1 September 2016
How new technologies are expanding the reach of cognitive assessment
Dr Jenny Barnett, Chief Scientific Officer for Cambridge Cognition reviews how the rise of touchscreen technology has enabled the expansion of cognitive assessment.
The first digital cognitive assessments
In the 1980s, University of Cambridge researchers Barbara Sahakian and Trevor Robbins and their teams started to designed cognitive tests where the participants’ responses could be measured directly through touchscreen computers. These first CANTAB tests were ahead of their time: touchscreen computers were large, expensive, and consequently, a relatively rare sight in psychology labs and clinics. Fast forward thirty years and touchscreens are all around us: ubiquitous in our personal computers and their descendants, the tablets and smartphones, but also on our TVs, fitness trackers, banking machines and supermarket checkouts.
CANTAB tests were originally designed for touchscreen administration so that cognition could be accurately assessed even among patients with impairments of fine motor control, such as those seen in Parkinson’s and other neurological conditions. The practicalities of touchscreen use were such that directly equivalent rodent and non-human primate assessments could also be developed. And touchscreen computers offered a step change in the measurement of speed and accuracy over and above that typically measured in traditional neuropsychological tests. 1980s touchscreen technology thus considerably expanded the range of participants for whom cognition could be rigorously assessed. The current decade’s explosion of touchscreen availability has similarly expanded the range of environments in which we can measure cognition.
The rise of the iPad
Our first insight into this came with the launch of the iPad. Although touchscreen tablets had been available for quite some time, the iPad was so well designed that people were essentially able to find their way around using it straight out of the box. With this user-friendly hardware, and automated voiceovers giving test instructions and feedback, we found for the first time that we could extend the use of CANTAB tests to settings such as primary care, where people with little to no neuropsychology training could now reliably administer CANTAB tasks. Second, by freeing the administration of cognitive tests from a one-to-one supervisory requirement, we could make cognitive testing more efficient and scalable in clinical trial settings, especially those where there is an increasing need to demonstrate the safety of compounds designed to treat non-CNS disorders.
Following the success of successive models of iPad, the reliability and price range of consumer grade tablets has proliferated. Each model differs in screen dimensions, touchscreen robustness and responsiveness, and operating system versions – all things that matter for accurate cognitive testing. As we continue to validate specific models that meet our stringent requirements for recommended hardware for clinical trials and academic research, we also experimented in thought and deed with the effects of loosening these parameters.
Harnessing the power of the web
Using internet-based services such as Amazon’s Mechanical Turk, which pays participants money to complete tasks on their home computers, we have experimented with the extent to which variations in people’s hardware, operating system, and home environment affects cognitive testing data. We conclude that while there are accuracy advantages to assessing cognition in a strictly-controlled environment, sometimes a much larger dataset collected in noisy conditions might be preferable, for example in the speed or ease of access to participants or the lower cost of remote testing requiring no travel to site. Similarly, for some tests the precision of hardware, and availability (or not) of a touchscreen seems more important that for others. You can read more on our web versus lab assessment experiments from Dr Francesca Cormack’s recent poster from the CINP World Congress for Neuropsychopharmacology.
This understanding has allowed us to develop CANTAB Recruit, a product which allows at-home pre-screening of cognitive function as part of a recruitment process for clinical trials. In the first instance, Recruit is being used to determine the presence or absence of a likely memory impairment in older adults interested in taking part in dementia treatment or prevention studies. This allows us to help clinical trials teams to better stratify, recruit or target potential participants before bringing them in for face-to-face site visits, reducing the number of at-site screen failures by increasing the probability that site-attending participants will meet both clinical and biomarker inclusion criteria.
Where will wearables help?
What of mobile and wearable devices? These forms of touchscreen have two distinct advantages over tablets and other computers. They are always with us, meaning that through them, we can collect data in a much more ecologically-valid way, as participants walk around in their daily life. This should allow us to collect data less intrusively but more frequently, allowing a more accurate picture of cognitive function or change over time to be built up with minimal hassle to participants. Second, these devices have two-way connectivity and a plethora of other sensors built in. This allows the intriguing possibility to prompt discreet cognitive assessments not only according to a fixed schedule but also in response to the participants’ individual behaviour: the place they are in, their physical activity or sleep pattern, or their current physiological state. This intelligent prompting and moment-by-moment assessment is the basis for Cognition Kit, a joint venture which we have launched to deliver rigorous cognitive assessment on devices including the Microsoft Band, Apple Watch, and both Apple and Android smartphone systems. Cognition Kit allows more rapid and contextual collection of participant data in clinical trials and, post-approval, provides a way for individual patients or doctors to monitor and manage their individual cognitive function, alongside other aspects of their health.
While tablets and smartphones are now near-universal, wearable devices are still relatively immature, with constant upgrades and improvements coming through in battery life, processing power and connectivity. Over the next year or two we expect these to become as usable and mainstream as smartphones and interest to shift instead to the implementation of touchscreens in ever more form factors, including perhaps your skin itself. We have come a long way in 30 years of touchscreen testing and look forward to the challenges and opportunities that such developments will bring.