Diagnostic accuracy of a self-administered digital cognitive test battery for differentiating early Alzheimer’s disease from controls

BACKGROUND: Paper-pencil assessments face limitations in accessibility, lengthy administration, and interrater reliability. Digital tools may overcome these by offering user-friendly, non-verbal, and scalable assessments. This study investigated the diagnostic accuracy of BraincheX, a novel digital test battery focused on accessibility and usability, by comparing its subtest performance to CERAD-Plus equivalents. METHODS: Fifty-six participants with early AD and amyloid negative cognitively healthy controls were recruited from the LMU hospital memory clinic. Diagnostic groups were defined by amyloid positivity and clinical expert consensus according to the IWG-2 criteria. Participants completed the BraincheX digital battery and CERAD-Plus. BraincheX total score was calculated by summing standardized subtest scores. Data were analyzed using ROC analysis and the Youden Index. Correlations between subtests were examined, and group differences analyzed via ANOVA. RESULTS: ROC analysis revealed strong diagnostic accuracy for the BraincheX total score in differentiating between early AD and controls (AUC = 0.86; optimal cut-off = - 1.51), with high sensitivity of 89.7% and moderate specificity of 68.0%. BraincheX total scores correlated with CERAD total scores (r = .73, p < .001). BraincheX demonstrated good internal consistency (Cronbach's α = 0.84). ANOVA confirmed significant group differences across all BraincheX subtests (p < .01). Individual subtests showed strong correlations with CERAD-Plus equivalents. CONCLUSIONS: Findings suggest BraincheX as a promising, accessible digital screening tool for cognitive decline, offering automated scoring, immediate feedback, and reduced examiner bias. Future research is required to validate BraincheX in larger and more diverse populations and to establish BraincheX's clinical applicability. TRIAL REGISTRY: This analysis was conducted within the framework of the prospectively registered German SCD-Q validation study (NCT06711952). The registered study protocol included digital cognitive testing as part of the assessment protocol. The present manuscript reports a secondary analysis focusing on the diagnostic performance of BraincheX.