Task switching: effects of practice on switch and mixing costs

In the task-switching paradigm, mixing costs indicate the performance costs to mix two different tasks, while switch costs indicate the performance costs to switch between two sequentially presented tasks. Applying tasks with bivalent stimuli and responses, many studies demonstrated substantial mixing and switch costs and a reduction of these costs as a result of practice. The present study investigates whether extensive practice of a task-switching situation including tasks with univalent stimuli eliminates these costs. Participants practiced switching between a visual and an auditory task. These tasks were chosen because they had shown eliminated performance costs in a comparable dual-task practice study (Schumacher et al. Psychol Sci 12:101–108, 2001). Participants either performed the tasks with univalent responses (i.e., visual-manual and auditory-verbal stimulus–response mappings) or bivalent responses (i.e., visual-manual and auditory-manual stimulus–response mappings). Both valence conditions revealed substantial mixing and switch costs at the beginning of practice, yet, mixing costs were largely eliminated after eight practice sessions while switch costs were still existent.     

Automatic motor activation by mere instruction

Previous behavioral studies have shown that instructions about stimulus–response (S-R) mappings can influence task performance even when these instructions are irrelevant for the current task. In the present study, we tested whether automatic effects of S–R instructions occur because the instructed stimuli automatically activate their corresponding responses. We registered the lateralized readiness potentials (LRPs) that were evoked by the instructed stimuli while participants were performing a task for which those mappings were irrelevant. Instructed S–R mappings clearly affected task performance in electrophysiological and behavioral measures. The LRP was found to deflect in the direction of the response tendency that corresponded with the instructed S–R mapping. Early activation of the instructed response was observed but occurred predominantly on slow trials. In contrast, response conflict evoked by instructed S–R mappings did not modulate the N2 amplitude. The results strongly suggest that, like experienced S–R mappings, instructed S–R mappings can lead to automatic response activation, but possibly via a different route.     

The dual implication of dual affordance: stimulus-task binding and attentional focus changing during task preparation

In task switching experiments, comparing performance with bivalent stimuli (affording both tasks) to univalent stimuli (affording one task) confounds the need to change focus between dimensions and stimulus-task binding, because bivalent stimuli require focusing (and refocusing) but also appeared in the competing task before. To separate these influences, participants switched between vertical and horizontal judgments performed on bivalent (e.g., up-left) or univalent (e.g., left) actual locations or location words. In a critical condition involving bivalence without stimulus-task binding, actual locations and location words were each linked to a different task. Bivalence increased switch costs and preparation reduced switch costs only with bivalent stimuli. Stimulus-task binding affected performance in task repetitions, especially when little preparation time was afforded.     

The role of task rules and stimulus–response mappings in the task switching paradigm

Switch costs occur whenever participants are asked to switch between two or more task sets. In a typical task switching experiment, participants have to switch between two task sets composed of up to four different stimuli per task set. These 2 (task sets) x 4 (stimuli) contain only 8 different stimulus-response (S-R) mappings, and the question is why participants base their task performance on task sets instead of S-R mappings. The current experiments compared task performance based on task rules with performance based on single stimulus-response mappings. Participants were led to learn eight different S-R mappings with or without fore-knowledge about two underlying task sets. Without task set information no difference between shifts and repetitions occurred, whereas introducing task sets at the beginning led to significant switch costs. Most importantly, introducing task sets in the middle of the experiment also resulted in significant switch costs. Furthermore, introducing task rules at the beginning of the experiment lead to slower RTs when simple stimuli (Experiment 1) had to be processed. This detrimental effect disappeared with more complex stimuli (Experiment 2). Results will be discussed with respect to cognitive control.     

