Modeling insect inspired mechanisms of neural and behavioral plasticity

Abstract

This thesis investigates neural activity underlying olfactory processing and associative learning in the insect brain. Chapter 2 presents a simplified version of a model of olfaction in the fly brain, that processes sensory input in order to generate appropriate motor commands that control the activity of a robot. This spiking neural network control architecture is tested in a simple conditioning experiment. Chapter 3 investigates the neural activity in synaptic terminals at the mushroom body input of the honeybee brain. For this, data from a classical conditioning experiment was analyzed. It is shown that individual behavior is correlated with learning-induces changes in neural responses towards rewarded odors. A possible explanation for this is described in detail in chapter 4 in form of a network model of the honeybee brain. Individual stages of olfactory processing are expressed in an abstract computation model, including a hypothesis on neural plasticity that is supported by the results of chapter 3. Subsequently, this hypothesis is evaluated based on a rich collection of data from elemental and non-elemental learning paradigms. Therefore, chapter 4 provides a link between behavior and neurophysiological knowledge about odor processing in the honeybee brain.