This application addresses broad Challenge Area (06) Enabling Technologies and specific Challenge Topic 06- DA-102 Tool Development for the Neurosciences. Nervous systems create behaviors as a result of the concerted activity of networks of interconnected neurons, analogous to the way that companies often design new products as the result of concerted activity of multiple small groups of workers. Understanding how these networks function of course requires knowing how the neurons that make them up are interconnected and what the activity of the network's neurons is, just as understanding how a group of humans functions requires knowing who talks to whom and what is being said. However, individual neurons also have intrinsic active properties that make them respond to identical input in neuron-specific ways. These differences in intrinsic active properties are analogous to differences in human personality and have analogous results: a person's response to a given input, e.g., a statement, will differ depending on the person's overall personality and the mood she or he is in at the moment. It is thus impossible to understand how neural networks function without knowing the intrinsic active properties of the network's neurons. These properties result from the concerted activity of electrical currents in the individual neurons;neurons with different intrinsic active properties having different complements of electrical currents. One way to define neuron intrinsic active properties would thus be to describe all these membrane currents, which would allow computer modeling to predict neuron intrinsic active properties and neuron response to input on the basis of those properties. Unfortunately, present experimental techniques do not allow measuring all a neuron's membrane conductances in an individual-neuron-by-individual neuron manner. This application proposes an alternative approach for characterizing neuron intrinsic active properties, and identifying a neuron's complete set membrane currents, individual-neuron-by-individual neuron. The key idea is that measuring the response of individual neurons to experimental perturbation (in the proposed research, to the injection of electrical current) allows neuron intrinsic properties to be described in great detail. What specific input is used does not matter provided the input is complicated enough that it mimics the full range of input that a neuron will ever receive, and thus elicits the full range of responses that a neuron will ever produce. It is therefore proposed to use random perturbation sequences, as random input contains the most information per time of any sequence. Another hypothesis is that, if the perturbation is complicated enough, only one set of neuron membrane currents would be able to reproduce the observed neuron responses. This technique would thus allow building models of neurons on the membrane current level on an individual-neuron- by-individual neuron basis from every neuron recorded from. These ideas have been tested on a preliminary level using computer models of neurons and experiments in a particularly experimentally-advantageous and well-investigated system, the pyloric neural network of the lobster. However, this early work has not addressed an issue of particular importance to this submission: whether complicated perturbation can distinguish between neurons in control saline and under the influence of substances such as modulatory neurotransmitters (and drugs of abuse;see Relevance, below) that change a neuron's membrane current make-up and thus its intrinsic active properties. They have also not been tested in less well-understood systems or in vertebrate nervous systems. This application proposes to address these lacks by further work in the pyloric network, in sensory and behavior-generating neurons in the crab, and in two types of sensory neurons in a vertebrate. Demonstration that these techniques can distinguish between neurons on the basis of their intrinsic active properties, distinguish between unmodulated (control) neurons and modulated neurons, and build accurate, current-based neuron models in both control and modulated states, in all these systems (in all cases key model predictions will be tested experimentally) would be strong evidence that these ideas and techniques are generally applicable across nervous systems.
Neuron intrinsic active properties are undoubtedly important in many aspects of nervous system plasticity, including learning and memory. Most important in the present context is that psychoactive drugs likely exert some of their effects by changing neuron intrinsic active properties. The proposed physiological and modeling techniques, if validated as suggested here, would provide a new method for measuring the effects of psychoactive drugs and the normal function of the neural networks these drugs affect. The data could suggest novel treatment approaches for drug addiction and drug-induced nervous system damage.
