- Mechanisms of action potential (AP) generation in neocortical pyramidal cells have been the focus of intense experimental and theoretical research over the last several decades. It has proven very difficult, however, to arrive at a consensus model which can satisfactorily account for all of its features. One of the still unresolved issues is lack of accurate description of Na+ channel kinetics in different neuronal compartments. Here, we measured kinetics of somatic Na+ channels using high temporal resolution (5-10 kHz, −3dB, low pass four-pole Bessel filter) cell-attached recordings from layer 5 pyramidal neurons in neocortical slices. The data were described by fitting different Markov models with differential evolution fit algorithms. The limited speed of voltage steps and the effect of current filtering were accounted for in the fit procedure. Hodgkin-Huxley-type models which assumed a number of independent activation gates were not the optimal description of the experimentally recorded currents. Activation kinetics was best described by Markov models with two sequentially activating gates, while inactivation was best described as a process that runs in parallel to activation. The best model described the channel data well enough to allow quantitative prediction of the somatic Na+ current during the somatic spike. To this end the AP waveform recorded in current clamp in the same preparation, was used to drive Na+ channels in the model. The resulting simulated current matched the second phase of the AP upstroke in the phase plot (dV/dt vs V). This is consistent with the long standing idea that somatic Na+ channels are the main current sink during this second phase of the AP upstroke but contribute little to its initial phase.Supported by the GIF and the BMBF (BCCN, Goettingen).