Faster, Clearer, Closer: A New Era of BCI for Paralysis

In the realm of neuroscience and medical technology, the dream of turning thought into action is no longer just theoretical. Brain-computer interfaces (BCIs) are fast evolving into viable, clinically meaningful tools, and transforming the lives of people with paralysis and neurodegeneration.

Below I share three powerful examples of how far we’ve come, and where we might be headed next.

When I began my PhD in 2015, the world of brain-computer interfaces (BCIs) was entering an exhilarating phase of growth. Three foundational advances were converging:

1. Computing power had finally caught up, capable of processing highly complex brain data in real time;
2. Neuroimaging and neural implants had become precise enough to detect signals at the millimeter, and even micromillimeter scale; and
3. Machine learning algorithms were no longer limited to manual input-output programing, they could automatically decode patterns and generate outputs.

It felt like something from a science-fiction novel: experimental, awe-inspiring, yet slightly out of reach. That’s when Miguel Nicolelis came to speak at my university, the Queensland Brain Institute (QBI), and absolutely blew my mind.

1. Pushing the frontiers of what BCIs can accomplish

Professor Nicolelis, based at Duke University, had been pioneering mind-controlled devices since the early 2000s; beginning with monkeys and advancing to humans. In his early work, monkeys would play games with brain-controlled bionic limbs. This was made possible by his seminal work developing the multi-electrode array. Later in 2008, his team demonstrated how one brain’s signals could be sent across the Atlantic ocean to control a different brain’s output. It was proof-of-principal telekinesis! Yet behind the bionics, Nicolelis was contributing much needed foundational research into neural processing, multi-sensory integration, subcortical motor function and clinical applications.

His team’s efforts culminated in an unforgettable moment at the 2014 FIFA world cup, when a quadriplegic man wearing a robotic exoskeleton kicked the opening ball. Even more astonishing was how, within a year, patients using the device began to re-gain some motor function. The addition of feedback signals to sensory cortex had led to clinical re-classification in some patients with paralysis.

BCIs were no longer just assistive – they were beginning to show therapeutic potential.

However, science is slow, and medicine is slower. The gap between lab success and the everyday patient remained – particularly for brain-to-speech communication.

2. Brain-to-speech: Communicating Through Thought

Assistive technologies for people with conditions like ALS or paralysis used eye tracking or residual muscle input. Think of the systems used by Stephen Hawking to produce speech – slow, mechanical, and tethered to 1990s-era interfaces. These systems are seldom used for conversation beyond “hi, how are you” delivered at a painstaking pace.

Here’s some context on communication speeds:

• Average speech: 125–160 words per minute (wpm)
• Professional speech (e.g., broadcasters): up to 200 wpm
• Average typing: ~40 wpm
• Professional typing: 80–100+ wpm
• Stephen Hawking’s system: ~1–2 wpm
• Older BCI spellers (pre-2020s): 5–10 wpm
• Recent non-invasive BCI spellers: 20–50+ wpm
• Cutting-edge invasive BCI speech decoders: 60–70 wpm with high accuracy

Delays of even 1 second in BCI response can disrupt natural conversation flow. For a smooth and intuitive dialogue, BCIs must operate with latency under ~100 milliseconds – and ideally much less.

BCIs generally fall into two categories:

• Invasive methods, like cortical implants and electrode arrays, offer higher precision due to their proximity to brain tissue — but they carry surgical risks.BCIs generally fall into two categories:

• Non-invasive methods, like EEG and TMS, offer lower signal fidelity but are safer and more scalable.

In 2019, a colleague of mine was working on a non-invasive brain-to-text speller using EEG. Unlike typical systems that relied on eye movements, this system leveraged covert attention signals and the more familiar QWERTY keyboard interface. The closed-loop system could decode intent in just 1.8 seconds, enabling a typing speed of over 20 words per minute. This was an improvement in speed and user-interface over existing devices. Though not commercially deployed, it demonstrated an elegant solution for intuitive, low-risk communication device.

3. Speed, Precision and Cortical Implants

While BCI research sped forward, I moved to Canada to pursue my postdoctoral research fellowship in neuromodulation at the University of Toronto. Since then, I’ve kept an eye on the emerging BCI work.

A recent BCI Award-winning device enabled brain-to-speech communication after just 1 day of training, with high accuracy. This is an improvement on the typical 2-weeks of training required for other invasive brain-to-speech devices.

💡 For patients with ALS – where speech and motor function can decline in a matter of months – a system that works in one day instead of 14 can mean significantly more human connection. More clinical trials are needed to bring this technology into patients’ homes and hospital rooms, particularly in the final months of life when communication is most precious.

I’m still in Canada, now working with ALS patients at Sunnybrook Hospital where I see first-hand the urgent need for tools that are fast, intuitive, and easy to integrate into daily life.

4. Where to next? Minimally Invasive BCIs

There are now over 700 companies globally working on brain-interface technology. (Its not just Neuralink!). The BCI market is projected to reach over $3.6 billion USD by 2030, fueled by applications in neurology, prosthetics, communication aids, gaming, and wellness.

What excites me most are the tangible, life-changing clinical applications we’re already seeing. For patients who’ve lost the use of their limbs due to stroke, spinal injury, or ALS, these devices offer not just hope, but actual progress.

A third innovation worth spotlighting is the Synchron Stentrode. A minimally-invasive alternative to neural implants. This BCI device bypasses traditional brain surgery altogether. It travels to the brain’s motor cortex through an artery in the neck, like a stent used for heart attacks, it is implanted via a catheter through the jugular vein, travels up the vasculature, and expands in place. This approach avoids penetrating brain tissue, making the procedure less risky, more scalable, and more accessible.

Designed by Tom Oxley and developed in Melbourne, the system has already undergone human trials in Australia and the U.S. It has allowed people with severe paralysis to text, email, and operate devices using thought alone — safely and reliably, without the need to open the skull.

Conclusion: A Signal of Hope

Whether it’s giving voice to someone silenced by ALS, reuniting a hand with its sense of touch, or offering neural access without brain surgery, the future of BCI is both astonishing and urgently needed.

But none of this progress happens in isolation. It’s made possible by brilliant researchers, bold funders, visionary engineers, and most importantly, brave patients who lend their minds to innovation so that others may benefit.

Here’s to closing the gap between thought and expression.
Here’s to the ties between science and care.
Here’s to the day these technologies are standard tools in clinics like ours at Sunnybrook – improving everyday acts of human connection.

1. What kicked it off: BCI Exoskeleton

Further Reading:

Lebedev MA, Nicolelis MA. Brain-Machine Interfaces: From Basic Science to Neuroprostheses and Neurorehabilitation. Physiol Rev. 2017 Apr;97(2):767-837.
doi: 10.1152/physrev.00027.2016. PMID: 28275048.

Donati ARC, Shokur S, Morya E, et al., Nicolelis MAL. Long-term training with brain-machine interfaces induces partial neurological recovery in paraplegic patients. Sci. Rep. doi: 10.1038/srep30383, 2016.

Mitchell, Peter et al. “Assessment of Safety of a Fully Implanted Endovascular Brain-Computer Interface for Severe Paralysis in 4 Patients: The Stentrode With Thought-Controlled Digital Switch (SWITCH) Study.” JAMA neurology vol. 80,3 (2023): 270-278. doi:10.1001/jamaneurol.2022.4847