Context-Sensitive Adjustments of Cognitive Control: Conflict-Adaptation Effects Are Modulated by Processing Demands of the Ongoing Task

Transcription

1 Journal of Experimental Psyhology: Learning, Memory, and Cognition 2008, Vol. 34, No. 3, Copyright 2008 by the Amerian Psyhologial Assoiation /08/$12.00 DOI: / Context-Sensitive Adjustments of Cognitive Control: Conflit-Adaptation Effets Are Modulated by Proessing Demands of the Ongoing Task Rio Fisher, Gesine Dreisbah, and Thomas Goshke Tehnishe Universität Dresden Dynami adjustments of ognitive ontrol in response to interferene from irrelevant stimulus attributes have repeatedly been shown. The purpose of the urrent researh was to investigate how these ontrol adjustments are modulated by the proessing demands of a primary task. To this end, the authors ombined a primary task (a number omparison task: lassifying digits as smaller or larger than 5) with a Simon task. Control adjustments were observed in the form of typial sequential modulations of the Simon effet. In addition, the authors found sequential modulations of the numerial distane effet and an interation of both effets. Results suggest that not only response onflit due to interferene from task-irrelevant features but also proessing demands of task-relevant features determine the level of ontrol adjustment in the subsequent trial. Keywords: ognitive ontrol, onflit, Simon task, sequene effets A fundamental ability of the ognitive system is to maintain urrent goal representations in the fae of distration. Seletive attention enables us to selet task-relevant stimulus features while suppressing distrating information, alternative goals, and ompeting response tendenies. Two entral, yet unresolved, questions are how attentional resoures are oordinated and how ognitive ontrol proesses, whih serve to shield the ognitive system from distrating information, are flexibly adjusted to hanging task demands (f. Goshke & Dreisbah, 2008; Miller & Cohen, 2001). To date most studies on the ontext-sensitive adaptation of ognitive ontrol have foused on the role of response onflit in triggering enhaned mobilization of ognitive ontrol (e.g., Gratton, Coles, & Donhin, 1992; Kerns et al., 2004). Response onflit ours whenever a stimulus omprises features that are assoiated with both the orret and the ompeting response. Aording to an influential theory (Botvinik, Carter, Braver, Barh, & Cohen, 2001; Botvinik, Cohen, & Carter, 2004), response onflit in a given task indues enhaned mobilization of ognitive ontrol on the subsequent trial of the task in order to prevent interferene. Consistent with this hypothesis, researh has shown that interferene from distrators in speeded hoie reation tasks is redued on trials following response-inompatible (high onflit) trials, ompared to trials following ompatible (low onflit) trials (Botvinik, Nystrom, Fissell, Carter, & Cohen, Rio Fisher, Gesine Dreisbah, and Thomas Goshke, Department of Psyhology, Tehnishe Universität Dresden, Dresden, Germany. This researh was supported by the German Researh Foundation (DFG) Grant Go 720/3-3. We thank Moritz Walser for data olletion, John Dunlosky and Mihael Masson for onstrutive omments on an earlier version of the artile, and Franziska Plessow for valuable disussions. Correspondene onerning this artile should be addressed to Rio Fisher, Department of Psyhology, Tehnishe Universität Dresden, Zellesher Weg 17, Dresden, Germany fisher@psyhologie.tu-dresden.de 1999; Gratton et al., 1992; Kerns et al., 2004). Suh onflitadaptation effets suggest that response onflits indue enhaned target proessing and/or distrator suppression on the following trial (but see Mayr, Awh, & Laurey, 2003). Here we investigated three hitherto open questions onerning the ontext-sensitive adjustment of ognitive ontrol in seletive attention tasks: (a) Is response onflit a neessary preondition for subsequent ontrol adjustments to appear? Or is it possible to obtain similar ontrol adjustments when proessing demands in the primary task inrease (f. Dreisbah & Haider, 2006)? (b) How do differenes in proessing demands in the primary task affet the ativation of (response) onflit-triggered ontrol adjustments? () How do proessing adjustments due to differential demands in a primary task interat with (response) onflit-indued ontrol adjustments? 