Cognitive Health Part 2: How to Measure Cognitive Health in Clinical Trials

This is part two of our three-part series on cognitive health. In part one, we made the case for why cognitive health matters as both a research and clinical endpoint—not just as an early warning sign. In this post, we lay out how cognitive health actually gets measured, from established site-based tools to the decentralized platforms, surveys, and devices changing what's possible in remote research. Part three will show you how to put these tools directly into an Alethios study.

How Cognitive Health Is Measured

Cognitive evaluation today splits fairly cleanly into two camps: site-based methods that require getting a specific person into a specific room, and decentralized methods that don't. Both have a role to play—the question is which one fits a given study's goals.

Site-Based Methods

• ‍Administered cognitive battery testing is still where most cognitive evaluation runs today. Clinicians rely on brief, in-person screens like the Montreal Cognitive Assessment (MoCA) or the Mini-Mental State Examination (MMSE) to screen for mild cognitive impairment and dementia in geriatric and primary care settings (MoCA validation study; MMSE in general medical practice), and for more detailed diagnostic work, hours-long neuropsychological batteries administered one-on-one by a trained specialist. All of these are well-validated and clinically useful for what they were built to do—but all of them also require a specific evaluator, in a specific room, on a specific day.

• ‍Participant-reported data in a site-based setting usually mean paper-and-pencil questionnaires filled out in a waiting room, asking someone to simply describe how their focus or memory has been. That's useful information, but it isn't the same as data you can act on: research has repeatedly found that subjective cognitive complaints don't reliably track with objective test performance—they're often driven more by mood and depressive symptoms than by what's actually happening cognitively (subjective complaints and depressive symptoms study; subjective complaints contributing to misdiagnosis of mild cognitive impairment). That's not a reason to ignore self-report—it's a reason not to rely on it alone.

• An EEG (electroencephalogram) is a test that offers a faster, cheaper window into what's happening in the brain itself rather than just how someone performs on a task. Because it directly measures electrical activity, EEG can capture things like attention and processing speed in real time, and it's increasingly feasible in remote and rural settings, though it still requires specific hardware and some technical setup (EEG and fNIRS in Alzheimer's digital assessment, review; portable EEG feasibility in rural communities). It earns its place when a study needs a direct physiological read on attention or processing speed that a behavioral test alone can't provide.

• An MRI (magnetic resonance imaging) is a diagnostic test goes a layer deeper still, showing brain structure rather than behavior or electrical activity. Functional and structural MRI has become a standard biomarker in Alzheimer's disease drug trials, used to track brain atrophy or confirm a drug is engaging its intended target, and neuroimaging biomarkers are now built into how disease-modifying therapies get evaluated in late-stage trials (neuroimaging biomarkers in Alzheimer's trials; imaging as a biomarker in drug discovery). It's expensive and requires specialized facilities, so it's best reserved for studies that need to confirm a biological mechanism or track structural brain changes directly, rather than routine cognitive tracking.

Why the Shift to Decentralized Matters

People have good days and bad days—cognitively as well as physically—but an in-person visit only ever captures one of them. To avoid overburdening participants with travel and scheduling, site-based trials typically test cognition at a small handful of fixed timepoints, with no way to tell whether a given score reflects someone's baseline or just a rough morning. Studies using ecological momentary assessment—repeated, brief cognitive check-ins throughout someone's normal day—have found significant day-to-day variability in cognitive symptoms, closely tied to factors like fatigue, mood, and time of day (EMA in multiple sclerosis study; JMIR reliability and validity study). A single site visit can't tell you which version of someone's cognition you're looking at. You get what you get.

‍

Cognition also isn't something that happens in a quiet room with a proctor and a stopwatch—it happens in traffic, at the grocery store, in the middle of a work meeting. A long-standing critique of traditional neuropsychological testing is that lab-based scores don't always predict how someone actually functions day to day, precisely because the testing environment looks nothing like real life (ecological validity review, Neuropsychology Review). That's part of why researchers have pushed toward "mobile cognition" approaches that try to measure cognitive function in the environments people actually use it in, rather than assuming a clinic visit is a fair stand-in for daily life (mobile cognition approach, PMC).

