Research reveals the neural computer system of communication reasoning

png">

Communication is one of the most important social behaviors of human beings.

How the brain encodes and decodes communication signals and conveys intentions according to local conditions is a generally recognized important but difficult research problem.

The Zhu Lusha research group of the School of Psychology and Cognitive Sciences at Peking University has recently combined decision-making neuroscience, computational linguistics, and game theory and other multidisciplinary research methods to open up new quantitative research ideas for this problem.

Using technical means such as functional magnetic resonance imaging and computational modeling, researchers have explored the neural computer system for inferring meaning in the classic "referential game".

In the example above, if a "cooperative" speaker wants to refer to the blue square, then he is more likely to say "square" to help the audience understand its meaning.

So, does the listener's brain also use a similar "reverse reasoning" algorithm to infer the speaker's intentions? Researchers have discovered that the ventromedial prefrontal lobe of the listener’s brain may be an "analog generator" of rational communication signals, which flexibly and quantitatively represents the conditional probability of various cooperative behaviors of the speaker (for example, in order to refer to the blue circle And the possibility of choosing the word "blue").

Interestingly, in some cases, the listener does not need to reason and can learn the intention of the other party from the literal meaning (as shown in the example below).