Assumptions on the underline spike activity from the calcium imaging

[Hananel-Hazan]

Adding to the question asked by @RandallJEllis on the issue #996.

We would like to use Suite2p to deconvolve spikes from calcium imaging of colon cells. We have a few questions regarding the spikes that Suite2p generates.

Base on the FAQ section Deconvolution means what?

Few spike deconvolution algorithms try to estimate single spike amplitude (look up “MLspike”), but we are in general suspicious of the results, and usually have no need for absolute numbers of spikes.

We would like to know your thoughts on using the spikes generated by the deconvolution algorithm for comparing between two colon cell culture. More specifically, we are interested in comparing the properties of spikes (e.g., spike size, height, and width) between two tissues of colon cells. We understand that the spikes produced by Suite2p are not "real spikes" but rather an indication of underlying activity.

• Do you think that the spikes are reliable enough for measuring their properties?
• Do the spike generator algorithm rely on ground truth neurons that may not be present in colon cells?

Moreover, its seems that some version of the OASIS algorithm are used in Suite2p to generate spikes from the fluorescent illumination values.

• Do the spike generation rely on any specific neuron type or parameters (e.g., rat, mouse, or area-specific neuron) as a ground truth model to generate the deconvolved spikes?