1 To address these questions, we manipulated the attentional resoure requirements in terms of different proessing modes in a primary task and investigated their diret effets on the effiieny of the onflit-indued ontrol adjustments. More speifially, we ombined a number size judgment task, whih served as the primary task, with a Simon task (Simon, 1990) that is known to indue robust response onflits. Partiipants had to deide whether a given number was smaller (left key) or larger than 5 (right key). There is broad evidene that task demands in the number omparison task vary as a funtion of the numerial distane to the omparison level (e.g., Dehaene, Dupoux, & 1 For reasons of larity, we use the term proessing adjustment whenever differenes in primary task proessing requirements lead to adjustments of proessing in the subsequent trial. In this respet, proessing demands are supposed to imply that the effort or resoure investment to suessful task aomplishment may vary even within the primary task irrespetive of additional task-irrelevant features. In ontrast, we use the terms onflit or ontrol adjustment when irrelevant stimulus features imposed onflits that triggered ontrol regulations, irrespetive of primary task demands. 712

2 BRIEF REPORTS 713 Mehler, 1990; Moyer & Landauer, 1967). Oriet, Tombu, and Jolioeur (2005), for example, assumed that the mental representation of digits from the omparison digit of 5 (e.g., 1 and 9) are tagged as small or large and are therefore preategorized as large or small in an automati fashion. In ontrast, size judgments for digits that are to 5 rely on a slower and more ontrolled omparison proess that requires resoures. At the same time, presenting digits either to the left or right of the sreen enter indues stimulus response ompatibility (so-alled Simon interferene), leading to response onflit, when the irrelevant stimulus attribute (loation) does not orrespond with the required response to the relevant attribute (e.g., a small number presented at the right). This allowed us to investigate whether and how different proessing modes (more easy and automati preategorization vs. more effortful and ontrolled omparison) in trial N 1 (i.e., the numerial distane [ND] effet) modulate response-onflitindued adjustment of ognitive ontrol on the subsequent trial and vie versa: whether and how a response onflit in trial N 1 modulates the ND-indued proessing adjustment on the subsequent trial. On a more general level, we thus are interested in how the ognitive system oordinates the assignment of attentional ontrol in the fae of distration from task-irrelevant stimulus features, on the one hand, and in the differential proessing requirements during the proessing of task-relevant stimulus features in the primary task, on the other. Given our general assumption that any ontrol and proessing resoures are oordinated to serve the goal of performing the primary task, we expet that a more effortful and ontrolled proessing mode in the ND task ( numbers in trial N 1) would bind attentional resoures, whih might be drawn from the pool needed for the suffiient ativation of ontrol adjustments following an experiened response onflit in Simon-inompatible trials N 1. We thus assume that these dynami adjustments of ognitive ontrol on trial N (refleted in redued Simon interferene) are modulated by proessing modes on trial N 1. That is, the sequential modulation of the Simon effet (Notebaert, Soetens, & Melis, 2001; Stürmer, Leuthold, Soetens, Shröter, & Sommer, 2002; Wühr & Ansorge, 2005) should be weaker when responding to than to digits on trial N 1 (this should show up in a ND N 1 Simon N Simon N 1 interation, whih we refer to as ND N 1 Simon-modulation N hereafter). Analogously, a response onflit due to Simon inompatibility in the previous trial N 1 typially triggers adjustments of ognitive ontrol that serve to redue interferene from irrelevant stimulus attributes (i.e., stimulus loation) in the subsequent task in trial N. Now, two possibilities are oneivable as to how these ontrol mehanisms might also alter the number omparison proess in the primary task of trial N. First, a heightened reruitment of ontrol, serving to redue interferene in trial N, might benefit all proesses that are related to target proessing (e.g., Egner & Hirsh, 2005). Aording to this possibility, the number omparison proess should espeially benefit from heightened ontrol when ategorizing numbers (ontrolled proessing mode). In ontrast, it is possible that both ontrol mehanisms ompete for the same resoures. From this perspetive, the (Simon-indued) enhaned ontrol, serving interferene redution, might even diminish the effiieny of the number omparison proess, espeially when proessing numbers, but less so when ategorizing numbers. In any ase, a response onflit in N 1 should modulate the ND effet in N (this assumption is refleted in a Simon N 1 ND N