BrainCallus Gaming Project

Games to Quantify Player Cognition

logo banner

Project Overview


About

The BrainCallus Gaming Project is a volunteer-based effort seeking to improve psychiatric decision-making by leveraging a combination of computational gaming, machine learning and cognitive science.

CONCEPT PAPER



Project Motivation

Cognitive deficits are a hallmark of many psychiatric disorders but they are difficult to evaluate outside the clinic. This is because evaluating a patient’s cognitive state is currently done by the physician, leveraging psychometric inventories that are expensive in both time and money.

The data obtained through traditional psychometric inventories provide temporal snapshots of the patient’s symptoms, with little insight into how symptoms dynamically change across patient context (e.g., time and location). Moreover, these data rely on the patient to consciously report on symptoms, which can be problematic in cases when the patient is unable or unwilling to self-report.

Our Solution

The BrainCallus Project seeks to overcome these limitations by developing games that are useful for evaluating patient symptoms outside the clinic. The proposed computational games are compelling to play, and produce data used to derive metrics of the patient’s perceptual, cognitive and motor performance. The metrics can be implicitly captured, allowing clinicians insight into symptoms the patient is not able to self-report.

Finally, patients of all ages find these games enjoyable, allowing for large amounts of data to be captured. These data are useful for deriving metrics to inform the physician’s clinical decisions, in addition to developing machine learning models that provide diagnosis and treatment recommendations.

Brain Barn


Computational games are designed to produce data that are useful for quantifying player performance. More specifically, they contain gaming modules that are created to quantify each component of the human perception-action cycle.

Brain Barn is a computational game being developed by the BrainCallus Gaming Project. It contains modules that quantify player perception, decision-making and movements. Our Explain the Game series details how gaming modules produce data useful for quantifying player performance.

Take a look at our current videos and check back for more, as we will be posting additional content frequently.

Chameleon Corral


Chameleon Corral is a computational game module in the Brain Barn series that measures low-level visual attention. In this Explain the Game episode, we detail the game, how it measures attention through its design, and the clinical significance of the measures it produces.

Watch this video and see if you're able to keep track of those crazy Chameleons as they try to hide.

Cash Cow


Cash Cow is a computational game module in the Brain Barn series that measures decision-making under risk. In this Explain the Game episode, we detail the game, how it measures risk- and loss-aversion through its design, and the clinical significance of the measures it produces.

Watch this video and see if you're able to make the decisions necessary to produce your very own cash cow.

Duck Dash


Duck Dash Logo

Duck Dash is a computational game module in the Brain Barn series that measures movement performance. In this Explain the Game episode, we detail the game, its ability to measure movement performance under uncertainty, and the clinical significance of the measures it produces.

Watch this video to see if your movements would be good enough to capture those speedy ducks.

[VIDEO COMING SOON]

Project Benefits


Benefits

Continuous Measurements

By leveraging games, it allows the patient to participate in measurement activities while at home. This enables powerful time series methods (e.g., LSTM RNNs) to be used to predict when the patient may realize an increase in negative symptoms between clinical visits. The physician can use this information to provide the patient with extra tools during periods that have a high-probability of a crisis, in addition to attempting to explore reasons for the increase in symptoms across time.

Implicit (Unconscious) Measures

Many times, patients are unwilling or unable to report when they are experiencing symptoms related to their disorder. Computational games allow for implicit measurements of patient symptoms to be obtained, providing the physician with otherwise unavailable metrics that are useful to inform diagnosis or treatment decisions.

Compelling Data Elicitation

The use of games to quantify patient symptoms provides an approachable and enjoyable measurement activity. Games will be developed so patients in all age groups (including children) find data elicitation compelling. This will likely result in larger amounts of high-quality data to be produced, compared to traditional psychometric approaches. The increase in high-quality data will allow for researchers to leverage data-driven approaches, such as AI and machine learning, to improve the diagnosis and treatment of psychiatric disorders across age groups.

Get Involved


Volunteer

Collaborators

We strongly encourage collaboration opportunities from researchers and developers in academia and industry. We are especially keen on securing volunteers who have expertise in one of the following domains:

  • Game Developers (Unreal Engine): volunteers with the ability to convert technical game specs into robust and scalable code that leverages secure cloud data storage.
  • Clinical Psychologists and Psychiatrists: volunteers will act as subject-matter-experts who assure the game is developed in a manner that provides information useful to practicing clinicians.
  • Business Development: volunteers will raise awareness of the project among potential investors.

Contact our team to explore how you may be able to help contribute to this project.

Investors

If you are considering investing in this project, you can support us by contributing to our Patreon page.

Project Leadership


Erik

Erik J. Schlicht, PhD (Founder, Machine Learning Algorithms, and Game Design)

Erik Schlicht conducted research at Harvard University, MIT, Caltech and the University of Minnesota, where he used machine learning and AI to quantify human performance under uncertainty and risk. During that time, he developed computational games that quantify player performance. For example, at MIT Lincoln Laboratory, he was a core member of a team who developed methods utilizing Serious Games to quantify operational decision-making under risk. While a postdocoral researcher between Harvard University and Caltech, Dr. Schlicht created a simplified poker task to quantify decision-making in a competitive (zero-sum) task. This effort resulted in a publication that ranks in the top 5% of all research output, according to metrics by Altmetric.