Digital technologies like the DCTclock are revolutionizing the way clinicians across all specialties can quickly assess for cognitive decline, an issue in today's aging society.
There are more than 6 million Americans are living with Alzheimer disease (AD) currently, and this number is projected to increase to nearly 13 million by 2050. Additionally, AD and other dementias are expected to cost the nation $345 billion in 2023, with costs potentially rising to nearly $1 trillion in 2050.1 These estimates raise the importance of early detection and diagnosis of cognitive impairment which can ultimately lead to earlier, appropriate planning for the future.
Above all, it is critical to the development and deployment of novel therapeutic interventions as more continue to be approved for in-clinic use. Clinical detection of cognitive decline has been largely dependent on observations from patients or their family, or impaired performance on standard cognitive screening tasks such as the Mini-Mental State Examination (MMSE) or the Montreal Cognitive Assessment (MoCA).
Most patients presenting with cognitive decline are first seen by primary care physicians, then by neurologists. In today’s age, with the rising number of AD cases and the lack of neurologists to treat, clinicians have begun to lean on the digital assessments at their disposal, one of them being the DCTclock. The DCTclock is part of Linus Health’s digital cognitive assessment platform, which has ties back to the original clock drawing test developed in the early 1900s.2
The Clock-Drawing Test
The clock-drawing test (CDT) is a simple, but effective tool that involves drawing a clock on a piece of paper with numbers, hour and second hands, representing a specific time. When asked to complete it, patients must possess adequate visuoconstructive and visuospatial skills, auditory comprehension, frustration tolerance, and executive functioning skills—including organization, planning, abstract reasoning, and parallel processing.3
Originally developed to evaluate solders afflicted with apraxia, the CDT became increasingly used in clinical practice throughout the early 20th century as a way to evaluate cognitive function. In the mid 1980s, Edith Kaplan, a clinical researcher at the time, introduced the Boston Process Approach (BPA), a process-based approach to neuropsychological testing later incorporated by Linus that changed the way the CDT was used.2
"There’s a large group of individuals, both in the scientific community and in the lay community, that associate changes in memory or decline as the earliest alteration in skills or cognition that might be related to an eventual diagnosis of dementia or AD; however, that may, in fact, not be the case," David Libon, PhD, told NeurologyLive "If you look at any of the diagnostic criteria for dementia, the first criteria that needs to be met is a decline in complex everyday activities, such as your ability to manage financial matters, shop, drive, etc."
Libon, a professor of geriatrics, gerontology, and psychology at Rowan University, added, “The scientific literature clearly shows that executive functioning, rather than memory, is related to the emergence and the decline of these executive abilities. Clock drawing does assess a wide number of neurocognitive operations and is a good assessment for a number of executive abilities."
The BPA essentially suggests that an analysis of the strategy or process by which tasks and neuropsychological tasks are completed, along with the errors made during test completion, conveys significant information regarding underlying brain cognition and is as important as the overall summary scores. “What is the process by which the patient arrived at that output? Those common errors, [those] deviations from the normal path—how informative are they at understanding the brain function of the individual?” David Bates, PhD, chief executive officer and cofounder of Linus Health, said.
The MoCA and the Mini-Cog, developed in 1995 and 2000, respectively, stand as 2 of the more prominent paper-based cognitive tests used in the field; however, although each of these incorporated the CDT, they do not include the BPA for analysis and scoring. It would take digital, artificial intelligence (AI)-powered innovation to offer a clock drawing test with the BPA. Building on Kaplan’s advances, collaborators at Massachusetts Institute of Technology created a digital version of the CDT around 2005 – later known as DCTclock™ – and, in 2015, that group formed Digital Cognition Technology, which was later acquired by Linus Health in 2020.
Following the acquisition, Linus Health’s medical team, led by Cofounder and Chief Medical Officer Alvaro Pascual-Leone, MD, PhD, were part of an effort to enhance the DCTclock with a more robust scoring algorithm, as well as a shift to an iPad-based platform. Later that year, in 2021, TIME awarded DCTclock a place on its Best Innovations of 2021 list, recognizing its ability to capture cognitive changes unseen to the naked eye.
"In memory clinics, the availability of skilled professionals who know the literature and the clinical science about dementia is very small," Libon, who is also an advisor to Linus Health, said. "These screenings that need to take place should be done in primary care. One of the challenges in the future is to work with our primary care colleagues, physicians in the primary care arena, and provide instruments which are efficient, validated, and easy to use, that can provide a real, adequate screening of cognition for folks coming in for their routine care."
In 2021, research on DCTclock using participants from the Harvard Aging Brain Study was published, with findings showing that the digital clock-drawing test was effective in identifying the early changes in AD pathology in cognitively normal individuals. The study included 300 participants, 264 of whom were cognitively normal (CN) and 36 with a diagnosis of mild cognitive impairment (MCI) or early AD dementia. Among CN participants with biomarkers, the DCTclock summary score and spatial reasoning subscores were associated with greater amyloid and tau burden and showed better discrimination (Cohen d = 0.76) between amyloid-ß-positive groups than the Preclinical Alzheimer Cognitive Composite (PACC) assessment.4
The modernized DCTclock is among other digital assessments designed to measure an individual’s performance on a variety of cognitive or functional tasks. Other testing tools, such as the Automated Neuropsychological Assessment Metrics (ANAM), Cambridge Neuropsychological Test Automated Battery (CANTAB Mobile), CognICA, Coognigram, and Cognivue, are also FDA-listed. In addition, the agency has also cleared Cognision, a medical device headset built with electrodes that are affixed to the scalp to measure electrical activity in the brain responsible for cognitive function.5
"The tools that specialists like neuropsychologists and neurologists would use are too lengthy, too vocal, [and] too cumbersome, ultimately, for a primary care physician to use," Pascual-Leone, said. "What they need is a tool that is sensitive, specific enough, quick, and can give you at the least, an appropriate triage."
Pascual-Leone also serves as a professor of neurology and medical director of the Wolk Center for Memory Health at Hebrew SeniorLife at Harvard Medical School. He believes that some of the other more commonly used tests like the Mini-Cog and MMSE have their role but are not well-suited to be used in a very time-constrained setting where speed and sensitivity of the assessment matters greatly.
"Broadly used, they all have their utility," he said. "They were developed for different reasons. The Mini-Cog was developed to characterize a population level assessment, not an individual assessment. The MoCA and MMSE are broadly used and are a nice snapshot of individual cognitive function, but they are not sensitive to detect early problems. It depends on a lot of the specific points that an individual loses."
Measuring Cognition Completely
In comparison with other traditional paper-based measures, the digital technologies being introduced require minimal training, are standardized and objective, and incorporate multilanguage support. Additionally, digital strategies include automated scoring and interpretation, fast online results, results tracking, and offer greater or comparable sensitivity in early stages of cognitive impairment relative to traditional methods.6
The MMSE, a commonly used cognitive test developed nearly 50 years ago, consists of 11 questions or tasks that require around 7 to 8 minutes to administer. Research comparing traditional cognitive assessments has shown that MMSE scores have a more pronounced ceiling effect than MoCA for distinguishing healthy controls and cases of MCI.7 The Mini-Cog, a shorter time-constrained test, has been shown to be more sensitive than MMSE in detecting MCI; however, it faces the same general limitations and subjectivity concerns of other paper-based tests and, similarly, only evaluates final test outputs.
In late 2021, Linus Health continued to build on the DCTclock, adding its Digital Clock and Recall (DCR™) solution to its arsenal of tools.2 By assessing verbal memory through immediate and delayed word recall tasks, clinicians using the solution are able to determine different subtypes of cognitive impairment, which could be critical in understanding the long-term trajectory of an individual. Amnestic MCI (aMCI), generally considered the prodromal stage of AD, has different and heterogenous patterns of progression, compared with nonamnestic forms of MCI.
In fact, a recently published retrospective study on patients who decline from aMCI showed a higher frequency of both visual and verbal memory dysfunction, late-stage aMCI, and multiple domain dysfunction, in comparison with those on a more stable trajectory.8 In addition, compared with the stable group, the slow decliners showed cortical thinning predominately in bilateral parietotemporal areas, whereas the fast decliners showed cortical thinning predominantly in bilateral frontotemporal areas.
Using cutting-edge AI, the platform developed by Linus Health combines objective analysis of cognitive performance through the DCR, with qualitative metrics from the company’s Life and Health Questionnaire (LHQ). Altogether, this provides a full snapshot of an individual, allowing providers a new level of both analysis and guidance.
"If you test in a certain way, you want actionable recommendations that may imply seeing further specialists or studying certain treatments," Pascual-Leone said. "Even when a patient tests well, they may require and desire an action plan that tells them what they should be doing to maintain that state of well-being. We’re committed to providing not just the results of the test, but the action of the recommendation for clinicians to guide their patients to sustain brain health and to make the choices they want to do."
The LHQ is a digital patient survey that identifies unique modifiable risk factors for dementia. Through patient responses, the platform assembles patient-friendly, personalized action plans aimed to promote brain health in daily life.
Above all, this type of system allows for easier sharing of data, an ability to assess day-to-day functioning and change strategies where needed, and detect particularly early, sensitive markers that summarize a patient’s health and the potential risks they are encountering. The ability to quantify a patient’s status and trajectory has become a more common theme in the AD field and across neurology. With the explosion of AI, some feel using these systems to "score" an individual’s condition will ultimately lead to treatment optimization and personalized medicine.
"When a lot of people realize, 'Hey, I could get dementia, I’m at risk,' they will do whatever their doctor tells them to do so they can avoid that," Bates said. "We’re putting all this thinking into our products to serve these providers in health systems and in private clinics."
Taking on the Challenges
In the new era of digital technologies, the emergence of these assessments coincides with the shift towards big data and the introduction of AI. Integrating large troughs of data from multiomics studies can provide the potential to explore the pathophysiological mechanisms of the entire biological continuum of AD.9 As it continues to grow, AI offers a wide variety of methods to analyze large and complex data in order to improve knowledge in the AD field.
Thus far, a majority of machine learning models for classifying AD from normal cognition are trained with neuroimaging data, which have the advantage of high accuracy, but limitations exist associated with their high cost and lack of diffusion in nonspecialized centers. Researchers have hypothesized that AI could potentially integrate data obtained through new technologies, such as devices designed for the evaluation of language and verbal fluency or executive functions in healthy or mildly impaired individuals.
In early 2022, data from the Framingham Heart Study using the DCTclock showed associations between the test and brain atrophy, an important physical marker of cognitive impairment. The study featured 1656 participants, with 23 diagnosed with MCI. Findings showed that all 18 DCTclock composite scores were associated with total cerebral brain volume, but only 1 composite score was associated with cerebral gray matter volume. The classification model for differentiating MCI and cognitively normal participants, which incorporated age, sex, education, MRI measures, and DCTclock composite scores, showed an area under the curve of 0.897.10
With the amount of funding granted in recent years, industry leaders in the AD field are under significant pressure to create and conduct clinical trials that accurately assess the efficacy of high potential agents. Despite the efforts, little regulatory success has been shown in the past few decades, with several still questioning whether certain strategies are worth it. In fact, according to a 2021 report, an estimated $42.5 billion in cumulative private expenditures has been sunk into clinical stage AD research and development since 1995. The greatest costs incurred during phase 3, totaling $24.1 billion.11
Improvements in the efficiency of these trials starts with patient selection and ensuring that it is the most optimal cohort of individuals to accurately assess an agent’s efficacy. Part of the issue stems from screening, as some methods have become out-of-date.
"There’s a real cost to bringing people in. You’re disturbing their life, there’s logistical issues, assessing them for hours, you’re running PET MRI, which can be expensive," Bates said. "If they know more quickly, very easily with low burden, who is likely going to be someone who has Alzheimer’s disease, that saves them a lot of money, lowers the costs of trials, speeds them up, and makes everything more efficient."
Like others in the field, Bates believes there’s a “renaissance” happening in neurology. He referred to not just the recent monoclonal approvals for AD, but the ability clinicians have to assess and understand brain health from both a drug and lifestyle perspective.
"It’s akin to the first chemotherapy, or the early days of cancer, early days of diabetes," he said. "All those cases were treated by specialists, and then it became more normal. The tools developed so that eventually they pushed the treatment of the common cases back into primary care. Eventually, it’s a shift back and forth. It is a combination of what drugs and understandings in knowledge are available, as well as what tools and capabilities are out there, both these digital tools and the emerging blood-based biomarkers."