Abstract

In the past few decades, the lobula plate of the fly has emerged as one of the leading models for the neural processing of optic flow stimuli that give rise to visual orientation behaviors (for recent reviews see Borst and Haag, 2002; Egelhaaf et al., 2002; Egelhaaf et al., 2002; Borst and Haag, 2007). The relative simplicity and accessibility of this neural system allows researchers to characterize the neural mechanisms that are thought to link the visual stimuli and the resulting behavioral responses. In the lobula plate, a set of 60 motion sensitive lobula plate tangential cells (LPTCs) integrate visual motion information from an array of local motion detectors, which form a retinotopic map of the fly’s visual space in the lobula plate. The selective pooling of local, direction selective inputs, together with a network of unilateral and bilateral interactions between LPTCs, shape and tune the response properties of LPTCs to behaviorally relevant optic flow stimuli. Over the years, lobula plate researchers assembled a formidable array of measurement and perturbation techniques that are usually available only in in-vitro systems. Additionally, the lobula plate and its presynaptic circuitry have been the subject of extensive and detailed modeling which allows a deeper synthetic understanding of the empirical results, as well as a more efficient and detailed way to generate hypotheses. In this work I used a selection of these tools to explore the role of intracellular processing of visual motion information in lobula plate neurons and the significance of spatial segregation and aggregation of these cells’ inputs in the context of their sensory function. Previous work on a network of ten LPTCs of the vertical system (VS cells) resulted in a prediction that due to lateral, gap-junction coupling of neighboring VS cells in their axon-terminals, the receptive fields of these cells should be broader in the axonal region than in the dendritic regions. I tested and confirmed this prediction using in-vivo calcium imaging and intracellular recordings. Using single-electrode voltage clamp I was able to perturb the flow of information in these cells and isolate the source of input responsible for this broadening, confirming that the coupling indeed takes place in the axon terminal. The separation of feed-forward, synaptic input in the dendrites from lateral, gap-junction coupling in the axon-terminals allowed me to experimentally ask what is the function of the receptive field broadening. Relying on model predictions, I showed that this broadening results in a more stable and smooth representation of optic flow in the output region of the cells than in their input region, when the fly is presented with naturalistic, patchy and non-uniform stimuli. I then showed, using a simplified compartmental model that the separation of axonal gap-junctions from the dendritic synaptic input makes the gap-junction coupling more effective, and is thus necessary to ensure the functionality of the lateral interactions.