Is Functional MRI capable of reading your mind?

USA: Scientists have developed a non-invasive brain-computer interface that can decode continuous language from cortical semantic representations recorded using functional magnetic resonance imaging (fMRI). The study was published in the Journal of Nature Neuroscience.

The objective of language decoding is to utilize recordings of brain activity to anticipate the words that an individual is listening to, speaking, or envisioning. They tested the decoder across the cortex and found that continuous language can be separately decoded from multiple regions.

The new decoder can generate intelligible word sequences that recover the meaning of perceived speech, imagined speech, and even silent videos, indicating that a single decoder can be applied to a range of tasks.

The predictions made by the decoder frequently captured the essence of an individual's internal speech, however, they did not reproduce an exact transcript. In approximately fifty percent of the instances, the decoder generated text that accurately or closely corresponded to the intended connotations of the words that a person was contemplating.

The study's findings are promising because brain-computer interfaces that decode continuous language from non-invasive recordings could have many scientific and practical applications, including communication for people with speech disorders and decoding thoughts. Currently, non-invasive language decoders can only identify stimuli from among a small set of words or phrases.

To ensure mental privacy, the study also tested whether successful decoding requires subject cooperation and found that subject cooperation is required to train and apply the decoder.

The researchers hope that their findings will pave the way for the development of non-invasive language brain-computer interfaces and lead to significant advancements in communication and neuroscience.

Dr. Jerry Tang stated that “Our research is the pioneering study that can decode continuous language, which surpasses the ability to recognize just single words or sentences, from non-invasive brain recordings. Presently, decoding language from neural activity primarily depends on invasive brain-computer interfaces that necessitate surgical implants.”

Reference:

Tang J, et al "Semantic reconstruction of continuous language from non-invasive brain recordings" Nat Neurosci 2023; DOI: 10.1038/s41593-023-01304-9