivity, dynamic range and speed of the new GECIs exceed those of the synthetic indicator OGB-1. The improved sensitivity further facilitated reliable measurement of synaptic calcium signals in the dendrites of pyramidal cells and parvabumin (PV)-positive interneurons in vivo. Hot spots of orientation-selective domains can be resolved both in single pyramidal cell spines and small segments of PV cell dendrites. These improved GECIs will permit a more complete description of neuronal circuit function and enable long-term functional imaging of single synapses.
Brain Windows 
“Still work to be done getting true 1AP resolution when simultaneously imaging large populations of neurons, but for single neuron imaging in vivo, this has 1AP resolution.”
Why is that? What’s the different, if I can scan fast enough, let say 6 hz full frame or faster with line scan, so I have nice resolution in time for imaging 20-30 neurons, why for single neuron I’ll be able to get 1ap resolution and for more I can’t? Is it just
Shot noise. It’s a simple issue of total number of photons you are collecting per cell. If you are scanning a large field of view (like a whole barrel column) with hundreds of neurons, your dwell time on any given neuron is much less than the total dwell time if you are scanning a single cell. Going from 1 neuron to 100 neurons will reduce your signal to noise about 10-fold.
Also, the distribution of dF/F per AP is not fixed across the population, so extrapolating the dF/F to # of APs is difficult without ground truth knowledge of what 1AP looks like in each cell.
Also, cellular calcium handling dynamics are non-linear following bursts of action potentials, slower extrusion following a burst, which further complicates extraction of APs from fluorescence traces.
Carefully and critically evaluate any claims of single AP resolution across a large population of neurons. What were the conditions, what were the AP rates they were looking at, what is the ground truth metric they use for confirmation.
Thanks! can we see some examples? 🙂