Validating AI-guided, rapid vision diagnostics in a remote setting—with Spark Fund Awardee Professor Peter Bex
Age-related macular degeneration (AMD) is a leading cause of visual impairment worldwide that affects around 15 million people in the US alone. AMD causes blind spots that progressively diminish the central visual field that allows people to see details of the world, such as reading text and recognizing faces. Current treatments for AMD aim to slow or stop the progression of the disease but there is currently no cure. Therefore, early detection and disease progression monitoring during treatment are critical to patient outcomes.
However, clinical diagnostics for visual impairment currently suffer a number of problems. For example, a patient must attend a clinic to see a trained clinician. Requiring a clinic visit is an inherent barrier due to socio-economic, geographical, or health-related reasons. Furthermore, the current tests are cumbersome, time-consuming, and unable to detect subtle changes in vision. They also do not measure important aspects of vision, such as color discrimination, motion perception, and face perception.
That’s why Professor Peter Bex and his team created PerZeption. PerZeption offers computer-based vision tests that are very fast and do not require a clinician for administration. These tests are also able to investigate a wide range of important aspects of daily vision, such as the ability to see colors, detect fine details, or recognize faces. These tests can be used in a clinic or remotely.
This work has earned Bex and his team selection as one of the Fall 2022 Spark Fund awardees.
Revolutionizing Remote Vision Tests with PerZeption
This project started when Bex and his team were looking for a new way to measure “multistable perception,” or the switching of perception when each eye sees a different image. This project eventually became the Indicate-Follow-Replay-Me or InFoRM platform.
Like much of the world, the team was forced to rethink their research in spring 2020 when they could no longer collect patient data in-person due to the COVID-19 pandemic. At the time, remote vision testing was insufficient for their purposes.
So, they pivoted to develop a prototype for a self-administered, rapid vision test for measuring one aspect of vision with a high degree of sensitivity. This eventually became the Foraging-Interactive-D-Prime (FInD) platform. Quickly, the team realized that the FInD technology was suitable for measuring many aspects of vision, but it was still inefficient for measuring factors, such as visual acuity.
So, the team adapted again. In their efforts to solve this testing problem, they developed the AIM platform, which combines the benefits of the
FInD platform with the principle of Angular Indication Measurement.
Now, their research investigates visual performance before and during an injection-based treatment that aims to stop or slow down the progression of AMD. To overcome the limitations of current vision tests, the team aims to find out which aspects of vision—such as the ability to see contrast, colors, and fine details—is most affected by vision loss and how those visual functions respond to treatment. Some of the vision tests also measure how well the visual system can discern visual signals in a visually noisy environment. These tests would serve as early indicators of visual impairment before actual vision loss occurs.
As the team gathers more detailed data, they will use modern machine learning technologies to increase the screening accuracy and build novel prediction models for specific patient groups in an effort to detect early signs of vision disorder before actual vision loss occurs.
Their work will improve understanding of the relationship between treatment and performance in AMD, but it will also help scientists better understand fundamental visual perception.
“I predict that the way vision screenings are done will fundamentally change in the coming decade,” says Bex. “I imagine a future in which more and better information about each patient is collected during a clinic visit and where remote testing will be used routinely as early detection screening or as follow-up after an intervention such as an operation. Moreover, I believe we can bring change to how vision care is perceived more generally and normalize routine at-home screenings as something that is engaging and even a fun thing to do,” says Bex.
Commercialization with the Spark Fund
Through commercialization, the ultimate goal of PerZeption is to prevent vision loss and enhance vision diagnostics by breaking down barriers to access while improving test efficiency and sensitivity.
In the near term, the team will bring their vision tests to the US market while doing the necessary scientific work around their technologies. Then, the team plans to expand its efforts abroad. They have already established a number of collaborations and contacts with researchers and clinicians in the UK, Germany, and New Zealand.
“With the Spark Fund, we will validate a battery of visual function tests both within a clinical setting and remotely at the homes of patients with AMD,” says Bex. “The CRI provided great opportunities to help us kick off fundraising and spread the word about our company and its exciting technologies. Therefore, we are grateful for all the help we have received from the CRI on our path to change the way we test visual perception.”
Learn more about Professor Bex’s research and the five other 2022 Fall Spark Award grantees here.