A brain-computer interface (BCI) reads signals from the brain and turns them into commands for a computer, a robotic arm or a speech synthesiser. The most advanced ones are implanted. In published studies, people with paralysis from ALS, spinal cord injury or brainstem stroke have used them to type, speak through a computer voice, move a robotic arm and, in one case, walk.
But these results come from small studies, often with a single participant. As far as we could find, no implanted BCI is approved for general use to restore communication or movement in Ireland, the UK, the EU or the US. They are available only through clinical trials. None of those trials was recruiting in Ireland when we checked.
This page explains what the published research shows, what it doesn't, and the questions it raises.
When you try to move or speak, nerve cells in the motor areas of the brain fire in patterns. That still happens after paralysis, even though the message no longer reaches the muscles. A BCI picks up those patterns and software learns what they mean.
There are four main ways to record the signals:
The closer the sensor is to nerve cells, the more detailed the signal, and the bigger the surgery. Scalp EEG is the same basic technology as a hospital EEG test. It cannot pick up the fine detail that implanted systems use for fast typing or speech.
Current implant trials are for adults with severe paralysis who can still think clearly. The BrainGate feasibility study, for example, enrols adults aged 18 to 75 with paralysis of all four limbs from spinal cord injury, brainstem stroke or motor neurone disease (Rubin et al., Neurology 2023). People with ALS who are losing the ability to speak are a major focus.
These are peer-reviewed results. Each involves one to a handful of participants, studied intensively.
For comparison, the Willett paper puts natural conversation at about 160 words per minute.
These are real advances. The ALS speech results in particular show people regaining conversation after losing their voice. But some limits are easy to miss:
In the Lorach study, the participant could walk with crutches even with the device switched off, after rehabilitation with it. The authors suggest the system supported some neurological recovery. That is one person.
The longest-running safety data come from BrainGate. Rubin et al. reported on 14 adults implanted between 2004 and 2021, with 12,203 days of combined implant time. There were 68 device-related adverse events, 6 of them serious. The most common was skin irritation around the connector on the head. There were no infections inside the skull, no device removals for safety reasons and no device-related deaths or permanent disability.
The endovascular approach avoids open brain surgery. In the SWITCH study, 4 people with paralysis (mainly from ALS) were followed for 12 months. None had a serious device-related adverse event, a blocked blood vessel or a device that moved. Each could control a computer. Four people is a very small safety study.
Safety for newer, wireless, fully implanted designs is still being gathered in trials.
Checked 3 November 2026.
Some details matter here:
If you or a family member is interested in a trial, ask your neurologist. Trials have strict criteria, involve surgery and are research, not treatment.
What helps now. People losing speech already have communication aids, such as eye-gaze and switch systems. Speech and language therapists assess for these. In Ireland, disability payments and HSE supports may help with costs.
Brain data is unusually personal. It may reveal intentions, emotions or health information. Questions that matter for anyone in a trial, and for regulators:
In Ireland and the EU, health data is a "special category" under the GDPR, with extra protection. The Data Protection Commission explains that processing it is prohibited except in limited circumstances.
In November 2025, UNESCO adopted the first global standard on the ethics of neurotechnology. The Recommendation calls for explicit consent and transparency. It warns against workplace monitoring and non-therapeutic use in children. UNESCO notes that consumer neural data can be shared without consent.
Implanted neurotechnology isn't new. A vagus nerve stimulator for epilepsy was first approved by the FDA in 1997 (PMA P970003). That device sends signals into the nervous system. BCIs mainly read signals out. We explain the difference between medical nerve stimulation and online claims in what the vagus nerve actually does.
Brainstem strokes are one cause of the "locked-in" paralysis these devices aim to help. For prevention, see our guide to stroke signs and modifiable risks.
Can I get a brain implant in Ireland? Not as a treatment. We found no approved implanted BCI and no recruiting BCI trial in Ireland when we checked. A neurologist can advise on trials elsewhere.
Can a BCI read my thoughts? Not in the everyday sense. Current speech BCIs decode attempted speech from motor areas of the brain, after training with the user. They do not read unspoken thoughts at will. They do, though, collect sensitive data, which is why consent and privacy matter.
Are consumer "brain-sensing" headbands the same thing? No. They use scalp sensors with much weaker signals. They are sold as wellness products, not as medical BCIs.
Is Neuralink approved? No. Neuralink is running clinical trials in the US and UK. Trial registration is not approval.
This article is general information, not medical advice. Talk to your neurologist about trials or communication aids.
Reviewed by: Proco Team
Last updated: November 2026