You can (literally) turn on the lights using your mind— Matilda is not science fiction anymore
--
In the past 100 years, we’ve made great advances of measuring and understanding the brain. Researchers and doctors have harnessed this knowledge to reverse the symptoms of Parkinson’s, decrease depressive episodes, help athletes run faster — and many other advancements we believed were only possible in science fiction. Now, many of these devices are making their way into our homes. Are we ready?
The tech world is filled with a cacophony of news on computer “brains” and AI development — but the advancements we are making with human brains has somehow missed the limelight. We’re moving into the world with human-led evolution. It’s not science fiction any more.
How can we see what’s happening inside the brain?
An electroencephalogram (EEG) is a method that records the electrical activity in the brain using electrodes attached externally to the scalp. Brain cells “communicate via electrical impulses and are active all the time” — even while sleeping.
Although, we’re still a long way from understanding all the fluctuations in the brain waves, one way to look at them is by “EEG Band.” Here’s a few examples when a certain range would occur.
Band (Frequency): Normal occurrence
- Delta (<4 Hz): Adult slow-wave sleep
- Theta (4–7 Hz): Drowsiness in adults and teens, associated with inhibitory tasks (e.g., avoiding a piece of cake)
- Alpha (8–15 Hz): Relaxed, reflecting, closing eyes, also associated with inhibition
- Beta (16–31 Hz): Active thinking
- Gamma (32+ Hz): Combining multiple senses (e.g., sight and sound), short-term memory matching
Challenge question: Which band do you think your brain is in right now?
EEGs are one of many tools to explore brain activity. What are the other techniques?
Traditionally, clinicians and researchers have used techniques like Functional magnetic resonance imaging (fMRI), which measures brain activity by detecting changes associated with blood flow. Another common method is Positron Emission Tomography (PET), which relies on injection of a radioactive isotope to measure cerebral blood flow.
New techniques are also starting to become more popular like Functional Near-Infrared Spectroscopy (fNIRS), which measures brain activity through hemodynamic (blood flow) responses associated changes in near-infrared light. It’s a form of optical brain imagine, and is relatively non-invasive, safe, and low-cost. There are many other brain imaging techniques, and the accuracy of these techniques are constantly improving.
Neuroimaging devices have been used for a wide range of clinical activities from confirming or ruling out conditions related to seizure disorders (e.g., epilepsy), traumatic head injuries, brain tumors, memory problems, sleep disorders, strokes and migraines.
Historically most of these imaging methods required huge and expensive machines that were only available in research labs and hospitals. But that’s starting to change.
EEGs, for example, are often used to determine if the root cause of a seizure is epilepsy or not. Rather than having a patient sit in a hospital waiting for a seizure to occur, ambulatory EEGs (aEEGs) allow the patient to return to their natural (and likely more stressful) environment — often resulting in a better diagnosis.
At-home devices have huge implications to reduce costs for hospitals and insurance companies, and — if the technology is up to par — results in a more natural set of data for the patient.
So what?
Neuroimaging devices are getting cheaper and more powerful — sensors are able to pick up more signals and the software is able to process even more data in real-time. This is driving much of the change in the neurotechnology community — because the real-time response is improving research and making this technology more accessable to a larger audience.
As brain-computer interfaces become more popular, high-quality brain-imaging technologies, like EEGs, are the linchpin. There are three stages to BCI technologies:
- Sense: Signal acquisition
- Infer: Signal processing
- Act: Effect device
Whether we call it “magic” in the case of Matilda or “information flow” in the case of a brain-computer interface — we need a way to take the signals from our brains and convert them into a desired action.
There’s a whole ecosystem of devices that’s coming online for you do to this in your own home. Want to control your TV without getting off the couch? You can do that. Want to play guitar even if you’re paralyzed with a severed spine? That’s possible, too.
NeuroTechX is creating a open-source repository of the tools you’d need to set up your own brain-computer interface. Let’s look at a few of the examples.
Muse is a brain-sensing headband designed to improve meditation practice by providing real-time audio feedback based on the state of your brain. If you’re calm, you’ll hear the sound of peaceful weather. If your mind wonders, the weather gets stronger — so you can guide your mind back to calm.
Muse also has an developer kit (SDK) that has inspired people around the world. Muse is one of the most popular consumer EEG, and it’s relatively high-quality hardware and software packages form the foundation of many new products. Judith Amores, from the MIT Media Lab, has hooked up her Muse with an Oculus rift to create a visual-based feedback loop to her meditation practice (and you can too!).
Neurosky, another commercial EEG with a strong developer kit, has been used for a range of features including mind-controlled lighting and TV and also video-gaming. Currently VR games require hand-held controllers for gameplay, but can you imagine a world where you can use your thoughts to change the game? It would be a fully immersive VR experience. Neurosky is already developing a set of games where this is possible. Ever wanted to telekinetically throw trucks at your enemies? It’s now possible.
Neurosky is not the only one looking at this gaming market. EMOTIV’s mobile EEG also has an integration with Unity, one of the leading game engines in the world.
There are many other EEGs on the market including OpenBCI (open-source, on GitHub), BrainCo, and OpenEEG, as well as research-level devices like g.BCIsys.
But really, how good is the data?
It depends what you want to use it for. In a medical setting, EEGs for clinicians will often have many more electrodes and will often use gel (“wet”) to lower the impedance and improve signal quality (i.e. data quality). Hair is one of the biggest challenge for outside-of-lab, easy-to-use, quick-to-setup EEG devices. Since nearly all of the consumer EEGs used at home are “dry,” meaning the electrodes do not use gel or a saline-based contact to the scalp, the quality of the recording is generally worse. Though, that goes back to the original question: what do you want to use the data for? For many BCI applications, in addition to price, there are a number of trade-offs that a consumer will consider when making a purchase:
- Number of electrodes
- Location of the electrodes
- Quality of the device
- Time to install / challenges to set up (e.g. having clean off gel from the scalp when using wet-electrodes)
- Accessibility to the RAW EEG signal via SDK (or not)
All of these factors impact what is possible to do with the device and influence the quality of the data. Cleaner data leads to better models (e.g., better analysis), and armed with these improvements, we’ll have sharper and more refined control of the devices.
Although the industry has made tremendous progress over the past decade, we’re still a limited in what we’re able to do with today’s devices. Though many devices can evaluate broad changes in brain state — and perhaps connect that change to a simple action like turning a light on/off — we have a long-way to go before BCI devices become what we see in the movies.
Are EEGs the only imaging technologies that will make it into our homes?
EEGs are popular for at-home use because devices like PET and fMRIs are large and expensive and don’t scale well to consumer use. Other technologies, like fNIRS will likely become more popular and powerful over the years — and will be able to measure other responses, such as bloodflow, to compliment the electrical measurements.
Moving beyond brain-imaging, from a biosensing perspective, EMGs, EKGs and more can sample the electrical activity of other parts of our bodies and will likely play a large role in future brain-computer systems.
What’s next?
All of this begs an even bigger question: if all of these devices measure the state of the brain to trigger actions in the outside world, what if we could change the state of the brain itself?
Put in other words, instead of measuring brain activity, what if there was a way to repair and improve it?
As you may have guessed, there is a way: it’s called neuromodulation. And we’ll dive into those technologies in Part 2.
Want more posts like these from NeuroTechX? Tap the ❤ below so more people can see this. Don’t forget the check out the open-source BCI GitHub Repository.
