Effects Of Task Probability In Context Processing:Spatiotemporal Analysis Of ERP

ObjectiveContext processing involves the ability to use task-relevant information to determine an appropriate behavioral response. It contains selective attention, working memory, executive control and other cognitive components. AX-type continuous performance test consists of cue (A or B) and probe (X or Y) and shapes four kinds of trial:AX, AY, BX and BY. One was required to make response only when the X probe follow the cue A and inhibit response in the other three trials. Therefore, the cue A and the probe X are related to executive (Go) and the cue B and the probe Y are related inhibition (Nogo).Task probability usually leads to a different executive strategy. For the task of high reactive probability, subjects are likely to use reactive strategy based on the perceived external information. In contrast, for the task of low reactive probability the proactive strategy is more adopted since the maintenance of cue information needs more working memory.Concerning the influence of task probability, this research reveals the neural assemblies with spatiotemporal analysis of ERP statistical parametric mapping (SPM) technology.MethodsEighteen college students (7females) participated in the experiment without neurological or mental disorders. Their ages ranged from21to27(24.83±1.82). All of them are right-handed and twenty-twenty vision.Four symbols with colors are regarded as cue and probe. Among them, the blue "OOO" and "XXX" represent Go cue and Nogo cue respectively. The green and the red "<<<" or ">>>" corresponding to Go probe or Nogo probe separately. Cues and probes present successive in the center of screen for200ms, stimulus onset asynchrony (SOA) are1200ms between cue and probe,1300ms between probe and cue. The subjects were asked to press the button only when the Go-probe followed the Go-cue and withdraw response when the Go-probe or a Nogo-probe followed the Nogo-cue.We set three groups with complementary probabilities (AX/AY:0.55/0.15,0.35/0.35,0.15/0.55) for AX task and AY task and kept the probability0.15for both BX and BY ones. In accordance with the AX type probability, three groups were named AX55,AX35and AX15.Each group has a total of400trials, and was divided into four blocks. Trials appear in a random, but the same type trials consecutive appear no more than3times. The duration of each block is approximately4minutes, and there is a30seconds break between blocks. Subjects had10minutes practice for button pressing. The formal experiment lasts54min.EEG was recorded using an ERP system developed in our lab. It was sampled from19electrodes mounted in an elastic cap according to international10-20system referenced to both earlobes. Prefrontal midline site was grounding. The ERP epochs were extracted off-line and included pre-stimulus activity of100ms and post-stimulus activity of1200ms for the cues and1300ms for the probes. Baseline was taken from-100ms to0ms in considering the stimulus onset. Artifacts above70μV were eliminated automatically.One-way ANOVA of repeated-measure was carried out with SPSS13.0and Greenhouse-Geisser correction was used for the factor has3levels. The ERP data of each channel at every moment was submitted to the same statistical comparison. Conservative lower-bound epsilon (0.5) was applied to adjust the degree of freedom of the task factor. Statistical parametric mapping F-values [SPM(F)] was gained from interpolation calculated by each channel’s F-values. The significant level was0.05.Results1. Behavior performanceTask probability effect of reaction time (RT) is significant:F(1.93,32.79)=18.155, P=0.000. The followed LSD analysis confirmed significant differences between all paired groups. The probability effect of false alarm (%) of BX type is significant:F(1.23,22.85)=8.642, P=0.005; LSD comparisons show AX55group was significantly higher than the other two ones.2. Spatiotemporal patterns of SPM(F)The ERP effects revealed by SPM(F) occurred in many regions during both cue and probe stages:Cue A-the centroparietal area (300-350ms) and the left occipital area (450-500ms); Cue B-the right frontal pole (350-500ms) and the left one (750-950ms); AX probe-the right frontoparietal area (200-250ms), the left frontal pole, the right dorsal prefrontal and bilateral parietal region (300-400ms), bilateral occipital area and parietal region (400-650ms); AY probe-the frontocentral area (350-500ms)-the P3effect; BX probe-the right temporal and the left occipital areas (300-350ms); BY probe-the left temporal area (150-250ms).ConclusionIn line with previous studies, our finding suggests that the reaction time of AX task was shorter as the go-cue probability increased. We also observed that the false alarm (%) of BX task increased in the higher go-cue probability, showing that higher response probability enhances the difficulty of inhibition.1. ERP task probability effect during the cue stage,In three task groups conditions, the probability of A and B cues do not change, but X and Y probes are different in probability, so we still observed probability effect of two kinds of cues. In the low reaction probability (AX15) group, the effects in left prefrontal, cingulate and the centroparietal area (350-500ms), reflect that cue A required more in encoding to response to a rare probe X. There is no ERP effect later, representing that task probability has no effect in cue A maintenance. Nogo-cue B means subjects are required no response to probes X and Y. The CNV effect of nogo-cue B in bilateral frontal pole (350-1150ms) is probably related to different expectancy level due to the changing rate of X/Y.2. ERP task probability effect during the probe stageIn the AX case, expectancy level and mental demand can be changed by task probability. More attention and vigilance support for probe X in a low reaction probability task, so a parietal go-probe P3effect appears(300-400ms). In addition to this, we find the activation in bilateral occipital area (400-650ms), which may be working memory supported by top-down system.In the AY trials, without a N2effects this study observes a frontocentral P3effect (350-500ms), indicating that the frontocentral cortex is likely to be involved in the cancel of a prepared response and planning a new reaction.Novelty-routinization theory can be used to explain our finding in BX and BY trials. In the BX task, ERP effect appear in the left temporal-occipital area, may related to the processing of familiar stimulus, and in the right temporal region, may related to probe-processing activation in the right ventral attention system or even the frontal-parietal network. It is supported by the higher false alarm rate in BX trials as increasing response probability. In the BY case, ERP effect only present in the left temporal area, related to the processing of familiar stimulusPrevious studies rarely reported BY type and BX type ERP results. Although the probabilities of BX and BY type are the same (each is0.15) in three groups, we observed the ERP effects, suggest that probe X/Y overall probability change is the external reason for ERP effect of BX and BY task. Probe X/Y probability appeared as a whole context factors involved in adjusting cognitive resources of context processing. The results suggest that not only frontal area but the others cerebral regions are also influence by the whole or local probability of task.In a word, the increasing of response-task probability may lead to a shorter RT and a higher false alarm rate in suppressive task. ERP effects show task probability influence not only encoding of the go-cue, but also expectancy of the nogo-cue and adjusting neural assembliess of context processing of four types probes.