We have been commercializing predictive attention software with the help of angel investors, commercial revenue, and small business innovation research (SBIR) awards. Our R&D efforts to improve health and licensed merchandising solution are showcased below the list of SBIR awards.

2020 — Visuotactile tests of mental domains
2016 — Assessing psychosis-related deficits based on gaze behavior
2014 — Cognitive testing based on visual paradigms
2010 — Automated Display Optimization Based on Attention Predictions


We aim to improve intervention decisions and rehabilitation efforts for patients with neurological and psychiatric conditions. Scientific studies indicate that associated symptoms arise from mental dysfunction across cognitive, emotional, and/or sensorimotor domains. However, clinical practice "algorithms" tend to manage symptoms through trial-and-error of available interventions without utilizing objective tests of mental dysfunction. In practice, patients often get mis-treated or under-treated, resulting in poor outcomes and high costs. We have been developing and validating behavioral test batteries to prioritize intervention targets by the type and severity of mental dysfunction in tested individuals. If better intervention decisions are made after integrating EyePredict reports with independent clinical information, such as patient history, the expected benefits are better patient outcomes and lower costs.

Value Proposition: Improve Decision Making

Health Value Proposition

EyePredict test batteries quantify attention deployment patterns that reflect mental dysfunction. This approach is motivated by the observation that paying attention to the right thing at the right time is critical for learning and survival in complex environments. Attention deployment is a precious resource tightly controlled by interactions between emotional, cognitive, and sensorimotor functions. Impairments in these mental functions can be quantified by challenging attention deployment and measuring associated patterns using precise yet noninvasive tools, such as eye trackers. See below representative eyegaze patterns from our case-control studies.

Our SBIR-reviewed case-control studies found 96% sensitivity [95% CI 83-100] and 87% specificity [95% CI 67-100], establishing feasibility and justifying larger R&D investments.


We leveraged predictive attention models to provide merchandising solutions. First, we established the feasibility of predicting what visual elements are likely to get noticed first by comparing our model predictions to visual engagement as measured by eye tracking and mouse clicks. This proof-of-concept led us to commercialize PlaceGage—a software-as-a-service solution that predicts visual engagement. To use it, customers upload an image, mark items of interest, and receive reports with PlaceGage scores:

Item Score Meaning
1. Face 100 Very likely to get noticed
2. Van 66 Likely to get noticed
3. Plate 49 May or may not get noticed
4. Magnum 40 Likely not to get noticed
5. Road 14 Very likely not to get noticed

PlaceGage is useful for evaluating the likelihood of “mental clicks” with visual items: products on store shelves or in catalogs, logos in print or outdoor ads, virtual goods, etc. Customers included a Fortune 500 packaging company that offered to license the underlying technology, a multinational market research firm, a multinational produce seller, a national sporting goods chain, a consumer research consultancy, and others.

Market demand for more actionable information than mere predictions led us to develop EPflow: a software solution that enhances viewer engagement with relevant items by re-positioning them on webpages based on attention predictions. Item relevance is defined by web merchants according to what consumers wish to buy (ex: as indicated by purchase history) and other considerations (ex: profit margin). EPflow performs directed search through a large number of candidate layouts, using PlaceGage as an internal engine to assess each layout and maximize the likelihood that the most relevant items will get noticed.

A/B tests found strong lifts in key performance indicators (ex: 175% higher click-through rate), and in 2016 we licensed our software to an international web merchant that has been in business for >20 years. More information is provided in the videos below and on VentureBeat.

Bonus Videos

DEMO Presentation

DEMO Interview

Where The Eyes Go


Software Developers — we want more of you! Join an inter-disciplinary team whose mission is to improve intervention decisions for people with emotional and cognitive symptoms. Your day-to-day activities will focus on developing a novel test battery based on visual stimuli and touch responses. We are open to candidates with diverse backgrounds, from graduating students to seasoned professionals.

Ideal Character

  • Open-minded
  • Self-motivated
  • Team-friendly
  • Responsible

Ideal Experience

  • Software startups
  • Unity/Godot, tablet apps
  • Python, PostgreSQL
  • Eye tracking, machine learning


We focus on deliverables and timeframes, so your schedule and work environment would be flexible.

  • Location: anywhere USA
  • Scope: part-time => full-time
  • Compensation: cash and equity


Email us two paragraphs describing your motivation and qualifications along with a one-page resume.