Computer vision and deep learning automates nocturnal rainforest ant tracking to provide insight into behavior and disease risk

Abstract

Determining how ant colonies optimize foraging while mitigating disease risk provides insight into how the ants have achieved ecological success. Fungal infected cadavers surround the main foraging trails of the carpenter ant Camponotus rufipes, offering a system to study how foragers behave given the persistent occurrence of disease threats. Studies on social insect foraging behavior typically require many hours of human labor due to the high density of individuals. To overcome this, we developed deep learning based computer vision algorithms to track foraging ants, frame-by-frame, from video footage. We found foragers can be divided into behavioral categories based on how straight they walk across the trail. Eighty percent of ants walk directly across the trail, while 20% wander or circle when crossing the trail. Departure from the main trail encourages exploration of new areas and could enhance discovery of new food resources. Conversely, results from our agent-based model simulations suggest deviation from a straight path exposes foragers to more infectious fungal spores. Consistency in walking behavior may protect most ants from infection, while the foragers with increased exposure due to their mode of walking could be a sufficient number of new hosts to sustain disease in this environment.