Portal de Periódicos da CAPES

Behavioral Neurobiology: A Vibrating Gyroscope Controls Fly Steering Maneuvers

2007; Elsevier BV; Volume: 17; Issue: 4 Linguagem: Inglês

10.1016/j.cub.2006.12.021

1879-0445

Mark A. Frye,

Visual perception and processing mechanisms

A clever ‘virtual reality’ experiment reveals that specialized mechanosensory organs, rather than the eyes, orchestrate the high-performance staccato turns that characterize the flight behavior of a fly. A clever ‘virtual reality’ experiment reveals that specialized mechanosensory organs, rather than the eyes, orchestrate the high-performance staccato turns that characterize the flight behavior of a fly. Imagine a fly's-eye-view of the world. It moves fast, very fast. Our own visual system would be useless if we moved at similar relative speeds. Even at the earliest step — phototransduction — in flies we find the fastest electrochemical kinetics yet measured [1Hornstein E.P. O'Carroll D.C. Anderson J.C. Laughlin S.B. Sexual dimorphism matches photoreceptor performance to behavioral requirements.Proc. Biol. Sci. 2000; 267: 2111-2117Crossref PubMed Scopus (59) Google Scholar]. Put simply, fly vision is built for speed. But as fast as the visual system is, recent work from Bender and Dickinson [2Bender J.A. Dickinson M. A comparison of visual and haltere-mediated feedback in the control of body saccades in Drosophila melanogaster.J. Exp. Biol. 2006; 209: 4597-4606Crossref PubMed Scopus (77) Google Scholar] shows that feedback from a gyroscopic sense organ sets the outer performance limits of a fly's high-speed world. While chasing mates, a male housefly can alter course a mere 40 milliseconds after its mark changes heading [3Wagner H. Flight performance and visual control of flight of the free-flying housefly (Musca domestica L.) III. Interactions between angular movement induced by wide- and small field stimuli.Phil. Trans. R. Soc. Lond. B. 1986; 312: 581-595Crossref Google Scholar]. The salient visual stimulus must be detected, processed by pre-motor networks, and transformed into a motor code for the muscle mechanics, wing kinematics and aerodynamics necessary to steer a turn. This entire neuro-mechanical cascade occurs within one third the time course of a human eye blink [4Fry S.N. Sayaman R. Dickinson M.H. The aerodynamics of free-flight maneuvers in Drosophila.Science. 2003; 300: 495-498Crossref PubMed Scopus (424) Google Scholar]. However, the remarkable steering behaviors of flies are not restricted to chasing sequences. Even species that don't pursue one another on the wing, such as fruit flies, reiterate a rhythm of straight flight paths interspersed with rapid 90° steering maneuvers called body saccades, named after our own gaze-stabilizing eye movements (Figure 1A,B). The fly visual system is fast, yet optical and physiological limits ensure that visual perception will be compromised to some extent during routine body saccades. Are these maneuvers evoked slowly enough to avoid the corrupting influence of motion blur? Or do flies have non-visual mechanisms for controlling the time course and amplitude of body saccades? All flies are equipped with elaborate neuro-mechanical ‘gyroscopes’ called halteres, which mediate powerful equilibrium reflexes during flight [5Dickinson M. Haltere-mediated equilibrium reflexes of the fruit fly, Drosophila melanogaster.Phil. Trans. R. Soc. Lond. B. 1999; 354: 903-916Crossref PubMed Scopus (143) Google Scholar]. The halteres look like tiny dumbbells that beat back and forth like the wings. Owing to the conservation of momentum, if the fly's body rotates, Coriolis forces distort the beating path of the halteres [6Pringle J.W.S. The gyroscopic mechanism of the halteres of Diptera.Phil. Trans. R. Soc. Lond. B. 1948; 233: 347-384Crossref Google Scholar]. The system acts like a vibrating gyroscope, the physical principle for which can be illustrated by the action of a tuning fork (Figure 1C). When rotated, the vibrating fork is subjected to Coriolis forces acting orthogonal to the primary vibrations. In principle, for halteres or any pair of oscillating test masses, the displacement from the plane of oscillation could be used as a feedback signal for the rate of body rotation. In the fly, Coriolis-imposed haltere deflections are detected by highly sensitive mechanosensory strain gauges. It is well known that the haltere and visual systems interact to keep the fly on a straight path between saccades [7Sherman A. Dickinson M. Summation of visual and mechanosensory feedback in Drosophila flight control.J. Exp. Biol. 2004; 207: 133-142Crossref PubMed Scopus (89) Google Scholar]. Bender and Dickinson [2Bender J.A. Dickinson M. A comparison of visual and haltere-mediated feedback in the control of body saccades in Drosophila melanogaster.J. Exp. Biol. 2006; 209: 4597-4606Crossref PubMed Scopus (77) Google Scholar] asked whether saccades themselves reflect ballistic commands that, once initiated, always run to completion. Alternatively, are saccades controlled by continuous sensory feedback? If feedback is