Current source density (CSD)

Current source density analysis converts recorded extracellular potentials into the underlying pattern of transmembrane current sinks and sources, removing the ambiguity and volume-conduction smearing of raw potentials. Under the assumption of a locally ohmic, quasi-static medium, Poisson's equation relates CSD to potential: the current source density is proportional to the negative divergence of conductivity times the gradient of the potential, which for a laminar recording with uniform conductivity reduces to the negative second spatial derivative of the potential along the depth axis.

In practice CSD is estimated from a linear array of contacts spanning cortical layers; a current sink marks net inward transmembrane current (for example an excitatory synaptic input onto a dendritic layer) and a source marks the return current. Because differentiation amplifies noise and is sensitive to electrode spacing and to the assumed conductivity, modern practice uses regularized or inverse (kernel / iCSD) estimators rather than the naive three-point second difference, and interprets sinks and sources as current, not directly as excitation or inhibition.