Motor patterns and sensory consequence: dynamic behavior and the neuronal processing of electric flow information.

点击翻译按钮获取中文摘要

Quantitative experiments on animal behavior concerning sensory-motor skills, learning and memory formation often miss much of the richness of natural animal behavior, and are likely to result in oversimplified models of the underlying neural control of behavior. Experiments under naturalistic conditions are therefore an important complement to the understanding of sensory-motor interaction, but they face unique technical challenges, including the problem of identifying relevant cues in the sensory flow. In Mormyrid weakly electric fishes this problem can be overcome by making use of their discontinuous electrical sensory sampling: since animals actively decide when the short electric signals for exploring the environment are generated, the sensory input that is being processed can be exactly defined. These fish thus offer an exceptional potential for a better understanding of the sensory-motor loops involved in active sensing in general.To exploit this potential we want to obtain the first quantitative analysis of sensory flow in unrestrained electrosensory behavior of Mormyrid fish (Gnathonemus petersii). To determine patterns in motor behavior and the associated sensory flow we will first compile an archive of quantitative prototypical behavioral sequences and based on these reconstruct the sensory input of natural behavioral sequences. By comparing the sensory flow of different behaviors one goal is to comprehend if and how animals actively structure their sensory input. To further quantify the information embedded in the sensory flow we will compare two extremes: the information available to the sensory array from a single sampling of the environment versus the information available when considering the temporal change of the input at a single receptor during electrolocation, i.e. a spatial versus a temporal analysis of the sensory input. From this we will determine the perceptual cues contained in the temporally sculpted sensory flow, addressing how spatiotemporal dynamics help in extracting or enhancing sensory features of importance.The hypothesis that weakly electric fish actively shape their sensory flow patterns to facilitate information extraction will be studied in the neuronal level as well: up to know it is unknown if and to what extend early stages of electrosensory processing treat electrosensory spatiotemporal patterns. This will be investigated by recording ELL single unit activity in a controlled closed-loop paradigm, in which the timing of sensing is controlled by the animal while being stimulated with moving objects. The main focus will be to study how object motion influences the encoding of localization cues in the first central processing area. As this has not been addressed in Mormyrids thus far, we will study responses to linear object trajectories, focusing on information-theoretically driven comparison between dynamical information encoding versus static information encoding for stimulus conditions.