The Effect Of Switching Between Neutral And Emotional Stimuli: An ERP Study Of "Switch Cost"

ObjectiveAdaptive behaviour requires an executive control system that allows us to respond flexibly to changing environmental events . The ability to alter behaviour on the basis of changing requires working memory. Working memory which can maintain and manipulate representations is the cornerstone of reasoning, decision making and high level cognitions ( Baddeley & Hitch 1974). The question how working memory deal with multiple internal representations is much concerned.. Garavan's (1998) Experiment is useful and important to study inner attention switching: as showing a sequence of figures , either rectangles or triangles, participants are asked to keep a running count of the two categories. Reaction times were larger following a switch of two figures than following a repetition. He assumed that the cost for time named " switching cost" is an index of the inner attention shifting from one counter to the other within working memory. The research suggested that the internal focus of attention was limited in capacity.The present study represents if switching cost exists when a sequence is comprised with emotional stimuli and alters when shifting orientation of emotional faces and neutral ones.Scalp - recorded event - related potentials ( ERPs) provide a method to explore the fast brain underlying attention and working memory processes in humans. In my study that was modified the Garavan task with expression, we can see a clear evidence of ERP activation.Methods SubjectsFourteen college students, seven male and seven females (age-. 21 ±0.7) with normal vision. All were right - handed and no history of neurological or psychiatric disorder.StimuliStimuli consisted of schematic lines of faces with emotional and neutral expression. They were presented on a background. Only two type counting trail lists: one is a random sequence of faces, either positive or neutral expressions; the other is a random sequence of faces, either negative or neutral expressions. The task was subdivided into four blocks in the each of which ten counting lists were randomized. The total number of stimuli presented in every trail list ranged from 16 to 20. The appearance of same stimulus was likely less than four times. The picture subtended a visual angle of 2. 4° . Stimulus duration is 200ms and the cue appears within an interval of 2500 - 3000ms.Task instructionSubjects maintained two running counts of emotional and neutral expressions during a trail lists. They were told to start each trail lists with both counts at zero and increment each counter upon the appearance of the stimulus, pressing the button to start the next trail only after they had updated the count. After looking at " counting over" participants report the total for the number of expression followed by the total for the number of the other one.ERP recordingEEG was recorded with 64 tin electrodes. Electrode impedances were kept below 5kΩ. EOG were recorded for eye blink correction. The EEG signal was amplified (band pass0.05 ~ 100Hz) digitized at 250/channel.Data analyzingEEG recordings were split up into time window of 1200ms including a 200ms baseline before the stimulus. Trails with EEG exceeding ±70μV in amplitude was discarded. We compute ERP averages for four types conditions: one (EE) is no-switching trails of emotional expressions; another (EN) is switching trails from emotional expressions to neutral ones; another ( NN) is no -switching trails of neutral expressions; last one ( NE) is switching trails from neutral expression to emotional ones. If we subtract EE from NE, we will get NEdifferent wave (NEDW). In the same way, ENDW will be gotten.N170, VPP, N2 and P3 amplitudes were subjected to a main analysis of ANOVA design with (EE/NE; NN/NE) as the repeated measures factor.ResultsBehavioral analysesBehavioral data was shown that the number of errors is quite low, averaging 1. 8 trail lists of incorrect counting in the whole task ( approximately 4. 1% ). The mean RTs show a significant switch cost; switch RTs ( 757 ±49. 7ms ) were slower than no switch RTs (631 ±43.4ms ). Furthermore, there was no apparent difference in behavioral performance when shifting from emotional expressions to neutral ones were compared with shifting from neutral expressions to emotional ones. ( F( 1,13 ) = 0.084, p = 0.7 )。ERP measuresWe first focused on four component of standard ERPs elicited by this task: the N170 ( temporal cortex) , vpp (parietal cortex) , N2, P3.The 2 ( channel) x 2 ( switch condition ) x 2 ( stimulus condition ) ANOVA showed that difference in N170 amplitudes ( F( 1,13 ) = 0. 558, p = 0.468) and latency(F( 1,13) =0. 389,p=0. 544) were not found by neutral and emotion faces. As were VPP amplitudes and latency.The analysis of a 2 (two types of difference waves) ×4(electrode side) repeated - measures ANOVA is: ①N2 amplitudes between EE and NE are different (F(l,13) =7.489,p =0.017) and electrodes have main effect(F( 1,23) = 7.957, p = 0.003)②NN of N2 amplitudes and EN of that have significant difference ( F( 1,13 ) = 9. 248, p = 0. 009) but interact with electrodes ( F( 1, 21) =6.874,p=0.008). ③P3 amplitudes was modified by an (EE/NE) two types x electrode interaction ( F( 1,24) = 3. 775 ,p = 0. 041)④The main effect of position was also observed ( F( 1,19 ) =15. 618, p < 0. 05 ) , with larger P3 amplitudes elicited by NN than EN(F( 1,13) = 8.833 ,p = 0.011) .In 350 - 400ms time window the main effect of two difference waves (NEDW and ENDW) is significant (F(1,13) =5.676,p =0.014). The scalp topography of the Nd370 activity is the center of parietal area.DiscussionThe tact that the N170 component, which reflects the structural encoding of faces, was entirely unaffected by facial expression in the present experiment also indicates that structural encoding and the analysis of facial expression may operate independently, and in parallel. Due to habituation and schematic lines of faces, subcortical pathway is not activated for emotional stimuli.The N2 of an ERP component is related to automatic processing of the physical mismatching between consecutive stimuli so it is possible that the difference of N2 is associated with mismatching between emotional expressions and neutral ones.One theory of the P300 posits that the P300 reflects " context updating" -executive control operation for updating the working memory representations that govern task performance. The amplitude of the P300 is proportional to the a-mount of working memory revision that is required. Therefore, it might also predict larger P300 on switching trails than no - switching trails. We did not see a relationship between the P3 amplitude and switching per se.In general, we observe the negative - going waves during 300 ~ 450ms. The results indicate that there is an negative component of ERP. The apparent peaking at 370 ms in NEDW and ENDW might be an index of " switch cost" shifting cognitive process. The nature of the counter - switching ERP effect lend some support to the conflict hypothesis. It is plausible that the negative - polarity activity ( Nd370) associated with counter switching. Many waves of ERPs effects for conflict processing such as N270 and N400. Compared with N270, N370 for delayed latencies is elicited by counting switch. However, N270 e-voked by information mismatching S1 - S2 during performance of vision memory task is the same to N2 of this study. N400 was originally discovered in response to processing incongruous semantic context. N400 generation is a kind of complicated processes with more profound information conflict. Nd370 is not related to semantic context but related to internal attention shifting. As to the scalp topography of the Nd370, the increase in parietal activation corresponding to the attention switching parameter suggests a parietal lobe involvement in executive...