Every time you blink, think or move, your brain generates electricity as individual neurons in the skull transmit information needed to make it happen. If we could detect the electrical signals produced by individual neurons, we could, in theory, read a person’s mind.
Amazing. And exceedingly difficult. The amount of electricity generated by an individual neuron transmitting a single piece of information is incredibly tiny. The brain, all 100 billion neurons of it, produces en masse about 20 watts—barely enough to power an incandescent light bulb. For decades, the best neuroscientists could do was use electroencephalography, or EEG, to detect the signals that characterized different stages of sleep, say, or the in-brain power surges brought about by epileptic seizures. And that wasn’t easy. They had to shave people’s heads, put them in a room far from any other sources of electricity and use conductive gel to stick several dozen electrodes to the skin atop their skulls.
Then, in 2007, Philip Low, while working on his Ph.D. at the University of California, San Diego, invented the Sleep Parametric EEG Automated Recognition System (SPEARS) algorithm. SPEARS gave Low the ability to create a cluster map of brain activity using only the information gleaned from one electrode—in industry terms, a “single-channel EEG.”
Before Low, EEG devices that took input from just a few channels weren’t considered good for much; to get really useful data, pre-SPEARS, you had to cover someone’s skull with electrodes, impeding most everyday activity. About the only use of single-channel EEG pre-SPEARS was in the toy market, where in 2007 the Silicon Valley–based brain-computer interface technology company NeuroSky released The Adventures of NeuroBoy, a simple video game in which players could use a low-cost single-channel EEG headset to control a telekinetic protagonist. Though relatively affordable ($199), the NeuroBoy headset couldn’t mimic the precision of Low’s algorithm. Yet it was ground-breaking for another reason; it was the first consumer EEG product to use “dry electrodes”—ones that could read signals sans conductive gel.
Dry electrodes soon proved of interest to the medical community. Because they didn’t require gel, which tended to dry or melt after a few hours (or even less time if the wearer was engaged in strenuous activity), they could stay applied indefinitely. Better yet, without the gel, users simply had to make sure the conductive part of the electrode was touching skin.
These advances in EEG technology stoked the ambitions of entrepreneurs like Tan Le, CEO of Emotiv, a company that manufactures EEG rigs meant for consumers. Emotiv’s older, more complicated setups were aimed at a hardcore hobbyist niche market. With new technologies rendering EEG relatively easy to use, Le now hopes to build a Fitbit or Apple Watch for the mind.
Reading healthy minds
“Anxiety, depression, schizophrenia, dementia, Alzheimer’s, Parkinson’s, autism…” Le ticks off a litany of neurological disorders, then continues: “Most of these conditions are developmental in nature. The markers probably exist decades before the symptoms manifest themselves. We need more early intervention and early monitoring.” According to Le, the way we do neuroscience nowadays is fundamentally flawed. “For the most part, we only study brains when something goes wrong.” People with supposedly healthy brains almost never have their brains scanned—in part because, until this decade, obtaining readable results from EEG has been time- and labor-intensive.
The rise in popularity of wearable health technology, Le says, has “opened up the opportunity for us to monitor, track and learn about the brain and to start to build better models of the brain across a broad spectrum of users,” not just people who are ill. In other words, if enough people start using the Insight (Emotiv’s new EEG rig) and let Emotiv collect data about their minds, maybe that amassed data could allow neuroscientists to finally know what a healthy brain looks like when it deals with various everyday stimuli. That information could allow neuroscientists to better identify “early biomarkers for a variety of neurological disorders,” says Le.
Low is on the same page. He founded his EEG-focused company, NeuroVigil, soon after creating the SPEARS algorithm. For years, the company has been working on small-scale projects for high-profile employers; the iBrain, his single-channel EEG device, has been used as part of a communications apparatus for Stephen Hawking and included in NASA’s field kit for astronauts. But now, says Low, it's time for NeuroVigil to start scaling up: “The senior care community has approached us to massively look at the brain activity of seniors before they're diagnosed with dementia. We're working with ASHA [American Seniors Housing Association], and we already have a contract with a number of operators to start delivering single-channel EEG to a number of senior centers this year.” The long-term plan is to make the iBrain an everyday device, for everyone. “People have their blood pressures checked as a matter of routine fairly frequently,” Low says. “I think one day the same will be true about the brain, and we will check it more and more frequently until it is continuous.”
