Simplest way to get the code working: download the anaconda python distribution (https://www.anaconda.com/products/individual). Then you can run the quantif_synapse.py and then analyse_synapse.py scripts from Spyder.

The provided scripts are as follows:

• quantif_synapse_cl.py: synapse counting with a command line interface.

• quantif_synapse.py: same as the previous one but with setting script variables (use it if you do not know how to use command line).

The input of this script consist in folders with the hierarchy:

analysisDir/batch1/excit/imageStack1.tif ... analysisDir/batch1/excit/imageStackN.tif analysisDir/batch1/inhib/imageStack1.tif ... analysisDir/batch1/inhib/imageStackM.tif analysisDir/batch2/.....

where "batchNumber" is an integer giving the id of the batch and imageStack1.tif ... imageStackN.tif are all the stacks for a given batch and synapse type.

The generate_RGB option allows to generate tiff stacks where synapses are individually color coded but this step takes extra time and storage space so is deactivated by default.

Note1: in any case, the script generate tif images containing the individual synapses, one number for each synapse but this image is not good for visualisation purposes.

Note2: by default the script does not re-run already generated analysis, if you want to do it, delete the pickle file or use the force option.

The output of these scripts are the cluster images (see Note1) as well as a pickle file (python data storage files) "clusts.pkl", to be further processesd by other python scripts (e.g. analyse_synapse_neurons.py)

Note2: the script also detects the different cells using the PV signal and assign synapses to a neuron if they overlap with any PV signal, but these information are not used at the moment.

• launch_repeat.sh: a bash script to run multiple times the counting script in case of memory issues

• analyse_synapse_neurons.py: output the plots.