Task switching and the measurement of “switch costs”

The measurement of “switch costs” is held to be of interest because, as is widely believed, they may reflect the control processes that are engaged when subjects switch between two (or more) competing tasks. [In task-switching experiments, the reaction time (RT) switch cost is typically measured as the difference in RT between switch and non-switch (repeat) trials.] In this report we focus on the RT switch costs that remain even after the subject has had some time to prepare for the shift of task, when the switch cost may be approximately asymptotic (so-called residual switch costs). Three experiments are presented. All three experiments used Stroop colour/word, and neutral stimuli. Participants performed the two tasks of word-reading and colour-naming in a regular, double alternation, using the “alternating runs” paradigm (R. D. Rogers & S. Monsell, 1995). The experiments were designed to test the hypothesis that RT switch costs depend on a form of proactive interference (PI) arising from the performance of a prior, competing task. A. Allport, E. A. Styles and S. Hsieh (1994) suggested that these PI effects resulted from “task-set inertia”, that is, the persisting activation-suppression of competing task-sets, or competing task-processing pathways. The results confirmed the existence of long-lasting PI from the competing task as a major contributor to switch costs. Non-switch trials, used as the baseline in the measurement of switch costs, were also shown to be strongly affected by similar PI effects. However, task-set inertia was not sufficient to account for these results. The results appeared inconsistent also with all other previous models of task switching. A new hypothesis to explain these between-task interference effects was developed, based on the stimulus-triggered retrieval of competing stimulus-response (S-R) associations, acquired (or strengthened) in earlier trials. Consistent with this retrieval hypothesis, switch costs were shown to depend primarily on the S-R characteristics of the preceding task (the task that was switched from) rather than the upcoming task. Further, the effects of the other, competing task were found to persist over many successive switching trials, affecting switch costs long after the stimulus overlap (and hence the principal S-R competition) between the current tasks had been removed. Switch costs were also found to be affected by recent, item-specific experience with a given stimulus, in either the same or the competing task. Finally, the results showed that switch costs were massively affected by the ratio of the number of prior trials, in response to the same stimuli, that had implemented either the currently intended or the competing S-R mappings. None of these effects are predicted by current models of residual switch costs, which appeal to the differences in control processes assumed to be engaged in switch versus non-switch trials. Received: 31 March 1999 / Accepted: 23 July 1999     

Implicit task sets in task switching?

In 2 experiments, the authors compare stimulus-based versus task-rule-based task performance. Participants practiced 8 stimulus-response mappings either with or without knowledge about 2 underlying task sets. After practice, 2 transfer blocks with 8 new stimuli were presented. Results show that rule knowledge leads to significant switch and transfer costs, whereas without rule knowledge neither switch nor transfer costs occur. However, significant Task Type x Response Type interactions occurred in both conditions. In a second experiment including only the no rule condition, half of the stimulus-response mappings in the transfer blocks were incongruent to the underlying task rule. Slower response times for these incongruent stimuli as compared with congruent stimuli and the absence of switch costs suggest that participants acquired (presumably implicit) knowledge about 4 different stimulus-response categories.     

Preparing or executing the wrong task: The influence on switch effects

In a previous study, it was proposed that executing a task leads to task strengthening. In other words, task activation at the moment of response execution determines subsequent switch effects (Steinhauser & Hübner, 2006). The authors investigated this issue by comparing switch effects after task and response errors. However, the use of bivalent stimulus–response mappings might have obscured some of the effects. Therefore, we replicated the experiment using univalent stimulus–response mappings. With this adjusted design, which overcomes some shortcomings of the original study, we were able to replicate the finding of switch benefits after task errors. Closer inspection of the data showed the importance of preexecution processes on subsequent switch effects. In a second experiment, we further elaborated on these preexecution processes. More precisely, we investigated the effect of task preparation on subsequent switch effects. Taken together, our data extend current accounts of task switching by showing that the preparatory processes occurring before the response on trial n influence the switch cost on trial n?+?1.     

