The Changes In Functional Brain Network Dynamics Underlying Task Switching

To adapt to the ever-changing environment, humans need to flexibly alternate between tasks in terms of an internal goal. Switching from a cognitive task to another is known as task switching which is a basic way of cognitive processing. Such process is commonly investigated in the task switching paradigm by requiring participants to continually shift between different tasks. It is constantly found that switching between tasks is slower and more error prone than repeating the same task, which is termed switch cost. When specified task demands are present, individuals are thought to enter a task-dependent cognitive state and configure a corresponding task-set where the task information is maintained as a configuration of perceptual, attentional, mnemonic and motor processes to perform the task. In task switching paradigm, a change of task results in a task-set updating which is associated with relatively stronger neural activity when compared with repeating the previous task-set. In the past two decades, neuroimaging studies have identified that task switching relies on the control processes for multiple task-set components, which are mediated through distinct nodes of a generic fronto-parietal control network. Furthermore, studies have proposed that control functions rely on transitory changes in patterns of cooperation and competition between large-scale brain networks. Specifically, there are two major cognitive control networks supporting the control of task-sets; namely, the fronto-parietal network(FPN) and cingulo-opercular network(CON). As a typical control process of task-sets, task switching should also rely on the interactions between large-scale brain networks. However, there is little known about the specific dynamical interactions between the large-scale functional brain networks during task switching. Furthermore, a series of studies have regarded the left inferior frontal junction(IFJ) as a crucial region in task switching, supporting the general task-set updating. However, the importance of left IFJ for the control of task-sets hasn’t shown in large-scale networks. The reason could be that the large-scale brain network dynamics show context-dependent nature and a direct investigation of the switching-dependent changes in the large-scale network dynamics is not yet undertaken. We think that the task-set updating may rely on the interactions between IFJ and other task-set regions.To characterize the changes in functional network dynamics during task switching, we utilized a typical task switching paradigm and conducted an event-related functional magnetic resonance imaging(fMRI) experiment to identify the switching-related regions, and then examined the switching-induced changes in connectivity between these regions. Switching-related regions were defined by the positive effect of switching condition minus repetition condition in the general-linear model(GLM) analysis, and the interactions between these regions were examined through the multiregional psychophysiological interaction(PPI) analysis which displayed superiority in exploring the context-dependent changes of functional brain networks. In the current study, both behavioral and fMRI data were recorded from twenty-six healthy participants when switching between two tasks. Participants showed slower response and lower accuracy in switching trials than repetition trials, suggesting a typical switch cost. The comparison between switching trials and repetition trials activated a set of brain regions encompassing left dorsolateral prefrontal cortex(DLPFC), bilateral anterior insula(AI), anterior cingulate cortex(ACC), bilateral dorsal premotor cortex(dPMC), left IFJ, bilateral inferior parietal lobule(IPL), bilateral intraparietal sulcus, bilateral superior parietal lobule(SPL) and bilateral occipital cortex. These results were in line with the activated regions in a majority of task switching researches and showed higher brain activations on the switching trials than the repetition trials, which was consistent with behavioral results. Then, we performed multiregional PPI analysis to examine the effective connective between 10 seed regions from the activation results. PPI analysis revealed a significant change in effective connectivity as the effect of switching in all these ten regions which overlapped with the large-scale functional networks of both the FPN and CON. In this connectivity pattern, the connections of CON were almost output toward other regions, and except the connections from CON to FPN, the connections of FPN almost reflect the interaction within these FPN nodes and motor regions. Furthermore, the left IFJ appears a target region for CON inputs from the ACC and bilateral AI; the left AI exerts influence on other regions including FPN nodes and motor regions. This pattern suggested the left IFJ played an important role in the interaction between networks. The correlation results can further illuminate the functional characteristics of the connectivity patterns. The behavioral switch cost was negatively correlated with the connectivity variation of all the connections and with the connectivity variation of the connections between IFJ and other regions. However, this negative correlation of the left AI’s did not appear. Results suggested that the CON is responsible for the detection of switching and the initiation of control, whereas the PFN receives the input from CON and leads the implementation of control during task switching. Task switching relies on the changes in the functional brain network dynamics where the left IFJ serves as a key hub for the interaction between networks. Our findings further provide novel insights into the understanding of the neural mechanisms of task switching.