This video showcases the tracking of a small group of three-spined sticklebacks using my custom written software AnimTrack (currently in beta). Different features of the tracking are dynamically shown and hidden, including background subtraction, social proximity network creation, vectoring, and individual trajectories. My tracking software enables me to keep track of the individuals in the group and thereby link individual characteristics of the fish to group-level outcomes.

A video taken while snorkelling in the Bodensee looking for nesting sticklebacks. This video shows some very brightly coloured males near their territories and larger shoals composed of mainly female sticklebacks swimming in the vicinity of the male breeding grounds.

This video shows some close-up shots of wild-caught three-spined stickleback swimming in their social housing tank. Our housing tanks contain both under gravel and external filters and have artificial aquarium plants and various structures to give the fish an optimal environment yet requiring minimal maintenance.

This video accompanied one of the papers arising from my PhD. By tracking the movements of interacting pairs of sticklebacks, we found that bolder fish led more and were less responsive to their partner and that pairs of bold fish had low cohesion. In contrast, pairs of shy fish showed high social attraction and strongly coordinated their behaviour. Read the paper here.

This is a short demo video of the manual tracking function ‘Mantrack’ of my tracking package AnimTrack. It shows its different functionalities, the majority which work via easy keypresses, currently including 14 unique keys. AnimTrack is build for Python and relies heavily on OpenCV and numpy and is currently still in beta.

This video shows the tracking and visual field reconstruction of a single stickleback in a large maze-like environment. Tracking and raycasting are performed using our Python package AnimTrack, custom-written for our experiments on individual differences and collective behaviour in fish.

Snorkling in the Bodensee in November. During the winter months visibility is best and during this particular day around 5-7m. Unfortunately, very few fish remain in the shallows and most move towards the deeper pelagic zone. Water temperature was ~ 11C.

This short video shows a very nicely aligned (polarised) school of sticklebacks swimming in the shallow waters of the Bodensee near the Limnological Institute. I was able to get relatively close while snorkelling in the ice-cold water in early spring. All sticklebacks in the lake are very large and easily reach 10cm. The individuals in this school are all very similar in size and are likely already 2 year old fish.

This video shows the tracking of an individual stickleback in a standard boldness assay. With my Python package AnimTrack I am able to automatically acquire the latency for a fish to go out of cover, the proportion of time out of cover, the distance moved out of cover, the maximum distance out of cover, as well as other detailed movement characteristics such as its speed, heading and orientation.

This is the classic assay for measuring ‘sociability’ in fish. Using my custom Python package AnimTrack I automatically track the fish and dynamically compute its average distance from the partition separating the conspecific shoal, providing a social proximity tendency of the fish. In recent work we have shown that this test is partly flawed as fish that are only faster may also appear as less social.

While snorkelling in the Bodensee, I found a large number of pikes stalking juvenile fish in the reeds along the shores. The pikes here can easily get 1m and are actually quite easy to approach to within 2m when quiet. Beautiful to see these hunters up close in the wild.

These wild-caught sticklebacks housed in one of my outside mesocosms were very interested in the flashing red light of the GoPro camera, likely due to its resemblance to their favourite food: bloodworms.

To investigate the role of individual differences in collective behaviour, we run agent-based simulations of self-organising, heterogeneous groups of agents varying in speed and goal-orientedness. This video shows a group of 5 and 20 agents moving in a boundary-less environment, with faster agents clearly ending up more in the front of the group, as is depicted in the dynamic graph. The results of these model simulations are integrated in our 2017 paper in Current Biology here.

Three-spined sticklebacks have a strong body armour and three (or sometimes two or four!) spines that protect them against predators. This wild-caught three-spined stickleback raises its spine as it feels threatened of the person standing next to the tank. This video nicely shows the membrane that is attached to the backside of the spines.

Tracking video of one experimental trial of a research project by Matt Grobis and I focused on understanding predator-prey dynamics and individual survival risk. Tracking of the pike was done with my custom tracking software AnimTrack. For analyses we cut the videos in short pieces so unfortunately the attack is not visible in this video.

This video shows a rook successfully pulling-up a string to retrieve a piece of food. It requires the rook to perform a complex foot-beak sequence a number of times, whereby the bird stands on the part of the string it had just pulled-up with its beak. The string-pulling task was used to investigate how dominance, boldness and pair-bonds affect social foraging tactics. The study, which I conducted during my Masters at the University of Cambridge, is published here.

Rooks and jackdaws form mixed-species flocks, especially during the winter months, when they aggregate in the thousands. They do not only forage and move together for most of the day, they also roost together during the night. But before going to roost, the mixed-corvid flocks often spent up to two hours congregating in pre-roosts, like shown here. I studied the mixed-corvid flocking and roosting behaviour during my research assistant job before my PhD. You can read more about their mixed-flocking behaviour in our paper here.