Selective attention, an essential part of daily activity, is often impaired in people with Alzheimer's disease (AD). Usually, it is measured by the color-word Stroop test. However, there is no universal agreement whether performance on the Stroop task changes significantly in AD patients; or if so, whether an increase in Stroop effects reflects a decrease in selective attention, a slowing in generalized speed of processing (SOP), or is the result of degraded color-vision. The current study investigated the impact of AD on Stroop performance and its potential sources in a meta-analysis and mathematical modeling of 18 studies, comparing 637 AD patients with 977 healthy age-matched participants. We found a significant increase in Stroop effects for AD patients, across studies. This AD-related change was associated with a slowing in SOP. However, after correcting for a bias in the distribution of latencies, SOP could only explain a moderate portion of the total variance (25%). Moreover, we found strong evidence for an AD-related increase in the latency difference between naming the font-color and reading color-neutral stimuli (r2 = 0.98). This increase in the dimensional imbalance between color-naming and word-reading was found to explain a significant portion of the AD-related increase in Stroop effects (r2 = 0.87), hinting on a possible sensory source. In conclusion, our analysis highlights the importance of controlling for sensory degradation and SOP when testing cognitive performance and, specifically, selective attention in AD patients. We also suggest possible measures and tools to better test for selective attention in AD.
Keywords: Aging; Alzheimer's disease; Stroop test; color vision; selective attention; sensory and cognitive interaction; speed of processing; visual perception.