Much of our daily life requires us to make inferences about the world around us. When you think about the direction in which your tennis opponent will hit the ball, or try to understand why your child is crying, your brain searches for answers about possibilities that are not directly accessible through sensory experiences.
MIT Associate Professor Mehrdad Jazayeri has spent most of his career exploring how the brain creates internal representations, or patterns, of the outside world to make intelligent inferences about the world’s hidden states. .
“The question that interests me most is how the brain forms internal models of the outside world? The study of inference is really a powerful way to gain insight into these internal models, ”says Jazayeri, who recently got a job in the Department of Brain and Cognitive Sciences and is also a member of the Institute. McGovern of MIT’s Brain Research.
Using a variety of approaches, including detailed behavioral analysis, direct recording of neural activity in the brain, and mathematical modeling, he discovered how the brain builds patterns of statistical regularities in the brain. environment. He also discovered circuits and mechanisms that allow the brain to grasp the causal relationships between observations and results.
An atypical course
Jazayeri, who has been on the MIT faculty since 2013, has taken an unusual path to a career in neuroscience. Growing up in Tehran, Iran, he was an indifferent student until his sophomore year of high school when he became interested in solving difficult geometric puzzles. He also began programming with the ZX Spectrum, a first 8-bit personal computer that his father had given him.
In high school he was chosen to train for Iran’s very first National Physics Olympics team, but when he failed to make the international team, he became discouraged and temporarily dropped out. the idea of going to college. Eventually, he took the national university entrance examination and was admitted to the electrical engineering department of Sharif University of Technology.
Jazayeri did not appreciate his four years of college education. Above all, the experience helped him realize that he was not destined to become an engineer. “I realized that I am not an inventor. What inspires me is the process of discovery, ”he says. “I really like understanding things, not building things, so these four years haven’t been very inspiring.”
After graduating from college, Jazayeri spent a few years working on a banana farm near the Caspian Sea, with two friends. He describes these years as some of the best and most formative of his life. He would wake up at 4 a.m., work the farm until late afternoon, and spend the rest of the day thinking and reading. One topic he read with great interest was neuroscience, which led him a few years later to apply for graduate school.
He immigrated to Canada and was admitted to the University of Toronto, where he obtained a master’s degree in physiology and neuroscience. There, he worked on building small models of circuits that would mimic the activity of hippocampal neurons.
From there, Jazayeri went to New York University to earn a doctorate in neuroscience, where he studied how signals from the visual cortex support perception and decision making. “I was less interested in how the visual cortex encodes the outside world,” he says. “I wanted to understand how the rest of the brain decodes signals in the visual cortex, which is, in fact, an inference problem.”
He continued to pursue his interest in the neurobiology of inference as a postdoctoral fellow at the University of Washington, where he studied how the brain uses time patterns in the environment to estimate time intervals and use knowledge. on these intervals to plan future actions.
Build internal models to make inferences
Inference is the process of drawing conclusions based on information that is not readily available. Making rich inferences from scarce data is one of the fundamental mental abilities of humans, one that is at the heart of what makes us the smartest species on Earth. To do this, our nervous system builds internal patterns of the outside world and those patterns that help us think about possibilities without experiencing them directly.
The problem of inferences arises in many behavioral contexts.
“Our nervous system creates all kinds of internal models for different behavioral goals, some that capture the statistical regularities of the environment, some that link potential causes to effects, some that reflect relationships between entities, and some that allow us to think about others, “Jazayeri says.
Jazayeri’s lab at MIT is made up of a group of cognitive scientists, electrophysiologists, engineers, and physicists with a common interest in understanding the nature of internal brain models and how these models allow us to make inferences in different behavioral tasks.
Early lab work focused on a simple timing task to examine the problem of statistical inference, that is, how we use statistical regularities in the environment to make accurate inference. First, they discovered that the brain coordinates movement over time using a dynamic process, similar to an analog timer. They also found that the neural representation of time in the frontal cortex is continually calibrated based on past experience so that we can make more accurate time estimates in the presence of uncertainty.
Later, the lab developed a complex decision-making task to examine the neural basis of causal inference, or the process of deducing a hidden cause based on its effects. In a 2019 article, Jazayeri and his colleagues identified a hierarchical and distributed brain circuit in the frontal cortex that helps the brain determine the most likely cause of failure within a hierarchy of decisions.
More recently, the lab has extended its research to other behavioral areas, including relational inference and social inference. Relational inference consists of locating an ambiguous observation using relational memory. For example, when exiting a subway in a new neighborhood, we can use our knowledge of the relationship between visible landmarks to infer which direction is north. Social inference, which is extremely difficult to study, involves inferring the beliefs and goals of others based on their actions.
Along with studies on human volunteers and animal models, Jazayeri’s lab is developing computer models based on neural networks, which helps them test different possible hypotheses about how the brain performs specific tasks. By comparing the activity of these models with neural activity data from animals, researchers can better understand how the brain actually performs a particular type of inference task.
“My main interest is how the brain makes inferences about the world based on neural signals,” Jazayeri explains. “All of my work involves looking inside the brain, measuring signals, and using mathematical tools to try to understand how those signals are manifestations of an internal pattern in the brain.”