IBN 97-28731 METZNER and KOCH The long term goal of this project is to determine how higher order nerve cells ("neurons") within the central nervous system extract behaviorally relevant features from their sensory input, a task common to all sensory systems. To address this question, we will focus on the question of how information is encoded at the primary sensory neuron level and subsequently processed by higher order neurons in the electrosensory system of weakly electric fish. These weakly electric fish produce a sinusoidally varying weak electric field in their environments using a specialized "electric organ." Electrosensory receptor cells located along the body of the fish detect modulations in the amplitude of the electrical field produced by the presence of nearby objects. Primary sensory neurons ("P-receptor afferents") transmit this information to the next level of the sensory pathway, the electrosensory lateral line lobe (ELL). There, behaviorally relevant features are extracted and encoded in the spike trains of pyramidal cells, the output neurons of the ELL. The mechanisms underlying this feature extraction will be investigated by combining neurophysiological, neuroanatomical and computational data analysis techniques. Major aims are: First, to characterize the accuracy and robustness of P-receptor afferent responses (the "nerve impulses" or "spikes" electrical fields cause) to time- varying modulations of electric field amplitude and second to perform simultaneous recordings from several pyramidal cells during presentation of the same stimuli. Accuracy and robustness of the impulse firing of primary sensory neurons will be studied using recently developed measures of "distance" between spike trains and standard signal processing techniques. The responses of pyramidal cells will be analyzed using signal detection techniques adapted to multiple spike train data. The proposed project is designed as a group proposal: experiments will be conducted in the laboratory of Dr. Metzner (UCR) and the computational analysis in the laboratory of Dr. Koch with participation of Dr. Gabbiani (both Caltech). This combination of neuroethological and computational/information theoretical approaches will clarify the nature of early sensory processing in a model sensory system ideally suited for this kind of analysis. In particular, this work is expected to lead to a quantitative understanding of the code used to represent and transmit sensory information across multiple neurons.