Effects of associative learning on age differences in task-set switching

Costs of switching between tasks may disappear when subjects are able to learn associations between tasks, stimuli, and responses (cf. Rogers, R. D., & Monsell, S. (1995). Costs of a predictable switch between simple cognitive tasks. Journal of Experimental Psychology: General, 124, 207-231). The first aim of this study was to examine this possibility by manipulating stimulus-set size. We expected that costs of switching between tasks would be strongly reduced under conditions of small stimulus-set sizes (n=4) as compared to large stimulus-set sizes (n=96) with increasing time on task. The second aim was to determine whether younger as well as older adults were able to create associations between task components. As age differences in task switching are often found to be larger when response mappings are incompatible we also investigated interactions with response compatibility. Results of our study indicated that practice effects on switch costs were much more pronounced for small than large stimulus-set sizes, consistent with the view that the strength of associations between task components facilitates task switching. Furthermore, we found that practice benefits on task switching for small stimulus-set sizes were sensitive to age and response compatibility. In contrast to younger adults, who showed a reduction of switch costs for both response mapping conditions, older adults showed a reduction of switch costs only when response mappings were compatible. That is, older adults showed less associative learning when the currently irrelevant task feature had to be suppressed, supporting the view that older adults have primarily problems in separating overlapping task-set representations.     

The bivalency effect: adjustment of cognitive control without response set priming

The occasional occurrence of bivalent stimuli, that is, stimuli with features relevant to two tasks, slows performance on subsequent tasks with univalent stimuli, including those which have no common features with bivalent stimuli (i.e., the “bivalency effect”). We have suggested that the bivalency effect might stem from an episodic context binding arising from the occasional occurrence of bivalent stimuli. However, as the same response set is used usually for univalent and bivalent stimuli, bivalent stimulus features may be negatively primed via response features. We investigated this possibility in two experiments, in which one group of participants used the same response keys for all tasks and another group used separate response keys. The results showed a comparable bivalency effect in both groups. Thus, it rather results from episodic context binding than from response set priming.     

The influence of cue-task association and location on switch cost and alternating-switch cost.

Task-switching performance is strongly influenced by whether the imperative stimulus uniquely specifies which task to perform: Switch cost is substantial with bivalent stimuli but is greatly reduced with univalent stimuli, suggesting that available contextual information influences processing in task-switching situations. The present study examined whether task-relevant information provided by task cues influences the magnitude of switch cost in a parallel manner. Cues presented 500 ms prior to a trivalent stimulus indicated which of three tasks to perform. These cues either had a preexisting association with the to-be-performed task (verbal cues), or a recently learned association with the task (spatial and shape cues). The results paralleled the effects of stimulus bivalence: substantial switch cost with recently learned cue-task associations and greatly reduced switch cost with preexisting cue-task associations. This suggests that both stimulus-based and cue-based information can activate the relevant task set, possibly providing external support to endogenous control processes. Alternating-switch cost, a greater cost for switching back to a recently abandoned task, was also observed with both preexisting and recently learned cue-task associations, but only when all tasks were presented in a consistent spatial location. When spatial location was used to cue the to-be-performed tasks, no alternating-switch cost was observed, suggesting that different processes may be involved when tasks are uniquely located in space. Specification of the nature of these processes may prove to be complex, as post-hoc inspection of the data suggested that for the spatial cue condition, the alternating-switch cost may oscillate between cost and benefit, depending on the relevant task.     

Is task switching nothing but cue priming? Evidence from ERPs

Recent findings suggesting that switch costs in the task-cuing paradigm are largely attributable to a change in the task-indicating cue have been interpreted in terms of a priming model of task-switch costs (Logan & Bundesen, 2003). According to this explanation, participants do not actually switch task sets, but merely use a cue-stimulus compound to disambiguate competing response tendencies associated with bivalent stimuli. Here, we report an event-related potential (ERP) experiment that provides evidence against this notion. In a paradigm with a 2:1 mapping between cues and tasks, we show that cue-switch and task-switch effects are dissociable on a neurophysiological level, indicating that task switching is more than a switch in the task-indicating cue. Moreover, a systematic analysis of the ERPs during the cue-stimulus interval suggests that updating processes can run in advance, before the stimulus is presented.     

Age equivalence in switch costs for prosaccade and antisaccade tasks

This study examined age differences in task switching using prosaccade and antisaccade tasks. Significant specific and general switch costs were found for both young and old adults, suggesting the existence of 2 types of processes: those responsible for activation of the currently relevant task set and deactivation of the previously relevant task set and those responsible for maintaining more than 1 task active in working memory. Contrary to the findings of previous research, which used manual response tasks with arbitrary stimulus-response mappings to study task-switching performance, no age-related deficits in either type of switch costs were found. These data suggest age-related sparing of task-switching processes in situations in which memory load is low and stimulus-response mappings are well learned.     