Low hopes eventually that this will take the trial and error out of how we study the psychological effects of drugs. “Before someone takes a drug, we want to use the iBrain and see: What does their brain look like? Then when they take the drug, we want to continue monitoring the brain. We want to see the changes the drug makes.” Low’s interest isn’t just scientific. On his 10th birthday, Low saw his father get thrown in jail for using a gun to threaten a banker who had defrauded him. His father was shortly afterward pardoned, on the basis that his behavior was caused by a sleeping pill he had taken, “which was apparently making a lot of people very aggressive,” Low says. As a teenager, Low had been dubious of the claim that the crime could be the sleeping pill’s “fault” and not his father’s—until he started his Ph.D. in computational neurobiology and a member of his thesis committee, J. Christian Gillin (best known as the late founder of the journal of Neuropsychopharmacology), cautioned him against running experiments with that very drug. “He said, ‘Don’t use that. It makes people crazy,’” Low recounts. “And I picked up the phone, called my father and apologized.”
Some indicators show that it won’t be too long before we have enough data to start to read our minds. Cost, long the sticking point for EEG makers, seems to be dropping sharply. In December 2009, Le, releasing an EEG rig called the Epoc for hobbyists, struggled to keep the cost under $500. But by December 2011, students at the National University of Computer and Emerging Sciences in Peshawar, Pakistan, published the specifications for a single-channel EEG-enabled brain-computer interface that lets people with Lou Gehrig’s disease communicate via a mind-controlled text-messaging apparatus for under $12 per device.
On the other hand, the success of large-scale projects like those Low and Le envision depends on getting healthy people to wear EEG rigs and share their everyday brains with anonymous scientists. Given that many people are prickly about companies collecting much less personal data—anonymous advertisers scooping information off our phones, for instance—one wonders if the nascent mania over everyday EEG may soon crash headfirst into privacy concerns. After all, it’s one thing to hook up your brain to electrodes and let doctors read the signal when you’re in a controlled setting, another to let scientists you may never see read your EEG day after day. (The Health Insurance Portability and Accountability Act, or HIPAA, requires doctors to keep patient EEG scans confidential, but consumer EEG devices face no such strictures.)
Still, Le has had some success in getting everyday EEG users—at least those who use Emotiv headsets—to share personal details above and beyond what one might expect. Turns out those early adopters already hooked on Emotiv are willing and often eager to share not just EEG readouts but also additional “neurologically relevant data” (age, gender, handedness, educational status, social data, languages spoken and musical skills) that can help contextualize the signal. And that’s crucial because in terms of technology the limiting factor is brain wave literacy. Even the scientists skilled at reading signals generated by EEG done as dozens of electrodes clamped to the skull with conductive gel in a controlled laboratory setting can’t necessarily interpret the signals generated by a person who has used an Insight or iBrain in the noisy, messy, untidy real world to perform a load of everyday tasks.
For the first time in history, the wonders of technology have allowed us to “hear” the noises our nerves make when they talk to one another. But before we can really read minds, we’re going to need to figure out what those noises mean—what the nerves are saying. And that task has only begun.
Le and Low say their companies are capable of matching signal to meaning and prepared to process EEG data at the enormous scale it would require. Of course, says Le, getting users to pitch in is crucial: Those who use Emotiv headsets are asked to “indicate” life events “such as moving house, getting married or divorced or suffering a loss,” as well as physical injuries, especially those that might affect the brain, like “head trauma or mild concussion.”
It won’t be just big companies breaking the code; the DIY and academic communities, while lacking Emotiv’s and NeuroVigil’s number-crunching infrastructure, will also be big players here. The same Pakistani team that built the mind-controlled communications device for under $12 has also found a way to amplify EEG so it can be read on a laptop sound card. In other words, if you want to engage in at-home EEG experimentation, all you need is to follow some basic instructions to build your own nonproprietary, startlingly inexpensive and simple EEG equipment.
And who knows? Maybe you’ll learn a little something about yourself.
Correction: An earlier version of this article incorrectly referred to Emotiv as Emotiv Systems.