Age-Related Alterations in Electroencephalography Connectivity and Network Topology During n-Back Working Memory Task

Abstract

The study of the healthy brain in elders, especially age-associated alterations in cognition, is important to understand the deficits created by Alzheimer's disease (AD), which imposes a tremendous burden on individuals, families, and society. Although, the changes in synaptic connectivity and reorganization of brain networks that accompany aging are gradually becoming understood, little is known about how normal aging affects brain inter-regional synchronization and functional networks when items are held in working memory (WM). According to the classic Sternberg WM paradigm, we recorded multichannel electroencephalography (EEG) from healthy adults (young and senior) in three different conditions, i.e., the resting state, 0-back (control) task, and 2-back task. The phase lag index (PLI) between EEG channels was computed and then weighted and undirected network was constructed based on the PLI matrix. The effects of aging on network topology were examined using a brain connectivity toolbox. The results showed that age-related alteration was more prominent when the 2-back task was engaged, especially in the theta band. For the younger adults, the WM task evoked a significant increase in the clustering coefficient of the beta-band functional connectivity network, which was absent in the older adults. Furthermore, significant correlations were observed between the behavioral performance of WM and EEG metrics in the theta and gamma bands, suggesting the potential use of those measures as biomarkers for the evaluation of cognitive training, for instance. Taken together, our findings shed further light on the underlying mechanism of WM in physiological aging and suggest that different EEG frequencies appear to have distinct functional correlates in cognitive aging. Analysis of inter-regional synchronization and topological characteristics based on graph theory is thus an appropriate way to explore natural age-related changes in the human brain.

Decreased connection density and modularity of functional brain networks during n‐back working memory paradigm

“…The RT of the subjects was found positively correlated to FC in the alpha and beta bands, however only during the control state. In contrast, subject accuracy was found negatively correlated with FC in the beta and gamma bands, although only in the most challenging setting (3-back).Our results are in line with earlier reports on decreased connectivity during the n-back WM paradigm(Dai et al, 2017;Ginestet & Simmons, 2011;Hou et al, 2018). However, in contrast to Dai et al (2017), we found a stimulation-related decrease in E and C…”

“…Conversely, no effect of task difficulty on FC could be identified. in both the theta and alpha bands and in the beta band, respectively, while Hou et al (2018) found an increase in the same measures. These seemingly contradictory results may be attributed to differences in the analysis pipeline, which can affect results significantly even when obtained from the same dataset (Jalili, 2016;Lindquist, 2020).…”

“…On one hand, findings in EEG studies are consistent in that functional segregation and connectivity strength decreases during tasks when compared to rest (Cohen & D'Esposito, 2016; Ginestet & Simmons, 2011; Hou et al., 2018). On the other hand, functional integration has been shown to decrease (Hou et al., 2018) and increase (Cohen & D'Esposito, 2016; Dai et al., 2017) alike while performing an n ‐back task. Furthermore, measures of global FC were found increased during n ‐back stimulation (Fishburn et al., 2014) when monitoring neural activity with functional near‐infrared spectroscopy (fNIRS).…”

“…On the other hand, functional integration has been shown to decrease (Hou et al, 2018) and increase (Cohen & D'Esposito, 2016;Dai et al, 2017) alike while performing an n-back task. Furthermore, measures of global FC were found increased during n-back stimulation (Fishburn et al, 2014) when monitoring neural activity with functional near-infrared spectroscopy (fNIRS).…”

“…This relatively simple concept was indeed demonstrated to be a powerful asset of neuroscience, which could be utilized for a better understanding of not only physiological but also pathological brain function (Bullmore & Sporns, 2009; Stam, 2014). Furthermore, the graph theoretical FC approach was successfully applied for characterizing the effects of various cognitive stimulation paradigms as well (Hou et al., 2018; Racz et al., 2017; Ren et al., 2017).…”