Decentralized Methods

Remote cognitive testing platforms bring real, objective data from controlled, validated tests directly to a participant's home. Remote, self-administered digital cognitive assessments have shown reliability comparable to in-person testing, including in people already experiencing subjective cognitive decline (PLOS Digital Health reliability study; Frontiers preliminary validation study), and direct comparisons between remote and in-person versions of standardized cognitive measures have found consistent performance across both formats (Mobile Toolbox and MyCog Mobile comparison). Because these tests don't require a drive to a site, researchers can also collect far more data points per participant than an in-person schedule would ever allow—turning a single test score into an actual trend line.

‍

Different digital platforms have found their footing in different types of research. Cambridge Cognition's CANTAB battery is widely used in CNS drug development and pharmaceutical trials, and has been deployed in some of the largest brain health studies conducted to date (Cambridge Cognition, CANTAB brain health study). Cogstate's digital battery, built specifically for clinical trials, has been directly compared against the NIH Toolbox—another computerized cognitive battery—in head-to-head validation research, giving trial sponsors evidence to choose between them for a given study (NIH Toolbox vs. Cogstate C3 comparison). BrainHQ, originally built for consumer brain training, now has a large published research base behind its assessment tools, making them relevant to cognitive wellness and healthy-aging research as well as clinical trials (BrainHQ published studies). And Creyos (formerly Cambridge Brain Sciences) offers a browser-based cognitive testing platform built for both clinical and research use, currently being validated in a dedicated clinical trial assessing its online cognitive assessments in older adults with major depressive disorder or mild cognitive impairment (Creyos validation trial, ClinicalTrials.gov; Creyos online cognitive tasks). None of them require a waiting room.

‍

Validated remote surveys let self-report do more without ever needing a clinic. The Everyday Cognition scale (ECog) is a validated informant- and self-report questionnaire that asks about cognitively relevant everyday activities, like managing finances or navigating a familiar route, and has been shown to track meaningfully with objective cognitive testing and neuroimaging findings, all completed remotely (ECog scale development and psychometric properties; ECog short form validation). The PROMIS Cognitive Function measure, built using modern patient-reported outcome methodology, has similarly been validated for capturing perceived cognitive functioning across both clinical populations, like cancer survivors, and the general population, and is increasingly used as a companion measure alongside objective testing rather than a replacement for it (construct validity of PROMIS Cognitive Function; PROMIS Cognitive Function Screener in Medicare Annual Wellness Visit). Delivered remotely, these become another continuous data point rather than a one-time form.

‍

Decentralized EEG devices are bringing electrophysiology out of the lab, too. NextSense, a company spun out of Alphabet's X, has built EEG sensors directly into earbuds, with clinical research already validating in-ear EEG for uses like predicting epileptic seizures and monitoring sleep, all without a cap full of electrodes or a lab visit (NextSense clinical evidence; in-ear EEG and seizure prediction). Emotiv makes a research-grade, multi-channel EEG headset that's been used in wearable-EEG benchmarking studies comparing consumer-accessible devices against lab-grade equipment (Emotiv EPOC X, research-grade EEG headset; real-world benchmarking of wearable EEGs). Zeto has built an FDA-cleared, dry-electrode wireless EEG headset positioned specifically for research and clinical trial use, backed by published technical validation data (Zeto EEG for research and clinical trials; technical validation of the Zeto wireless EEG system). And Muse, originally built as a consumer meditation headband, has a growing published research base behind it, including studies on cognitive functioning in children practicing mindfulness with the device (Muse research; mindfulness with a brain-sensing device and cognitive functioning in children).

Access to Cognitive Health Evaluation

This benefits more than research logistics—it changes who actually has access to cognitive health evaluation in the first place. Remote neuropsychological testing has been shown to be a feasible, valid way to reach populations who live far from the nearest specialty clinic, including dedicated programs built specifically for rural older adults (University of Iowa rural remote neuropsychological assessment program). Video-supported remote cognitive assessment has also proven usable and acceptable in general practice settings, which is a sign this isn't just a research convenience—it's a way to get real evaluation to people, and researchers, who'd otherwise go without it entirely.