Working memory: what does research say about oscillation and functional connectivity?

Abstract

Working memory is a crucial cognitive function for processing and storing information in short-term memory during purposeful activities [1]. Numerous studies using EEG, MEG, and stereo-EEG were conducted to identify the neurophysiological correlates of working memory, including oscillations in specific brain regions in the absence of a memorized stimulus. The studies explored multiple aspects such as encoding, maintenance, processing, ordering, updating, and inhibition and various working memory modalities. Additionally, the studies highlighted the significance of examining the functional connectivity between different brain regions. As a result, researchers have amassed a large amount of data that is often contradictory and lacks organization [2]. The aim of this study is to perform a methodical review of literature published from 1980 up to the present period, cataloged in the Scopus, Web of Science, and Pubmed databases. The systematic review follows the PICo structure and addresses three main questions: which neuronal networks are involved in different working memory modalities and detect synchronizations between and within frequencies in cognitively safe subjects; are there disparities in the characteristics of neuronal networks between adults and the elderly; and can various interventions be employed to align the characteristics of neuronal networks in the elderly with those of adults [3]. The systematic review follows PRISMA-P 2015 protocol [4]. Main inclusion criteria focus on investigating working memory, analyzing visual, verbal and nonspecific modalities, using EEG/MEG/stereo-EEG, assessing adult and elderly subjects, and providing quantitative results from behavioral and neurocognitive research. Exclusion criteria: The study excluded emotional stimuli and assessments of long-term memory as well as the use of fMRI. Additionally, unigender studies and studies with subjects who had cognitive or other impairments were not included. In addition, the study required a sufficient number of subjects and time-frequency analysis. Moreover, physical or dietary interventions were excluded, and experimental research was necessary for inclusion. The study did not assess working memory in a foreign language for the subjects. A systematic review was conducted on the platform nested-knowledge.com. Based on the given criteria, a literature search was performed, yielding 2,828 articles meeting the inclusion criteria. These were further screened by two experts, resulting in the identification of 234 relevant articles, and upon full text review, 89 articles were deemed relevant. The resulting pool of articles is characterized by the following parameters: 69 articles describe adult subjects, 5 articles describe elderly subjects, and 12 articles compare both groups. Additionally, 42 articles focus on oscillation, 15 on functional connectivity, and 32 on both metrics. Concerning working memory, 22 articles describe all stages, 8 focus on encoding, 11 on encoding and maintenance, 4 on maintenance and recall, 23 on maintenance, 9 on maintenance and processing, 7 on processing, 1 on recall, and 3 without separating stages. Furthermore, 28 articles pertain to verbal memory, 51 to visual memory, 8 articles contain comparisons of modalities, and 2 articles discuss modally non-specific tasks. Preliminary findings from the systematic review indicate distinct functions of various frequency bands for working memory. Specifically, theta rhythms (4–8 Hz) in the frontal lobe were predominantly associated with information maintenance, while demonstrating a propension for increased synchronization during information processing. The alpha rhythm (8–14 Hz) was associated with inhibiting irrelevant data and protecting the current content of working memory, aligning with the widely accepted paradigm [5]. The beta rhythm (14–28 Hz) was most frequently noted to occur when maintaining and recalling information, and its power was associated with memory accuracy indicators. The gamma rhythm (28 Hz and higher) was seen during the process of encoding and maintaining information, and exhibited greater power for more complex stimuli. The analysis of connectivity revealed that encoding visual and verbal stimuli involves interhemispheric frontal-temporal and frontal-central connections that interact through theta rhythms. Additionally, storing information relies on the interplay of theta and gamma rhythms between the frontal and parietal networks. Notably, when dealing with verbal stimuli, the connectivity of the frontal and temporal brain lobes through theta rhythm enhances with load during storage. The alpha rhythm facilitates communication across posterior and frontal divisions during information storage, whereas beta rhythm is associated with frontal-temporal connections. During visual information storage, theta rhythm mediates communication across frontal-postcentral connections. As storage load increases, theta rhythm leads to strengthening of frontal-parietal and frontal-frontal connections until the threshold of working memory capacity is reached, after which these connections weaken. When processing information, connectivity of the right frontal-occipital network, right prefrontal and left occipital regions, right frontal and occipital-parietal areas for visual memory, and frontal-parietal regions for verbal memory were observed in the theta band. Connectivity of the occipital brain regions primarily occurred in the alpha spectrum, and temporal regions exhibited high frequencies. No specific differences were observed between the elderly and adults, aside from the suppression of low frequencies and the prevalence of high frequencies in the connectivity analysis.