interation). Note, however, that the diretion of this interation will distinguish between both possibilities. Furthermore, if our assumption proves right and adaptations of ognitive ontrol within seletive attention tasks are not restrited to experiened response onflits, then we should find sequential modulations of the ND (analogous to the Simon effet), resulting inand N 1 ND N interation. A repeated appliation of the slower ontrolled omparison proess, for example, should be effetive when the subsequent target is again but should lead to performane osts when the subsequent target is from 5. This would result in a smaller ND effet. By ontrast, when ategorizing digits that are from 5, a repeated appliation of the fast and automati ategorization proess would result in benefits if the subsequent target is again from 5, whereas osts should emerge when digits follow. In onlusion, we expet a larger ND effet when the preeding trial ontains rather than digits (i.e., an ND N 1 ND N interation). Partiipants Method Fifteen students (9 women, 6 men; mean age 23.2 years) at the Tehnishe Universität Dresden took part in the experiment. All had normal or orreted-to-normal vision. Partiipants attended a single experimental session lasting about 50 min and reeived a 5 payment. Apparatus and Stimuli The numerals 1 to 9 (exept 5) served as stimuli and were displayed on a 17-in. (43.18 m) monitor. Stimuli were presented in Arabi notation 2.8 m either to the left or to the right of a fixation (plus sign), whih resulted in a visual angle of 2.7 to the left and right, respetively (viewing distane 60 m). The digits were presented in white against a blak bakground and subtended a visual angle of about 0.48 to Responses were made with the index finger of the right hand pressing the period (.) key or with the left index finger pressing the y key of the standard German (QWERTZ) omputer keyboard. Proedure Partiipants were asked to ategorize digits as either smaller or larger than 5. Instrutions emphasized speed and auray to the same extent. Small numbers were responded to with the left index finger and large numbers with the right index finger. The stimulus response mapping was kept onstant aross partiipants. Eah trial began with the presentation of the fixation sign for 1,000 ms. Following a blank sreen of 600 ms, a digit was presented for 200 ms. In the ase of orret responses, a blank sreen was provided as feedbak for 300 ms. If no response was given 1,800 ms after stimulus onset, the feedbak zu langsam (too slow) was shown. In the ase of an inorret response, an error feedbak Fehler (error) was provided. Both kinds of feedbak were displayed for 300 ms. Following either kind of feedbak (blank, too slow, or error) the sreen went blank for another 700 ms before the next trial started with the presentation of the fixation.

3 714 BRIEF REPORTS The 16 ombinations of digits and loations were repeated four times per blok (64 trials). Trial transition probabilities were ontrolled for a priori. The experiment onsisted of 12 bloks that were separated by short breaks in whih partiipants were enouraged to relax. Exept for the first blok, eah blok started with the presentation of the last trial of the previous blok in order to keep transition probabilities equal and in order to provide an appropriate N 1 history of the first trial (now the seond trial) in the present blok. For Bloks 2 12 this inreased the number of trials to 65 per blok. However, the first trial in eah blok was eliminated prior to analyses, whih again resulted in 64 trials for Bloks 2 12 and in 63 trials in Blok 1. Before the experiment, partiipants performed 16 pratie trials. Results To minimize the possibility that the Simon modulation was driven by repetitions of idential trial events (Hommel, Protor, & Vu, 2004; Mayr et al., 2003), we exluded all idential repetitions from the analyses (6.4%). 2 Erroneous trials (3.4%) and all trials in whih reation times (RTs) did not fit the outlier riterion (2.5 SDs; 2.3%) were also exluded prior to statistial analyses. A repeated measures analysis of variane (ANOVA) inluding the fators Simon N (ompatible [C] vs. inompatible [IC]), Simon N 1 (C vs. IC), numerial distane (ND N, vs. ) and ND N 1 ( vs. ) was onduted on RTs and error rates alike. The fator numerial distane represented a two-level fator in whih the numbers 1, 2, 8, and 9 were onsidered from 5, whereas the numbers 3, 4, 6, and 7 were lassified as to 5 (see also Oriet et al., 2005). Results are shown in the Appendix in Table A1. As expeted, we found a reliable Simon effet of 22 ms, F(1, 14) 51.34, MSE , p.001, ε.786, as well as a typial effet of numerial distane (30 ms), F(1, 14) 43.93, MSE 1,243.98, p.001, ε.758. Both fators did not interat (F 1). The Simon effet was larger after Simon-ompatible trials (36 ms) and was virtually eliminated, although signifiant (8 ms), t(14) 2.26, p.05, after Simon-inompatible trials. This sequential modulation, Simon N Simon N 1, F(1, 14) 47.91, MSE , p.001, ε.774, (see Figure 1A, right panel) may be attributed to adaptive ontrol mehanisms regulating behavior after the detetion of response onflit (Botvinik et al., 2001; Gratton et al., 1992; Wühr, 2005; Wühr & Ansorge, 2005; but see Hommel et al., 2004). 