Working Memory Capacity, Memory Load, and Cognitive Control Network

INTRODUCTION An important function of working memory (WM) is to integrate incoming information into an appropriate cognitive model by using two executive functions (EFs that govern the coordination and manipulation of information in WM (Baddeley, 2012). Three core EFs are commonly differentiated: updating, shifting, and inhibition (Scharinger et al., 2015). In order to design an augmented cognition program that identifies and extracts meaningful neurophysiological features related to a human operator’s cognitive state as well as the level of mental workload, we must first understand neural mechanisms underlying EFs, which are building blocks of cognition. Traditionally, augmented cognition uses neurophysiological measurements (e.g. P300 amplitude and ERP) to evaluate the user’s cognitive state and changes the human-system interface in order to enhance performance (Schmorrow & Kruse, 2004). However, these traditional measures focus on frequency and temporal correlation within the EEG signal. This only provides insights into relationships between the stimulus and activity in specified brain regions. Conversely, effective connectivity measures, such as Granger Causality, can be used to analyze time-varying interactions between brain areas in order to draw conclusions regarding causal relationships. Therefore, we explored working memory capacity (WMC) as an individual difference that can impact performance and connectivity. Niendam et al. (2012) has shown through a systematic review of studies investigating brain activation during multiple executive functional tasks that specific regions, termed the cognitive control network (CCN). Niendam et al. (2012) identified activation of the CCN in the frontal and parietal areas for both inhibition and updating tasks. In the present study, we investigated how individual differences in WMC influence the participants’ performance and cognitive brain network activation. A dual-task paradigm combining flanker and n-back stimuli was used to modulate updating load and evaluate inhibitory control. All subjects were categorized by the WMC, as measured by the operational span task (OSPAN). We hypothesized that participants with low WMC would show inefficient brain connectivity because of their inability to maintain a stable activation of their WM representations and brain activity will be localized to sources within the CCN and that the WM load will influence the pattern of information transfer between these brain regions depends on WMC.