Research ReportDiversity of the P3 in the task-switching paradigm
Highlights
► Task set switching increases the cue-locked centro-parietal P3 regardless of persisting response inhibition. ► Task set switching decreases the target-locked centro-parietal P3 regardless of persisting response inhibition. ► Persisting response inhibition reduces the frontal P3. ► The same pattern of centro-parietal and frontal P3 occurs in the in target- and response-locked ERPs. ► Task-set switching and response inhibition represent distinct functional mechanisms involving different brain areas.
Introduction
The task switching paradigm is a very useful tool for investigating a number of cognitive control processes in humans (Monsell, 2003). It has been used to examine the ability to alternate between two or more tasks, to prepare for a forthcoming task and to maintain a number of tasks in working memory. In recent times behavioral studies were increasingly complemented by electrophysiological recordings which allow systematic analyzing neural correlates underlying these cognitive mechanisms with a high time resolution (see Karayanidis et al., 2010 for overview).
In the task switching paradigm participants are asked to classify the same stimuli according to different rules on a trial by trial basis. The crucial outcome are the so-called switch costs reflecting longer reaction times and higher error rates in switch relative to repetition trials (Allport et al., 1994, Rogers and Monsell, 1995). In a cueing paradigm (Meiran, 1996) the relevant task is conveyed by an explicit cue stimulus, that is presented prior to target onset to allow sufficient task preparation. Although the task can be fully prepared in advance, significant residual switch costs are usually observed (Allport et al., 1994, De Jong, 2000, Rogers and Monsell, 1995), which were originally attributed to “task-set inertia”, that is persisting activation of a competing task set from the previous trial. In the current trial some additional time is needed to resolve the interference and select the response (Allport et al., 1994; see Kiesel et al. 2010 for overview).
Mayr and Keele (2000) showed that this interference involves an inhibitory component which persists from a previously inhibited task. Schuch and Koch (2003) investigated the inhibitory process using occasional no-go trials and observed no residual switch costs after no-go trials, suggesting that competition during response selection and/or activation triggers switch costs (see Koch et al. 2010 for overview). In our recent report (Gajewski et al., 2010a) we investigated the impact of response inhibition on residual switch costs using the same paradigm. We focused on the frontocentral negative ERP-component the N2 and found a relationship between the amplitude and latency of the N2 on the one hand and residual costs on the other. We concluded that proactive task set interference has to be resolved during response selection, leading to an enhanced and delayed N2 and consequently enhanced residual switch costs. Interestingly, two earlier studies conducted by Hsieh and Yu, 2003, Hsieh and Liu, 2005 investigated stimulus-locked LRPs (S-LRP) that have been also related to central response-selection processes before the motor response. Both reports showed delayed S-LRP and RTs for task switch relative to task repetition. This pattern was interpreted in terms of a carry-over effect existing at the response selection stage which accords with our proposal.
Two recent ERP studies investigated inhibitory processes in the task switching go no-go paradigm and focused mainly on the P3 like positive waves. Astle et al. (2006) addressed the question whether response execution in the previous trial (go vs. no-go) differently influences task preparation. They replicated the behavioral findings obtained previously and observed an increase of the parietal P3 (termed late parietal positivity; LPP) during task preparation in switch relative to non-switch trials regardless of whether the previous trial was a go or no-go, and a late frontal negativity (LFN) on go following go trials only. The authors proposed that the inhibition from the previous trial was overcome already during preparation for the following task as reflected in the LFN. The second ERP study conducted by Jamadar and coworkers (2010) used basically the same paradigm and replicated the behavioral results. Additionally, the authors analyzed cue-locked, target-locked and response-locked ERPs. In the cue-locked data again a larger LPP was found for switch than non-switch trials and no impact of go vs. no-go in n-1 was found, corroborating the finding yielded by Astle et al. (2006). Thus, the authors proposed that no-go in n-1 cannot contribute to the sequence effects as the inhibition in n-1 should affect the task activation already in the preparation interval. Finally, in contrast to previous findings (e.g. Karayanidis et al., 2003, Nicholson et al., 2005) only a marginal switch effect on the LPP was found in the target-locked data which was also not modulated by go or no-go in n-1. However, the crucial difference between the study provided by Astle et al., 2006, Jamadar et al., 2010 on the one hand and Schuch and Koch, 2003, Gajewski et al., 2010a on the other hand was the usage of different no-go stimuli. Whereas the former used unspecific no-go stimuli which did not indicate a response, the latter used no-go signals which were accompanied by specific target stimuli (digit) that enabled response selection.1 Thus, differences in the response selection process in a previous no-go trial may be crucial for the divergent explanation of residual switch costs provided by Schuch and Koch, 2003, Koch and Philipp, 2005, Astle et al., 2006, Jamadar et al., 2010.
