Spatial navigation as a digital marker for clinically differentiating cognitive impairment severity

Spatial navigation impairments emerge early in Alzheimer's disease, but assessments targeting these deficits remain underutilised or impractical for cognitive screening. The Spatial Performance Assessment for Cognitive Evaluation (SPACE) is a newly developed digital tool that evaluates spatial navigation deficits associated with cognitive impairment.

We assessed spatial navigation ability using SPACE in 300 older adults recruited from memory clinics and the general community. Participants were classified across different levels of cognitive impairment using the Clinical Dementia Rating (CDR) scale. Performance in SPACE was compared with clinical diagnosis, standard cognitive assessments, and demographic models using Area Under the ROC Curve (AUC), sensitivity, and specificity.

We show that SPACE reliably distinguishes CDR levels, exceeding the accuracy of demographic models and matching or surpassing most traditional neuropsychological tests. Including SPACE significantly increases the AUC for distinguishing between no dementia from mild dementia (0.76 to 0.94), no dementia from moderate dementia (0.79 to 0.95), and questionable dementia from mild dementia (0.70 to 0.91), all with consistently high sensitivity and specificity. A shortened version of SPACE, lasting less than 11 minutes, reduces administration time by 40% while maintaining high diagnostic accuracy. Cross-validation analyses confirm the reliability and robustness of these models.

These findings highlight the potential of digital spatial navigation assessments to advance early detection, contributing to scalable and accessible healthcare.