Abstracts

Rationale: Computational simulations have been used in a variety of cellular and electrophysiological formats to explore the time-frequency and spatial properties of seizures and therapeutic interventions. We present here a multicompartment Hodgkin-Huxley based model of neocortex that exhibits realistic epileptiform behavior. The model has a three dimensional structure, and allows the user to explore local field potential recordings within it from a variety of current sources including total synaptic currents as well as specific synaptic currents between various cell classes. The model exhibits spiral wave behavior spontaneously as observed in experimental preparations. Methods: The model represents a region of cortex of dimension 1.0 mm X 1.0 mm, and includes 12 neuron classes organized by cortical layer, inhibitory or excitatory properties, and spiking characteristics. There are a total of 33,600 modeled Traub-like multicompartment neurons (range 59-137 compartments) that operate on a modified version of Hodgkin-Huxley dynamics [1]. The intercellular wiring is based on our previous modeling studies and is patterned after histological data [2]. A spatially uniform distributed Poisson background activity representing inputs from neighboring cortex affects a given percentage of the modeled neurons. Time-frequency analysis of the spontaneous bursting patterns is performed using the FFT. Local field potentials are recorded from the ongoing activity with a variety of simulated microelectrode recording elements and positions, and can be focused on specific sets of synaptic currents. The simulations are performed on a 16-node distributed 64-bit processor system in the parallel Genesis (pGenesis) environment [3]. Results: The activity in the model is characterized by a decreasing frequency spectrum in the LFP, with peaks at the primary network oscillation frequency and harmonics. Varying the global spatial extent of connections by all cell classes quickly alters the dynamics of the model. The LFP derived strictly from synaptic currents or subsets of synaptic currents continue to demonstrate the higher frequency components in the time-frequency spectrum indicating their likely importance in microelectrode recording efforts. Spatially evolving spiral wave dynamics are readily and spontaneously produced by the model. Conclusions: This multicompartment simulation of neocortex demonstrating epileptiform behavior has been created in the freeware parallel Genesis (pGenesis) environment and currently runs on 16 computational nodes. It allows the user to selectively record LFPs from various synaptic current sources, or LFPs from arrays of superficially simulated microelectrodes. The time-frequency behavior of the modeled seizures is presented, and demonstrations of spiral wave activity similar to experimental preparations are shown. References: 1. Traub RD, et al. J Neurophys 2005;93:2194-2232. 2. Anderson WS, et al. Biol Cybern 2007;97:173-194. 3. Bower JM, Beeman D. The Book of GENESIS. Springer, NY, 1997. Support: Charles H. Hood Foundation (WSA), NIH-NINDS K08 (1K08NS066099-01A1) (WSA)