Digital Cognitive Assessment as a Digital Biomarker
Digital cognitive assessment delivers validated cognitive tasks on a screen. Its advantage is frequency, not novelty, and the distance from brain training apps is the whole point.
Evidence maturity
Graded with the V3 framework: whether the sensor measures accurately, whether the algorithm has been validated against a reference standard, and whether the measure has been shown to matter clinically.
Individual tasks often have strong technical validation and a long research record. The weaker step is clinical validation: showing that change on the test corresponds to change that matters to the person is missing for many available products.
What is Digital Cognitive Assessment
Digital cognitive assessment means administering cognitive tests on a computer, tablet or phone rather than with pencil, paper and a trained administrator. It covers established computerised batteries with decades of use as well as newer app based tests designed for remote self administration.
The advantage is not that the tasks are better. It is that they can be given far more often. A paper battery requires a clinic visit and an administrator, which limits it to a few timepoints; a self administered digital battery can be repeated weekly at home, producing a trajectory instead of two distant points and averaging out the day to day variability that makes single assessments noisy.
The category also has a boundary worth defending. A validated digital cognitive assessment is developed to detect specific impairment and to withstand repetition; a consumer brain training product is not, and the visual similarity between them does not make them interchangeable.
How it is measured
Tasks are drawn from established neuropsychology and translated to a screen: paired associate learning for memory, set shifting and trail making for executive function, simple and choice reaction time for processing speed. Computerised batteries with a long translational research record, and publicly funded batteries developed for large scale use, provide the validated task pool most digital assessments draw from.
Two technical issues dominate. Practice effects mean the same task cannot simply be repeated, so batteries use alternate forms or adaptive item generation. Device dependence means timing sensitive tasks give different values on different hardware, so studies should fix the device or record it as a covariate.
Remote self administration adds a third: nobody is watching, so the testing environment, distractions and effort are uncontrolled in a way clinic testing is not.
Clinical use
The largest application is Alzheimer's disease and mild cognitive impairment research, where digital assessment is used for screening, enrichment of trial populations, and repeated measurement of decline. Reviews of digital technologies in this field have covered their use as biomarkers, as clinical outcome assessments and as recruitment tools, which is a wider role than most digital measures occupy.
Other uses include monitoring cognitive side effects of medication, tracking recovery after stroke or concussion, and assessing cognitive change in conditions with known cognitive involvement such as multiple sclerosis. Measures are reported alongside established instruments such as the Addenbrooke's Cognitive Examination, the Trail Making Test and informant questionnaires, which supply the clinical anchors and the perspective of someone who knows the person.
Regulatory status
No blanket qualification. Digital cognitive assessments appear in registered trials as outcome measures and screening tools, and individual products differ substantially in the evidence supporting them.
Limitations
Technical validation is often good, and clinical validation is often the weak link. Demonstrating that a digital task correlates with a paper test is not the same as demonstrating that change on it corresponds to change that matters to the person, and that second step is missing for many available tests.
Self administration introduces uncontrolled conditions and raises questions about who actually took the test. Performance is affected by education, language, familiarity with technology and sensory impairment, and digital delivery can widen rather than narrow those gaps in older populations.
Practice effects are managed rather than eliminated, and residual improvement across repeated sessions can mask real decline in a study that assumes stability.
References
- Koo BM, Vizer LM. Mobile technology for cognitive assessment of older adults: a scoping review. Innov Aging. 2019. pubmed.ncbi.nlm.nih.gov
- Gold M, et al. Digital technologies as biomarkers, clinical outcomes assessment, and recruitment tools in Alzheimer's disease clinical trials. Alzheimers Dement (N Y). 2018. pubmed.ncbi.nlm.nih.gov
- Barnett JH, et al. The Paired Associates Learning (PAL) test: 30 years of CANTAB translational neuroscience from laboratory to bedside. Curr Top Behav Neurosci. 2016. pubmed.ncbi.nlm.nih.gov
- Ho EH, et al. NIH Toolbox cognition performance in older adults with normal cognition, mild cognitive impairment and dementia. Arch Clin Neuropsychol. 2025. pubmed.ncbi.nlm.nih.gov
Direct counterparts of the supervised cognitive examinations this measure digitises, including the Addenbrooke's examination, the Trail Making Test and the Frontal Assessment Battery. The informant questionnaire is included because it captures the perspective of someone who knows the person, which no test does.
Run cognitive studies with WeGuide, the all in one patient engagement platform
Deliver repeated cognitive testing alongside wearable data and patient reported outcomes in one workflow.