necessary, do the signals originate with the visual system, the mechanosensory system, or some combination of both? Meeting these challenges is not as simple as posing them. How to systematically and independently vary visual and mechanical stimuli? Bender and Dickinson [2Bender J.A. Dickinson M. A comparison of visual and haltere-mediated feedback in the control of body saccades in Drosophila melanogaster.J. Exp. Biol. 2006; 209: 4597-4606Crossref PubMed Scopus (77) Google Scholar] came up with a clever device, a visual-mechanical ‘virtual reality’ system. In this system, a fruit fly glued to a small steel wire (itself a delicate undertaking) is suspended vertically between two powerful magnets spaced about an inch apart. The wire holds fast to the bottom edge of the upper magnet, but the fly dangles just above the lower magnet without touching it. The magnetic field keeps the pin oriented vertically while providing a near-frictionless pivot joint. The end result is that, by beating its wings the tethered fly can spin freely about the horizontal (yaw) axis, distorting the haltere beating plane and activating mechanosensory feedback as in free-flight. This whole apparatus is surrounded by a cylindrical array of light emitting diode panels, such as those used to make digital readouts like scrolling stock tickers. The fly's changes in flight orientation are tracked in real-time with an infrared video system. Under these conditions, it is fairly straightforward to manipulate the fly's visual environment. In response to moving a striped ‘drum’, the fly pivots back and forth in an attempt to minimize retinal slip — a classical opto-motor response [8Götz K.G. Flight control in Drosophila by visual perception of motion.Kybernetic. 1968; 4: 199-208Crossref PubMed Scopus (212) Google Scholar]. Owing to the physics of haltere action, the mass of the oscillating appendage is directly proportional to the magnitude of the induced Coriolis forces (Figure 1C). Bender and Dickinson [2Bender J.A. Dickinson M. A comparison of visual and haltere-mediated feedback in the control of body saccades in Drosophila melanogaster.J. Exp. Biol. 2006; 209: 4597-4606Crossref PubMed Scopus (77) Google Scholar] either added some harmless epoxy to the tiny organs to increase their mass, or cut off the bulbous end-knob to reduce mass. The fly's saccade dynamics changed in the manner predicted by a Coriolis-dependent sensory feedback mechanism: weighty halteres resulted in overestimated body rotation and hence smaller saccade angles, whereas truncated halteres resulted in underestimated rotation and larger saccade angles (Figure 1D). To determine whether visual feedback has an additional influence on saccade dynamics, Bender and Dickinson [2Bender J.A. Dickinson M. A comparison of visual and haltere-mediated feedback in the control of body saccades in Drosophila melanogaster.J. Exp. Biol. 2006; 209: 4597-4606Crossref PubMed Scopus (77) Google Scholar] devised a way to vary the velocity of the visual display in real-time depending upon the fly's own steering dynamics. This enabled them to manipulate the magnitude and direction of the visual feedback that a fly experienced during a saccade. This clever feat of engineering was for naught, because no combination of syndirectional or counterdirectional visual stimuli had significant impact on the time course or amplitude of saccades. Taken together, these results show, first, that fly body saccades are not ballistic motor programs but rather are controlled by continuous sensory feedback, and second, that the mechanosensory haltere system contributes the relevant feedback signals. These findings are fascinating, in part because they displace a common presumption that vision is the most significant sensory modality contributing to the staggering ecological success of the winged insects. Is it not opto-motor responses that enable behavior as robust and sophisticated as fly flight? In the case of Diptera, the answer is no, not entirely. Indeed, genetically blinded fruit flies, such as photopigment-defective ninaE17 mutants, can fly — they crash about like drunken sailors, but they can fly. But remove a fly's halteres and there is no hope whatsoever for controlled flight. A fly's astonishing behavioral repertoire relies upon the rapid integration of visual and mechanosensory feedback signals to remain airborne, on course, and clear of obstacles [9Frye M. Gray J. Mechanosensory integration for flight control in insects.in: Christensen T.A. Methods in Insect Sensory Neuroscience. CRC Press, Boca Raton2006: 107-129Google Scholar]. Disclosing the cellular mechanical mechanisms by which relatively sluggish tonic visual signals descending from the brain are integrated or ‘fused’ with phasic wing-beat-synchronous mechanosensory signals remains to be explored. The results of these analyses will undoubtedly extend beyond the realm of fly neurobiology and shed valuable light on the general mechanisms of multisensory fusion and sensory-motor integration controlling high-performance behaviors across animal taxa.