Scientists at Stanford University report breakthroughs in neuroscience that could enable people who are unable to communicate more naturally. The research team has developed a brain implant that can decipher “inner speech.” Survey results published in cell In August 2025, it represents one of the first demos that can reliably translate imaginary words into text via brain and computer interfaces.
Four participants were involved in the study, each of whom had lost their ability to speak due to conditions such as motor neuronal disease or brainstem stroke. These individuals were equipped with small arrays of electrodes implanted in the motor cortex, the area of the brain that controlled movements that contain movements required for speech. The implant recorded patterns of neural activity as participants imagined words and sentences that they spoke quietly. The artificial intelligence system then analyzed these signals to predict the intended words.
Researchers found that brain activity generated during internal speech was significantly weaker than when participants tried to move their mouth or tongue. Still, the patterns contained enough information for the system to decode them. Beyond the vocabulary of about 125,000 words, the system achieved an average accuracy of about 74%. This takes into account great advances, given the difficulty in detecting signals that do not produce physical movement.
Previous brain-computer interface studies focused primarily on deciphering speech attempts. Participants try to clarify the words even if they are unable to produce sounds. This approach produced a higher accuracy rate, as the attempted speech included stronger signals in the motor cortex. The Stanford team’s work expands this advance by demonstrating that fully imagined words leave no outward traces. For those who may not even be able to attempt a speech, decoding of internal monologues may provide an alternative route to communication. Researchers placed particular emphasis on privacy and consent. To address the concern that implants could decipher thoughts without human intent, the team designed a safety mechanism. Participants were asked to consider a specific “password phrase” before decoding began. Only after creating this phrase internally will the system activate and start translation of imaginary words. In the exam, this password mechanism was extremely reliable, triggering over 98% correctly. Researchers emphasized that systems cannot continuously eavesdrop on thoughts. It only works if the user is explicitly involved.
It’s promising, but the work is still in its early stages. The accuracy of decoding is not unprecedented, but it is not accurate enough to allow for fluent conversations. Statements can be reconstructed with errors, and the decoding process requires intensive computing power and calibration. The implant itself is invasive and involves surgery to place electrodes in the brain. The long-term stability of the implant and the durability of the signal quality need to be further tested.
Nevertheless, this study offers a glimpse into what will be possible in the next few years. When refined, internal speech decoding can give people with paralysis or advanced neurodegenerative diseases a way to communicate more directly and spontaneously than current assistive technologies. Researchers imagine future systems that can produce naturalistic speeches and synthetic speech outputs in real time on-screen, allowing conversations similar to normal dialogue.
The findings also raise broader questions. Deciphering silent thinking touches deep and personal aspects of human experience, and ethical safeguards are essential. Stanford’s team emphasizes that their systems are designed to be strictly commanded for clinical purposes only. However, they acknowledge that the possibility of reading inner monologues presents new responsibility for how such technologies are developed, regulated and used. For now, the results exist as a scientific milestone. They show that even the quietest expressions of language in the mind leave detectable traces in the brain, and that appropriate tools can utilize these traces for communication. For those who have lost their voices, the possibility of regaining the means to express themselves more naturally is no longer in the realm of science fiction, but in the steady advancement of neuroscience.
New research can be read here
