NEUROTECH STATE OF THE SECTOR
PART TWO — COSTS TRENDS AND ENABLING TECHNOLOGIES
INTRODUCTION
The Neurotech category within TransTech has been experiencing rapid growth in recent years. It comprises brain-computer interface (BCI) technology such as EEG and fMRI brain measurement and monitoring, along with a range of platforms and applications aimed at diagnosing, maintaining or treating mental health and wellbeing. Part 2 in this series focuses on cost and technology trends driving the explosive growth in this sector.
COST DRIVERS
Sensors/Electrodes
Sensor technology is becoming cheaper and more powerful. The broader Internet of Things sector in recent years has transformed the smart home, connected mobility, supply chain and other sectors. Bio-sensors, in the wake of the popular FitBit debut, are slowly entering the health market. Some recent examples include Redeye, a hemoglobin sensor, Senseye, an ocular scan for workforce applications such as a “vigilance test”, and Bello — portable device that measures body fat. And several consumer EEG products combine EEG sensors with pulse rate oximeters or heart rate variability measures to enable broader contextual insights.
Consumer EEG headset retail prices seem to scale in a linear fashion with the number of EEG sensors (channels) enabled in each model. This is counter to the situation in the mass market for sensors, i.e. research and medical uses, where one would expect discounts based on large volume orders. This indirectly suggests that the consumer EEG companies are still in very early stages in terms of volume sales. Some market analysts expect the overall market for electrodes to grow as much as 8% per year, which could exert downward pressures on price as well.
The quality and performance of EEG electrodes is being actively researched, according to some recent interviews.
- NextMind in conversation has stated that improved electrode design and selection of materials has led to better signal quality in their devices, and that improvement has resulted in better ability to control external devices.
- OpenBCI has noted that they make their own EEG electrodes, and observe that there is “lots of room to improve on the dry electrode side of things.”
- Colorado State University is engaged in ongoing research and a potential spin-off to improve electrode performance.
The choice of electrode type also varies. Medical EEG technology typically uses gel applied to each electrode, which reduces impedance and improves quality of data. As far back as 2012, a G.Tec-authored publicationnoted “The quality of scalp electroencephalogram (EEG) recordings critically depends on the connection between the amplifier input and the skin surface. Wet electrodes that rely on conductive gel to guarantee low impedance levels (<10 KOhm) remain the gold standard for clinical recordings. However, EEG recording with wet electrodes requires skin abrasion, gel application, impedance optimization and cleaning after recording, all of which are time consuming.”
A more recent research article concluded that “the signal quality, ease of set-up and portability of the dry electrode EEG headset used in our study comply with the needs of clinical applications.”
And some companies try to split the difference. The MindAffect headset uses sponge covered electrodes; when used with water, it provides somewhat better conductivity than dry electrodes, without the slow process of applying gel to dry electrodes during set-up.
Bio-sensing Circuit Boards
Bio-sensing circuit boards are components that enable translation and analysis of brain activity. Here’s an example with some detail from OpenBCI:
The OpenBCI Cyton Board is an Arduino-compatible, 8-channel neural interface with a 32-bit processor. At its core, the OpenBCI Cyton Board implements the PIC32MX250F128B microcontroller, giving it lots of local memory and fast processing speeds. The board comes pre-flashed with the chipKIT™ bootloader, and the latest OpenBCI firmware. Data is sampled at 250Hz on each of the eight channels. The board communicates wirelessly to a computer via the OpenBCI USB dongle using RFDuino radio modules. It can also communicate wirelessly to any mobile device or tablet compatible with Bluetooth Low Energy (BLE).
These boards may constitute as much as half the price of a commercial headset package, and right now seem to follow a linear pricing curve along with the number of electrodes/channels in the device. But the chips on these boards will obviously follow Moore’s law improvements in price/performance over time, and so it’s not clear whether that linear trend is related to price/performance of the chipsets, or in the nature of the connection to the sensors as a system.
Software
Software is the third major component, and software advances are an indirect outcome of Moore’s law. But exponential advances in artificial intelligence and machine learning are clearly providing huge gains in analytical capabilities that are growing at well beyond a linear rate. These advances will show up in the discussion of applications below, but a full analysis is beyond the scope of this article.
