Brown Researchers Model Animal Gaits with Single Neural Network

Study finds new approach to modeling how four-legged creatures move.

Apr. 14, 2026 at 2:24pm

Got story updates? Submit your updates here. ›

A highly textured, abstract painting featuring sweeping geometric arcs, concentric circular motions, and precise botanical spirals in earthy tones, conceptually representing the complex biomechanical forces and rhythmic patterns underlying the diverse locomotion of four-legged creatures.A new unified neural network model aims to capture the diverse gait patterns of quadruped animals in a more efficient and versatile way.Greeley Today

Researchers at Brown University have developed a new neural network model that can simulate the distinct gait patterns of various quadruped animals using a single unified system, rather than requiring separate models for each type of four-legged creature. This marks a significant advancement in the field of biomechanics and animal locomotion research.

Why it matters

Traditionally, researchers have had to build separate neural network models to capture the unique walking, running, and movement patterns of different quadruped species. This new unified approach could lead to more efficient and versatile simulations that better reflect the natural diversity of animal gaits, with potential applications in fields like robotics, prosthetics, and the study of animal behavior.

The details

The research, led by Professor and Associate Chair of the Department of Mathematical Sciences at the University of Northern Colorado Katherine Morrison, demonstrates how a singular neural network can be trained to model the distinct gait patterns of multiple quadruped animals. This marks a significant advancement from previous methods that required separate neural networks for each type of four-legged creature.

  • The research was published on April 14, 2026.

The players

Katherine Morrison

Professor and Associate Chair of the Department of Mathematical Sciences at the University of Northern Colorado, who led the research on modeling quadruped gaits with a singular neural network.

Got photos? Submit your photos here. ›

What they’re saying

“In the past, when researchers modeled quadruped gaits — how four-legged organisms walk, run and move — gaits have been modeled with separate neural networks.”

— Katherine Morrison, Professor and Associate Chair of the Department of Mathematical Sciences

The takeaway

This breakthrough in modeling animal locomotion with a single neural network could lead to more efficient and versatile simulations, with potential applications in fields like robotics, prosthetics, and the study of animal behavior.