By Heather Allen

The ability to control an object by thought alone sounds like the stuff of science fiction. Yet here we are in the 21st century, a time period often used as a setting for classic sci-fi novels, so one would expect that some of their predictions would have come true by now. In fact, they have. As promised, we now have self-driving cars, mobile phones, 3-D holograms and virtual reality; not to mention the wonderful World Wide Web, all of which feel miraculous to my Gen-X self. Nevertheless, one thing I am eager to see is a thought processor – a way to convert brainwaves into computer data, which will translate into instructions enabling us to manipulate matter without using our physical bodies. Anyone who remembers me from my University days will know how I used to be obsessed with this topic, and much mirth was had at my expense as a result.
In fact, research on brain-computer interface technology (BCI) using electroencephalography (EEG) waves began back in the 1970s with the work of UCLA Professor Jacques Vidal. Since that time, experimentation has largely involved the use of invasive implants directly into the brain, starting with monkeys and rats, and only progressing to humans in the early years of this century. The uses for this technology are wide ranging, particularly for people with disabilities.
Unsurprisingly, the race to bring BCI technology to market has stepped up since 2021’s publication of FDA guidance on nonclinical testing and study design related to BCI devices. A number of organisations are jostling for first place, including Synchron, BrainGate, and Elon Musk’s recently launched Neurolink. However, not every potential beneficiary of a brain-machine interface would be willing, or able, to tolerate the implantable technology many of these organisations are trialling.
The solution to this dilemma is the use of non-invasive brain interface technology. This method is used by the authors of a new study on how a mind-controlled wheelchair can help paralysed people gain mobility. The study by researchers at the University of Texas in Austin involved a test group of three tetraplegic people, who wore a skullcap that detected brain activity through EEG waves. Their thoughts were translated into mechanical commands for the wheelchair via a brain machine interface device, which enabled them to negotiate a natural, cluttered environment after training for an extended period.
“Our research highlights a potential pathway for improved clinical translation of non-invasive brain-machine interface technology,” José del R. Millán, the study’s corresponding author, said. “We show that mutual learning of both the user and the brain-machine interface algorithm are both important for users to successfully operate such wheelchairs.”
Dr Millán and his colleagues recruited three tetraplegic men who were all wheelchair users following similar spinal cord injuries. Each participant underwent training sessions three times per week for between two and five months. The participants were asked to control the direction of the wheelchair by thinking about moving their body parts. Specifically, they needed to think about moving both hands to turn left and both feet to turn right.
In the first training session, all three participants had similar levels of accuracy (when the device’s responses aligned with users’ thoughts) of around 43-55 per cent. Over the next few months, the team saw significant improvement in accuracy in participant 1, who reached an accuracy of over 95 per cent by the end of his training. The team also observed an increase in accuracy in participant 3 to 98 per cent halfway through his training before the team updated his device with a new algorithm.
The improvement seen in participants 1 and 3 is correlated with an improvement in ‘feature discriminancy’, which is the algorithm’s ability to discriminate the brain activity pattern encoded for ‘go left’ thoughts from that for ‘go right’. The team found that better feature discriminancy is not only a result of the device’s machine learning, but also learning in the brain of the participants – in other words, the participants and the device were learning from each other in tandem. The EEG of participants 1 and 3 showed distinct shifts in brainwave patterns as their accuracy in mind-controlling the device improved.
Compared with participants 1 and 3, participant 2 had no significant changes in brain activity patterns throughout the training. His accuracy increased only slightly during the first few sessions, which remained stable for the rest of the training period. According to Dr Millán, this suggests that machine learning alone is insufficient for successfully manoeuvring a mind-controlled device. Interestingly, the most successful participant in the experiment, participant 1, was also the most severely disabled, with no mobility at all below the neck and relying on assisted ventilation.
By the end of the training, all participants were asked to drive their wheelchairs across a cluttered hospital room. They had to go around obstacles including a room divider and hospital beds, set up to simulate the real-world environment. Participants 1 and 3 both finished the task, while participant 2 failed to complete it.
“We believe there is a cortical reorganisation that happened as a result of the participants’ learning process,” Dr Millán said. “It seems that for someone to acquire good brain-machine interface control that allows them to perform relatively complex daily activity like driving the wheelchair in a natural environment, it requires some neuroplastic reorganisation in our cortex.”
The longitudinal study is one of the first to evaluate the clinical translation of non-invasive brain-machine interface technology in tetraplegic people. In the next phase, the team aims to establish why participant 2 failed to experience the learning effect. They hope to conduct a more detailed analysis of all participants’ brain signals to understand their differences and possible interventions for people struggling with the learning process in the future.
Source: Tonin et al., Learning to control a BMI-driven wheelchair for people with severe tetraplegia, iScience (2022), https://doi.org/10.1016/j.isci.2022.105418