SwarmSight: Real-time Tracking of Insect Antenna Movements and Proboscis Extension Reflex Using a Common Preparation and Conventional Hardware

The overall goal of this procedure is to rapidly measure the movements of insect antennae and proboscis from video of a common animal preparation using open-source swarm site software and conventional camera and computer hardware. These methods can be used to rapidly obtain high resolution measurements of antenna and proboscis movements. Application of these methods can elucidate how antenna and proboscis movements change in response to chemical, developmental, and genetic manipulations.

The main advantage of this technique is that it can rapidly measure the movements using free software and does not require sophisticated video or computer hardware. Though this method can provide insight into odor responses in honey bees, it can also be applied to other insect systems. To begin, use vials to collect worker bees from the entrance of the hive, and briefly place the vials into an ice slurry.

With duct tape, secure each bee in a harness tube and then apply tape over the top of the tube to hide the legs. Ensure the legs cannot be seen from the top of the harness tube. Next, under a dissection microscope, use a low temperature soldering iron to apply heated wax to the back of each bee's head to immobilize it.

Use a small rod to hold the antenna away from the wax and ensure that the antennae can move freely. Use a web cam holder to position a camera directly above the bee's head, and use the recording software to inspect the video and zoom into the head. After the tape and wax are applied, preview the scene using the camera and look for anything that might be moving or changing that is not the antenna or proboscis.

Then adjust tape and wax as needed. Then, use the recording software settings to maintain constant camera exposure and adjust the exposure slider to the desired value. Adjust contrast, brightness, and sharpness sliders to enhance antenna detection.

Adjust the ambient lighting to reduce antenna shadows. Use a diffuse light source, such as a laboratory goose neck lamp or a flashlight with paper to illuminate the bee. Next, place an LED light placed within the camera view to indicate when the stimulus odor is delivered.

In a fume hood, prepare odor cartridges, which will be used to stimulate the bees during video recordings. Position the odor delivery source near the harness tube, taking care not to obstruct the camera view. Position a vacuum source opposite the odor source to remove stimulus odors, and turn it on prior to odor presentation.

Once bees are restrained and camera preparations are complete, film each individual bee under each experimental condition as a separate video file. Next, in the swarm site antenna tracking module, click on the browse button to open a video file. After the video file loads, place the antenna sensor widget over the bee's head.

Use the rotation and scale icons to adjust widget alignment. The widget should cover the full range of motion on the bee's antenna. Next, place the treatment sensor widget over the LED light that indicates when stimulus odor is being delivered.

Click the play button in the bottom left corner to commence frame by frame display of the video file. The likely antenna and proboscis points will be highlighted in yellow, and the yellow rings will indicate the appendage tips. Adjust the sliders in the filters section to optimize filter sensitivity.

When the ideal sensitivity is achieved, only the bee's appendages will be highlighted in the frame. Fast forward through subsequent parts of the video to ensure filter sensitivity is optimized throughout the entire video. Test your set-up by making a short sample video of the insect and analyze it with the software, looking for extraneous moving objects that the software detects.

Adjust the scene, lighting, and camera angle until the software isolates the antennae and proboscis movements. After setting up the widgets and filters, pause the video, restart it, and play it from the beginning to the end. Finally, expand the save section, click the save tracking data button, and select a folder in which to save the file.

To validate the appendage tracking software, two human raters located the tips of the antennae in 425 video frames, which were then analyzed by the software. The distance between the two human raters is demonstrated in this figure, with inter-human distance and distance between locations detected by the software and the human raters is demonstrated with software-closest human distance. The inter-human distance was small for frames at the beginning and then increased during the second half of the video.

Software-closest human distance levels remain constant throughout the video. 23 female honey bees were exposed to five different odor conditions, and videos of the trials were analyzed using the appendage tracking software. With the exception of the control condition, the software was able to detect significant differences in mean antennae angle changes after exposure to experimental odor conditions.

Additionally, antenna angle density maps revealed antenna location preference clusters. Red clusters in the map indicate that the antennae prefer angles away from the odor source shortly after odor onset and prefer angles toward the odor source shortly after odor conclusion. While attempting this procedure, it is important to remember to carefully inspect each video scene prior to recording for optimal bee placement and lighting.

Additionally, it is important to mark the odor delivery frames with a visual indicator. After performing this procedure, other methods, such as odor conditioning, can be performed. This procedure is also useful for exploring innate responses in honey bees and other insects.

Don't forget that working with bees and certain chemicals can be hazardous. Appropriate personal protective equipment should always be worn while performing this procedure. Once mastered, this procedure can be used to efficiently obtain antenna and proboscis movement data for a large number of insects and experimental conditions.

Swarm Sight, the open-source software used in this procedure can be obtained from swarmsight.org.