Deciphering mechanisms regulating information processing in neurons and neural networks

Understanding of mechanisms underlying animal behavior requires probing of neuronal function at multiple scales, from single neurons to complete neuronal networks. Here we first discuss how a single olfactory neuron in C. elegans translates environmental information into sensory activity. A variety of different coding strategies have been proposed to mediate sensory detection across organisms and systems, but differentiating between models is challenging without a deliberate systematic experimental design. We measure neuronal responses over a wide range of stimulus conditions, and find that previous sensation models could only match subsets of experimental observations. We formulate an alternative adaptive concentration threshold model in which sensory activity is regulated by an absolute signal threshold that continuously adapts to odor history. The model fits measured sensory responses over a broad stimulus range and accurately predicts sensory activity and probabilistic behavior during animal navigation in odor gradients. The model generality was demonstrated by predicting activity of larval zebrafish optic tectum neurons in response to visual stimuli. Theoretical analysis shows that our model unifies previous sensation models under one mechanism and demonstrates an efficient, accurate and fast encoding strategy. Next, we present a systems approach for identifying mechanisms of neural information processing in complete neural networks. We measure the responses of all neurons in the worm head to controlled environmental stimuli, and use it to decipher their potential role in decision making and behavior plasticity.