Nature

A mindreading brain implant that comes with password protection

Inside Precision Medicine

Decoding Inner Speech in Real Time With AI and BrainComputer Interfaces

The New York Times

For Some Patients, the Inner Voice May Soon Be Audible

Stanford

Study of promising speech-enabling interface raises hopes

Gizmodo

New Brain Interface Interprets Inner Monologues With Startling Accuracy

Interesting Engineering

Mind-reading tech from US decodes inner thoughts with 74% accuracy

National Institutes of Health (NIH)

Brain-computer interface restores natural speech after paralysis

Earthling News

Things Are Starting to Feel Dystopian

Harvard Gazette

Hope for life-changing therapies comes with a chilling caveat  Harvard Gazette

Stanford University researchers developed a brain-computer interface that can decode inner speech with up to 74% accuracy by recording neural activity from microelectrodes implanted in the motor cortex of four participants with severe paralysis. The system translates imagined sentences from a vocabulary of 125,000 words using artificial intelligence models trained to recognize phonemes and stitch them together in real time.</span

Stanford University researchers developed a brain-computer interface that can decode inner speech with up to 74% accuracy by recording neural activity from microelectrodes implanted in the motor cortex of four participants with severe paralysis. The system translates imagined sentences from a vocabulary of 125,000 words using artificial intelligence models trained to recognize phonemes and stitch them together in real time.

The study involved four participants with severe speech and motor impairments, including one person affected by stroke and three with motor neuron disease, who had microelectrode arrays surgically placed in their motor cortex. Researchers found that attempted and inner speech activate overlapping brain regions, with inner speech producing similar but weaker neural signals compared to attempted speech.</span

The study involved four participants with severe speech and motor impairments, including one person affected by stroke and three with motor neuron disease, who had microelectrode arrays surgically placed in their motor cortex. Researchers found that attempted and inner speech activate overlapping brain regions, with inner speech producing similar but weaker neural signals compared to attempted speech.

The brain-computer interface includes a password-protection system that prevents decoding of inner speech unless users first think of a specific keyword, with researchers using the phrase "Chitty Chitty Bang Bang" in their experiments. The system recognized this mental password with 98.75% accuracy, ensuring users maintain control over when their thoughts are translated into text or speech.</span

The brain-computer interface includes a password-protection system that prevents decoding of inner speech unless users first think of a specific keyword, with researchers using the phrase "Chitty Chitty Bang Bang" in their experiments. The system recognized this mental password with 98.75% accuracy, ensuring users maintain control over when their thoughts are translated into text or speech.

Three participants with severe dysarthria reported that the inner speech BCI required less physical effort and improved comfort compared to attempting actual speech, bypassing physiological barriers such as breathing control issues. The technology addresses fatigue problems associated with current BCIs that require users to physically attempt speaking, which can be exhausting for people with limited muscle control.</span

Three participants with severe dysarthria reported that the inner speech BCI required less physical effort and improved comfort compared to attempting actual speech, bypassing physiological barriers such as breathing control issues. The technology addresses fatigue problems associated with current BCIs that require users to physically attempt speaking, which can be exhausting for people with limited muscle control.

During testing, the system occasionally detected unintended thoughts, such as when participants silently counted shapes on a screen while performing visual tasks, demonstrating the potential for accidental decoding of private mental activity. Researchers developed two solutions to address privacy concerns: training the BCI to ignore inner speech when decoding attempted speech, or implementing the password system for inner speech-based communication.</span

During testing, the system occasionally detected unintended thoughts, such as when participants silently counted shapes on a screen while performing visual tasks, demonstrating the potential for accidental decoding of private mental activity. Researchers developed two solutions to address privacy concerns: training the BCI to ignore inner speech when decoding attempted speech, or implementing the password system for inner speech-based communication.

The research represents the first successful demonstration of decoding inner speech from motor cortex signals, published in Cell journal on August 14, with researchers noting that current accuracy levels remain insufficient for fluent conversation. Future improvements are expected through enhanced hardware with wireless capabilities, better algorithms, and exploration of additional brain regions beyond the motor cortex that may contain higher-fidelity speech information.</span

The research represents the first successful demonstration of decoding inner speech from motor cortex signals, published in Cell journal on August 14, with researchers noting that current accuracy levels remain insufficient for fluent conversation. Future improvements are expected through enhanced hardware with wireless capabilities, better algorithms, and exploration of additional brain regions beyond the motor cortex that may contain higher-fidelity speech information.

Nature #

Inside Precision Medicine #

The New York Times #

Stanford #

Gizmodo #

Interesting Engineering #

National Institutes of Health (NIH) #

Earthling News #

Harvard Gazette #

The technology raises serious privacy red flags since it can accidentally pick up unintended thoughts like counting. A 74% accuracy rate means one in four words could be wrong, making real conversations frustrating. The system works only with preset phrases, not natural thinking, limiting its practical value for genuine communication needs.

The technology raises serious privacy red flags since it can accidentally pick up unintended thoughts like counting. A 74% accuracy rate means one in four words could be wrong, making real conversations frustrating. The system works only with preset phrases, not natural thinking, limiting its practical value for genuine communication needs.

Brain Implant Decodes Inner Speech With 74% Accuracy

Neurological disorders

Artificial Intelligence

This breakthrough offers genuine hope for people trapped by paralysis and speech disorders. The 74% accuracy rate proves the technology works, and the password system elegantly solves privacy concerns while giving users complete control. For someone with ALS who finds attempted speech exhausting, being able to communicate through thought alone could be life-changing.

This breakthrough offers genuine hope for people trapped by paralysis and speech disorders. The 74% accuracy rate proves the technology works, and the password system elegantly solves privacy concerns while giving users complete control. For someone with ALS who finds attempted speech exhausting, being able to communicate through thought alone could be life-changing.

Brain Implant Decodes Inner Speech With 74% Accuracy

Brain Implant Decodes Inner Speech With 74% Accuracy

The Spin

Articles on this story

Sign Up for Our Free Newsletters
Sign Up for Our Free Newsletters

Brain Implant Decodes Inner Speech With 74% Accuracy

Brain Implant Decodes Inner Speech With 74% Accuracy

The Spin

Articles on this story

Sign Up for Our Free NewslettersSign Up for Our Free Newsletters

Sign Up for Our Free Newsletters
Sign Up for Our Free Newsletters