Time to Reassess Cognitive Testing: Using Simulated Environments to Assess Real-world Function
Intellectual ability, such as thinking, reasoning, and remembering, is tricky to measure. This may be the reason why no major advancement has taken place in this field for decades, until recently.
I am a neuroscientist, and I spent the past four and a half years researching whether virtual reality (VR) can be used to assess cognitive functions at the Institute of Psychiatry, Psychology & Neuroscience. My PhD focused on developing and validating a novel VR grocery shopping task — VStore. I am passionate about VR because of its potential for the treatment and assessment of mental health conditions.
To put my research into context, I will briefly explain what cognition is, why and how it is measured, and finally, I will share what I have learned from my PhD.
What is cognition?
The word “cognition” is derived from the Latin word cognoscere, meaning “to get to know”. In neuroscience, it refers to the mental processes involved in the acquisition, storage, manipulation, and retrieval of information. For example, the speed with which information is processed, attention, learning and memory, reasoning and problem solving, planning, language, and motor skills.
Why and how cognition is assessed?
We measure cognition because it is essential for our day-to-day living. Your cognition is at work throughout your waking hours: when you plan your day, learn new information or skills in your job or studies, travel from one location to another, shop, or socialize.
Impaired cognition has a dramatic impact on quality of life, as people may no longer be able to carry out their daily activities. This is often the case in traumatic brain injury, neuropsychiatric conditions, such as psychosis (a condition when you perceive or interpret reality in a very different way from others and you may experience hallucinations or delusions), depression, dementia, and even healthy ageing. For this reason, the accurate measurement of cognitive performance is crucial, so we can detect impairment, track changes, and inform treatment decisions.
Traditionally, cognition is assessed with pen-and-paper or computerized test batteries that include several tasks — each measuring specific intellectual functions. One example is the Stroop test, which is thought to engage processing speed, selective attention, and cognitive control. When you complete this assessment, you have to name the colour of the word, instead of the word itself. Sometimes the word matches the colour, other times it does not. People can name the colours of words that match the word faster and more accurately, than the ones that do not.
This approach to cognitive testing has been in place for a century. However, it is not without limitations. While cognition and successful everyday functioning are closely linked, the measures we get from cognitive tests are poor, and, arguably, how well people manage their daily tasks is more important and informative than some abstract test results.
VR cognitive testing
VR technology now enables scientists to mirror the challenges of everyday life in a simulated (and controlled!) environment, so real-world or functional cognition (as I like to call it) can be measured accurately.
Despite VR being around for a while, we know very little about how people interact with these environments — could the sort of people we want to test complete these assessments, do they cause side effects, do they even measure what we think they are measuring?
As part of my PhD, I set out to answer a number of these questions using VStore.
First, I wanted to know if doing shopping in VR was something people were able to do easily and whether it caused any unwanted effects associated with VR (i.e. nausea, eyestrain, or headache). I also wanted to test how VStore relates to traditional cognitive assessments and if it can be used to identify people with poorer cognitive performance.
We recruited three groups of people and had them complete VStore alongside other measures. The first study included all participants — two groups of healthy volunteers from across the age spectrum (20–79) and a group of people with psychosis. In this study, we found that 99.95% of participants were able to complete VStore, and the task did not cause any unwanted side effects. Hence, we concluded that VStore was safe and viable to use.
At this early stage of research, we also found evidence that ageing participants completed the task less efficiently than younger participants. Indeed, volunteers aged from 20 to 29 took ten minutes, while volunteers aged from 70 to 79 took fifteen minutes on average to complete VStore.
Our prediction that VStore can be used to spot people with declining cognition was confirmed in the second study, where we successfully sorted participants into young (20–30) and ageing (65–79) groups based on their VStore performance alone. We were also able to predict the participants’ exact age fairly accurately using the VR task.
Additionally, we were pleased to find that VStore engages the same cognitive functions as a commonly used computerized cognitive battery. While grocery shopping in VStore, participants had to pay attention, navigate the environment, and learn and process information from it as quickly as possible, use their verbal and visual memory, and plan ahead, just like during standard cognitive testing.
You may wonder at this point if VR assessments truly measure cognitive performance or if they may measure something else, such as technological ability. The answer is that they probably measure both. Nonetheless, we found that, while age was certainly important, the difference in VStore performance between younger and older participants was not simply because younger people were more used to high-tech devices. As a matter of fact, most young participants never used VR before taking part in the study.
We found similar results in people with psychosis, who took sixteen minutes to complete VStore, while healthy volunteers of the same age and gender completed the task in ten minutes. This helped us differentiate participants with psychosis from participants who did not have psychosis in a similar way we did with age groups in the second study.
Why does this matter?
The goal of treatment for cognitive impairment is not increasing a score on the cognitive test. Instead, patients, their friends and family, and clinicians hope to improve or even restore people’s everyday functioning so that they can lead fulfilling lives. And, ironically, the simulated VR world can help us assess real-world cognitive functions with a high degree of accuracy and control.
VR is an immensely flexible environment where you can put people in different situations. Just with a simple grocery shopping scenario, you can test people in small shops, and larger shops, and control other factors such as the number of people around, the length of the shopping list, or having a shopping list in the first place. These small changes alter what aspects of cognition are measured with a few clicks.
The adaptable nature of VR means that it can be practically used for anything, from surgical training through physical therapy to the treatment of mental health conditions. What an exciting future!