A cross-sectional investigation conducted in Japan has established the first age- and sex-stratified normative reference values for the tablet-administered Brief Assessment of Cognition app across a cohort of healthy Japanese adolescents and adults. The measurement battery, translated from an established digital platform, is designed to evaluate neurocognitive functions frequently impaired in neuropsychiatric conditions such as schizophrenia. Because standardized composite scoring requires comparison against healthy general population baselines, the collection of performance parameters across distinct demographic tiers represents an essential prerequisite for interpreting patient assessments. The evaluation analyzed data from one hundred and twenty participants across twelve distinct subgroups divided by sex and age bracket, revealing stable distributional performance across most nonverbal and executive subtests while uncovering significant performance discrepancies between alternate verbal memory test forms.
Cognitive impairment is recognized as a fundamental manifestation of schizophrenia that exerts a substantial, persistent influence on everyday social and vocational functioning. Across clinical settings, deficits commonly emerge across multiple interrelated functional domains, encompassing attention, executive processing, working memory, verbal learning, and expressive verbal fluency. Studies indicate that approximately seventy to seventy-five percent of individuals diagnosed with schizophrenia exhibit measurable neurocognitive impairments across the course of the disorder, beginning during premorbid stages and persisting following psychotic manifestation. Historically, the paper-and-pencil Brief Assessment of Cognition in Schizophrenia and its validated Japanese edition provided brief, standardized metrics to monitor these cognitive domains, relying on demographic-specific norm tables to calculate standardized z-scores and composite summaries.
The transition from traditional paper-and-pencil evaluation instruments to automated digital applications, such as the Brief Assessment of Cognition app, addresses several clinical and operational challenges. Digitized cognitive assessments standardize task administration using pre-recorded automated narration, automate raw response calculations, and mitigate measurement error arising from inter-rater variability across diverse clinical centers. Nevertheless, methodological guidelines established by the National Institute of Mental Health Measurement and Treatment Research to Improve Cognition in Schizophrenia initiative emphasize that cognitive assessment instruments must possess reliable normative benchmarks derived from the general population. In the absence of native normative baselines, clinicians and researchers in Japan were previously unable to standardize individual subtest metrics or calculate composite z-scores comparable to traditional paper-based standards.
The clinical trial enrolled healthy native Japanese individuals aged twelve to sixty-nine years, recruiting participants through Buzreach, Inc., between April 2024 and September 2025 under protocol approval from the National Center of Neurology and Psychiatry Ethics Committee. To ensure an unimpaired baseline sample, eligibility criteria mandated screening via the Structured Clinical Interview for DSM-5-Research Version to rule out psychiatric disorders, the absence of any outpatient psychiatric care history, and an estimated intelligence quotient of at least seventy determined through the Japanese Adult Reading Test-25. Prospective volunteers who were non-native Japanese speakers, individuals who had taken any prescription or over-the-counter medications within one week prior to providing consent, or those deemed unsuitable by study investigators were excluded from participation.
The enrolled analytical cohort comprised one hundred and twenty participants who completed all digital testing without missing entries, displaying an overall mean age of 38.6 years, an average formal educational attainment of 13.6 years, and a mean reading test score of 104.7. The planned study cohort was systematically stratified across six defined age tiers, encompassing individuals aged 12 to 17 years, 18 to 29 years, 30 to 39 years, 40 to 49 years, 50 to 59 years, and 60 to 69 years, with each tier populated by exactly ten male and ten female participants. Statistical power considerations grounded in theoretical chi-squared distributions demonstrated that allocating ten participants per sex and age stratum restricted the relative error of estimated standard deviations to 23.2 percent, offering an effective trade-off between logistical feasibility and statistical precision.
Trained examiners administered the software using a shared tablet computer passed between examiner and participant in a uniform, non-varying testing order. The sequence began with the Verbal Memory Task, followed by the Digit Sequencing Task, Token Motor Task, Verbal Fluency Task, Symbol Coding Task, and the Tower of London Task. Subtest rules and instructional scripts were read aloud automatically through the application narration feature to maintain standardized conditions, with raters repeating instructions manually via touchscreen controls whenever necessary. The Verbal Memory Task required participants to recall a list of fifteen spoken words presented across five identical auditory trials, while the Digit Sequencing Task presented spoken digit strings of increasing length that participants had to rearrange and repeat in ascending order.