3 Most important for the present study, the Simon adaptation effet (the interation Simon N Simon N 1 ) was modulated by the ND-indued proessing demands in trial N 1, as suggested by the reliable three-way interation between Simon N Simon N 1 ND N 1, F(1, 14) 15.52, MSE , p.005, ε.526. Consistent with the expeted ND N 1 Simonmodulation N interation, Simon-indued onflit adaptation was less pronouned when digits to the riterion were ategorized in trials N 1 (inreased ND-indued task demands for the primary task) ompared to the ategorization of digits from the riterion in trials N 1 (see Figure 1B). The ativation of ontrol adjustments after Simon onflit also affeted the effiieny of the subsequent proessing mode, whih was onfirmed by the interation between Simon N 1 and ND N, F(1, 14) 33.17, MSE 50.32, p.001, ε.703. Interestingly, the ND effet was larger after trials with response onflit (35 ms) than after trials with no response onflit (25 ms). As an be seen in Figure 2A (right panel), previous onflit slowed primarily the ontrolled omparison proess, whih eventually resulted in a three-way interation between ND N ND N 1 Simon N 1, F(1, 14) 12.16, MSE 81.87, p.01, ε.465 (see also Figure 2B). Irrespetive of response onflit, a sequential modulation was also observed for the ND effet. That is, the ND effet inreased after ategorizing numbers that were from the standard (40 ms) ompared to onditions, in whih numbers were ategorized in trial N 1 (20 ms), resulting, as hypothesized, in an interation between ND N 1 ND N, F(1, 14) 29.19, MSE , p.001, ε.676 (see Figure 2A, left panel). Further signifiant results inlude an effet of Simon ompatibility in trial N 1 on RTs in trial N, F(1, 14) 23.02, MSE 90.35, p.001, ε.622. That is, previous response onflit prolonged RTs in trial N. Finally, a signifiant interation between all fators (Simon N, Simon N 1,ND N, and ND N 1 ) was found, F(1, 14) 53.61, MSE 75.04, p.001, ε.793 (see Appendix). A repeated measures ANOVA onduted on the error data ontained the same fators as the ANOVA on RTs. More errors were ommitted in Simon-inompatible onditions (4.6%) ompared to Simon-ompatible onditions (2.2%), F(1, 14) 16.42, MSE 20.78, p.005, ε.540. Also, the error data mirrored the sequential modulation of the Simon effet found for the RT data. That is, the fators Simon N and Simon N 1 interated signifiantly, F(1, 14) 23.34, MSE 14.19, p.001, ε.625. Furthermore, the Simon effet in the error data was larger when numbers (3.2%) rather than numbers (1.6%) were ategorized in trial N 1, F(1, 14) 9.58, MSE 3.77, p.01, ε.406. Far numbers in the urrent trial, however, led to a smaller Simon effet than did numbers, F(1, 14) 5.46, MSE 17.49, p.05, ε.281. Similarly, the sequential modulation of the Simon effet was less pronouned for numbers in N (2.5% to 0.3%) than for numbers in N (6.9% to 0.3%). This was shown in the three-way interation between Simon N, Simon N 1, and ND N, F(1, 14) 17.26, MSE 3.13, p.005, ε.552. At the same time, no interation was found between the fators Simon N, Simon N 1, and ND N 1 (F 1). ND also affeted error rates. That is, error rates were higher for numbers to the riterion (5.2%) than for numbers from the riterion (1.7%), F(1, 14) 30.37, MSE 24.13, p.001, ε.684. Like in the RT data, the size of the ND N effet was modulated by ND N 1, F(1, 14) 8.02, MSE 8.67, p.05, ε.364. Furthermore, ND N 1 affeted error rates in general, with more errors being ommitted when the previous number was (4.0%) 2 None of the results hanged when idential repetitions were inluded in the analyses. 3 This sequential modulation of the Simon effet in partiular (as well as all other effets) was neither hanged when exluding all stimulus repetitions, F(1, 14) 65.14, MSE , p.001, ε.823, nor affeted by the additional fator response repetition (repetition vs. hange; F 1). Although it is hard to safely exlude any involvement of repetition priming, these analyses suggest that suh an involvement, if any, is rather negligible for the present study.

