Elon Musk’s new brain-machine interface startup, Neuralink, is just one of several recent announcements that have sparked discussion about the potential of letting computers read our brains. Whether it’s Facebook’s newly revealed Building 8 project, or even more out-there experiments like “neural dust,” developments in the field of mind-reading are becoming almost routine.
There are some potentially troubling implications to Musk’s descriptions of a “neural lace” implant’s robust set of abilities, though. One is its ability is to create neural activity, in addition to reading it.
Neural lace is an injectable implant, a rolled-up flexible net of conductive filaments that will travel to the brain and unfurl to distribute a network of controllable electrodes over much of its surface. It was first tested in mice in 2015, and this study found the mouse brains had fully accepted the neural lace and had begun to incorporate it, growing around them over time without any negative effect.
All this means that neural lace has far better access to neurons from within the skull than, say, the non-invasive system in development at Facebook. By injecting its electrodes beneath the scalp and the skull, Neuralink has given itself a potential platform from which to insert new electrical signals into the brain.
We know that Neuralink fully plans to make use of that platform — in fact, brain stimulation with neural mesh underlies the very first projected applications, including the elimination of seizures and the application of “electroconvulsive therapy” with far fewer and less severe side-effects than currently possible.
But research into neural control has also scratched at the surface of other possible applications for stimulating neural electrodes that might give pause to anyone considering the implant for pure medical use.
In 2012, researchers from the University of Kansas published a study in the journal Neuroscientist titled Neural Hijacking, which focused on the capacity for outside stimulation of neurons to interfere with proper functioning of the brain and mind. They found that precisely modulated patterns and intensities of electrical signaling can not only block but override natural signaling in the motor cortex.
University of Kansas professor Paul Cheney, one of the authors of the 2012 paper, tells Inverse that while project is intriguing, it will need careful modulation to avoid doing more harm than good.
“A double-edged sword”
“A network of electrodes or even a concentration of electrodes appropriately placed in or on the cerebral cortex might be capable of shutting down seizure related activity,” Cheney says. “However, this type of electrical stimulation is a double-edged sword.
“It is also very easy to produce seizures, even with a single microelectrode, if sufficient frequency and current are applied.”
In terms of development, this simply means the medical applications have to be thoroughly tested to avoid causing negative side-effects, but makers of connected (and purportedly wireless) modern technology should be far more worried about the possibility of malicious interference. The potential dangers of a hacked brain implant generally stop at surveillance of a user’s thoughts, but given the abilities of neural lace, those dangers could soon extend to seizure and sudden unconsciousness, as well.
Right now, understanding of the brain doesn’t allow electrodes to create intentional movement in the motor cortex. An array of electrodes could be used to intentionally control a person’s left arm, for instance, but the movement of that arm couldn’t be directed toward any one point in space. If researchers do ever map the brain thoroughly enough to reveal how all the data of a sensible movement is created, however, implants like neural lace would be an optimal, non-invasive way of doing it.
“The effects you see with stimulation will be related to the sites in the cerebral cortex you stimulate. Motor cortex produces movements, visual cortex flashes of light, somatosensory cortex sensations related to the skin surface, etc.”
To see that neural lace could have worrying consequences, we don’t need to imagine neuro-hackers or even go beyond Neuralink’s own press releases. Remember that one projected use was “electroconvulsive therapy” — a treatment for extreme clinical depression.
In a single breath, this upstart tech company is suggesting both that it wants to develop this implant’s ability to powerfully modulate emotional well-being with as few side-effects as possible and that it wants to push this implant to as many consumers as it can.
Still, Musk’s vision of a “merger of biological intelligence and digital intelligence” doesn’t seem to include much in the way of restraint, or genuine consideration of the potential second- and third-order consequences of introducing these capabilities.
Luckily, Cheney said that researchers still have “a long way to go” to develop anything like a consumer-ready neural lace implant. And in any case, his own work on neural hijacking aside, he’s optimistic about the future for this technology.
“Real potential for the future”
“I think the idea of melding biologic and machine intelligence through electrical stimulation has real potential for the future not only for repairing damaged and diseased brains but also, potentially, for improving brain function in general.”