Abstract

We aimed to investigate the activity within and the connectivity between resting state networks (RSNs) in healthy subjects and patients with Alzheimer’s disease (AD) or mild cognitive impairment (MCI). Magnetic resonance imaging (MRI) and resting-state MRI were performed on patients diagnosed with AD (n=18) or MCI (n=16) and on healthy subjects (n=18) with matching demographic characteristics (age, sex, and education level). Independent component analysis and Granger causality analysis (GCA) were used during image postprocessing. We calculated ‘In + Out degree’ for each RSN. Then, we investigated the relationships between “In + Out degree” of each brain network and the cognitive behavioural data. RSNs were obtained using the optimal matching method. The core areas of the five RSNs were similar between the AD, MCI, and healthy control groups, but the activity within these five RSNs was significantly lower in the AD and MCI groups than in the healthy control group (P<0.01, false discovery rate corrected). The GCA results showed that the connectivity between the five RSNs, particularly the connectivity from the default mode network (DMN) to the other RSNs, was slightly lower in MCI patients and was significantly lower in AD patients than in healthy subjects. In contrast, increased connectivity was evident between the memory network and the executive control network in the AD and MCI patients. The “In + Out degree” of the DMN negatively correlated with the Montreal Cognitive Assessment score in AD patients (R=-0.43, P<0.05). In conclusion, the activity within RSNs and the connectivity between RSNs differed between AD patients, MCI patients, and normal individuals; these results provide an imaging reference for the diagnosis of AD and the measurement of disease progression and reveal insight into the pathogenesis of AD.

Keywords: Alzheimer’s disease (AD), mild cognitive impairment (MCI), dementia, resting-state networks (RSNs), resting state functional magnetic resonance imaging (rs-fMRI), independent component analysis (ICA), granger causality analysis (GCA).