News // Technology

How we decipher thoughts with the help of artificial intelligence

Researchers at Baylor College of Medicine and Rice University have developed artificial intelligence models that help them better understand brain calculations to decipher thoughts.

This is a first step. So far there has been no way to measure thoughts. Scientists initially developed a new model that can estimate thoughts by assessing behavior, and then tested the model on an artificial brain .

Eventually, they discovered neural activity associated with those thoughts. The theoretical study is to be published in the Proceedings of the National Academy of Sciences.

“For centuries, scientists have studied how the brain works by analyzing brain activity. For example, when studying the science of movement, researchers measure muscle movements as well as neural activity, and then make a connection between those measurements. But to study cognition in the brain, we have nothing to compare neural activity with, ”said Dr. Xaq Pitkow, one of the study’s authors, according to MedicalXpress .

Sometimes animals’ thoughts or assumptions are wrong about what is going on around them

To understand how the brain produces thoughts, researchers must first measure a thought. Thus, they developed a method called “Inverse Rational Control” that analyzes behavior and deduces the thoughts that best explain that behavior.

Researchers in this field have worked with the notion that animals solve tasks optimally, behaving in such a way as to maximize their benefits. But when scientists studied animal behavior, they learned that this does not always happen.

“Sometimes animals have wrong thoughts or assumptions about what is going on around them, but they are still trying to find the best long-term results. That could explain why the animals seem to behave in a less optimal way, “added Pitkow.

The study may provide new perspectives on neurological diseases

For example, an animal hunts and hears noises that it associates with prey. If a potential prey produces those noises, the optimal behavior for the hunter would be to focus their movements on a single noise. If the hunter mistakenly believes that the noises are coming from several animals, he could choose a “suboptimal” behavior, constantly scanning the surroundings to try to find one of the animals. By acting according to what he believes, the hunter behaves in a way that is both “rational” and “suboptimal”, according to ME.

“We can look at the dynamics of thoughts and the dynamics of the brain’s representations of those thoughts. If those dynamics take place in parallel with each other, we are thus confident that we capture aspects of the brain calculations involved in those thoughts. By developing methods for estimating thoughts and interpreting associated neural activity, the study can help scientists understand how the brain produces complex behaviors and provide new insights into neurological diseases, ”Pitkow explained.