Characterization of the neural codebook in an invertebrate sensory system

Abstract

An outstanding problem in neuroscience is to describe the relationship between various stimulus sources in the environment and how they are represented by patterns of activity in nervous systems, a problem generically referred to as 'neural coding'. Most previous methods developed to address this problem have assumed a linear relationship between environmental stimuli and neural responses, and generally relied on measures of the mean state of the environment preceding neural activity to characterize the stimulus-response transformation. The goal of this thesis is to develop new methods of characterization that extend earlier work, and to demonstrate the utility of these new methods through application to an invertebrate sensory system. All applications of the methods developed in this thesis were carried out in the cercal system of crickets. The cercal system mediates the detection and analysis of low velocity air currents, and is implemented around an internal representation of air current direction that demonstrates the essential features of a continuous neural map. The stimulus feature selectivity, timing precision and coding characteristics of two bilateral pairs of primary sensory interneurons of the cercal system were characterized using three novel techniques. First, estimates of the cells' feature selectivity that take the natural variance in stimulus-response latency (i.e., spike 'jitter') into account were derived. Second, the cells' stimulusresponse relationship was probed for specific non-linear aspects that could constitute 'temporal' encoding. Third, an iterative stimulation paradigm was used to test and refine the predictions of the cercal system's stimulus selectivity. Compared to earlier characterization of this system, these new analytical procedures yield significantly different estimates of the stimulus feature selectivity of these cells. A 'code book' for the stimulus-response characteristics of these cells is presented, with emphasis on demonstrating instances where a cell represents different stimuli with distinct spike 'code-word' patterns.