Predictions tutorial

Inferring neuronal activity

The classifier takes as inputs the motion corrected calcium imaging movie and spatial footprints of the sources (cells).

The outputs are float values between 0 and 1 for each frame and each source, representing the probability for a cell to be active at that given frame.

The classifier we provide was trained to consider a cell as active during the rise time of its transients.

For more information, check-out our demo code.

On google colab

you can run this notebook.

On your local device

You can follow the steps described in this demo file.

Predicting cell type

The classifier takes as inputs the motion corrected calcium imaging movie and spatial footprints of the sources (cells).

The outputs are float values between 0 and 1 for each cell type, representing the cell type probability of a given cell.

We have trained a classifier on two cell type interneurons and pyramidal cells. For training, interneurons were identified using GadCre mouse while pyramidal cell were putative.

A .yaml file allows to configure the cell types you want to use.

We are currently improving the classifier.

For more information, check-out our demo code.

On google colab

you can run this notebook.

On your local device

You can follow the steps described in this demo file.