A recent study published in “Neuron” provides insights into the interplay between the cellular properties of single neurons, their connections and the network architecture.
Cortical neurons receive thousands of excitatory and inhibitory signals, both from local and long-range circuits. To keep the neurons in a functional dynamic range, near their firing threshold – for example, to detect new sensory stimuli – the excitatory and inhibitory signals have to be balanced. Substantial perturbation of this balance between excitatory and inhibitory signals results in inadequate responses of the neuron to stimuli – either the neuron remains completely quiescent or it is active in an uncontrolled way, as during an epileptic seizure. Commonly, models of the cortex assume that circuit architecture is homogeneous. For example, the number of excitatory and inhibitory connections to each neuron is assumed to be highly coordinated. This allows each neuron to remain in balance between excitation and inhibition. However, anatomical studies have now shown that this is not the case! Excitatory and inhibitory connections do not follow uniform patterns. The functional effect of such structural heterogeneity has been investigated by scientists of the group “In Silico Brain Sciences” headed by Marcel Oberlaender at research center caesar in Bonn in collaboration with a group of scientists headed by Haim Sompolinsky, and was supported by the Max-Planck-Hebrew Center, the Max Planck Society, and the Hebrew University in Jerusalem. The study focused on circuits in the somatosensory cortex of rats, which processes touch sensations. Using 3D reconstructions of the neural networks and numerical simulation of the activity of neurons in this brain region, the scientists found that structural heterogeneity can have substantial effects on cortical function. For example, in networks with realistic structural heterogeneity, very few neurons would react to sensory signals – a prediction that cannot be reconciled with activity measurements. The study therefore describes how properties of individual neurons can compensate structural imbalance, resulting in a network state where neurons can again adequately respond to stimuli. For example, individual neurons can reduce their excitability and thereby dynamically adapt their activity. When these properties were taken into account in the simulations, the network returned to a balanced state with activity patterns as observed in living animals. “Our detailed model of rat cortical circuits provides insights into the interplay between the cellular properties of single neurons, their connections and the network architecture, and how realistic activity patterns result from their interactions,” says Oberlaender. “We also demonstrate strategies of how to test theories about brain function by simulating models of the neuronal circuitry.”
Landau, I.D., Egger, R., Dercksen, V.J., Oberlaender, M. & Sompolinsky, H. (2016) “The Impact of structural heterogeneity on excitation-inhibition balance in cortical networks” Neuron 92, 1106-1121
Caesar ist associated with the Max Planck Society und runs a research center for neurosciences.
Dr. Marcel Oberlaender,
Head of Max Planck Research Group “In Silico Brain Sciences”
Website: In Silico Brain Sciences