(c) 2015-2021 W.P.Karel Zapfe Zaldivar México.
The present code is distributed under GNU licence v03. This code is presented as is, no guarantee whatsoever.
This code presents some tools to obtain the CSD from 2D recordings of Electrophysiological Activity. The code was developed with High Density Microelectrode Arrays (MEAs) in mind.
There are two main strategies, meant for different use cases. The first one is the common finite diference approach, for signals that have been recorded with small errors or errors outside the region of interest. The advantages of this approach are speed and conceptual simplicity. The main disadvantage is that amplifies noise and errors, so it requires denoising mechanisms in order to give acceptable results.
The second approach is an inverse problem approach, named kernel current source density analysis, or kCSDA (Potworowski, bla bla). This second approach is meant for dealing for recordings that have many failing electrodes near or in the region of interest. It is also quite usefull for non rectangular grid matrices. This approach does not need denoising mechanism, as, by construction, ignores much of the noise. Its main disadvantage is that is conceptually more sofisticated and it requires implementation in parallel in order to compute the necesary oparators in acceptable times.