Reading People's Minds with AI
The ability to read minds seems like something that is only possible in superhero or science fiction movies. Many technological advancements have made this a possibility. While the ability to understand exactly what a person is thinking is a concept that may only see the light of day in the far future, AI has made associations between brain waves and words possible. This technology has many possibilities but it has especially promising applications in the medical field. People who are paralyzed after stroke or neurological disorders have reported being able to remember conversations that have occurred around them (1). With one instance of a man hearing his wife and doctor discuss removing his life support (2). Being able to understand the thoughts of these patients would be imperative for communication, wellness checks, and signs of life in paralyzed victims. A study published in Nature Neuroscience showed that this was possible (3). They would expose volunteers to 16 hours of spoken narrative stories and take fMRI (functional magnetic resonance imaging) scans of the brain, showing neural activity. However, they would recognize similar neural scans for a pool of about 20 candidate words. This is where they utilized artificial intelligence, to score the likelihood of each word and predict how likely it would be in normal grammatical flow. For example, when exposed to the phrase “Adam disappeared and I cleaned up alone crying”, the AI decoded it as “then he stormed off, I thought he left, I started to cry”. While not a direct translation of what was heard, the ability to get the general concept of the sentence as well as key words like “crying” and “he left” is truly unremarkable. Some phrases translate better then others as the AI has made incorrect translations but as a whole, AI was able to correctly interpret the gist of sentences. With further training and proper rules made about its use, brain decoding can be a valuable resource to interpret and assist paralyzed victims.
References:
Flam, F. D. (2023, December 30). The 10 Most Intriguing Science Breakthroughs of 2023. Bloomberg.com; Bloomberg. https://www.bloomberg.com/opinion/articles/2023-12-30/the-10-most-intriguing-science-breakthroughs-of-2023
Hill, A. (2012, August 7). Locked-in syndrome: rare survivor Richard Marsh recounts his ordeal. The Guardian; The Guardian. https://www.theguardian.com/world/2012/aug/07/locked-in-syndrome-richard-marsh
Tang, J., LeBel, A., Jain, S., & Huth, A. G. (2023). Semantic reconstruction of continuous language from non-invasive brain recordings. Nature Neuroscience, 26(5), 858–866. https://doi.org/10.1038/s41593-023-01304-9