Facilitation in dragonfly target motion detecting neurons.

Abstract

Dragonflies are masters of visually guided aerial pursuit. Their visual systems overcome the challenges presented by the environment in which they live and behave. These adaptations come from multiple stages of processing. Perhaps the most important adaptations are those that allow for the detection, identification, pursuit and capture of small moving features in a complex visual environment. Perching Libellulid dragonflies execute prey capture flights with a 97% success rate from a stationary start point (average pursuit time, 1B4ms, Olberg et al,, 2000). The Australian hawking dragonfly, Hemicordulia tau, performs similar prey capture flights on the wing. H. tau also engages in longer duration territorial and courtship pursuits of conspecifics and other anisopterans, which can last tens of seconds. These chase sequences include changes in velocity and direction, as well as the possibility of target occlusion by background features. The dragonfly brain has been found to contain neurons specialised to respond to small moving target-like objects (O'Carroll, 1993; Olberg, 1986). These small target motion-detecting (STMD) neurons likely play an important role in some or all of these pursuit flights. One of these neurons, the centrifugal STMD L (CSTMD1) (Geurten et al., 2007), has recently been shown to respond more robustly following periods of slow continuous target motion [Nordström et al., 2011). This enhancement of response following target motion or self-facilitation is fascinating and was the focus of enquiry throughout this thesis. I probed this facilitation by quantifying neuronal response to stimulation from intracellular, in vivo recordings from dragonfly small target motion detecting neurons, mostly CSTMDl and the newly identified and characterised neuron BSTMD1 (work from this thesis, published in Dunbier el a1.,2012). The key results of the thesis confirmed that the slow onset time course first observed in CSTMD1, (Nordstrom et al., 2011) was in fact due to a facilitation rather than slow kinetics (using a new computational model). This facilitation effect is most pronounced when targets move at velocities slower than what was previously thought to be optimal in CSTMD1 (Geurten et al., 2007). Further, the facilitation is evident in other wide-field STMD neurons like BSTMD1. The facilitation is disrupted if targets are instantaneously displaced a few degrees from their current location. I investigated this mechanism more fully in the next publication. Presenting single target stimuli that moved along an initial 'priming’ path before undergoing spatial, temporal or combined discontinuities in their trajectories, I found that facilitation is initially spatially localized. When larger spatial displacements were combined with a delay in reappearance, however, responses were significantly elevated, even for a 20' displacement with a 500ms delay in reappearance. Backward displacements (i.e. across previously traversed location) yield strongly inhibited responses. This suggests that facilitation is mediated by a process of local gain modulation that actively spreads from the last seen location of a stimulus and in the approximate direction of travel. Such predictive modulation of local target salience may be a key mechanism for selective attention during target tracking.