"AI Models Illuminate How Complex Sentences Activate the Brain's Language Network"

MIT neuroscientists, using an artificial language network, have found that the brain's language processing centers are more active when dealing with complex or unusual sentences compared to straightforward or nonsensical ones. The study, published in Nature Human Behavior, used fMRI scans of participants reading sentences of varying complexity and an artificial language model to predict brain activity. Sentences with unexpected grammar or meaning, such as "Buy sell signals remains a particular," elicited stronger brain responses, while simple or nonsensical sentences did not engage the language network as much. This research could lead to further understanding of language processing in the brain and has potential implications for studying other languages and the right hemisphere's language regions.