From neuronal networks to behavior: dynamics of spontaneous activity and onset of movement in the leech

Animal behavior was once seen as a chain of reactions to stimuli from the environment. From chemotaxis in bacteria to mammals withdrawing from painful stimuli, most of the actions taken by animals are clearly driven by external inputs. Reflexes were among the first phenomena to be studied to have an insight on the dynamics of the nervous system. Later, a step forward was the discovery of central pattern generators: once a behavior is started by a stimulus, some neuronal networks are able to maintain it without further inputs from the environment. The nervous system of all animals, however, is so complex that is displaying a rich dynamics even in the absence of external inputs or, in a more realistic situation, when no single input is able to drive a clear-cut reaction. In the same way, at the motor output level, animals keep moving in the absence of evident stimuli. These spontaneous behaviors are still far from being understood. Difficult problems are often easier to solve in simple systems. The leech has a relatively simple nervous system, composed of ~103 neurons disposed in a regular structure, but at the same time displays a variety of different behaviors. It seems then a good preparation to approach the spontaneous dynamics problem. The aim of my PhD research is to describe the spontaneous behavior of the leech and the spontaneous activity of its nervous system. A first, necessary step for this study was to develop a method of automatic classification and analysis of the leech movements. Thanks to this method we described accurately the properties of the different behaviors: we focused particularly on the largely unknown irregular exploratory behavior, which is found to display a broad range of oscillation frequencies and displacement speeds, but with some recurrent movement patterns. Finding the complete list of the leech spontaneous behaviors, and the probability of the transitions between them, it was possible to demonstrate that decision making in the leech is a Markovian process. The spontaneous activity in the isolated leech ganglion was found to be characterized by long-term correlations and a large variability in bursts size and duration. The same dynamics was observed in dissociated culture of rat hippocampal neurons, despite the difference in the structure between the two networks. We studied the effects of pharmacological modulations of inhibitory and excitatory processes on the spontaneous activity, and the role of single identified motor neurons in spontaneous bursts. Finally we proposed a simple statistical model accounting for experimental results. We studied then the spontaneous activity of the leech ganglion when it was connected to the other ganglia and in the semi-intact moving animal. Inputs received from the head and tail brain caused a drastic change in the activity of the ganglion, increasing synchronization among neurons and leading to a regime dominated by very large bursts. By recording at the same the movements of the leech and its nervous activity it was possible to have a better understanding of the relationship between the motor neuron bursts and the onset of movements.