The nonverbal and motor components tested manual motor speed, cognitive processing speed, and executive planning capabilities. The Token Motor Task instructed participants to use both index fingers simultaneously over a sixty-second interval to swipe displayed tokens into a central tablet container, recording the total count of successfully transferred tokens. The Symbol Coding Task assessed processing speed by prompting users to match numerals to geometric symbols via a digital onscreen keypad within ninety seconds, based on an associated reference key. Executive functioning was measured through the Tower of London Task, which presented pairs of three colored balls positioned across three pegs, requiring participants to calculate the minimum physical transfers needed to replicate the target pattern across two alternate test forms, designated Form A and Form B.
Across the complete sample of one hundred and twenty participants, overall mean performance metrics and standard deviations were documented as follows: 51.36 points with a standard deviation of 9.63 on the Verbal Memory Task; 23.02 points with a standard deviation of 3.64 on the Digit Sequencing Task; 46.44 points with a standard deviation of 8.28 on the Verbal Fluency Task; 55.98 points with a standard deviation of 10.98 on the Symbol Coding Task; 108.47 points with a standard deviation of 29.86 on the Token Motor Task; and 18.83 points with a standard deviation of 2.06 on the Tower of London Task. Scores on the Symbol Coding and Token Motor tasks exhibited noticeable declines among individuals in their fifties and sixties, whereas performance on the Tower of London, Digit Sequencing, and Verbal Fluency measures showed minimal variation across age categories.
A critical divergence emerged during alternate-form equivalence testing between the nonverbal and verbal measures. In the Tower of London Task, mean scores between sixty participants completing Form A and sixty completing Form B were nearly identical, yielding means of 19.08 and 18.58 respectively, with a non-significant mean difference of 0.5 points. Conversely, the Verbal Memory Task revealed a pronounced discrepancy between its two alternate word lists among adult participants. Fifty-nine participants administered Form 1 achieved an average score of 46.68, whereas sixty-one participants administered Form 3 achieved an average score of 55.89, representing a statistically significant mean difference of 9.2 points, accompanied by a ninety-five percent confidence interval ranging from 6.1 to 12.3 points.
Detailed secondary analyses in participants aged eighteen and older demonstrated that every individual word in Form 3 exhibited numerically higher correct recall rates than the corresponding words in Form 1. Inter-word correlation analyses using phi coefficients revealed that Form 3 contained numerous adjacent word pairs with strong co-recall associations, whereas Form 1 exhibited essentially random inter-word correlations, with only a single adjacent pair exceeding a coefficient of 0.3. Hierarchical agglomerative cluster analysis using Ward’s method similarly identified four adjacent presentation pairs clustered closely together in Form 3, compared to only two adjacent pairs in Form 1. Researchers noted that structural characteristics in Form 3 likely facilitated associative memory strategies, reflecting Japanese linguistic features such as homophones, polysemy, and unspoken word frequency differences across translated lists.
The authors highlighted several methodological considerations and constraints when interpreting the normative dataset. Because testing was conducted at a single institutional center, caution remains warranted when extending these norms across geographically wider populations without larger multicenter validations. Furthermore, the protocol did not evaluate test-retest reliability over repeated sessions, and the modest subgroup sample size of ten participants per stratum yields standard errors that could potentially bias standardized z-score conversions if applied rigidly in clinical contexts. The observed cross-sectional age trends on the Token Motor Task also suggest that touchscreen familiarity and digital literacy in younger generations may influence motor test performance alongside underlying motor speed, necessitating periodic recalibration of digital reference standards as technological adoption changes across aging populations.
Subject of Research: Normative baseline data collection and task-equivalence evaluation for the Japanese Brief Assessment of Cognition tablet app
Article Title: Establishment of age- and sex-stratified normative data for the Brief Assessment of Cognition (BAC) app in healthy Japanese individuals
Article References: Kubota, R., Izumi, S., Oba, M. S., Sanji, S., Tsuno, R., Miwa, M., Takai, K., Shima, H., & Nakagome, K. (2027). Establishment of age- and sex-stratified normative data for the Brief Assessment of Cognition (BAC) app in healthy Japanese individuals. Schizophrenia Research: Cognition, 47, Article 100464. https://doi.org/10.1016/j.scog.2026.100464
Image Credits: AI Generated
DOI: 10.1016/j.scog.2026.100464
Keywords: Brief Assessment of Cognition, BAC App, BACS, Cognitive Assessment, Schizophrenia Research, Neuropsychological Testing, Normative Data, Japanese Adult Reading Test, Verbal Memory Task, Digit Sequencing Task, Token Motor Task, Symbol Coding Task, Verbal Fluency, Tower of London Task, Digital Health, Cognitive Impairment, Touchscreen Usability
News Source: Ophelia Keating. (November 22, 2026). Normative Baseline Data Established for Japanese Tablet-Based Cognitive Assessment App. Scienmag.



