Fly eyes are not still: a motion illusion in Drosophila flight supports parallel visual processing

Abstract

Most animals shift gaze by a 'fixate and saccade' strategy, where the fixation phase stabilizes background motion. A logical prerequisite for robust detection and tracking of moving foreground objects, therefore, is to suppress the perception of background motion. In a virtual reality magnetic tether system enabling free yaw movement, Drosophila implemented a fixate and saccade strategy in the presence of a static panorama. When the spatial wavelength of a vertical grating was below the Nyquist wavelength of the compound eyes, flies drifted continuously and gaze could not be maintained at a single location. Because the drift occurs from a motionless stimulus - thus any perceived motion stimuli are generated by the fly itself - it is illusory, driven by perceptual aliasing. Notably, the drift speed was significantly faster than under a uniform panorama, suggesting perceptual enhancement as a result of aliasing. Under the same visual conditions in a rigid-tether paradigm, wing steering responses to the unresolvable static panorama were not distinguishable from those to a resolvable static pattern, suggesting visual aliasing is induced by ego motion. We hypothesized that obstructing the control of gaze fixation also disrupts detection and tracking of objects. Using the illusory motion stimulus, we show that magnetically tethered Drosophila track objects robustly in flight even when gaze is not fixated as flies continuously drift. Taken together, our study provides further support for parallel visual motion processing and reveals the critical influence of body motion on visuomotor processing. Motion illusions can reveal important shared principles of information processing across taxa.

Keywords: Control; Feedback; Motion vision; Saccade; Stability.

Fig. 1.

Magnetic tether paradigm and control framework. (A) Flies were suspended within a magnetic field, surrounded (360 deg) by LED panels, and were free to rotate about the yaw axis. A high-speed camera recorded the fly's bottom position. (B) Closed-loop control diagram of flight in the magnetic tether. With a static panorama, flies produce body motion that generates visual reafference. The difference between motion and reafference generates some error (retinal slip). (C) Left: diagram of compound eye ommatidia mosaic. The separation distance between each ommatidium defines the inter-ommaditial angle Δφ. The distance about the horizontal axis is considered for vertical gratings. Right: grating defined by spatial wavelength λ. (D) Contrast ratio (actual divided by perceived contrast) as a function of spatial wavelength for Drosophila melanogaster. Acceptance angle Δρ=5 deg for the simulation. At λ=7.5 deg, the contrast ratio is ∼1%. (E) Closed-loop control diagram. Inset: proposed parallel visual motion processing pathway for object tracking and background stabilization.

Fig. 2.

Gratings of spatial wavelength below Nyquist wavelength destabilize the gaze stabilization reflex. (A) Top: example 25 s trials for the same fly presented with a static 7.5 deg (left) and 15 deg (right) spatial wavelength pattern. Bottom: angular speed data. The dashed line is the calculated threshold for saccade detection. The inset shows the drift generated by the 7.5 deg static background. Arrowheads indicate inter-saccade intervals, with marked differences between 7.5 deg (yaw drift) and 15 deg (no yaw drift) spatial wavelengths. (B) Simulation of two-dimensional flight trajectory from fly heading data by prescribing a fixed flight speed (30 cm s⁻¹). For visual clarity, a randomly selected subset of trials is shown (gray lines) and three trials are highlighted in red. (C) Angular heading data (with saccades removed) for six static gratings of different spatial wavelength and a randomly textured and uniform grating. Trials are shown for flies that drifted predominantly in the clockwise (CW, left) and counter-clockwise (CCW, right) direction. (D) Box plot of net heading angles for data in C. (E) Speed of flies for data shown in C and D. (F) Drift speed in magnetic tether with a paper drum of λ=7.5 deg (n=15 flies, 75 trials) and 9 deg (n=12 flies, 60 trials). (G) Drift speed in the magnetic tether with higher spatial resolution (each pixel subtending 1.875 deg; n=5 flies, 25 trials). The drift speed is statistically significant between 3.75 and 7.5 deg (P<0.001). (H) Spontaneous saccade dynamics. For C−E and H, n=36 flies.

Fig. 3.

Modeling of perceptual aliasing. (A) Proposed interpretation of perceptual aliasing in closed loop. A mismatch between the sign of the perceived motion direction (V_p) and the actual body velocity (V_f) elicits a non-zero body velocity due to a non-zero error, corresponding to the observed drift in the magnetic tether. (B) Hassenstein–Reichardt elementary motion detector (EMD) model with spatial filter (S), first-order, low-pass filter (LP), multiplication non-linearity (×), summation (Σ) and inter-ommatidial distance (Δφ). (C) EMD steady-state response of the analytical model as a function of spatial frequency for a fixed temporal frequency of 2 Hz. Shaded region: aliasing of visual input. (D) EMD steady-state response of the analytical model for distinct spatial wavelengths λ. For visual clarity, the 3.75 and 15 deg EMD responses were offset as they fully overlap. (E) Same as D but for a computational EMD model with a discrete low-pass filter and spatial filter simulating Drosophila optics. For all simulations, we used Δφ=4.5 deg.

Fig. 4.

Gaze fixation is not necessary for object detection and pursuit. (A) Sample 25 s trials for (top) a bar moving counter-directionally over a randomly textured background and (bottom) a bar moving over a λ=7.5 deg static background for the same fly. Top: flies generate bouts of smooth gaze stabilization (black arrowhead) interspersed with object-tracking saccades (green arrowhead). As a wide-field stimulus, the background absolute angle is arbitrary but is shown here for reference. Bottom: flies drifted in the presence of a static background and generated tracking between bouts of drifting. (B) Left: tracking saccade count for a textured bar moving counter-directionally to a randomly textured ground. Right: tracking saccade count for a textured bar moving on a λ=7.5 deg ground. n=32 flies, 18,189 saccades total, 3,195 tracking saccades total.

Fig. 5.

Rigid-tether paradigm indicates that aliasing effects are induced by body motion. (A) A fly is suspended within a virtual reality arena and wing motion is tracked to infer steering effort via changes in wing-beat amplitude (ΔWBA). (B) Open-loop control diagram of the rigid-tether paradigm. (C) Wing steering responses (ΔWBA) to the static random (left) and λ=7.5 deg grating (right). The thick black line and gray area indicate mean±1 s.d. Colored lines represent the mean for each individual fly. (D) Top: pseudo-random sequence of object position. Bottom: wing steering response of one fly to the sequence. (E) Example impulse response function between visual stimulus and steering for one fly tested at one azimuthal location. The unit of response amplitude on the scale bar is uncalibrated ΔWBA (volt degree second or V deg s). (F) Impulse responses to pseudo-random object motion measured at 24 azimuthal locations and assembled into a spatio-temporal action field (STAF) for n=12 flies.