PLATFORMS AND ENABLING TECHNOLOGIES
An adjacent category of consumer EEG is less about selling headsets or service but more on providing a platform for developers of commercial products, or providing tools and services in support of product development. There are many software development shops, but not a lot that focus on the emerging Neurotech market.
Neurosity is more of a platform company, seed funded, but is preparing a first product characterized as a “thought-powered wearable computer,” at a $1,099 price point.
MindAffect is a Dutch company that debuted at CES 2020. Their focus is software. MindAffect has developed a Brain Computer Interface that is marketed as part of a development kit. They are seeking collaborations with this kit to further develop use-cases specific to other companies’ needs. The platform is based on OpenBCI (what I saw was a 6-sensor model.) They demonstrated calibration of an individual who could then play a video game by applying focus to different areas of the screen. The company’s headset was using sponge covered electrodes; when used with water, it provides enough conductivity without the slow process of applying gel to dry electrodes during set-up.
Novela Neurotech offers a “smart platform” called Neureka, designed for clinical discovery and validation.
iMotions is a platform that integrates multiple biosensors including Eye Tracking, EEG, and Facial Expression Analysis, helping product and service providers quantify engagement and emotional responses.
Senso Medical is a services company that exhibited at TransTech 2019. They are a software development shop with experience in the BCI space. They will work at no charge with potential partners in developing RFP responses in exchange for a contract if the partners win a grant or contract.
neuromore is a Neurotech research and services company. They have developed a set of tools that are used in university and government research. Neuromore products are open source, and a free license is available to non-profits.
Neurocort is an interesting market entry from France. The company is focused on detecting and cataloging neural signatures, sometimes called neuromarkers, for early diagnosis of neurodegenerative disease.
EEGworks provides a toolkit of “All Things EEG” and is targeting the neurofeedback and neuromodulation markets.
Intheon is a data analysis and app development platform and middleware. It incorporates machine learning to enable “neuroscale insights.”
With tongue-in-cheek (partially), I predict that many of these platform and data companies will soon be owned by Apple, Microsoft or Google.
A NOTE ABOUT ADJACENT SECTORS AND ACTIVITY
Non-EEG Neuroscience — There is interesting work going on that leverages non-EEG devices such as HRV (for example Immersion,) or computer vision (for example Okaya,) to derive neuroscience insights. Mindstrong is a well-funded ($160 M) entrant in this category. The service allows remote monitoring of digital bio markers in order to, for example, prevent unneeded emergency room admissions. The software platform can record smart phone usage — typing, swiping, scrolling — to measure things stress or mental health symptoms.
Neurostim Tech — Coverage of the BCI sector typically incorporates not only EEG/fMRI but also the adjacent neuromodulation or Neurostim sector, but that will be the focus mostly of a future article. But it is worth briefly commenting, if only to draw the distinction. Elon Musk’s Neuralink especially tends to draw press attention to the broader BCI sector, even though Neuralink represents not only the more invasive Neurostim category, but a fairly unique technology direction within that sector.
The term invasive is used in medical parlance to indicate surgical incisions through the skin and skull and into the brain. Some Neurostim companies call their companies “non-invasive” because there are no incisions, even though sending electromagnetic signals into the brain is certainly a physical intervention.
As such, companies such as Medtronic, who develop clinical Neurostim products, are generally more deeply engaged with regulatory authorities and walk a careful line. Indeed, I noticed in recent years the presence of a booth at the Consumer Electronics Show of the North American Neuromodulation Society, representing the need to educate audiences on the legitimate uses of Neurostim. The NANS organization president also led a CES Digital Health panel on neuromodulation’s contributions to fight the opioid crisis “by addressing the root cause of pain.” These are the types of conditions in which large medical and pharmaceutical companies see a major market opportunity, and Neurostim solutions seem to be slowly achieving legitimacy.
From a TransTech perspective, the broader availability of cheaper interventions also is worth support and tracking. A follow-up article will allow a fuller and more nuanced treatment of the sector.
