Sensory coding in an identified motion-sensitive visual neuron of the locust (Locusta migratoria)
McMillan, Glyn Allan
Visual environments may contain a complex combination of object motion. Animals respond to features of complexity by generating adaptive behavioural responses. One important feature of a complex visual environment is a rapidly expanding object in the visual field (looming) which may represent an approaching predator or an object on a collision path. Many animals respond to looming objects by generating avoidance behaviours (Maier et al. 2004; Santer et al. 2005; Oliva et al. 2007) and neurons involved in the detection and relay of looming stimuli are present in birds (Sun and Frost 1998) and many insects (Simmons and Rind 1992; Hatsopoulos et al. 1995; Wicklein and Strausfeld 2000). One of the most widely studied visual pathways is found in the locust. This visual pathway, which includes the lobula giant motion detector (LGMD) and its post-synaptic target, the descending contralateral motion detector (DCMD), signals the approach a looming visual stimulus (Schlotterer, 1977; Simmons and Rind, 1992; Hatsopoulos et al., 1995). The DCMD descends through the ventral nerve cord and synapses with motorneurons involved in predator evasion and collision avoidance (Simmons, 1980; Simmons and Rind, 1992; Santer et al., 2006). Previous studies have suggested that this pathway is also affected by more complicated movements in the locust’s visual environment. For example, Guest and Gray (2006) demonstrated that the approach of paired objects in the azimuthal position and approaches at different time intervals affect DCMD firing rate properties. In my first objective of this thesis (Chapter 2), I tested locusts with computer-generated discs that traveled along a combination of non-colliding (translating) and colliding (looming) trajectories and demonstrate how distinctly different DCMD responses result from different trajectory types. In addition to estimating the time of collision and direction of object travel, the presence of a discernable peak associated with the time of object deviation suggests that DCMD responses may contain information related to changes in motion. Previous studies suggest that LGMD/DCMD encodes approaching objects using rate coding; edge expansion of approaching objects causes an increased rate of neuronal firing (Schlotterer, 1977; Hatsopoulos et al., 1995; Judge and Rind, 1997; Gabbiani et al., 1999). Based on observations of DCMD responses to simple looming objects that showed oscillations in DCMD responses (for example, Fig. 1D Santer et al., 2006) and the fact that bursting occurs in many other sensory systems (Yu and Margoliash, 1996; Sherman, 2001; Krahe and Gabbiani, 2004; Marsat and Pollack, 2006), it was hypothesized that the DCMD may show bursting activity. In my second objective of this thesis (Chapter 3), I tested locusts with simple looming stimuli known to generate behavioural responses in order to identify and quantify bursting activity. Results show that the highest frequency of bursts occurred at intervals of 40-50 ms (20-25 Hz). The behavioural significance of this frequency is related to the average wingbeat frequency of the locust’s forewing during flight (~25 Hz; Robertson and Johnson, 1993). Based on previous evidence of DCMD flight-gating (see, for example, Santer et al., 2006), bursting may gate information into the flight circuitry, thereby providing visual feedback that may be modified to generate an avoidance response during flight. Single spiking and bursting occurred throughout object approach up until the late stage of approach, where burst frequency rapidly increased. Results predict that the DMCD may use a bimodal coding strategy to detect looming visual stimuli, where single spiking at the beginning of approach may result in subtle course changes during flight and bursting near the time of collision may initiate an evasive glide. Taken together, these results illustrate that the encoding of visual stimuli in single neurons is dynamic and likely much more complicated than previously thought.
DegreeMaster of Science (M.Sc.)
CommitteeFisher, Thomas; Niyogi, Som; Campanucci, Veronica; Chivers, Doug
Copyright DateAugust 2009