Researchers developed a computational model that managed to find out How the brain turns sounds into conversations and predicts human lines . Analyzing 100 hours brain activities during conversations, the group understood how the sequence of actions to form dialogues .
The research result was published this Friday (7) in the magazine Nature Human Behaviourdetailing the use of the Whisper tool – which turned the conversations recorded into three parts: simple sounds, speech patterns and the meaning of words.
With this information in hand and the use of the computational model, the researchers found which order used by the brain to create dialogues.
First, the mind organizes in the words that will be used and then forms the sounds of the sentences. When we hear the other part of the conversation, the organ makes the opposite way, reconstructing the meaning of what was said.
“Our findings help us understand how the brain processes conversations in real -life scenarios,” said Dr. Ariel Goldstein, a researcher at the Department of Cognitive and Brain Sciences and the Jerusalem Hebrew University Business School.
“By connecting different layers of language, we are discovering the mechanics behind something we all do naturally – speak and understand each other.”
In tests conducted during the study, the group of scientists obtained potential practical applications of discoveries – such as improving speech recognition technology until developing better tools for people with communication difficulties.
Dolphin brains refrigerator helps to understand disease in humans
This content was originally published in researchers create computer model capable of predicting human conversations on CNN Brazil.
Source: CNN Brasil
Charles Grill is a tech-savvy writer with over 3 years of experience in the field. He writes on a variety of technology-related topics and has a strong focus on the latest advancements in the industry. He is connected with several online news websites and is currently contributing to a technology-focused platform.
