Investigating Social Communication of Mormyrid Weakly Electric Fish using Machine Learning Algorithms

Rachel Miller
Rachel Miller

Rachel Miller is a rising senior (‘22) from Los Angeles, California. She is majoring in Neuroscience and Behavior and minoring in Informatics and Modeling. After graduating from Wesleyan, she plans on attending medical school.

Abstract: Mormyrids are pulse-type weakly electric fish that emit and receive electric organ discharges (EODs) in order to communicate and locate objects in their environment. They generate stereotyped EOD waveforms, but are able to vary the interval between EODs (inter-pulse-intervals – IPIs). The IPIs differ depending on social context, and the timing of EODs are coordinated between pairs of fish during behaviors such as mating, hunting, and aggression. Much of the past research on Mormyrid social behavior has examined pairs of fish, finding that a single fish will coordinate its EODs with a dummy fish emitting electrical signals into its tank and that pairs of fish will synchronize pulsing during mating and aggression. In order to fully understand Mormyrid social communication, we need to understand their behavior in larger groups, as large group interactions are principal in overall social life and structure. Two metrics are necessary for investigating these group dynamics: the positions of the fish and the timing of each fish’s EOD. Having these metrics allows the analysis of environmental interaction and of social communication and the coordination of EODs. We plan on investigating group dynamics of Mormyrid fish in a novel environment, observing how different sized groups of fish interact with each other and with their surroundings when being shifted from a known to an unknown environment. As group size increases, we expect to see increased communication shown by shorter IPIs and coordination of EOD pulses. Past investigations on the social dynamics of larger groups of fish have been done by manually identifying EOD fish pairs in recordings, but our research utilizes machine learning algorithms to synchronize tracking of the fish with EOD recordings. This streamlined approach to the analysis of movement and EOD pulses will provide insight on large group social communication and on the coordination of EODs in Mormyrid fish.

Lab-Poster-Rachel

Video:

Rachel Miller (Biology Department)

Live Poster Session:
Thursday, July 29th 2:45-3:45pm EDT

https://wesleyan.zoom.us/j/97141221131?pwd=SmlVRm1KWWNpa3VMOW41RkE5aW1Udz09