Hi everyone,
So I’m not just a newbie to MNE. I’m also new to electrophysiology and python. I am a lab technician who primarily handles behavioral testing, surgical implants, VERY basic data analysis, etc. But my PI is interested in looking into new ways to visualize/analyze our data, and I’d like to try and use the opportunity to delve into the worlds of python, statistics, and electrophysiology. I apologize if some of this has been answered already, I wasn’t sure if some of the specific conditions would be important.
So my big question is: Is MNE right for the data I’m working with? (skip to next paragraph for description).
Smaller questions: What kinds of subjects and concepts (perhaps some statistics books/classes, certain foundations of computer science/data structures, etc.) do I need to begin studying in order to reach some sort of proficiency - even marginally - with this package? On the bright side, I have some great resources at my disposal (my mother teaches python/software engineering, and I work with an institution full of talented scientists). On the not so bright side, my educational background is strictly neuroscience with pretty much no math/CS at all.
Our data:
So the data I have consists of a series of LFP recordings from 8 different brain regions (After removing junk recordings) from a sample of ~10 rats. These rats performed an episodic memory test where each “trial” is a nose-poke and each rat gets up to ~20 trials per “session.” We are interested in the recordings during the time periods -3-0 seconds and 0-3 seconds (where “0” is the nose-poke). We would then like to differentiate between the first 4 trials (nose-pokes) and the last 4 for each session.
What we’ve done so far:
So far, we have generated a list of “pathways” between each LFP region (about 28 pathways), and we used Spike2 to calculate coherence between each pathway. So now, I have datasets containing both broad-spectrum and frequency-band-specific (theta, beta, low gamma, high gamma) coherence in each of these 28 pathways, pre-nose-poke and during-nose-poke, first 4 trials and last 4 trials. We also have these numbers expressed as difference in coherence from baseline.
My PI would like to look into something like the circular graph shown here.
So does this sound possible? Would I have to start over from the raw-inputs, or is there some sort of way I can arrange the calculations we already have into a valid matrix? What we’re seeing from the data so far is very interesting. But finding a way to analyze it and try to get an intuitive understanding of how the whole network is shifting to accommodate the task is a real head-scratcher.
Any help is appreciated, and I am happy to provide more information/context if it helps!