Notion Brain Computing Platform: Interview with AJ Keller, CEO of Neurosity

By | March 3, 2020

Neurosity, a neurotech company based in New York, has developed the Notion brain computing platform. The system has a huge range of potential medical applications, including monitoring a patient’s mental health, to diagnosing concussions, to allowing paralyzed patients to control electric wheelchairs. The headset can do this because it incorporates sensors that detect brain activity and translates those signals into specific commands.

Notion is the first “brain reading computer,” according to the people behind the technology, and despite being able to connect to the internet it is largely self-reliant, incorporating a CPU, RAM, sensors, and apps within the headset itself. The computing processes take place within the headset, a device which is as powerful as a Macbook Air, rather than on a remote computer, which sets the device apart from other neuro devices that can read brain waves.  

The developers claim that the system will be a game changer for people who use an electric wheelchair. They would only need to think about the wheelchair moving in order to control it. This would be particularly useful for people who struggle with conventional controls for electric wheelchairs, or those who can’t use them at all.

Neurosity are developing apps for the system themselves, but have also made it possible for independent app developers to create their own apps that can take advantage of the hardware. The company hopes that there will eventually be a wide array of apps to choose from on a platform similar to a smartphone app store.

Medgadget had the opportunity to talk to AJ Keller, co-founder and CEO of Neurosity, about the technology.

Conn Hastings, Medgadget: Please give us an overview of the current state of the art with regard to devices that can read brain waves. What are their limitations?

AJ Keller: In terms of current devices that read brain waves, they typically require large nets of sensors, preparation on each sensor location, and require constant attention from the person administering the brain wave reading. These sensor nets feed into large expensive amplifiers, with up to 256 sensor connections, that are hardwired into a computer for data collection. Most of the time the data processing and signal analysis occurs after the experiment is conducted. These systems are inaccessible to the public, can’t leave the lab, and the research performed using them can’t be implemented in the real world.

Medgadget: How does Notion differ from other pre-existing neurological monitoring devices?  

AJ Keller: Notion is the first and only “brain-reading computer” which combines computers, brain sensors, and internet connectivity all in one device you wear on the back of your head. Notion has eight sensors that allow for a preparation free recording, enabling high-quality EEG recordings in seconds versus other neurological monitoring devices that can take dozens of minutes to set up. Also, Notion includes computing power. Neurosity will allow third-party developers to create never-before possible applications to run right on the ‘headwear.’

As powerful as a MacBook Air, Notion is certified by the FCC, uses the Linux OS and includes 8 sensors that image the brain at 250 times per second. Notion’s power comes from its quad core 1.2 GHz CPU, 4 GB flash storage, 1 GB RAM plus internet connectivity, creating the first and only all in one device you wear on the back of your head.

Medgadget: Please give us an overview of the potential medical applications of Notion, including wheelchair control, seizure and concussion detection, and mental health monitoring/diagnosis. Would users need a period of training, to learn how to use the device to control a wheelchair, for example?

AJ Keller: The first medical application Notion can enable is real-time concussion detection. This can be extremely helpful in scenarios from the NFL to the military. It works like this: A player or soldier takes regular scans and evaluations with the Notion to build a typical neurological profile. Then after a large impact, a scan would determine if the detrimental activity resulted in an impairment of typical neurological ability.

We also foresee a range of applications around depression treatment which would require a period of training sessions. Notion has the ability to run machine learning locally and store that data in its Secure Enclave, so it’s the safest way to monitor brain changes over time. For example, a doctor may have a patient wear the device for a few minutes a day to assess their state of mood and emotion. The data would be saved and reviewed by the doctor over time to see subtle changes in an emotional state. This would allow a doctor to make changes to medications sooner.

For wheelchair control, users would go through a training period that uses machine learning to learn how the user’s brain activity changes while imagining different body parts such as arms and legs. For example, a user imagines moving their left foot and the left wheel moves forward and imagining the right foot makes the right wheel move forward.

Medgadget: When will this technology be widely available for these applications? Can the system currently be used for such medical applications, or is this in development?

AJ Keller: For the next couple years we’ll be working to improve signal quality and comfort. This will lead to an FDA clearance for EEG monitoring in 2–3 years. We have begun working with a range of application developers who will write and verify efficacy for its use in these medical applications. We plan for a wide general availability in 2023 for the first use in medical applications.

Medgadget: How does the system protect user data? Do people have unique brainwave patterns? Is this a privacy/security concern?

AJ Keller: Privacy is built into every part of Notion. It’s the only device that features authentication and industry-standard data encryption. We have several applications which require the user to provide their usernames and passwords to grant authorization to their brain wave metadata. The device continuously processes raw brain wave data to produce metadata, and the raw brain wave data is deleted once it’s no longer being processed. And raw brain wave data is always being overwritten as new raw brain wave data is collected. Research shows us that people’s reaction to certain images and sounds is unique, so yes, they have unique brain wave patterns.

Medgadget: Where do you see this type of technology going in the future?

AJ Keller: Our technology will help to solve some of the biggest questions humans have about the way they interact with people and the world. Why do we react a certain way when someone triggers us? Why are some days mentally tougher than others? How can we prevent ourselves from wandering into moments of depression through continuous brain activity monitoring? These questions are only really answered when we democratize and provide billions of people with the tools to become masters of their own subconscious. A future with Neurosity is private, secure, and filled with more moments of happiness through the control of emotional and mental states.

Here are a couple videos that highlight the possibilities of the Neurosity. The first shows how a light switch can be controlled and the second how the scroll feature can be used on a tablet by simply thinking about it:

Link: Neurosity homepage…


Medgadget