EN
INTRODUCTION: Current Source Density (CSD) is spatially smoothed transmembrane activity of the neurons. Local Field Potential (LFP) is the electric potential generated by ionic currents in the neural tissue and it is directly related to the CSD. LFP is relatively easily accessible experimentally but due to the long range of electric field, it is difficult to interpret. CSD needs to be calculated but it reflects the local neural activity directly. Since the currents directly reflect neuronal computations, using electric potentials (LFP’s) to infer performed computation may lead to misinterpretations. AIM(S): 1) Discuss challenges arising in multielectrode LFP and CSD analysis, in particular case where direct analysis of LFPs can lead to misinterpretation. 2) Show that the kernel Current Source Density reconstruction method (kCSD) gives a better insight into the underlying phenomena than the observed potentials, and to show the limits and uncertainties yielded by the method. 3) Present the kCSD-python toolbox for CSD analysis. METHOD(S): All of the modeling and computations was done in Python and tested on model data. Potentials were calculated using assumed physical models of tissue. The kCSD library in Python is available at: https://github.com/ Neuroinflab/kCSD‑python. RESULTS: We show examples where the LFP’s can ‘hide’ more complex underlying CSD patterns and how the kCSD can reconstruct those sources, depending on the number and configuration of recording electrodes. CONCLUSIONS: Complex CSD patterns studied at the resolution of few electrodes can be obscured if only direct LFP analysis is used. The kCSD method can help to recover them. The main limiting factors are the number of recording electrodes and their configuration. FINANCIAL SUPPORT: This work was supported by EC-FP7-PEOPLE sponsored NAMASEN Marie-Curie ITN grant 264872, Polish Ministry for Science and Higher Education grant 2948/7.PR/2013/2, Narodowe Centrum Nauki grants 2013/08/W/NZ4/00691 and 2015/17/B/ ST7/04123.