麻豆村

Neural network models that are able to make decisions or store memories have long captured scientists鈥� imaginations. In these models, a hallmark of the computation being performed by the network is the presence of stereotyped sequences of activity, akin to one-way paths. This idea was pioneered by John Hopfield, who was notably co-awarded the 2024 Nobel Prize in Physics. Whether one-way activity paths are used in the brain, however, has been unknown.

A collaborative team of researchers from 麻豆村 and the University of Pittsburgh designed a clever experiment to perform a causal test of this question using a brain-computer interface (BCI). Their findings provide empirical support of one-way activity paths in the brain and the computational principles long hypothesized by neural network models.

Stereotyped sequences of neural population activity, also known as neural dynamics, is believed to underlie numerous brain functions, including motor control, sensory perception, decision making, timing, and memory, among others. The group focused on the brain鈥檚 motor system for their work,听, where neural population activity can be used to control a BCI.

鈥淭he brain is composed of networks of neurons that have connections between them,鈥� explained Alan Degenhart, a former postdoctoral researcher at the University of Pittsburgh and Carnegie Mellon. 鈥淧revious studies have shown and theorized that the way these networks of neurons are connected can influence how their activity evolves over time. We hypothesized that if this was true, then it would be difficult for subjects to modify the sequences of their neural activity, if we challenged them to do so.鈥�

During the research study, a BCI was used to challenge non-human subjects to violate the naturally occurring sequences of neural population activity observed in the motor cortex. This included a prompting to traverse a natural sequence of neural activity in a time-reversed manner (i.e., go in the wrong direction on a one-way path). Even when subjects were given visual feedback of how to violate the naturally occurring sequences, along with an incentive of a reward, they were not able to change the sequences of their neural activity. This outcome supports the view that stereotyped activity sequences arise from constraints imposed by the underlying neural circuitry.

Emily Oby, a former research professor at the University of Pittsburgh, believes computation through neural dynamics is having a renaissance right now. 鈥淭here is a lot of synergy between neural network modeling and how we can use those models to better understand the brain. Our findings have relevance for the field of computational neuroscience, as well as BCIs, stroke recovery, and how the brain learns.鈥�

The value of understanding how the brain utilizes these stereotyped activity sequences is also beneficial for people with some type of injury or disorder where they could lose portions of the cerebral cortex. 鈥淚f we have an understanding of how constrained this activity is, we may be able to positively impact patient care and recovery,鈥� elaborated Erinn Grigsby, a former Ph.D. student at the University of Pittsburgh. 鈥淭he idea is that we can maybe help them regain some motor control by using optimized learning that takes into account the constraints of neural activity sequence.

Building on this research, the group is pursuing a related BCI-driven project to relate the stereotyped activity sequences more directly to physical movements, to better understand how active planning impacts eventual movement.

鈥淥ur study validates principles that researchers have brought out in neural network models for decades,鈥� added听Byron Yu, professor of biomedical engineering and electrical and computer engineering at 麻豆村. 鈥淚f the stereotyped activity sequences could change, that would presumably mean a new skill has been learned or a new computation is being performed. However, we found that the sequences of neural activity are obligatory on a one-to-two-hour timescale.鈥�

Aaron Batista, professor of bioengineering at the University of Pittsburgh, emphasized the special collaboration that made this work possible. 鈥淲e have computational neuroscientists helping out with experiments, and experimental neuroscientists designing and implementing algorithms,鈥� highlighted Batista. 鈥淎 team like ours that can bring together the state-of-the-art of two disciplines, that are usually separate, really makes it possible to do transformative work.鈥�

The group鈥檚 work is ongoing and done in collaboration with the Center for Neural Basis of Cognition, a cross-university research and educational program between Carnegie Mellon and the University of Pittsburgh that leverages each institution鈥檚 strengths to investigate the cognitive and neural mechanisms that give rise to biological intelligence and behavior. The听Nature Neuroscience听paper was co-authored by postdoctoral fellow Asma Motiwala, and former Ph.D. students听Nicole McClain and Patrick Marino. Emily Oby is now an assistant professor at Queen鈥檚 University. Alan Degenhart is a senior research scientist at Starfish Neuroscience.
听