Most people take the ability to communicate for granted. Yet for individuals with neurodegenerative conditions such as amyotrophic lateral sclerosis (ALS) or those who have suffered brain trauma, this fundamental capacity can vanish entirely.
Implantable brain-computer interfaces (BCIs) offer a path to restoring that connection by translating neural activity into text. “These devices are providing some of the most exciting examples of the translation of fundamental neuroscience and neuroengineering toward clinical care,” said Leigh Hochberg, a neuro-intensive care physician at Mass General Brigham, during a press conference at the 2026 Society for Neuroscience meeting.
At the meeting, three researchers presented advances in implantable BCIs that could one day restore voice to these patients.
Making BCI Technology Mobile
Although implanted BCIs enable conversation, many current systems depend on bulky hardware and suffer from noticeable latency between a user’s intent and the system’s output. “These slow and variable response times can impede natural communication,” said Chethan Pandarinath, a neuroscientist at the Georgia Institute of Technology and Emory University. He added that the need for large computers with sufficient memory and processing power “really restricts them from being mobile, on-the-go, always available systems.”
To address these limitations, Pandarinath and his team explored running BCIs on smartphones, leveraging their portability and processing power to reduce latency. They adapted existing smartphone dictation technology to decode neural signals, developing an adaptor that converts brain activity into speech signals recognizable by standard dictation engines.
The team trained their model using neural data from two BrainGate 2 trial participants. The smartphone-based system produced responses faster and with a smaller memory footprint than conventional computer-based setups.1 The platform is now undergoing real-time evaluation with trial participants. One participant described the difficulty of ordering food at a restaurant when forced to rely on lip-reading, expressing enthusiasm that this technology could offer a viable alternative.
BCIs as Lifelines for Communication in Locked-In States
BCIs serve as a critical lifeline for patients with severe motor impairments who cannot write or speak. While some patients retain limited eye or lip movement for basic communication or eye-tracking devices, total paralysis results in locked-in syndrome, severing all motor output.
Daniel Rubin, a neurologist at Mass General Brigham and Harvard Medical School, worked with an ALS patient who had lost most motor function but retained eye movement. Rubin’s team implanted a BCI that decoded attempted movement signals, enabling the patient to spell and select words. Within a year, the patient lost all eye movement, leaving the implanted BCI as his sole communication channel with the research team.
“We quickly set out to make sure that he could use this system anytime he needed, and importantly, that he could use this system when our team wasn’t present,” said Rubin.
Rubin’s team trained caregivers to set up and calibrate the system. A major challenge emerged: the BCI could not be calibrated during sleep, yet the patient’s locked-in state made it impossible to discern wakefulness from rest behaviorally. The team initially attempted to use electroencephalography (EEG)—the standard for sleep staging—to monitor his state.
Surprisingly, while the patient was awake and communicating via BCI, his EEG patterns resembled sleep stages and differed markedly from typical awake baselines. “This suggests there may be many individuals diagnosed as unconscious via EEG who are actually awake and locked in, unable to communicate due to paralysis,” Rubin said.
Extending BCI Technology to the Voice
Finally, Maitreyee Wairagkar, a neuroscientist at the University of California, Davis, presented a system that translates neural activity directly into audible speech—a brain-to-voice interface. Her team implanted probes in the speech motor cortex—which governs the lips, jaw, tongue, and larynx—of a single participant. Neural signals fed an AI decoder that converted them into words, which a speaker then vocalized in real time, enabling natural conversational flow.2
In comprehension tests, 94 percent of listeners understood the synthesized speech.
“But our speech is more than just words,” Wairagkar said. “Pauses, intonation, and pitch all influence a person’s meaning when they speak. We emote through our voice.”
To capture this nuance, the team investigated decoding pitch and intonation from neural signals. Working with a former trained singer who lost her voice to ALS, the system decoded intended musical notes with 84 percent accuracy.
Although implanted BCIs remain investigational, the researchers agreed they are progressing toward broader regulatory approval. “The hope is that in a few short years, these devices will become part of standard clinical care,” Rubin said.
Also Read
- Consumer Watchdog Says Big Tech’s DMA Compliance Is All Show, Urges Harsher EU Penalties
- VMware Ships Two Critical Advisories as KVM Nested-Virtualization Flaw Sparks Cloud-Vendor Standoff
- Apple Launches Hiring Drive for Senior Engineer to Lead In-House AI Agent Platform
- Solo Meerkats Successfully Navigate Home From Over a Kilometer Away