Voluntary task switching: chasing the elusive homunculus

In the voluntary task switching procedure, subjects choose the task to perform on a series of bivalent stimuli, requiring top-down control of task switching. Experiments 1-3 contrasted voluntary task switching and explicit task cuing. Choice behavior showed small, inconsistent effects of external stimulus characteristics, supporting the assumption of top-down control of task choice. Switch costs were smaller when subjects chose to switch tasks than when instructed by an external cue. Experiments 4-6 separated choice costs from switch costs. These findings support models of task switching that incorporate top-down processes in accounts of switch costs. The degree to which task switching procedures capture top-down versus bottom-up processes may depend on the extent of environmental support provided by the procedure.     

Keys and seats: Spatial response coding underlying the joint spatial compatibility effect

Spatial compatibility effects (SCEs) are typically observed when participants have to execute spatially defined responses to nonspatial stimulus features (e.g., the color red or green) that randomly appear to the left and the right. Whereas a spatial correspondence of stimulus and response features facilitates response execution, a noncorrespondence impairs task performance. Interestingly, the SCE is drastically reduced when a single participant responds to one stimulus feature (e.g., green) by operating only one response key (individual go/no-go task), whereas a full-blown SCE is observed when the task is distributed between two participants (joint go/no-go task). This joint SCE (a.k.a. the social Simon effect) has previously been explained by action/task co-representation, whereas alternative accounts ascribe joint SCEs to spatial components inherent in joint go/no-go tasks that allow participants to code their responses spatially. Although increasing evidence supports the idea that spatial rather than social aspects are responsible for joint SCEs emerging, it is still unclear to which component(s) the spatial coding refers to: the spatial orientation of response keys, the spatial orientation of responding agents, or both. By varying the spatial orientation of the responding agents (Exp. 1) and of the response keys (Exp. 2), independent of the spatial orientation of the stimuli, in the present study we found joint SCEs only when both the seating and the response key alignment matched the stimulus alignment. These results provide evidence that spatial response coding refers not only to the response key arrangement, but also to the—often neglected—spatial orientation of the responding agents.     

Positive affect facilitates task switching in the dimensional change card sort task: Implications for the shifting aspect of executive function

Using the modified Dimensional Change Card Sort task, we examined the influence of positive affect on task switching by inspecting various markers for the costs, including restart cost, switch cost and mixing cost. Given that the executive-control processes that underlie switching performance—i.e., inhibition or shifting—are distinct from the component processes that underlie non-switching performance—i.e., stimulus evaluation, resource allocation or response execution—we hypothesised that if positive affect facilitates task switching via executive-control processes, rather than via component processes, positive affect would reduce both switch and restart costs, but not mixing cost, because both switch and restart costs rely on executive processes, while mixing cost imposes only minimal demands on executive processes. We found beneficial effects of positive affect on both restart and switch costs, but not on mixing costs. These results suggest that positive affect improves switching abilities via executive processes rather than via component processes.     

Instruction-based response activation depends on task preparation

An increasing number of studies have demonstrated that a response in one task can be activated automatically on the basis merely of instructed stimulus–response (S–R) mappings belonging to another task. Such instruction-based response activations are considered to be evidence for the formation of S–R associations on the basis of the S–R mappings for an upcoming, but not yet executed, task. A crucial but somewhat neglected assumption is that instructed S–R associations are formed only under conditions that impose a sufficient degree of task preparation. Accordingly, in the present study we investigated the relation between task preparation and the instruction-based task-rule congruency effect, which is an index of response activation on the basis of instructions. The results from two experiments demonstrated that merely instructed S–R mappings of a particular task only elicit instruction-based response activations when that task is prepared for to a sufficient degree. Implications are discussed for the representation of instructed S–R mappings in working memory.     

