Laminar Population Analysis: Estimating Firing Rates and Evoked Synaptic Activity From Multielectrode Recordings in Rat Barrel Cortex

Abstract

We present a new method, laminar population analysis (LPA), for analysis of laminar-electrode (linear multielectrode) data, where physiological constraints are explicitly incorporated in the mathematical model: the high-frequency band [multiunit activity (MUA)] is modeled as a sum over contributions from firing activity of multiple cortical populations, whereas the low-frequency band [local field potential (LFP)] is assumed to reflect the dendritic currents caused by synaptic inputs evoked by this firing. The method is applied to stimulus-averaged laminar-electrode data from barrel cortex of anesthetized rat after single whisker flicks. Two sample data sets, distinguished by stimulus paradigm, type of applied anesthesia, and electrical boundary conditions, are studied in detail. These data sets are well accounted for by a model with four cortical populations: one supragranular, one granular, and two infragranular populations. Population current source densities (CSDs; the CSD signatures after firing in a particular population) provided by LPA are further used to estimate the synaptic connection pattern between the various populations using a new LFP template-fitting technique, where LFP population templates are found by the electrostatic forward solution based on results from compartmental modeling of morphologically reconstructed neurons. Our analysis confirms previous experimental findings regarding the synaptic connections from neurons in the granular layer onto neurons in the supragranular layers and provides predictions about other synaptic connections. Furthermore, the time dependence of the stimulus-evoked population firing activity is predicted, and the temporal ordering of response onset is found to be compatible with earlier findings.