| The rapidly processing of emotional stimuli is important, because emotional stimuli mean threat or reward in the environment. There are two viewpoints about emotion processing. On the one hand, some researchers proposed that the emotional stimuli are automatically processed, i.e. independent of attentional resources. On the other hand, some other researchers think that the processing of emotional stimuli is modulated by the cognitive load of current task. Working memory (WM) is the center of cognitive function, because the WM is the hub across the perception, long term memory, as well as human behavior. According to the WM model proposed by Baddeley, the WM system is constituted by four sub-systems, including central executive function, phonological loop, visuo-spatial sketchpad, and episodic buffer. Lots of studies were focus on the interaction between WM and emotional processing. Most of those studies were investigated the distractor effect of emotional stimuli on the information maintain, as well as the encoding of stimuli. The results of those studies, however, are contradictory. Some researchers have shown that emotional distractor improve the WM performance, whereas other researchers have shown that emotional distractor impair the WM performance. Furthermore, only few studies were investigated the emotional distractor effect on the central executive function of WM, including information refresh. To this end, the present thesis investigates the interaction between emotional distractor processing and information refresh in WM via four event-related potentials studies.In the first research, we investigated how the task irrelevant emotion interfere information refreshing in WM using the digital N-back task combine with event-related potentials recording. A mixed design emotional N-back was implemented with the WM load as block factor and the emotional distractor as trial-varied factor. The emotional distractor was presented behind the digit in each trial. In the 0-back condition, participants were asked to detect whether the current digit was a "1" or not. In the 2-back condition, they were asked to detect whether the current digit had appeared two positions back or not. Participants were asked to focus on the digits and ignored the emotional picture. On the behavioral level, a significant WM load effect was observed. The reaction time and error rate for the 2-back task was larger than for the 0-back task. No significant interaction between WM load and valence was observed. On the neural mechanism level, a significant WM load effect was observed for the P3 component which was related to the information refresh in an N-back task. The average amplitude of P3 was larger for the 0-back task than for the 2-back task, which was in line with previous studies. Moreover, a significant interaction between WM load and distractor valence was found. The negative distractor condition elicited larger P3 comparing to the neutral distractor during the 0-back task, whereas there were no significant differences between two kinds of distractor during 2-back task. These results support the dual competition model which posited that the emotional interference effect is modulated by the cognitive load.In the second study, we investigate how the acute stress affects the information refreshing in WM. The stress was induced by showing short movie clips containing scenes with extremely aversive content. Two groups, i.e., stress group and control group, of participants were asked to perform a digital N-back task which was same as paradigm used in study one. The stress group showed significantly increased negative emotion and heart rate relative to control group, which suggested that the stress movie was successfully induced stress. On the behavioral level, similar with study one, a significant main effect of WM on reaction time and error rate was found. Furthermore, a significant main effect of WM was found on the P3 component, the amplitude for the 0-back task was larger than for the 2-back task. A significant interaction between WM load and stress was found. The amplitude for stress group was larger than for the control group during 0-back task, whereas this difference was not significant during 2-back task. Those results were in line with results found in the first study. All the results support the dual competition model of cognition and emotion interaction.In the third study, we investigate the interaction between emotion and WM during an emotional WM task with fearful and neutral face expression as stimuli. Participants were asked to perform 0-back or 2-back task for the expression. During the 0-back condition, participants were asked to detect whether the current facial identity was the target face which presented in the beginning of each block or not. During the 2-back condition, participants were asked to detect whether the current facial identity was showed two positions back or not. The event-related potentials were recorded during the task. The behavioral results showed that there was a significant main effect of WM load, the reaction time for the 2-back was longer relative to 0-back task, and the error rate was for 2-back task was higher comparing to the 0-back task. No significant interaction between WM load and facial valence was found. On the ERP level, there was a significant main effect of WM load on N170; the 2-back task elicited larger N170 than the 0-back task. There was also a significant main effect of facial valence on N170; the fearful expression elicited larger N170 than the neutral expression. Furthermore, there was a significant interaction between WM load and facial valence. The fearful expression elicited larger N170 than the neutral expression during both 0-back and 2-back condition, while the differences during 0-back condition was larger than during the 2-back condition. For another component, i.e., VPP, we found significantly main effect for the WM load and facial valence. The amplitude elicited by the 2-back task was larger than by the 0-back task, the fearful expression elicited larger VPP than the neutral expression. Moreover, for the WM related P3 component, there was a significant main effect of WM load, the average amplitude for the 2-back task was smaller than for the 0-back task. There was also a significant main effect of facial valence; the fearful expression elicited larger P3 than the neutral expression. However, we did not found any significant interaction between WM load and facial valence, which was contradicted with our prediction. These results suggested that the early visual representation of facial expression was modulated by the cognitive load. The WM was modulated by the facial valence under both WM load. In the last study, we investigated whether the positive and negative facial expression interact with the WM load in a similar pattern or not. Participants were asked to perform 0-back or 2-back task with fear, happy and neutral face as stimuli. The facial valence was manipulated as a block variant. So, there were six kinds of blocks including (1) 0-back neutral, (2) 2-back neutral, (3) 0-back fear, (4) 2-back fear, (5) 0-back happy, and (6) 2-back happy. The block order was counterbalanced across participants. On the behavioral level, a significant main effect of WM load was found, the reaction time for the 2-back was longer than for the 0-back task. The error rate for 2-back task was higher than for the 0-back task. No other significant main effects or interaction were found. On the ERP level, for the N170 component, we found a significant main effect of WM load, the amplitude for the 2-back task was larger than for the 0-back task. We also found a significant main effect of facial valence; the fearful expression elicited larger N170 than for the happy, as well as neutral face. There was a significant interaction between WM load and facial valence, fearful elicited larger amplitude of N170 than the happy and neutral face under both WM load, and the difference across three kinds of expression was larger in the 2-back task than in the 0-back task. For the VPP component, we found significant main effect of WM load. The amplitude was larger for the 2-back task than for the 0-back task. There was also a significant main effect of facial expression. The amplitude of VPP was larger for the fearful expression than for the happy face and neutral face. For the WM related P3 component, we found significant main effect of WM load. The P3 was larger for the 0-back task than for the 2-back task. A significant main effect of facial valence was observed. The happy face elicited smaller P3 than for the fearful and neutral face. No significant interaction between WM load and facial valence was observed. These results suggested that the representation of facial expression is modulated by the WM load.The results from these four study indicated that the interference effect of task irrelevant emotional distractor on WM is modulated by the WM load. The interference effect is higher when WM load is low, whereas the interference effect is lower when WM load is high. Furthermore, the interaction between WM load and emotion was also modulated by the emotion intensity. |
PDF Full Text Request