Researchers at Harvard and Google DeepMind have developed a virtual rat with an artificial brain that can mimic the movements of real rats. This innovative project aims to study how real rats control movement and could revolutionize neuroscience and robotic control systems. The virtual rat’s brain activity closely matches that of real rats, offering new insights into brain function.
The virtual rat was created by training an artificial neural network using high-resolution data recorded from real rats. This network was then used to control the virtual rat’s movements in a physics simulator called MuJoco, where gravity and other forces are present. The researchers found that the activations in the virtual control network accurately predicted neural activity measured from the brains of real rats producing the same behaviors.
This collaboration between Harvard and Google DeepMind represents a new approach to studying how the brain controls movement. By leveraging advances in deep reinforcement learning and AI, as well as 3D movement-tracking in freely behaving animals, the researchers were able to create a biomechanically realistic digital model of a rat. The virtual rat was able to imitate a diverse range of behaviors, even ones it hadn’t been explicitly trained on, by learning to produce the forces to generate the desired movements.
The implications of this research are significant. The virtual rat could be used to study neural circuits and how they are compromised in disease, potentially leading to new treatments. Additionally, the platform could be used to engineer better robotic control systems. The researchers plan to continue developing the virtual rat, giving it autonomy to solve tasks akin to those encountered by real rats, and using it to test ideas about how real brains generate complex behavior.
Source NeuroScienceNews