A systematic analysis of the well established late positive ERPs (P3 family) as a function of previous or current informative no-go trial should help to disentangle task and response selection processes from response inhibition that may shed light on the still unresolved relative contribution of these functions to task switching.
Therefore, in the present study, we reanalyzed the data presented in Gajewski et al. (2010a) and focused on the impact of response selection and/or execution on the late positive ERPs during task preparation, implementation and execution. In order to do this, we analyzed frontal and centro-parietal positivity, the P3 (see Footnote 1 in Jamadar et al., 2010 regarding different labels) in task-set switch and repetition trials as a function of response mode (go vs. no-go) in n-1 in cue-, target- and response-locked ERPs in trial n.
As outlined above, in the cue-locked ERPs previous research consistently found a prominent parietal positivity (LPP or P3) which was more pronounced for task-switching than task-repetition trials (Barceló et al., 2000, Barceló et al., 2002, Jost et al., 2008, Karayanidis et al., 2003, Karayanidis et al., 2010, Kieffaber and Hetrick, 2005, Lorist et al., 2000, Nicholson et al., 2005, Nicholson et al., 2006, Rushworth et al., 2002). This cue-locked positivity consists of a number of overlapping potentials reflecting reconfiguration of stimulus sets, response sets and the mapping between them (see Karayanidis et al., 2010 for an overview), which was also interpreted in terms of task-set updating in working memory (e.g. Barceló et al., 2000, Barceló et al., 2002) whenever a model of the environment requires revision (Donchin and Coles, 1988).
In contrast, the late parietal positivity in target-locked ERPs was found to be consistently smaller in switch than in non-switch trials (e.g. Barceló et al., 2000, Barceló et al., 2002, Gajewski et al., 2010b, Hsieh and Liu, 2009, Karayanidis et al., 2003, Kieffaber and Hetrick, 2005, Lorist et al., 2000, Rushworth et al., 2002). Lorist et al. (2000) suggest that during switch trials cognitive resources are addressed to a greater amount than in repetition trials; as the P3 amplitude is believed to decrease when target processing involves more effort, i.e. when working memory resources are more heavily demanded (Kok, 2001, Wickens et al., 1983). Another explanation is proposed by Barceló et al. (2000) who observed a progressive increase of the parietal P3 in repetition trials following switch trials. They concluded that the representation of a task is initially weak after a switch and is strengthened by successive repetition trials. Finally the amplitude reduction of the parietal P3 can be explained in terms of its sensitivity to a response-related process (Verleger et al., 2005), which would propose that a complex cognitive operation (e.g. task switch) produces more variability in the data than a simple operation (e.g. task repetition). The greater the variability in the P3 the broader and smaller should be its average waveform (Falkenstein et al., 1993, Falkenstein et al., 1994).
The P3 sensitivity to response-related processes can be investigated in response-locked ERPs that are related to the point of time of a correct response (Verleger et al., 2005). There are only a few reports investigating response-locked ERPs in the context of task switching. Friedman et al. (2001) investigated the midfrontal negativity (MFN) appearing shortly after a response. Interestingly, this negativity was preceded by a clear positivity which was more pronounced for repetition than switch trials. Hsieh and Yu, 2003, Hsieh and Liu, 2005 investigated lateralized readiness potentials (LRP) that reflect an index of hand-specific response preparation. They showed that there was no impact of task switching on the response-locked LRP when participants switched between simple stimulus–response mappings (Hsieh and Yu, 2003); but significant differences occurred when the response demands were increased (Hsieh and Liu, 2005). Jamadar et al. (2010) obtained a positive wave in the response-locked ERPs that clearly differentiated between go and no-go trials in n-1 already 600 ms prior to an overt response that tended to be less positive in task switch than repetition trials in the Exp. 2.
The present study aims at systematically investigating late positive ERPs occurring in the task-switching paradigm. These positive waves, termed P3 in the following, are analyzed as a function of three events, namely preparation (cue-locked ERP), implementation (target-locked ERP) and execution (response-locked ERP) of a task switches and repetitions. Additionally, effects of previous response inhibition on the P3 in the current trial will be assessed. We hypothesize that according to previous reports (Astle et al., 2006, Jamadar et al., 2010) the cue-locked ERP should reveal a large, long lasting parietal P3 that is enhanced when a cue indicates a task switch than repetition. Moreover, this parietal P3 should remain unaffected by persisting response inhibition from the previous trial. At frontal electrodes, however, the P3 should be abolished after response inhibition in the previous trial (cf. LFN, Astle et al. 2006).
