The Goalkeeper Game, a tool for massive data collection and experiments

NeuroMat's Goalkeeper Game has evolved as two concomitant research, development directions. Firstly, the game, that had a first version launched in early 2016, is being developed as a tool for massive data collection. Secondly, the game remains a resource for efficient diagnosis and allows for changes in settings for experiments. Challenges in these two directions have been advanced by a multidisciplinary team within NeuroMat's innovation area.

The reference website for the game is Versions for Android and iOS will be launched in the coming weeks, at Google Play and App Store. The game was created by NeuroMat director Antonio Galves, initially as a dissemination strategy and has evolved as a scientific project of the team.

In both development directions, the goal of the game is to collect data from users, specifically the strategy that was used, the response time and the number of successes. The strategy is described as a context tree, an extension of Markov chain models, in which a random variable may be predicted in a given sequence based on the minimum necessary previous actions. The time at which the game was played will also be recorded. The development of the game has been approved in research committees, so data storage strictly complies to ethical research recommendations.

The Goalkeeper Game is at the same time a computer game and an ambitious transdisciplinary research project. In this multi-level computer game, an individual takes up the role of a goalkeeper that faces a penalty taker who is ready to shoot. As the game evolves, the expectation is that the individual will be able to make sense of the strategy of the penalty taker and have a high rate of success in thwarting scores. The scientific questions that are associated to this project revolve around the learning process and the decision-making model of the goalkeeper.

Widespread data collection

A project is undergoing to launch a Goalkeeper Game applet and have the game available worldwide. The goal here is to collect massive data. This goal is associated to devising and designing game features to sustain widespread interest, including visual and sound effects, a system of specific rewards and rounds and continuing updates of rounds and trees. A new frontpage for the game will also be developed, with a new layout for game rules and settings. University of São Paulo students who specialize in game development as well as a professional artist are involved in this work, being led by Marco Dimas Gubitoso, a NeuroMat associate investigator and professor at IME-USP.

"People will likely be attracted by the idea that they can play a game and contribute to a scientific enterprise. At the scale at which we are working this will be a premiere in Brazil. It is, however, hard to guess how many people we will attract to the game, as this is strongly dependent on dissemination," said Gubitoso.

The research agenda that is associated to the massive use of the game relates to making sense of what are characteristics of trees that make them easier or harder. Preliminary results show that entropy and height are important but not sufficient features. This question is associated to modeling learning strategies. The expectation is that the dataset that results from the game use is large enough to allow for robust analysis of these strategies.

Controlled version

The second development direction involves the creation of a controlled version of the game for experiments. In this case, the expectation is that there are as few features as possible, as these features could be distractive. Experiments are being conceived --and had had pilot uses-- in the context of NeuroMat's initiatives on the brachial plexus injury and Parkinson's disease, respectively AMPARO and ABRAÇO.

"A relevant feature for experiments is that researchers are able to establish their own features, thus allowing for controlling variables," said Gubitoso.

As reported previously, a clinical research associated to the Goalkeeper's Game looks at how to create less time-extensive application and assessment for the diagnosis of Parkinson's Disease.

This piece is part of NeuroMat's Newsletter #53. Read more here

Featuring this week:

Stay informed on our latest news!

Previous issues

Podcast A Matemática do Cérebro
Podcast A Matemática do Cérebro
NeuroMat Brachial Plexus Injury Initiative
Logo of the NeuroMat Brachial Plexus Injury Initiative
Neuroscience Experiments System
Logo of the Neuroscience Experiments System
NeuroMat Parkinson Network
Logo of the NeuroMat Parkinson Network
NeuroMat's scientific-dissemination blog
Logo of the NeuroMat's scientific-dissemination blog