Stabilizing responses to sideslip disturbances in Drosophila melanogaster are modulated by the density of moving elements on the ground

Abstract

Stabilizing responses to sideslip disturbances are a critical part of the flight control system in flies. While strongly mediated by mechanoreception, much of the final response results from the wide-field motion detection system associated with vision. In order to be effective, these responses must match the disturbance they are aimed to correct. To do this, flies must estimate the velocity of the disturbance, although it is not known how they accomplish this task when presented with natural images or dot fields. The recent finding, that motion parallax in dot fields can modulate stabilizing responses only if perceived below the fly, raises the question of whether other image statistics are also processed differently between eye regions. One such parameter is the density of elements moving in translational optic flow. Depending on the habitat, there might be strong differences in the density of elements providing information about self-motion above and below the fly, which in turn could act as selective pressures tuning the visual system to process this parameter on a regional basis. By presenting laterally moving dot fields of different densities we found that, in Drosophila melanogaster, the amplitude of the stabilizing response is significantly affected by the number of elements in the field of view. Flies countersteer strongly within a relatively low and narrow range of element densities. But this effect is exclusive to the ventral region of the eye, and dorsal stimuli elicit an unaltered and stereotypical response regardless of the density of elements in the flow. This highlights local specialization of the eye and suggests the lower region may play a more critical role in translational flight stabilization.

Keywords: Drosophila melanogaster; dot field density; flight control; insect vision; optomotor response; regional optic flow.

Figure 1.

(a) Projection arena used to present stimuli to specific regions of the visual field on flies. (b) Tethered flies held under an IR light at the centre of the arena cast a shadow over a sensor below. Changes in wing-beat amplitude (ΔWBA) alter the size of the shadow of each wing and are registered by the sensor as voltages. Differences in amplitude between both wings represent steering attempts. (c) Responses to regional stimuli were elicited by exposing the flies to laterally moving dot patterns projected on the upper or lower faces of the projection arena. (d) Spatial and temporal frequencies in the stimuli. Coloured regions represent limits of perception for spatial (blue), and temporal (red) frequencies in flies. (e) Mean steering response (solid lines) and standard error of the mean (s.e.m.) (colour shading) of 50 flies, elicited by a moving dotfield with 4.6% element density, presented to the dorsal and ventral region of the eye at t = 0. The shading represents the interval over which responses were compared between treatments.

Figure 2.

Relative steering responses (ΔWBA) to sideslip disturbances in coherently moving dot fields as a function of visual element density in dorsal (blue) and ventral (green) visual regions (FOV = 90°, n = 50). (a) Means (black) steering responses to dorsal (i) and ventral (ii) sideslip disturbances across five levels of element density. Violin plots represent the distribution of data within interquartile ranges, along with the median (grey). Mean responses labelled with different letters are significantly different when compared within the ventral region of the eye (Bonferroni post hoc analysis, p ≤ 0.05). (b) Heatmaps of mean temporal series at each level of element density in the dorsal (i) and ventral (ii) regions of the eye. Dark tones represent relatively stronger responses (ΔWBA). (c) Mean time series of steering responses across levels of element density. Envelopes represent s.e.m. and coloured rectangles the time window within which responses were analysed. (d) Post hoc pairwise t-test for multiple comparisons among both regions of the eye and all levels of element density. Bonferroni-adjusted probabilities are represented as shades of grey according to their value. Not significant differences (NS) are represented in light grey.