No-go trials can modulate switch cost by interfering with effects of task preparation

It has recently been shown that the cost associated with switching tasks is eliminated following ‘no-go’ trials, in which response selection is not completed, suggesting that the switch cost depends on response selection. However, no-go trials may also affect switch costs by interfering with the effects of task preparation that precede response selection. To test this hypothesis we evaluated switch costs following standard go trials with those following two types of non-response trials: no-go trials, for which a stimulus is presented that indicates no response should be made (Experiment 1); and cue-only trials in which no stimulus is presented following the task cue (Experiment 2). We hypothesized that eliminating no-go stimuli would reveal effects of task preparation on the switch cost in cue-only trials. We found no switch cost following no-go trials (Experiment 1), but a reliable switch cost in cue-only trials (i.e., when no-go stimuli were removed; Experiment 2). We conclude that no-go trials can modulate the switch cost, independent of their effect on response selection, by interfering with task preparation, and that the effects of task preparation on switch cost are more directly assessed by cue-only trials.     

Logical recoding of S-R rules can reverse the effects of spatial S-R correspondence

Two experiments investigated competing explanations for the reversal of spatial stimulus—response (S—R) correspondence effects (i.e., Simon effects) with an incompatible S—R mapping on the relevant, nonspatial dimension. Competing explanations were based on generalized S—R rules (logical-recoding account) or referred to display—control arrangement correspondence or to S—S congruity. In Experiment 1, compatible responses to finger—name stimuli presented at left/right locations produced normal Simon effects, whereas incompatible responses to finger—name stimuli produced an inverted Simon effect. This finding supports the logical-recoding account. In Experiment 2, spatial S—R correspondence and color S—R correspondence were varied independently, and main effects of these variables were observed. The lack of an interaction between these variables, however, disconfirms a prediction of the display—control arrangement correspondence account. Together, the results provide converging evidence for the logical-recoding account. This account claims that participants derive generalized response selection rules (e.g., the identity or reversal rule) from specific S—R rules and inadvertently apply the generalized rules to the irrelevant (spatial) S—R dimension when selecting their response.     

A new look at “the hard problem” of bilingual lexical access: evidence for language-switch costs with univalent stimuli

Considerable work has used language-switching tasks to investigate how bilinguals manage competition between languages. Language-switching costs have been argued to reflect persisting inhibition or persisting activation of a non-target language. However, these costs might instead reflect the use of bivalent stimuli (i.e. pictures or digits that can be responded to in either language). That is, language-switching costs may simply reflect a cost of selecting the task-appropriate response for a given item and so may not be reflective of bilingual lexical access [Finkbeiner, M., Almeida, J., Janssen, N., & Carramaza, A. (2006). Lexical selection in bilingual speech production does not involve language suppression. Journal of Experimental Psychology: Learning, Memory, and Cognition, 32(5), 1075–1089]. The present study addresses this concern by having Chinese/English bilinguals switch between languages in response to inherently univalent stimuli (English words and Chinese Characters) as well as lexically univalent, but orthographically bivalent, stimuli (English words and Chinese Pinyin). Speakers showed switch costs when naming both univalent and orthographically bivalent stimuli, showing that switch costs can be found even with inherently univalent stimuli.     

More than a surprise: The bivalency effect in task switching

When switching tasks, if stimuli are presented that contain features that cue two of the tasks in the set (i.e., bivalent stimuli), performance slowing is observed on all tasks. This generalized slowing extends to tasks in the set which have no features in common with the bivalent stimulus and is referred to as the bivalency effect. In previous work, the bivalency effect was invoked by presenting occasionally occurring bivalent stimuli; therefore, the possibility that the generalized slowing is simply due to surprise (as opposed to bivalency) has not yet been discounted. This question was addressed in two task switching experiments where the occasionally occurring stimuli were either bivalent (bivalent version) or merely surprising (surprising version). The results confirmed that the generalized slowing was much greater in the bivalent version of both experiments, demonstrating that the magnitude of this effect is greater than can be accounted for by simple surprise. This set of results confirms that slowing task execution when encountering bivalent stimuli may be fundamental for efficient task switching, as adaptive tuning of response style may serve to prepare the cognitive system for possible future high conflict trials.