A lay summary written by Ludmila Kucikova, PhD Student, Artificial Intelligence and Computational Neuroscience Group, University of Sheffield
Background
Games and digital tools are becoming very popular in the health sector for diagnosis, prevention, awareness, training or rehabilitation. Some of those tools that monitor our daily routine are already common. Some of us might have watches that monitor our heart rate or physical activity and tell us when to move for a little bit. Or our phones can track how many steps we do in a day and remind us to do some exercise. Some of us might play games on our phones that are calming and help to relieve anxiety.
Games can offer an insight into how people behave and provide researchers with easily accessible data from real life. They are available as smart devices or on computers and can be accessed when it is suitable for patients. Some of such games have the potential to distinguish early changes related to some disorders, for example, dementia.
Why is the study important?
Early identification of Alzheimer’s Disease offers opportunities to treat its symptoms before the disorder damages the brain too much. If the brain is damaged too much, treatments become ineffective, and damages are irreversible. Many ways to identify Alzheimer’s Disease that are used in clinical practices nowadays are not sensitive enough for early changes related to the disorder. Therefore, researchers are constantly trying to find new ways of identifying Alzheimer’s Disease earlier to increase the chances of successful treatment. In this study, the authors are trying to find a way to identify Alzheimer’s Disease by using a game.
What did the authors do and how did they do it?
The authors designed a game called AlzCoGame which aims to monitor a range of functions in older people. This game allows users to train specific kinds of memory that are linked with Alzheimer’s Disease, as well as other functions such as their concentration or attention. All games are based on virtual scenarios that occur in daily lives, for example, a shopping task or a cooking task.
While users engage in a game, their data are collected and used for statistical analysis. This analysis is based on artificial intelligence which can predict whether users might develop Alzheimer’s Disease based on their performance in a game. The researchers test several artificial intelligence methods to see which one is the most accurate.
What are the results?
The results revealed that data collected from the game could identify early stages of Alzheimer’s Disease with 83-92% accuracy depending on what artificial intelligence method was used for the analysis. Based on these encouraging results, the authors recommend this game to be included in clinical trials. This means that more patients would test how useful this tool is and whether it could be used in standard clinical practices.
What do the findings mean going forward for people with the disease?
Using games is increasingly popular among healthcare workers. Such games offer a tool to engage patients, while also gaining data for researchers to learn more about disorders. With games based on real-life scenarios, researchers and clinicians could identify key problems in day-to-day living directly and accurately. Patients in the early stages of the disorder could also use the game to continue engaging in activities that are increasingly difficult to do in a real life like shopping or cooking.