Brain implant that decodes intention will let us probe free will

日期:2019-02-28 04:17:02 作者:郇帕躐 阅读:

By Helen Thomson (Image: Spencer Kells) IMAGINE a world where you think of something and it happens. For instance, what if the moment you realised you wanted a cup of tea, the kettle started boiling? That reality is on the cards, now that a brain implant has been developed that can decode a person’s intentions. It has already allowed a man paralysed from the neck down to control a robotic arm with unprecedented fluidity. But the implications go far beyond prosthetics. By placing an implant in the area of the brain responsible for intentions, scientists are investigating whether brain activity can give away future decisions – before a person is even aware of making them. Such a result may even alter our understanding of free will. The implant was designed for Erik Sorto, who was left unable to move his limbs after a spinal cord injury 12 years ago. The idea was to give him the ability to move a stand-alone robotic arm. People with similar injuries have previously controlled prosthetic limbs using implants in their motor cortex – an area of the brain responsible for the mechanics of movement. But this is far from ideal because it results in delayed, jerky motions as the person thinks about all the individual steps of a longer action. Richard Andersen at the California Institute of Technology in Pasadena and his colleagues hoped they could achieve more fluid movement by placing an implant in the posterior parietal cortex – a part of the brain used in planning movements. “We thought this would allow us to decode brain activity associated with the overall goal of a movement – for example, ‘I want to pick up that cup’, rather than the individual components,” Andersen told delegates at the NeuroGaming Conference in San Francisco earlier this month. The two tiny electrodes implanted in Sorto’s posterior parietal cortex were able to record the activity of hundreds of individual neurons. Then, with some training, a computer was able to match patterns of activity with Sorto’s intended movement. “We found there was amazingly specific activity for specific gestures,” says Andersen. For example, certain neurons fired when Sorto imagined moving his right hand to the back of his head, and others when he wanted to move his left hand to his lips. Once this neuronal information had been collected, a computer translated Sorto’s intentions into movements of a robotic arm. This enabled him to control the speed and trajectory of the arm, so he could shake hands with people, play rock, paper, scissors and switch on a blender to make a smoothie. Most importantly to Sorto, he was able to pick up a beer and swig it at his own pace (Science, “It’s awesome, it’s awesome,” Sorto repeated, after drinking his beer. “I hope some day that people with these conditions will have a robotic arm and regain some sort of independence,” he says. Andersen’s team is also working on giving people like Sorto back their sense of touch by placing a similar implant in the somatosensory cortex. People who have had this region stimulated during brain surgery have reported feeling things such as “a wind rushing over my hand” or “my finger being wrapped in something”. At the conference, Andersen announced that his team had placed such an implant in their first volunteer and is working out how to replicate realistic perceptions of touch by stimulating this brain area. Another tantalising possibility is using other intentions decoded from brain activity to control our environment. For example, could we identify the pattern that corresponds to the thought of wanting to watch a film, then have that switch on the television? “You could identify the brain activity for wanting to watch a film and have that turn on the TV” To investigate the feasibility, Andersen’s team had a person with a similar implant to Sorto’s play a version of the prisoner’s dilemma, where players can either collaborate or double-cross one another. The team was able to predict the volunteer’s decision based on the neural activity the implant recorded. This showed that more abstract decisions such as, in this case, the intention to snitch on a hypothetical partner, can indeed be decoded from the posterior parietal cortex. Ori Cohen from the Advanced Virtuality Lab at the Interdisciplinary Center in Herzliya, Israel, says that using abstract commands for brain-computer interfaces is a promising idea. “After all, this is how we control our body – we have a goal such as getting coffee and our brain kick-starts a range of processes involving complex geometrical computations in order to achieve it,” he says. Eventually, he believes that a person with paralysis could imagine themselves making a cup of coffee and have a humanoid robot automatically carry out the action. He is hopeful that such approaches could one day be achieved using non-invasive techniques, such as recording brain activity with an EEG headset, rather than having to embed electrodes in the brain. Others are not so sure.”It’s hard to get really high-quality brain signals with non-invasive technology,” says Jörn Diedrichsen, a neuroscientist at University College London. He thinks this might be off-putting to those who have no medical need to link up with their environment. “You have to ask whether you’d want to have invasive surgery to not have to press a button on a remote control,” he says. “It might be technically possible in five to 10 years, but would you do it?” The most intriguing aspect of Andersen’s work, he says, is that we are now able for the first time to record the brain activity underlying intentions while asking about a person’s conscious experience. For example, Andersen’s team has already started to repeat classic free will experiments in which researchers try to use brain activity to predict a person’s decisions before they are consciously aware of making any. “We will be able to look carefully into big philosophical questions of whether a person’s future decisions can be decoded from their neural activity before the individual is aware of having formed them – and what that all means for our ideas on free will,” says Diedrichsen. “It really captures the imagination.” “If decisions can be decoded before we’re even aware of them, what does this mean for free will?” As our ability to decode brainwaves improves (see main story), the market for mind-reading devices that can be used at home is growing. Some estimates predict this market could be worth as much as $6 billion by 2020. At the NeuroGaming Conference in San Francisco, companies were showing off a range of devices – scientifically validated to varying degrees. An example is NeuroMage – one of the first computer games to harness the power of the mind to control a character using an EEG headset. Players must use certain kinds of thoughts to build up an armoury of spells to cast on opponents. If you have trouble concentrating, MUSE might be more up your street. This app, combined with an EEG headset, teaches you to focus your attention using brain training exercises, and promises to improve your emotional well-being. The system is being tested by several US universities. Then there’s EDGE – the world’s first brain stimulation device aimed at improving your fitness. It involves a portable headset that sends a low current to the temporal lobe. The idea comes from a study in which professional cyclists showed a 4 per cent improvement in stamina when they received a similar kind of stimulation. The researchers involved – who have not endorsed any product – suggest that stimulating the temporal lobe can affect how difficult a workout feels (British Journal of Sports Medicine, While there is no evidence to suggest that short bursts of stimulation can damage the brain, many researchers at the conference were still wary of such devices, since they do not need to be approved by medical regulators. This article appeared in print under the headline “I think, therefore I can” More on these topics: