Peer Reviewed Study Validates Creyos as a Faster, More Sensitive Approach to Dementia Detection in Primary Care

• As dementia cases rise globally, more physicians and researchers are pointing to early intervention to improve patient outcomes
  
  
• Published in the Journal of Alzheimer's Disease, results show the brief digital screener built from two Creyos cognitive tasks can identify individuals at risk for cognitive impairment in minutes
  
  
• Researchers analyzed cognitive performance data from thousands of individuals, accurately detecting Alzheimer's cases 100% of the time

TORONTO, April 8, 2026 /PRNewswire/ -- Creyos, a leader in digital cognitive and behavioral health assessment, today announced the publication of new peer-reviewed research in the Journal of Alzheimer's Disease validating its two-task digital screener as highly accurate in detecting cognitive impairment linked to Alzheimer's disease. The study, conducted by researchers at Western University, demonstrates that the screener can accurately detect Alzheimer's disease with 100% sensitivity and 86% specificity in approximately six minutes, without the need for clinical supervision at point of administration.

Traditional cognitive screening tools, including the Mini-Mental State Examination (MMSE) and the Mini-Cog, remain the most widely used assessments in clinical practice, but both carry well-documented limitations. They are subject to ceiling effects, influenced by educational and cultural factors, and require trained administrators, creating bottlenecks in an already overburdened primary care system. At the 2025 Alzheimer's Association International Conference (AAIC), emerging clinical perspectives signaled growing limitations of traditional tools like the MoCA in primary care, alongside increased recognition of digital assessments as a scalable approach to cognitive detection and triage.

More than half of individuals with dementia never receive a formal cognitive evaluation during the early stages of their disease. Cognitive deficits linked to Alzheimer's disease can emerge more than a decade before a formal diagnosis, and waitlists for healthcare specialists often delay the diagnostic process further. Without fast, accurate, and accessible screening tools, early detection is routinely missed in Alzheimer's and related dementias, resulting in missed opportunities to slow decline, treat symptoms, and help patients maintain autonomy for longer.

The study used the Creyos cognitive assessment platform, an online neuropsychological battery with over 20 million task completions accumulated over more than 15 years. Unlike traditional screening tools that rely on simple pass/fail thresholds, Creyos's machine learning models are trained on a deep dataset to recognize complex patterns in cognitive performance and able to capture subtle signatures of impairment that conventional scoring methods can miss.

The dementia screener was constructed by selecting two tasks from the platform: a visuospatial working memory task (Number Ladder) and an attention task (Feature Match). Tasks were carefully chosen based on their demonstrated association with patterns of cognitive change associated with Alzheimer's disease and mild cognitive impairment (MCI). An AI-driven machine learning model was then trained on a Creyos data set to establish accurate criteria for a positive result — translating years of accumulated cognitive data into a clinically actionable signal. This model was subsequently evaluated against two large, independently published datasets comprising a total of 9,350 individuals aged 21 to 97, as well as a smaller cohort of patients independently diagnosed with Alzheimer's disease.

Key findings from the study include:

• 100% sensitivity: Every patient with a clinical diagnosis of Alzheimer's in the validation cohort was correctly identified by the Creyos screener.
  
  
• 86% specificity: The screener generally did not flag healthy individuals, with a small subset flagged for follow-up evaluation.
  
  
• Outperforms the MMSE: One patient with Alzheimer's who scored above the standard MMSE impairment threshold—and would not have been flagged by that tool alone—was correctly identified by the Creyos screener.
  
  
• Dementia-specific: Across ten health conditions including hypertension, concussion, depression, and anxiety, the Creyos screener uniquely identified dementia-related impairment. While individuals with non-dementia conditions sometimes showed lower cognitive scores, they were significantly less likely to be flagged as positive by the screener, demonstrating the screener's ability to isolate dementia-related cognitive impairments and reduce false positives.
  
  
• Mirrors real-world prevalence of MCI: The screener was accurate enough that positive results in the general population, which increased significantly after age 50, aligned with the estimated 20% prevalence of mild cognitive impairment.
  
  
• Consistent across populations: Screener outcomes were not significantly affected by gender, education level, or study population. Creyos's assessments are designed to be accessible for all, regardless of language, culture, literacy, or ability. This is supported by the platform's use of visual tasks and shapes that are universal, leaving no room for misinterpretation or confusion with instructions.

Together, these findings suggest that the Creyos screener enables earlier identification of cognitive decline and can serve as a complementary tool for clinical assessment, while its accessible, digital-first format supports greater uptake in routine clinical care.

"Primary care providers are on the front lines of cognitive health, but they've lacked the right tools. For decades, the science underpinning Creyos's tasks has demonstrated strong sensitivity and specificity when it comes to measuring core domains of cognition," said Marc Lipton, CEO and President of Creyos. "What this study shows is how that scientific foundation can support better clinical decision making by leveraging a screener fast enough for a routine appointment, sensitive enough to catch impairment that the MMSE might miss, and accessible enough to work for every patient regardless of background."

The findings support the use of the Creyos screener within a two-step evaluation approach aligned with leading clinical guidelines, serving as an initial signal to guide timely follow-up. By helping identify which patients may benefit from further assessment and reducing unnecessary specialist referrals, Creyos offers a practical way to help providers triage cognitive concerns in primary care. With dementia-related health and long-term care costs in the United States projected to reach $1 trillion by 2050, scalable approaches that enable earlier identification of potential impairment and more efficient care pathways represent a critical investment in population health.

About Creyos
Creyos, formerly Cambridge Brain Sciences, is redefining how brain health is assessed. Founded by renowned neuroscientist Professor Adrian Owen, PhD, OBE, FRSC, FCAHS, the platform is built on three decades of research and one of the world's largest cognitive datasets, leveraging machine learning techniques to develop its battery of digital cognitive assessments long before AI became mainstream. Designed as a clinical decision support tool, Creyos combines validated online cognitive tasks with digitized behavioral health questionnaires to equip providers with objective, scalable insights to inform clinical decision making and care. The FDA-registered platform has been used in over 400 peer-reviewed studies and is trusted by tens of thousands of clinicians and some of the largest health systems and organizations across North America. Visit www.creyos.com to learn more.

Media Contact
[email protected]

SOURCE Creyos