QUICK SEARCH:   [advanced] Author: Keyword(s):
Home     Help     Feedback     Subscriptions     Archive     Search     Table of Contents

This Article Summary Full Text Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Related articles in JEB Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Tammero, L. F. Articles by Dickinson, M. H. Search for Related Content PubMed PubMed Citation Articles by Tammero, L. F. Articles by Dickinson, M. H.

Collision-avoidance and landing responses are mediated by separate pathways in the fruit fly, Drosophila melanogaster

Lance F. Tammero^1,* and Michael H. Dickinson²

¹ UCB/UCSF Joint Bioengineering Graduate Group, University of California, Berkeley, CA 94720, USA
² Department of Integrative Biology, University of California, Berkeley, CA 94720, USA

View larger version (22K): [in a new window]   Fig. 1. Schematized experimental setup for measuring a fly's response to image expansion. During tethered flight, the fly's wingstroke amplitude and frequency are measured by optically tracking the shadows cast from an infra-red (IR) diode by each of the wings on an optical wing-beat analyzer (Dickinson and Lighton, 1995; Lehmann and Dickinson, 1997). During closed-loop experiments, the difference between the amplitude of each wing stroke controls the visual display, allowing the fly to orient actively toward the position of the 15°x15° square. At periodic intervals, the square symmetrically expands, eliciting a behavioral response.

View larger version (25K): [in a new window]   Fig. 2. Wing and leg responses elicited by an expanding object (recorded as V). In response to a square expanding at a rate of 500° s-1, the fly generates both wing and leg responses. The time course of stimulus expansion is shown in the bottom traces. If the object is displaced laterally, the inside wing (that on the side of the stimulus) shows a transient increase in wing-beat amplitude, while the outside wing decreases in stroke amplitude. (A) If the object is to the left of the fly, the left wing-beat amplitude (blue) increases while the right wing-stroke amplitude (red) decreases, causing the square to move to the rear of the fly's field of view. In contrast, expansion of centrally positioned objects elicits smaller changes in wing motion, causing little change in the position of the object (B). Image expansion in the frontal field of view elicits leg extension as well as an increase in wing-beat frequency, both indicative of a landing response. When the stimulus is to the right of fly, the sign of the change in both wing-beat responses is reversed, again causing the object to move to the rear of the fly's field of view (C). Laterally positioned image expansion elicits a transient increase in wing-beat frequency but does not evoke a leg response.

View larger version (51K): [in a new window]   Fig. 3. The effect of stimulus position on behavioral response. (A) A single fly's response to multiple presentations of a square expanding at 500° s-1 varies with stimulus position. Each individual trace shows the response of the left (blue) and right (red) wing to a presentation of the expansion stimulus. The bold and dotted lines represent the mean response ± S.D. for stimuli between given positions. Expansion in lateral positions evokes the largest change in wing-beat amplitude (WBA), with responses decaying for more frontal and caudal stimulus presentations. (B) Results from multiple flies. The individual traces are the mean left and right wing-beat amplitude response taken from 12 individuals. The bold and dotted lines represent the mean ± S.D., respectively, across individuals.

View larger version (24K): [in a new window]   Fig. 4. Collision-avoidance and landing responses vary with the position of stimulus expansion. (A) The maximum change in value of the wing-beat amplitude (WBA) from the baseline level of both the right (R; red) and left (L; blue) wings varies sinusoidally with the position of the stimulus. (B) A similar variation occurs for the maximum change in the difference between the left and right wing signals. (C) The percentage change in wing-beat frequency (WBF) was largest for expansion occurring in front of the fly and decreases slightly for lateral positions. (D) The probability of eliciting a landing response is greatest for frontal positions. Data points represent the mean value of maximum change ± S.E.M. The number of trials at each position is different because it was determined by where the fly happened to position the object at the onset of expansion. Data are taken from 300 presentations of a square expanding at a rate of 500° s-1 to a single fly.

View larger version (35K): [in a new window]   Fig. 5. The effects of expansion rate on collision-avoidance and landing responses. Each column represents the wing-beat amplitude (WBA), and the landing response probabilities plotted against stimulus position as described in Fig. 4 for a different rate of expansion. The functions shown in Fig. 4 were determined for each fly, with each data point representing the mean ± S.E.M. taken over all the flies. The numbers of flies tested were 8, 11, 11, 8, 12, 11, 8, 10, 7, 7 and 5 for expansion rate in ascending order (starting at 100° s-1). The total numbers of stimulus presentations, again in ascending order, were 1945, 2905, 2584, 1746, 3237, 2529, 216, 2223, 1771, 1391 and 1132. The sinusoidal shape of the wing-stroke amplitude responses holds for all expansion rates, with the amplitude of the response being largest for an expansion rate of 1000° s-1. The probability of landing is high over the greatest range of positions at an expansion rate of 1430° s-1. L, left; R, right; WBF, wing-beat frequency.

