The American Academy of Neurology (AAN) Annual Meeting took place in Chicago this April, bringing together neurologists, researchers, and trainees for five days of sessions spanning Alzheimer's disease, movement disorders, cognitive care in neurodiverse populations, women's neurology, and the growing role of AI and digital tools in clinical practice.
One question that sits at the center of our work at Creyos kept surfacing across the week, sometimes directly and sometimes more subtly in unrelated discussions: are the cognitive assessment tools neurologists rely on keeping pace with what the field is asking of them?
The limitations of traditional cognitive screeners came up repeatedly. Presenters noted that the Mini-Mental State Examination (MMSE) and Montreal Cognitive Assessment (MoCA) are not sufficient for clinical decision-making at the mild cognitive impairment (MCI) stage, and that some centers are already moving away from using these traditional tools in both research settings and clinical practice.
The concern extended beyond Alzheimer's. In discussions about cognitive assessment and neurodiverse populations, most standard tests, including the MoCA, MMSE, and Saint Louis University Mental Status (SLUMS), were noted to be inadequately normed for intellectual disability. The MMSE in particular requires spoken language, receptive language, expressive language, and motor control, limiting its applicability for many patients and for non-verbal autistic individuals. These discussions pointed consistently toward the need for digital cognitive assessments that reduce dependence on spoken language.
A related concern was raised in a session on women's neurology. Neuropsychological testing in menopausal brain fog most often comes back normal or reveals only mild deficits, even when patients report notable real-world impairment. Presenters emphasized the need for sensitive cognitive assessment coupled with subjective patient experience to differentiate menopausal brain fog from early dementia, a distinction that traditional screening tools are not always equipped to make.
These limitations take on new urgency in the context of Alzheimer's disease, where blood-based biomarkers are changing what early detection looks like. Among blood-based biomarkers, p-tau217 was highlighted as one of the most promising for detecting Alzheimer's pathology, alongside two other modalities: amyloid PET (which can cost $10,000 per scan in the U.S.) and cerebrospinal fluid biomarkers. The consensus was that blood-based testing is a more accessible alternative that makes earlier detection feasible, and that biomarker tests should be done at the same time as cognitive testing, at the first reported sign of objective cognitive impairment.
A more complex conversation was about what to do with this earlier detection. There was discussion about how current treatments are more effective when initiated at earlier stages of the disease, and how late identification may be limiting their impact. This prompted a broader question: if the clinical symptoms of MCI are already present, why wait for a biomarker test? Early symptoms are often brushed off in primary care, and physiological testing such as brain scans and cerebrospinal fluid markers has historically been expensive and difficult to access. The general consensus was that testing at the first sign of impairment is when clinicians can identify who will actually benefit from treatment, and that waiting for full dementia means missing that window.
The field has not yet settled on exactly what the best course of action is at the MCI stage, whether that means pursuing further cognitive testing, physiological testing, or initiating anti-amyloid therapy. A live audience poll produced a 50-50 split on whether attendees would prescribe lecanemab to a patient with MCI, illustrating how much ambiguity related to available therapy remains, even as the tools for earlier detection continue to improve.
If the biomarker conversation raised the stakes for Alzheimer's assessment specifically, other sessions made clear that the demand for better cognitive measurement extends well beyond dementia.
The menopause and cognition discussion described earlier is one example. Beyond the screening sensitivity question, presenters went deeper into the biological mechanism, noting that estrogen receptors are expressed throughout the brain and that the drop in estradiol is associated with impaired deep sleep, low mood, vasomotor symptoms, and measurable cognitive changes. The discussion called for gender-based norms in cognitive testing and for research to account for gender differences and health disparities as outcomes.
Pain management was another area where cognition surfaced. Cognitive effects in patients with conditions like cervical radiculopathy are most often attributable to gabapentin and other medications rather than the underlying condition itself, and pain patients more broadly are frequently on drugs that may induce cognitive impairment. Ruling out medication effects is an important step before attributing symptoms to a neurodegenerative process.
In the neurodiverse population discussions, cognitive decline in patients with intellectual disabilities was described as being frequently missed because it is dismissed as part of their pre-existing condition. Speakers called this "diagnostic overshadowing" and emphasized the importance of incorporating caregiver perspectives and accounting for meaningful change from the patient's personal baseline when assessing for new or worsening cognitive changes.
Across these conversations, a consistent principle emerged about what kind of cognitive measurement the field actually needs. In neurodiverse populations, change from baseline is the most important metric for detecting decline in patients whose function has never aligned with standard population norms.
The Alzheimer's discussions reinforced this from a different angle. Cognitive testing at the first sign of impairment, rather than at a single later point, is what allows clinicians to track progression and determine who will benefit from treatment while still in the early stages of disease. Meaningful clinical decisions depend on understanding how a patient's cognition is changing over time, not necessarily on a single score.
Presenters also called for clinical outcomes to be captured digitally with electronic health record (EHR) integration, noting that pen-and-paper approaches are no longer adequate for longitudinal data. And in discussions about AI and clinical data collection, the observation was that while more data is being collected than ever before, the field is still largely missing standardized tools that allow multiple data sources to be integrated for research purposes.
For those of us working on cognitive assessment at Creyos, these conversations felt familiar. The measurement gaps described across MCI, neurodiverse populations, and pain management are the same gaps we're working to address, both in how we design the platform for clinical use and in the research studies we conduct and support.
For MCI, our dementia screening and assessment protocol is built to detect the subtle, early cognitive changes that traditional screeners miss, providing domain-specific profiling to support early intervention decisions. For neurodiverse populations, our tasks are self-guided, do not require spoken language, and can be completed on any device, addressing several of the accessibility barriers that were raised. For pain management, Creyos can measure the impact of pain and pain treatment on cognition across domains key to quality of life, including short-term memory, reasoning, concentration, and verbal ability. The platform also includes validated questionnaires like the Pain Medication Questionnaire (PMQ), which adds context by assessing medication use behaviors and risk factors for misuse.
The emphasis on longitudinal tracking also resonated. Creyos reports include a meaningful change indicator that uses our normative database to determine whether a change in performance between assessments is clinically significant or within the range of normal variation. For clinicians tracking patients over time, this is the difference between knowing that a score moved and knowing whether that movement is clinically meaningful. Creyos assessments are also structured for EHR integration and can be administered on any device, supporting the digital infrastructure that presenters called for throughout the week.
Conferences like the AAN Annual Meeting are useful for pressure-testing whether the problems we're focused on in our research match what clinicians are encountering in practice. This year, they did. If any of these themes resonate with challenges you're navigating in your own practice or research, we'd welcome the conversation.
Written by Sydni Paleczny, Staff Scientist
Sydni earned her MSc in Neurosciences at Western University under Dr. Adrian Owen. Her research explores neuropsychological outcomes after cardiac surgery, with interests in cognitive neuroscience, critical care, and brain health. At Creyos, she supports scientific validity, health technology, and ongoing research.