Temporal Aspects in Collective Motion

Cite This

Files in this item

Checksum: MD5:4bc49b91a1c683357682a8c7b1538bad

AMICHAY, Guy, 2021. Temporal Aspects in Collective Motion [Dissertation]. Konstanz: University of Konstanz

@phdthesis{Amichay2021Tempo-56404, title={Temporal Aspects in Collective Motion}, year={2021}, author={Amichay, Guy}, address={Konstanz}, school={Universität Konstanz} }

2022-02-01T08:17:56Z 2021 eng terms-of-use Amichay, Guy Self-organisation is fundamental to biology and is evident across diverse taxa and scales. Broadly speaking, we can divide self-organizing processes into different forms of coordination: for example, the coordinated movement of flocks of birds through space and time; or, the coordinated chirping of crickets in time. In this thesis, I focused on collective motion, studying schooling fish. A common thread throughout the thesis is highlighting temporal aspects that play a role in these behaviours, that have, to date, generally been overlooked - characterising them, explaining them, and revealing their importance. Juvenile zebrafish, like many other fish species, exhibit a swimming pattern that can be characterised by short accelerations followed by longer decelerations, often termed ‘burst-and-glide’ motion. We found, that when paired, they exhibit what could be described as alternating behaviour, where one individual accelerates the other tends to decelerate and vice versa. By further analysing these movement data, we were able to reveal their phase response curve, which defines how they couple. This revealed that the fish tend to be oblivious to the tailbeats of their neighbour when the neighbour beat up to ~100 ms after them. When the neighbour’s delay was >100 ms after the focal, the focals’ response (timing its’ next tailbeat) can be approximated as simply doubling that delay. In other words, if the neighbour beat 150 ms after the focal, the focal would end its’ own period after 300 ms. We then derived, based on this, a simple model of this temporal interaction, and tested it with a closed-loop virtual reality experiment, whereby a real fish swam and interacted with a virtual avatar, that responded to the real fish’s behaviour according to our model in realtime. By doing so we recover similar temporal dynamics as seen between two real fish. Moreover, we further characterised these ‘oscillators’ as irregular oscillators, due to the inconsistent period lengths they show over time, which is (experimentally) relatively unexplored. We continue this line of work, and ask what functional role this coupling might play in a collective. To do so, we utilised the virtual reality system, but with open-loop experiments, that render a more controlled setting. We show that real fish temporally couple with non-responsive virtual fish, and that they aim to maintain a fixed lag after the virtual fish (their bursts relative to the virtual fish’s bursts), no matter the applied period length of the virtual fish (time between bursts). Next, we tested what a real fish would do when swimming with a virtual fish that unexpectedly turns away. We show that if the real fish was temporally coupled with the virtual fish prior to the turn, it is more likely to continue swimming with it afterwards, suggesting that the real fish is more responsive to its’ neighbour when temporally coupled. Next, we take a different direction and use a bottom-up approach. We implement a simple self-propelled particle model, with the minor addition that the particles have oscillating speeds – accelerating and decelerating following a sine-wave pattern. Each individual is assigned a random starting speed on this curve, which can be considered as its’ ‘phase’. We then ask what might emerge from adding these simple dynamics. We revealed nontrivial behaviour, whereby individuals with similar phases tend to ‘couple’ with each other (they were connected for long durations) - which is a new form of self-sorting. Finally, the last part of the thesis is dedicated for a broader look at the future of the field, suggesting a more ‘holisitic’ view on different forms of self-organization, aiming to better unify collective motion and temporal synchronization (and similar) under the same umbrella, owing to their known analogous behaviour. While this has been known and studied theoretically, experimenters have mostly ignored this possible comparison. We suggest using network theory (specifically, mainly temporal networks) as a common language, and argue that we are at a time, both experimentally (due to technological advancements) and theoretically (due to more attention being payed to temporal networks in the last decade), for a fruitful interplay. Amichay, Guy Temporal Aspects in Collective Motion 2022-02-01T08:17:56Z

Downloads since Feb 1, 2022 (Information about access statistics)

Amichay_2-1g3cbcf983ky64.pdf 78

This item appears in the following Collection(s)

Search KOPS


Browse

My Account