View larger version (11K): [in a new window]   Fig. 6. Summary of changes in collision-avoidance and landing responses with rate of image expansion. The collision-avoidance response for a given expansion rate (open circles) is the sinusoid amplitude best fitting the maximum change in the difference between wing-beat amplitudes (see Fig. 5, second row), normalized by the maximum mean amplitude. The width of the range of positions for which the probability of landing is greater than 0.5 characterizes the landing response for a given expansion rate (filled circles). This response is normalized by the maximum mean width value. Values are means ± S.E.M. for each fly.

View larger version (22K): [in a new window]   Fig. 7. Effect of stimulus position and expansion rate on the time course of the wing response. Responses to stimuli presented within ± 10° of the position were pooled. Each trace represents the mean ± S.D. (shaded area) of the average responses taken from multiple flies. The time course of the responses does not vary with stimulus position but does vary greatly with rate of expansion. The number of flies at each expansion rate is given in Fig. 5. L, left; R, right.

View larger version (14K): [in a new window]   Fig. 8. Collision-avoidance and landing response latencies depend on stimulus position and expansion rate. Latency is measured as the time interval between the onset of image expansion and the initiation of the landing or collision-avoidance response. (A) Latency in response to expansion at a rate of 500° s-1 is relatively constant over lateral portions of the fly's field of view and increases for positions to the front and rear. Data points represent mean latency ± S.E.M. for 12 flies. (B) Landing response latency to a square expanding at 500° s-1 is constant at the stimulus positions at which landing response probability is high. At this expansion rate, the collision-avoidance latency is approximately half that of the landing response. (C) Response latencies plotted as a function of expansion rate. For a given rate of expansion, the minimum of the mean delay functions (such as the two plotted above) was determined. Filled circles represent the minimum mean delay in the landing response, while empty circles represent the minimum mean delay of the collision-avoidance response. The landing response latency decreases with the rate of expansion, whereas for most expansion rates the delay of the collision-avoidance response is constant.

View larger version (22K): [in a new window]   Fig. 9. Open-loop versus closed-loop responses to image expansion. (A) During open-loop presentation the position of the square was controlled externally, as opposed to the closed-loop paradigm in which the fly maintains control over the position of the square. The closed-loop wing responses (filled circles) are repeated from Fig. 4. The open-loop responses (open circles), also generated in response to an expansion at a rate of 500° s-1, vary roughly with stimulus position as a square wave, in contrast to the open-loop responses, which vary sinusoidally. Thus, the ability to control the position of the square during the collision-avoidance reaction does affect the amplitude of the response. (B) The probability of landing is slightly reduced for open-loop presentations. (C) The latency of the collision-avoidance response is qualitatively similar for the closed- and open-loop stimuli, with slightly larger latencies in response to open-loop image expansion. (D) The open-loop landing response latencies were qualitatively similar to those seen during closed-loop presentations. Again, the latency is slightly shorter during closed-loop presentations. WBA, wing-beat amplitude.

View larger version (22K): [in a new window]   Fig. 10. Comparison of the time course of responses for closed-loop and open-loop presentations. The changes in wing-beat amplitude (WBA) in response to a square expanding at 500° s-1 positioned between -140° and -120° followed a similar time course for open- and closed-loop presentations. The responses elicited by closed-loop presentation of the square were slightly smaller in magnitude than those in response to open-loop presentations. Closed-loop responses were taken from Fig. 3B; open-loop responses were taken from 5 flies in a manner analogous to the data plots in Fig. 3B.

View larger version (20K): [in a new window]   Fig. 11. Model for eliciting collision-avoidance and landing responses. A fly estimates the optic flow experienced during flight using a two-dimensional array of motion detectors (i). Local motion information is then spatially pooled such that the image expansion in both the lateral and frontal fields of view is calculated (ii). The outputs of each of these three expansion calculations are then temporally integrated (iii) and passed through a threshold detector (iv). Expansion detected in a lateral field of view triggers a collision-avoidance response in the opposite direction, while frontal image expansion causes a landing response (v). Lateral expansion on one side inhibits the opposite expansion pathway, preventing a saccade from being immediately followed by another saccade in the opposite direction.