The views expressed here are solely those of the author and do not represent positions of IEEE Spectrum or the IEEE.
Scientists have spent decades trying to build machines that talk to the brain. Robust and reliable neural interfaces have long been a holy grail in the field of neuroscience. The hope is that a new wave of research programs, including the BRAIN Initiative and Human Connectome, among others, could lead to groundbreaking technologies for helping people with brain-related diseases. If effective, these new therapies could even, some argue, bring about the end of disability.
Some believe that developing such interfaces will require advanced brain implants that are still a decade or more away. More recently, though, neuroscientists—as well as a legion of “brain hackers”—have turned to powerful new sensing, processing, and prototyping tools to explore a host of non-invasive techniques to stimulate the brain. Some of these methods, proponents say, could benefit not only patients who suffer from disease or injury, but also healthy individuals, who would be able to learn faster, acquire better math skills, improve their memory capabilities, and even boost their creativity.
So, is this all still science fiction or are consumer brain products nearing commercialization? Should entrepreneurs and brain researchers consider starting companies right now? I believe the answer is yes—now is the time for you to start a brain tech startup. Belgian company STX-Med has recently received FDA approval to market its Cefaly headband device to people suffering from migraines. At my firm, Lux Capital, we have invested in Halo Neuroscience, which is building a non-invasive brain-stimulating device capable of enhancing cognitive performance.
Lux and other VC firms are on the prowl to back more entrepreneurs aiming to build promising companies around our brains. We believe this is just the beginning of what will become a huge industry. To see why, consider these four trends contributing to make neuroscience a fertile ground for exciting new startups.
1. Better Science
Johnny Depp, as Professor Will Caster in the futuristic science fiction thriller Transcendence, would drill holes into the skull to model the brain. This is what scientists often have to do in real life to examine individual neurons. They use sophisticated electronics that capture high-fidelity signals from hundreds of neurons, and apply complex algorithms to interpret these signals. This requires complex, expensive custom electronics to be implanted into the brain—an approach that still has a lot of technical and safety hurdles to solve until it can be applied to patients. Engineers, however, are starting to explore more pragmatic approaches, embracing the reality that we don’t, and probably won’t, fully understand how the brain works for many, many years. As a result, they are building empirical models of how the brain functions by mapping out the relationship between 1) simple physical/emotional stimuli, and 2) easy-to-measure physical/emotional responses. One advantage of this approach (at least from a VC point of view) is that it is more conducive to building products, and we’re already seeing a first generation of such devices hitting the market.
2. Better Sensors
Implanting a microelectrode array, similar to a miniature bed of nails, entails a potentially life-threatening surgical procedure involving drilling holes into the skull. The alternative is surface sensors that use “wet” hydrogel-prepared electrodes requiring direct contact with the scalp. As you would expect, the more invasive electrodes can monitor small groups of neurons (and potentially even individual neurons), offering more accurate and refined signals. The surface electrodes, on the other hand, capture the activity of larger groups of neurons. But here’s the thing: the non-invasive techniques are getting better at a fast pace. Thanks to modern machine learning and other AI methods, researchers are learning how to make sense of non-invasive sensing data. Recently, neuroscientists managed to build a device durable enough to capture high-fidelity neural signals from surfers—while they are surfing! The research not only proved that off-the-shelf sensors can work in harsh environments, but also showed that neural activity in the surfer’s “peak performance” state resembles that of a meditative state. Coupled to the ubiquitous sensors in our phones, and soon in jewelry and clothing, we will be able to capture ever larger amounts of data to infer intent, mood, and treat disease.
3. Better Data
Rather than build a deterministic “state machine” for the brain, researchers are attempting to use big data and analytics tools to build sophisticated models to link stimulus, intent, and action together. For example, in the most basic case, our neurons fire to signal intent to our peripheral nerves, and eventually muscles, resulting in movement. But instead of linking each firing pattern to a movement, from which to infer intent, why not measure the signals from our peripheral nerves, or even muscles, which are far easier to capture and simpler to interpret? Thalmic Lab’s Myo demonstrates this approach by fusing data from our muscles and measuring motion from an arm band. By combining signals from an array of cheap, passive sensors using advanced machine-learning algorithms, we can potentially infer intent with high fidelity—and as a result, a host of commercial applications would become possible.
4. Better Wearables
Until the middle of the last century, consumers typically refused to adorn their bodies with objects unless they were studded with precious stones, or assembled by craftsmen. Cheap watches from Asia in the late ’70s began the trend of function over form. Today, the helmet-mounted GoPro camera is becoming standard equipment for bikers, skydivers, and snowboarders. Not only are they accepted—they are cool. Cognition-enhancing games such as InterAxon’s Muse and the ECG headband Melon are being embraced by professionals hoping to reduce anxiety and heighten focus. In the future, devices we wear and carry on an everyday basis may communicate with each other and with our smartphones. These devices can form ad-hoc networks, and perform certain functions based on the quality/completeness of the data they are collecting from sensors around us, towards improving models and performing more complex tasks. We’re in the early days, but next-generation wearables might also react to sensor data to provide brain-stimulating responses. This means they could potentially help us be more calm, productive, and happy in our daily lives. They might even give us new sensory experiences, or ways to share and express ourselves beyond text, images, and videos.
Shahin Farshchi is a partner at Lux Capital where he invests in hardware and robotics companies. He is based in Palo Alto, Calif.