Response-related effects in task switching.
Sunday, October 02nd, 2011 [09:30 - 11:10]
SY_26. Response-related effects in task switching
University of Konstanz, Germany
Switching between tasks usually produces costs. Meanwhile, it is known that these costs are not caused by a single factor, but are due to various sources of interference such as task-set cueing, stimulus congruency, rule congruency, etc. Moreover, they are controlled by mechanisms such as backward inhibition, competitor rule suppression, response inhibition, etc. For most of these factors responses play a crucial role. This is not surprising, because task switching requires individuals to flexibly combine a fixed set of responses with a variable set of stimulus features. Especially if stimulus features related to different tasks are simultaneously present, selection conflicts are produced at the response level, the rule level, and even at the task level. The talks of this symposium are concerned with these conflicts and corresponding mechanisms. Their topics focus on effects closely related to responses. One effect is response-repetition (RR) costs, which are observed on task-switch trials and which contribute considerably to the task-switch costs. Several accounts have been proposed for their explanation. In the talks of Koch and Schuch and of Druey studies are reported that were designed to distinguish between two of these accounts: response-inhibition and the associative strengthening. In the study reported by Grzyb and Hubner the details of the previous-trial congruency effect were investigated, i.e. of the phenomenon that the size of RR costs depends on the response activation on the previous trial. Hubner and Grzyb take response inhibition as granted and show how the RR costs can be considerably increased by incongruent and rule-incongruent stimuli. They also discuss several possible mechanisms. Finally, Meiran and Hsieh extend the focus by considering negative repetition effects caused by the competition between whole task rules. They provide behavioral as well as electrophysiological evidence for the idea that that irrelevant task rules that generate response conflicts are inhibited.
SY_26.1 - Response representations and task representations in task switching
Koch, I. & Schuch, S.
RWTH Aachen University, Germany
Cognitive control mechanisms underlying flexible action can be examined using task switching methodology. There are (at least) two robust findings in task switching. First, a switch of tasks incurs performance switch costs. Second, repeating a response is beneficial in task repetitions but leads to costs in task switches. This interactive pattern of switching tasks and responses may be informative with respect to the underlying representations. Task representations ("task sets") probably include the mapping of stimuli to responses, suggesting task-specific response representations (e.g., pressing a left key may be used to represent a color feature in one task but a shape feature in another task). However, theoretical accounts differ in terms of whether they assume associative strengthening of task-specific response meaning or persisting inhibition of representations of just executed responses, which has different implications for the nature of task representations. In this presentation we discuss existing evidence and report new experimental data. We conclude that both accounts are not mutually exclusive, and we present new evidence for inhibitory mechanisms acting on response representations.
SY_26.2 - Dissociating learning and inhibition based accounts for response repetition effects in task switching
University of Zurich
In many studies it has been shown that response repetitions produce benefits in task repetition trials, but costs in task switch trials. So far, three main types of accounts have been proposed to explain this interaction of task and response sequence: First, models referring to reconfiguration mechanisms, second accounts based on the learning of associations and, third, priming and inhibition accounts. In the present study, the last two of these accounts were directly contrasted in one experiment by implementing all possible types of stimulus and response category repetitions. Whereas from both accounts costs are predicted for response category repetitions, they differ with respect to the predictions derived for stimulus category repetitions. These should also produce costs according to the learning accounts, but they should result in benefits according to the priming and inhibition accounts. The results show a clear pattern of benefits in case of stimulus category repetitions, and of costs in case of response category repetitions. The present results thus provide clear evidence for an explanation of the interaction of task and response sequence in terms of priming and inhibition.
SY_26.3 - The previous-trial congruency effect in task switching: Which property of response activation on trial n-1 determines the amount of response-repetition costs on trial n?
Grzyb, K. R. & Hubner, R.
University of Konstanz, Germany
Under task switching response repetitions (RR) are typically associated with costs compared to response shifts. Moreover, when the stimulus on the previous trial was congruent RR costs are usually larger than when it was incongruent. This previous-trial congruency effect has been explained by assuming that a response is generally inhibited after its execution and that the amount of inhibition depends on the activation of the response on the previous trial. Here, we investigated which property of the response activation on the previous trial is crucial in determining the amount of inhibition: the absolute activation of the correct response or the activation difference between the alternative responses. To differentiate between these two possibilities we compared RR costs after congruent, neutral, and incongruent trials. In two experiments we found similar RR costs after congruent and neutral trials, whereas the RR costs were smaller after incongruent trials. These results support the hypothesis that the amount of response inhibition is determined by the activation differences between the response alternatives on the previous trial.
SY_26.4 - The amplification of response-repetition costs by response and task competition
Hubner, R. & Grzyb, K. R.
University of Konstanz
It is a well-known phenomenon that response repetitions (RR) produces costs on task-switch trials. However, the size of these costs varies considerable between different experiments. This suggests that several factors contribute to these costs. Unfortunately, up to now these factors are largely unknown. Therefore, the goal of the present study was to examine the extent to which different factors affect the size of RR costs. Specifically, we investigated the role of congruency, i.e. whether the stimuli activate only the correct response or also the incorrect one, and of valence, i.e. whether the stimuli activate only the relevant task or also the irrelevant one. We report a series of experiments in which we show that these conflicting stimulus properties increase RR costs. With simple univalent stimuli RR costs are relatively small but still reliable, which indicates that responses are generally inhibited after their execution. However, by adding incongruent stimulus items or by using bivalent stimuli the small response inhibition effect can be amplified. Moreover, if both factors are combined, then RR costs can increase dramatically. This shows that response conflict and task conflict interact. Possible accounts of this interaction are discussed.
SY_26.5 - Resolving task rule incongruence during task switching by competitor rule suppression
Meiran, N. 1 & Hsieh, S. 2
1 Ben-Gurion University of the Negev, Beer-Sheva, Israel
2 National Cheng Kung University, Tainan, Taiwan
Task switching requires maintaining readiness to execute any task of a given set of tasks. However, when tasks switch, the readiness to execute the now-irrelevant task generates interference, as seen in the task rule incongruence effect. Overcoming such interference requires highly fine-tuned inhibition that impairs task readiness only minimally. In experiments involving two object classification tasks and two location classification tasks, the authors show that irrelevant task rules that generate response conflicts are inhibited. This Competitor Rule Suppression (CRS) is seen in response slowing in subsequent trials, when the competing rules become relevant. CRS is shown to operate on specific rules without affecting similar rules and to operate on the competing responses only when generated by the specific competing rule. CRS and backward inhibition, which is another inhibitory phenomenon, produced additive effects on reaction time, suggesting their mutual independence. Using Event Related Potentials, the authors show CRS to operate during the cue epoch, suggesting that it involves the rules rather than the responses. Implications for current formal theories of task switching are discussed.