Â鶹´«Ã½

News — Understanding how the in the brain function is the key to both treatment of disease and advancing artificial intelligence. Thanks to advances in brain-computer interfaces (BCI), researchers at the University of Pittsburgh and Carnegie Mellon University now have a better understanding of how these pathways function.  

Aaron Batista, professor of bioengineering at Pitt’s Swanson School of Engineering, with long-time collaborator Byron Yu, professor of electrical and computer engineering and biomedical engineering at CMU, expanded their BCI research to provide empirical evidence that neural population activity is constrained to follow specific sequences. Their results, “Dynamical constraints on neural population activity,” were published today in

Batista and Yu, leading researchers in BCI, explained that a fundamental principle in neural dynamic models assumes that neural activity sequences in the brain are as rigidly defined as a roller coaster. Like a roller coaster ride, neural activity can only move forward along its path, apparently without the ability to readily go backward or jump the rails. Although this theory has laid the groundwork for the development of complex artificial intelligence models in use today, it had never been directly tested— until now.

“Our findings are a direct causal test of ideas that have shaped the field of neural dynamics,” Batista said. “We used BCI technology to challenge our subjects to deviate from these neural sequences and found that it was nearly impossible for them to do so.”  

Batista’s team leveraged BCI technology to provide Rhesus monkeys with real-time feedback of their own neural activity. The subjects were then prompted to use their minds to move a cursor on the screen in front of them in different directions, in ways that both followed the natural sequence of activity and in ways such as backwards or sideways that diverged from these standard sequences. Even after extended practice, the team found that the subjects lacked the flexibility to move the cursor in different ways, which demonstrated that neural activity in the brain is in fact constrained to predefined computational paths.

According to Yu, interdisciplinary research between engineering and neuroscience has helped to advance our understanding of the brain. BCI in particular has enabled researchers to examine in real time how its complex neural structure enables human and non-human primates to function.

“We combined experimental and computational neuroscience together every step of the way in this project.” Yu said. “Starting off with a common scientific goal from day one and then working together on all levels of the project leads to science that could not have been done with either of our groups alone.”

Since these neural activity sequences are fundamental to countless functions in the brain, the team sees their discovery as a launching point for leveraging this knowledge in a broad scope of applications in the future.

“If the brain operates in this particular way, we can then understand its mechanisms to leverage them.” Batista said. “This knowledge can help us shape future therapies, stroke rehabilitation, or build better brain-computer interfaces.”