Eyes are windows to the...brain?

BERKELEY, Calif.—Eye tracking is a common element in annual vision exams and as a way to gauge concussions (and sobriety), but the traditional “follow my finger with your eyes” approach is being significantly upgraded by neurotech/AI company C. Light Technologies. The company is advancing retinal eye-tracking technology combined with machine learning to evaluate and predict neurological health in a new, non-invasive manner.

 

This approach is centered on the use of a Tracking Scanning Laser Ophthalmoscope (TSLO) developed by Dr. Christy Sheehy, founder of C. Light. The TSLO images the retina and can track eye motion on a cellular scale. The process takes less than a minute. Patients place their chin on the chinrest, similar to an eye exam, then fixates on a target for 10 seconds while a video is recorded of their eye. Blinking will not disrupt the process, and no pupil dilation is necessary.

 

“The back of your eye is actually the front of your brain. We use AI paired with eye tracking to create a digital fingerprint of your neurological health, with unprecedented speed and sensitivity,” explained Dr. Zachary Helft, co-founder of C. Light. “Other technologies use the pupil to track eye motion, but our technique images the retina for 120 times more sensitivity than the other tracking systems available today. In other words, C. Light measures eye motion that has been otherwise invisible through existing pupil-tracking technologies.”

 

“The sensitivity of the tracking that we have using the retina is higher than that of the pupil due to resolution with which we can resolve the structures,” Sheehy and her team tell DDNews. “The human pupil is generally 2-4 mm in size, while the structure that we can resolve on the retina with our proprietary device is about 1 micron - or roughly 1/100 the size of a human hair. By tracking microscopic structures as they move, we are able to resolve movement down to the cellular scale. For neurodegenerative diseases, where the loss of every neuron counts, we want to be able to see even the smallest of changes to help inform patient care.”

 

In terms of which neurodegenerative disease C. Light intends to work with first, the company has selected multiple sclerosis. Future disease targets will include amyotrophic lateral sclerosis, Parkinson’s disease and Alzheimer’s disease.

 

According to Sheehy and her colleagues, MS represents a good proving ground for this technology due to the nature of the disease and its effects on patients. “[A]s a patient becomes more clinically disabled with the disease, the more microsaccades (small, involuntary quick movements of the eye during fixation) will occur,” they explain. “This type of subclinical information can serve as granular progression feedback that can help to inform drug decisions and patient care. Our ultimate goal is to be able to provide prognostic, or predictive, feedback as well as early detection to clinicians and drug developers.” C. Light published a paper on this concept recently: “Fixational microsaccades: A quantitative and objective measure of disability in multiple sclerosis,” which appeared in the Multiple Sclerosis Journal.

 

“Utilizing advanced statistical models and machine-learning approaches give us the framework to enable mass data collection and point-of-care analytics,” adds Dr. James Golden, data science lead at C. Light. “Right now, clinicians use a series of physiological and behavioral tests to estimate the severity of disease progression in MS. The same latent neurological factors that determine the outcomes of those tests can also affect eye movement patterns that we record. We use machine-learning techniques to untangle the complex statistical relationship between those eye movements patterns and disease progression, enabling us to make estimates approaching the accuracy of clinicians.”

 

In addition to microsaccades providing an obvious ophthalmic metric that can be measured, Sheehy and her team note that “MS is the only neurodegenerative disease that has viable therapeutics on the market.” As such, once a diagnosis is secured with C. Light’s technology, it can be acted upon, and treatment effectiveness and/or disease progression can be tracked in an ongoing effort with this approach.

 

C. Light has big plans for the future, as well. Sheehy says the company hopes to move into other neurodegenerative diseases in two years, and currently has “active data collection pipelines in Alzheimer’s disease and in concussion recovery.” To support this, she reports that C. Light is undertaking a $1.25-million seed round, with their eyes on two immediate goals: to advance C. Light’s machine-learning algorithms past the current standard of care, and to transition the prototype for this technology into a design for manufacture.

 

The market potential for this technology is extensive. Current statistics estimate that roughly one million individuals have multiple sclerosis, six million are diagnosed with Alzheimer’s disease, and 60,000 are living with Parkinson’s disease—just in the United States. As with any disease, earlier intervention can improve patient outcomes, and a quick, non-invasive monitoring approach such as C. Light’s retinal-tracking technology could be a significant boon in this diagnostic space.