author: David Shorten dpshorten@gmail.com
Instructions:
1.) Install CaImAn as per the instructions on github. We were using the June 11 2019 commit
(https://github.com/flatironinstitute/CaImAn/commit/c26be44cafc61c3fc5d77434ad0fce5f6efc9cac)
3.) Find where CaImAn is installed on your system. In the file caiman/source_extraction/cnmf/estimates.py
Insert the following line on line 512 (after the line "mov = caiman.concatenate(...")
mov.save("foo" + str(frame_range[0]) + ".tif")
2.) Run the file demo_caiman_basic2.py. I'm not sure whether or not it has to be inside the
demos/general/ directory.
This file is an edited version of the original demo_caiman_basic.py
The changes that have been made are as follows:
1.) A few parameters
2.) file names
3.) Turning off the use of the CNN classifier
4.) Calling the function detrend_df_f() to normalise the fluorescence values
5.) Dumping the resulting df_f values into a pickle file
6.) Calling nb_view_components which (if I remember correctly) is what lets you view
the results in a browser.
7.) Changed the play_movie call such that it was called multiple times on fewer frames
to prevent crashes due to lack of memory.
Tips:
If you find that you are running out of memory, consider:
1.) Reducing K
3.) Decreasing rf (although this will incur a time penalty)
4.) Dividing your recordings into smaller files