The Amplitude of Low Frequency Fluctuation (ALFF) proposed by Zang YF et al in 2007 is a fundamental index charactering spontaneous fluctuations in the resting brain. The ALFF has been widely used in cognitive and brain disease researches. It is generally believed that the ALFF is a representative research based on theory of functional separation that mainly reflects local oscillation and function of brain regions. Recently, several studies have revealed that the task-evoked regional activations can be predicted by interareal activity flow (AF) across the brain. Moreover, task activation and ALFF may have a common "source", suggesting that the resting-state ALFF may not only represent the functional activities of local brain regions but also reflect the exchange of information between interconnected regions in the brain. Using exploratory analysis and validation analysis, this study demonstrated that resting-state ALFF flows through intrinsic functional connectivity (FC) pathways based on the fMRI data of Nathan S. Kline Institute for Psychiatric Research (NKI-Rockland Sample) and the data from 1000 Functional Connectomes (FC1000Beijing_Zang). We hypothesized that the ALFF in local brain region can be predicted by estimated AF over resting-state FC networks. The fMRI data was first used to calculate the ALFF for the 160 brain regions of interest (ROIs) from the Dosenbach-6 network and 400-1000 ROIs from the Yeo-7 network, and the FC between these regions using Pearson correlations and multiple regression methods. The AF in each ROI was calculated as the FC-weighted sums of the ALFF of all other regions. Finally, spatial correlation between the ALFF and AF in whole brain and network levels was performed. The results showed that, in the whole brain and also in each functional network, there was a significant correlation between the ALFF and multiple regression-based AF, and the ALFF of the local brain region can be predicted by the AF from other regions. It is suggested that the ALFF represents the information interaction between the brain regions: not only of the whole-brain network but also within the local functional networks. This result can be reproduced and verified on different brain network maps and different independent samples. We also found that the FC calculated by multiple regression could improve the predictive effect of AF. Furthermore, this paper also explored the correlation of resting-state AF and ALFF with intelligence quotient (IQ) measured by the Wechsler Abbreviated Scale of Intelligence (WASI). The results showed that the correlation patterns of AF and ALFF with IQ were similar. Specifically, these regions were mainly concentrated in the default-mode network (DMN) and the frontal-parietal network, especially in the DMN areas such as the posterior cingulate/ precuneus and the bilateral inferior parietal lobule, suggesting the importance of DMN in gathering information for the whole-brain network. At the same time, the results supported the core theory of DMN and the parieto-frontal integration theory. The brain regions with significantly negative correlations between AF, ALFF and IQ scores were mainly concentrated in the sensorimotor cortex areas such as the lateral temporal lobe and the occipital lobe. This result suggests that the resting-state DMN may inhibit the sensorimotor cortext, and form two competing functional systems with the sensorimotor networks. In summary, ALFF not only characterizes the activity of local brain regions but also reflects that local spontaneous neural activities spread across brain regions and brain networks in the form of activity flows. At the same time, AF should be used as an index to describe the information flow attribute of the brain regions. It provides a new perspective for us to explore the mechanism of information communication between brain regions, to study the mechanism of spontaneous brain activities, and to explore the cognitive process and the pathological mechanisms of brain diseases.
修改评论