search

brainhack-school2020 / haqqeez_bhs2020_project

Determine relevant statistical and machine learning tools best suited to consolidate neural data with behavioural data.
0 stars 0 forks source link

Method for analysing the calcium data ? #2

Open ltetrel opened 4 years ago

ltetrel commented 4 years ago

Hi @haqqeez,

This is an exciting and challenging project!

Have you decided on what methods you will use to analyse the calcium images ? We already talked about nilearn tools that could definitively help since they rely on vectorized data.

haqqeez commented 4 years ago

Thank you very much! As for analyzing the video files frame-by-frame, I already have a MATLAB script that does that to tag groups of pixels are neurons and then measure their overall brightness changes during the video to create a raster plot of activity.

This generates a big matrix, about 30,000 rows (time series, 30ms each) by ~100 columns (neurons). This can be synchronized and filtered through the behavioural data so I will only be looking at frames when the animal was doing something of interest (i.e., making a decision). This is also accomplished via a MATLAB script, but I will need to modify it to search for events that interest me in particular.

Next, I need to import the MATLAB files to python. I know there are some tools available and have had another user also give my some tips on how to do this.

Finally, I will take a close look at the nilern annd scikitlearn packages to see how I can generate correlation matrices from my data, then use machine learning to see how accurately the neural data can model the animal's behaviour. This will be very similar to the intro to machine learning tutorial we had last week, but for predicting categorical data (correct vs incorrect) instead of ratio data (age).