About the Author

The Dynamics of Rhythm Synchronization in Acorn Ants

By Jillian Kunze

Figure 1. An acorn ant colony can comfortably live within a small cardboard cutout. Figure courtesy of Grant Doering.
Acorn ants—which comprise over 500 species—are quite useful for experimental studies. These insects are tolerant of laboratory conditions and live in small colonies — an acorn ant colony can be comfortably housed within a cardboard cutout the size of a credit card (see Figure 1). During a minisymposium presentation at the 2023 SIAM Conference on Applications of Dynamical Systems, which is taking place this week in Portland, Ore., Grant Doering of Texas Tech University described his experimental and simulation work on the dynamics of these ants. 

“One of the main things that I study is rhythm synchronization in ants,” Doering said. Ant colonies exhibit endogenous, periodic cycles of collective activities; that is, they sleep, wake, and work together in phases that are not caused by environmental conditions. This activity is spread by physical contact between the ants. “When ants are not moving, they are blocking access to larvae,” Doering explained. “When they start to move, synchronization removes these local traffic jams.” Furthermore, individual ants that are separated from the colony will also exhibit intrinsic oscillations with probabilistic refractory periods. 

To find the synchronized activity rhythms, Doering recorded videos of ant colonies and used basic image recognition algorithms to measure and plot their periodic and regular collective activity over time (see Figure 2). This data can help characterize two basic features—the periodicity and dominant rhythm of the oscillations—and investigate how heterogeneity in collective behavior evolves among different ant species.

Figure 2. A time series of the oscillating activity of an acorn ant colony. Figure courtesy of Grant Doering.
In terms of anatomical evolution, the framework of mosaic evolution states that heterogeneity in traits is important because it widens the phenotypical space. However, this framework is more difficult to study for collective, nonphysical traits. Ants provide a way to address this difficulty, since it is relatively easy to gather large volumes of data on ant colonies via image recognition.

“I collected hundreds of time series for both colonies and isolated individuals, creating more than 1.5 years’ worth of time series data,” Doering said. He found a lot of variation between ant species, and wavelet analysis in particular demonstrated a very strong difference in the regularity of cycles between species. But how did these traits evolve? The data did not indicate a correlation between rhythm and periodicity, so these attributes do not appear to have coevolved; and perhaps the heterogeneity in both qualities served to widen the phenotypical space in a manner similar to mosaic evolution.

Next, Doering addressed the effects of noise and heterogeneity in individual ants. He decided to use Myrmica punctiventris—an ant species that is not technically an acorn ant, but still lives in acorns—to look for repeatable differences among individual’s oscillations, since this species is larger and easier to track. When Doering separated out 42 workers from eight colonies and measured their oscillations on their own, he found strong statistical support for variation in their rhythms.

“Having established these differences in rhythms, I then gathered 16 different pairs,” Doering said, adding that he tracked each individual ant by giving them a dot of paint. There turned out to be no consistency in the synchronization between the different pairs, and pairs with dissimilar intrinsic rhythms were able to synchronize just as well as more similar pairs.

There is noise in the length of ants’ social refractory periods (that is, the resting time in which they are less likely to respond to social stimuli can vary quite a lot), so Doering turned to simulations to investigate the effects. “If we have a model of this behavior, can noise impact the synchronization of ants?” he asked. He implemented a random walk that determined when the simulated ants became active or inactive, and also allowed them to be activated or inactivated by social stimulation.

Figure 3. A comparison of the ant activity in real data and a simulation. Figure courtesy of Grant Doering.

This simulation output oscillations that appear similar to those observed in real ants (see Figure 3). Doering found that increasing the refractory period also increased the period of the oscillations up to a certain point — but at sufficiently large values of the refractory period, the oscillation period split, leading the ant colony to potentially switch between faster and slower cycles. Furthermore, adding more noise to each ant’s refractory period reduced this birhythmicity and caused the colony to start favoring the longer period.

“To summarize, there is substantial behavioral heterogeneity in ants’ activity rhythms,” Doering concluded. The individual heterogeneity between ants does not necessarily affect their collective behavior, and the periodicity and dominant rhythm of the oscillations do not appear to be phenotypically linked.

  Jillian Kunze is the associate editor of SIAM News