Objectives Early detection of cognitive decline is critical for timely intervention, yet traditional neuropsychological assessments are time-intensive, require trained examiners, and may introduce administration bias. Digital platforms can offer scalable and accessible alternatives by enabling automated scoring and self-administration. To address this gap, we developed Cognosis, a tablet-based cognitive battery tailored for older adults, and aimed to (1) provide preliminary demographically stratified reference data and (2) examine the effects of age, education, and gender on cognitive performance in cognitively healthy Korean older adults. Methods We conducted a community-based normative study using a cross-sectional design. A total of 381 cognitively healthy adults aged 55 to 84 years were recruited from community centers, hospitals, and other local settings across Seoul, Gyeonggi, and Jeolla provinces in South Korea. Cognosis assesses four cognitive domains: attention/executive function, language, memory, and social cognition. To evaluate demographic influences, we applied penalized regression models (Least Absolute Shrinkage and Selection Operator) and stratified the preliminary normative data by age (55–64, 65–74, and 75–84), education (0–6, 7–12, 13–16, and ≥17 years), and gender. Results Age and education significantly influenced cognitive performance, with younger age and higher education associated with better scores. However, memory performance was not associated with these demographic variables. While gender did not have a main or interaction effect in regression analysis, subgroup analyses revealed that highly educated women outperformed men in several cognitive domains, including attention/executive function, fluency, and verbal/visual memory. Conclusions This study provides preliminary, demographically stratified reference data for Cognosis, improving its clinical utility for cognitive assessment in Korean older adults. Future research should validate its applicability in clinical and cross-cultural populations.
Kim et al. (Sun,) studied this question.