BACKGROUND: Alzheimer's disease (AD) is associated with selective attention impairments, which could contribute to cognitive and functional deficits. Using visual scanning parameters, selective attention toward novel stimuli, or novelty preference, can be measured by a non-verbal, non-invasive method that may be of value in predicting disease progression.

OBJECTIVE: In this longitudinal study, we explored whether novelty preference can predict cognitive decline in AD patients.

METHODS: Mild to moderate AD patients viewed slides containing both novel and repeat images. The number of fixations, the average fixation time, and the relative fixation time on the two types of images were measured by an eye-tracking system. Novelty preference was estimated by the differences between the visual scanning parameters on novel and repeat images. Cognition and attention were assessed using the Standardized Mini-Mental Status Examination (sMMSE) and the Conners' Continuous Performance Test (CPT), respectively. Cognition was re-assessed every 6 months for up to 2 years.

RESULTS: Multivariate linear regressions of 32 AD patients (14 females, age = 77.9±7.8, baseline sMMSE = 22.2±4.4) indicated that reduced time spent on novel images (t = 2.78, p = 0.010) was also associated with greater decline in sMMSE scores (R2 = 0.41, Adjusted R2 = 0.35, F3,28 = 6.51, p = 0.002), adjusting for attention and baseline sMMSE.

CONCLUSION: These results suggest that novelty preference, measured by visual attention scanning technology, may reflect pathophysiological processes that could predict disease progression in the cognitively-impaired.