According to previous results, in the target locked ERPs, a reduced parietal P3 during task switching relative to task repetition should be observed. Moreover, persisting inhibition from previous trial may affect the P3 at frontal but not parietal electrodes.
If the P3 is associated with response-related processes during task switching, it should also occur in response-locked ERPs and show similar morphology and functional properties as the target-locked P3. Moreover, if the switch-related reduction of the P3 is merely due to latency jitter in the target-locked ERPs, the amplitudes in task-set switch and non-switch should not differ in the response-locked ERPs. Additionally, we expect that persisting response inhibition from n-1 affects the response-related P3 also at frontal but not parietal sites.
Finally, we expect a modulation of the parietal P3 in task switch trials also in a no-go trial when the accompanying target stimulus indicates a particular response. As the task-set switch effect on the P3 is mostly parietally distributed, we did not expect any modulation of the frontal “inhibitory no-go P3”.
Section snippets
Behavioral data
In 1% of all trials participants responded faster than 100 or slower than 2500 ms. The ANOVA revealed a main effect of task-set transition (F(1, 16) = 16.8, p < .001) and an effect of response mode in n-1 (F(1, 16) = 7.3, p = .015). Moreover, response mode in n-1 and task-set transition interacted significantly: while there were substantial costs after go trials (90 ms), these were absent (6 ms) after no-go trials (F(1, 16) = 27.7, p < .0001). The RTs in switch trials were longer than in repetitions after go
Discussion
The aim of the present study was to investigate the functional properties of the P3 as a function of task-set switching and response inhibition during task preparation, implementation and execution. The behavioral data replicated previous observations indicating reliable switch costs in trials following go-, and no switch costs after no-go trials (Astle et al., 2006, Jamadar et al., 2010, Kleinsorge and Gajewski, 2004, Koch and Philipp, 2005, Schuch and Koch, 2003). These behavioral results go
Conclusions
Taken together, in the current study we replicated well established findings and obtained a number of new findings regarding the P3 in the task-switching paradigm. In the preparation interval, the parietal, long-lasting P3 enhancement for switching relative to repeating cues was observed both in trials following go as well as no-go trials, that is, in trials which lead to a loss of a repetition benefit. The cue-locked P3 was more delayed after go than no-go trials. Thus, we assume that this
Participants
Seventeen undergraduate students (6 male), aged 20–35 (mean age 24 years) from the University of Dortmund participated in the study for course credit. All subjects were right handed and had normal or corrected-to-normal vision. They had no prior exposure to the paradigm and were not informed about the aim of the study.
Stimuli and tasks
The experiment was conducted with subjects seated in a dimly illuminated, electrically shielded room in front of a computer monitor at a distance of 80 cm. The stimuli were
Acknowledgments
We thank Claudia Wipking for running the experiment, Ludger Blanke for providing the software and technical support and Ines Mombrei for assistance with manuscript preparation. The research reported in this article was supported by grant Kl 1205/3-3 of the Deutsche Forschungsgemeinschaft to Thomas Kleinsorge (Ifado).
References (40)
- et al.
Dissociating neural indices of dynamic cognitive control in advance task-set preparation: an ERP study of task-switching
Brain Res.
(2006) - et al.
Attentional set shifting modulates the target P3b response in the Wisconsin card sorting test
Neuropsychologia
(2000) - et al.
Late visual and auditory ERP components and choice reaction time
Biol. Psychol.
(1993) - et al.
Effects of choice complexity on different subcomponents of the late positive complex of the event-related potential
Electroencephalogr. Clin. Neurophysiol.
(1994) - et al.
The novelty P3: An event-related brain potential (ERP) sign of the brain's evaluation of novelty
Neurosci. Biobehav. Rev.
(2001) - et al.
ERP-correlates of response selection in a response conflict paradigm
Brain Res.
(2008) - et al.
Electrophysiological correlates of residual switch costs
Cortex
(2010) - et al.
Effects of aging and job demands on cognitive flexibility assessed by task switching
Biol. Psychol.
(2010) - et al.
A new method for off-line removal of ocular artifact
Electroencephalogr. Clin. Neurophysiol.
(1983) - et al.
Switching between simple response-sets: inferences from the lateralized readiness potential
Brain Res. Cogn. Brain Res.
(2003)