Dynamical behavioral aging trajectory of C. elegans

Aging is not a single event but rather a continuous trajectory that evolves over a lifetime. However, aging is typically quantified by focusing on single time points, obscuring the essential dynamics linking changing physiology to changing behavior. In this talk, I will describe our approaches for measuring the aging trajectory of the nematode C. elegans across its entire lifespan within a custom-built microfluidic apparatus. Specifically, we use methods from unsupervised machine learning to construct a low-dimensional state space and a novel stochastic differential equation modeling framework to reveal how the topology and stability of locomotion dynamics alter with age. To test the effects of diet on aging, we analyze data from worms in multiple feeding conditions, including an intermittent fasting (IF) condition where the worms are given only periodic access to food. We find age-dependent bifurcations and deepening potential wells in the worms’ dynamics that are common across conditions, and that IF extends the worms’ lifespan by delaying these alterations. Moving forward, this framework enables testing of how genetic and neural perturbations influence aging trajectories, linking behavioral change to the physiological mechanisms that drive it.