Dissociable neural mechanisms in selective attention across different mild cognitive impairment subtype

Mild cognitive impairment (MCI) is a neurodegenerative disorder found in more than 30% of the elderly population. Patients with MCI exhibit a variety of cognitive deficits, including working memory and attention. Classifying MCI into meaningful subtypes based on these functional deficits will thus help determine selective brain targets for effective interventions for Alzheimer’s disease and other dementias. Recently, we have found that early selective visuospatial attention is generally impaired in MCI. We used a model-based approach (i.e., inverted encoding model; IEM) to reconstruct spatially selective representations of the attentional focus based on the activity of slow-going EEG oscillations at ~9-12Hz (i.e., alpha band activity) recorded while subjects performed a variant of the attention-cueing Eriksen Flanker tasks and compared the fidelity of the alpha-based spatial representations across the MCI and healthy aging populations. In general, we found the reduced fidelity of alpha-based spatial reconstructions in the MCI compared to the healthy aging groups. In the present study, we examined further how the neural deficit in early selective visuospatial attention is related to working memory function. Here, we divided MCI patients into different subtypes based on their behavioral performance on subsets of the Montreal Cognitive Assessment (MoCA), which require WM. We found that MCI patients with intact WM functions were slower than those with WM deficits. Moreover, these working memory-intact MCI patients showed the reduced fidelity of alpha-based reconstructions compared to the patients who exhibit substantial WM impairment. On the other hand, the MCI patients with WM deficits exhibited the increase in the activity of the late frontal negative-going wave in response to the incongruency between targets and distractors. Together, these results suggest dissociable neural deficits underling dysfunctions in selective attention in different MCI subtypes categorized based on their WM function, which could be used as potential brain targets for the future development of personalized treatments for MCI.