Are cognitive subtleties too subtle to see? A diffusion tensor imaging validation of the subtle cognitive impairment test and other psychometric assessments in a normative sample.

Dementia care and research is limited by difficulties identifying patients in its prodrome. Existing cognitive screening forms have poor sensitivity with respect to subtle changes. The Subtle Cognitive Impairment Test has been proposed as a superior assessment, but remains largely unproved. We investigate this and other cognitive tests by establishing their sensitivity to brain diffusion tensor imaging metrics in a sample of healthy, older adults. Healthy participants (N = 84) aged 50–70 were recruited. All subjects underwent a holistic cognitive test battery. Diffusion tensor imaging data were processed to create maps including mean diffusivity, fractional anisotropy, radial and axial diffusivity. Cross-sectional correlations compared mean diffusivity and fractional anisotropy against all cognitive test outcomes using tract-based spatial statistics, with any significant results being re-run with age correction. Models which survived this were further inspected with regards to axial and radial diffusivity as appropriate. Only the subtle cognitive impairment test (“tail percentage errors”) and trail making (“A” condition) provided areas where voxels reached p < 0.05 in univariate models, while only the former survived age correction. Post-hoc analyses using the subtle cognitive impairment test variable found a small cluster of significant voxels with respect to axial diffusivity. It appears unaffected by participant level of education, mood or quality of life. The subtle cognitive impairment test appears to be a superior cognitive marker of subtle shifts in neurophysiology and may therefore be a desirable tool in the investigation or triage of early neurodegeneration.