Skip to main content
Blog Article

Cognitive Flexibility in Artificial Intelligence

The Innovators in Science Award Honorees are Breaking New Ground in Neuroscience: Dr. Michael Halassa’s research on AI systems could impact our perception of reality.

Albert Einstein reportedly once said, “Not everything that can be counted counts, and not everything that counts can be counted.” Though the 2017 honorees of the Innovators in Science Award have plenty of countable achievements, their stories reveal a common thread — creative approaches to their work and the development of disruptive tools that transformed scientific understanding in their discipline.

Biological Underpinnings of the Mind

Michael Halassa

Michael Halassa, MD, PhD, an Early-Career Scientist Finalist, has traced the neural correlates of cognition from the thalamus to the cortex and beyond. But his interests in neurocomputational frameworks trace back even farther — to the first time he watched “The Matrix.”

As he watched the film’s characters grapple with a simulated reality, Halassa began wondering how something as intangible as the mind can perceive reality in the first place. If we were to look inside the brain, he wondered, where would we find the mind? How do we make decisions and solve problems?

“If we can understand how these functions are normally accomplished by the physical device we call the brain, then we’ll have a better understanding of how these functions go awry in conditions such as schizophrenia, autism or ADHD,” says Halassa, an Assistant Professor of Brain and Cognitive Science at Massachusetts Institute of Technology (nominated while at New York University in New York).

Computational Frameworks

Halassa abandoned the traditional tactic of studying the molecular and electrical properties of individual cells. Instead, he assembled computational frameworks that could map physical features, such as synapses, onto abstract processes such as thought. His approach revealed that the thalamus, a brain region long assumed to relay simple sensory input to the cortex, actually streams detailed instructions that allow the cortex to shift between tasks.

“From moment to moment, your brain reconfigures on the fly to perform different types of tasks. That reconfiguration is what defines things like intelligence, productivity and performance.” Glitches in this network configuration may contribute to psychiatric diseases, he says.

His findings could lead to artificial intelligence systems that display similar cognitive flexibility. Such “neuromorphic computing” could lead to a greater understanding of how we perceive reality.

Read more about Innovators in Science Award Honorees:


Author
Image
Academy Staff
This article was written by a member of the Academy staff.