Coupled Electromagnetic-Electrophysiological Modeling of Neural Circuits

Much of our knowledge about the brain is derived from extracellular recordings of the electric fields produced by neural activity. However, interpretation of these signals is nontrivial. In par- ticular, it is difficult to link macro-scale observations of electric fields with micro-scale neural activity. In silico models of neural tissue may help to bridge these scales, shedding light onto the biophysical basis of voltage signals from extracellular recordings. Simulating extracellular fields produced by neural activity is nevertheless a computationally difficult problem. In this thesis, we use a reciprocity-based approach to efficiently simulate neural recordings. To do so, we have developed a state-of-the-art software package, called BlueRecording, to rapidly and flexibly calculate voltage signals from large-scale simulations of neural circuits. BlueRecording is extensively tested, and is shown to outperform previously available software in both speed and accuracy. We then use BlueRecording, in conjunction with the Blue Brain Project model of the rat somatosensory cortex, to investigate the mechanisms underlying the somatosensory evoked response to a whisker flick stimulation. We find that the initial positive deflection in the signal is the result of direct thalamic input onto pyramidal cells, primarily in cortical Layer 2/3 and 5, while the initial negative deflection is driven largely by cortico-cortical inhibition, though thalamic inputs to Layer 6 also play a role. We also use BlueRecording and the Blue Brain Project somatosensory cortex model to study the inverse current source density technique to estimate the distribution of neural currents on the basis of voltage signals recorded from penetrating electrodes. We find that as the density of the recording array increases, the estimated current source density diverges from the expected distribution. We suggest two possible explanations for this phenomenon, both related to the violation, in the simulated neural circuit, of the assumptions underlying the inverse current source density technique. Finally, we apply the reciprocity-based approach to calculating neural signals to develop a simplified model of the evoked compound action potential signal generated by vagus nerve stimulation. We determine that in multifascicular nerves, incomplete activation of the fiber population in a fascicle has a strong impact on the recorded signal. We find that the signal is sensitive to the orientation of both the stimulus and recording electrodes relative to the fascicles, and that, contrary to popular belief, the amplitude of the recorded signal is not monotonic with respect to fiber activation. Our approach of simulating extracellular recordings of neural circuit models is therefore able to generate insights into the relationship between micro-scale activity and macro-scale signals. In addition to explaining the biophysical basis of observed phenomena, which may lead to clinically relevant insights, this