%0 Figure %A Gomez-Marin, Alex %A Iyengar, Balaji %D 2012 %T A simple method to study Drosophila leg movements during free climbing %U https://figshare.com/articles/figure/A_simple_method_to_study_Drosophila_leg_movements_during_free_climbing/96615 %R 10.6084/m9.figshare.96615.v1 %2 https://ndownloader.figshare.com/files/101013 %K climb %K Drosophila %K locomotion %K tracking %K Behavioral Neuroscience %X

Exquisite new automated methods are now available to track the leg movements of Drosophila during walking and grooming behaviors (Kain et al. 2012). The tethered approach is most suitable for parallel functional-imaging of the brain (Seelig et al. 2010). Simpler, but yet effective automated protocols for analyzing walking kinematics of any insect, crustacean or arachnid at the level of leg swing and stance states will be very useful not only for data-mining towards bio-robotics applications but also for high throughput screening for neurological effects. With regard to Drosophila walking behavior (Strauss and Heisenberg 1990), accurate automated measurements of gait parameters will facilitate a better characterization of the genetic basis of nervous system disorders (for example, White et al. 2010; Wosnitza et al. 2012).

Here, we attempted to utilize the leg-shape-feature along the body contour to automatically track the identity and movement of each leg during a free-climbing episode inside a transparent curved glass tube. We used a Basler A602 monochrome video camera controlled by a LabView program to capture the climbing episode at approximately 100 fps. We noted that inside a curved glass tube (stem portion of a 5.75″ pasteur pipette, Fisher, Cat. no. 13-678-20A) the fly adopts a broader stance as compared to a planar surface, this facilitated better leg-contour-tracking (LCT) in MatLab.

Figure legend:

A. Morphometric curvature profile along the fly’s body contour and leg identification. The curvature scan along the periodic contour of the animal starts and ends at the peak corresponding to the 2R leg. This computer-vision method allows automatic leg detection during free-climbing locomotion via the most prominent curvature peaks. 

B. Raw image of the fly seen from below overlaid with automated leg position and identity tracking during climbing behavior. The contour curvature of panel A corresponds to the current fly image in panel B. In the path traces, the leg stance and propulsion points while in contact with the ground stand out as peaks, most noticeable for the 2R and 2L legs, while the dips represent “air-time” or the swing phase (see supplementary video http://youtu.be/YvUXRZ27VLY).

C. Leg movement pattern. Standing phase of each leg tip is represented as dark blue segments. Swing-phase motion is coded with a jet colormap, where higher speeds are shown in yellow towards red. This actogram clearly demonstrates the rhythmic, stereotypical and coordinated pattern of leg motion.

D. Lateral displacement of each color-coded leg (as annotated in B) as a function of time. The traces show the actual leg kinematics. These traces can be taken to indicate the stability of the leg-tip upon landing, during the stance phase. Notice a slight slip of 2L and 2R legs while in contact with the ground in the stance phase before the swing-phase begins. Yellow line represents the stability of the geometric center of mass, a reference trace.

E. Forward displacement of each color-coded leg as a function of time. Yellow line represents the geometric center of mass. This trace allows us to appreciate the pattern of leg engagement during the climb. For example, the coordinated movements of 1L-2R-3L and 1R-2L-3R sets appear very closely linked together. Panels C, D and E correspond to the trajectory sequence displayed in B.

Conclusion:

We have developed a simple protocol, called LCT, for automated tracking of each fly leg during climbing using computer vision techniques. This was accomplished by tracing the contour profile of the fly’s body, extracting the shape-feature of each leg and marking their contour curvatures as local maxima. Leg identity is annotated based on invariant geometric symmetries of the leg arrangements. Our initial results suggest that the leg coordination pattern during climbing involves a tripod-like state. In the “climbing gait” the stance phases are represented by reciprocating surface contacts made by three legs namely 1L-2R-3L and 1R-2L-3R. We plan to further apply this method to study insect locomotion in several different contexts. The first version of our MatLab LCT-code is available upon request.

References:

1.    Kain et al. 2012. Leg-tracking and automated behavioral classification in Drosophila. 2012. arXiv:1210.4485v1

2.    Seelig et al. 2010. Two-photon calcium imaging from head-fixed Drosophila during optomotor walking behavior. Nat Methods: 535-40.

3.    Strauss and Heisenberg. 1990. Coordination of legs during straight walking and turning in Drosophila melanogaster. J Comp Physiol A. 167:403-12.

4.  Wosnitza et al. 2012. Inter-leg coordination in the control of walking speed in Drosophila. J Exp Biol. doi:10.1242/jeb.078139

5.   White et al. 2010. The dopaminergic system in the aging brain of Drosophila. Front Neurosci. 8;4:205.

%I figshare