4 BRIEF REPORTS 715 A ND (N-1) Simon (N-1) i IC in N C in N B N ND (N-1) x Simon (N-1) i i Figure 1. Modulations of the Simon effet in trial N, represented by mean reation times (RTs). Panel A shows the size of the Simon effet as a funtion of previous numerial distane (ND; left side) and as a funtion of previous Simon ompatibility (right side). Panel B shows the sequential modulation of the Simon effet as a funtion of both previous response onflit and previous ND. Error bars represent standard errors of the mean. ompatible; i inompatible. than (2.8%), F(1, 14) 19.18, MSE 4.53, p.01, ε.578. Finally, the interation between Simon N 1 and ND N 1 in the present trial shows that the most errors were produed when in the previous Simon-ompatible trial a number was judged (4.6%) and that the least errors were ommitted when in the previous trial (irrespetive of Simon ompatibility) a number was judged (2.8%), F(1, 14) 8.74, MSE 2.58, p.05, ε.384. Note, however, that espeially the results for the higher order interations should be taken with aution given the very low error proportions per partiipant and ondition. Disussion The aim of the present researh was to study different ontrol regulation mehanisms that support the effiient proessing of the urrently relevant task goal and the shielding of this goal from distrating information. More speifially, we investigated the interplay between ontrol adjustments eliited by previous response onflit from task-irrelevant stimulus information, on the one hand, and proessing adjustments on the basis of previous variations in task-relevant proessing demands, on the other.

5 716 BRIEF REPORTS A ND (N-1) Simon (N-1) i B in N in N Simon (N-1) x ND (N-1) onflit no onflit Figure 2. Modulations of the numerial distane (ND) effet in trial N, represented by mean reation times (RTs). Panel A shows the size of the ND effet as a funtion of previous ND (left side) and as a funtion of previous Simon ompatibility (right side). Panel B shows the sequential modulation of the ND effet as a funtion of both previous response onflit and previous ND. Error bars represent standard errors of the mean. ompatible; i inompatible. In this respet, the most important results of our study are the interations between the two kinds of dynami adaptations of ognitive ontrol. First of all, we found typial effets of Simon ompatibility and its sequential modulation, whih has often been taken as evidene for onflit-indued ontrol adjustments (Botvinik et al., 2001; Gratton et al., 1992; Wühr, 2005; Wühr & Ansorge, 2005; but see Hommel et al., 2004; Wendt, Kluwe, & Peters, 2006). In addition, we found that these ontrol adjustments were redued in onditions of inreased ontrol demands for the primary task in the previous trial (ND N 1 Simon-modulation N ). This result (see Figure 1B) shows that inreased proessing demands in trial N 1 (lassifying numbers) affeted the effiient ativation of ontrol mehanisms that serve to redue interferene from task-irrelevant stimulus features in the subsequent trial N. This extends previous studies by showing that not only the quantity of onflit modulates subsequent adjustments of ognitive ontrol but, moreover, that suh adjustments of ognitive ontrol are also sensitive to speifi attentional requirements of proessing task-relevant stimulus features (without onflit) in trial N 1 (e.g., Freitas, Bahar, Yang, & Banai, 2007).

6 BRIEF REPORTS 717 Seond, our results show not only that Simon-indued onflit in trial N 1 leads to heightened ontrol in the subsequent trial (onflit adaptation) but moreover that this ontrol, serving to redue interferene from irrelevant stimulus attributes, negatively affets the effiieny of the number omparison proess (see the Simon N 1 ND N interation). Espeially the ontrolled proessing mode for ategorizing numbers is slowed after onflit, whereas the ategorization of numbers is not affeted (see Figure 2A, right panel). To aount for this result, one ould assume that both mehanisms ompete for the same resoures. Alternatively, one ould also assume that in onditions of heightened ontrol for interferene redution, swithing from judging numbers in N 1 (automati preategorization proess) to a slow and ontrolled omparison proess in N is espeially diffiult. 4 Although further studies might be needed to disentangle these possibilities, the present results learly show that the effortful number omparison mode annot benefit from inreased ontrol for interferene redution in trial N. Third, as predited, we found that the size of the ND effet in trial N was modulated by the numerial distane in trial N 1 (ND N 1 ND N ; see Figure 2A, left panel). ND effets were smaller when rather than digits were ategorized in trial N 1, indiating dynami adjustments in ognitive ontrol that are not related to previous response onflit. This finding is onsistent with our assumption that proessing demands in the number omparison judgment are inreased when the judgments are based on an effortful omparison proess (e.g., in the ase of numbers), whereas proessing demands are dereased when size judgments are based on an automati ategorization proess (e.g., in the ase of numbers; f. Oriet et al., 2005). A repeated appliation of a partiular ategorization proess will speed up the number lassifiation only if the stimulus lass is also repeated (e.g., N 1, N ). A repeated automati ategorization proess on trials in whih a number is followed by a number, for example, will prolong RTs and inrease the probability of wrong responses. Together, this results in a sequential modulation of the ND effet, whih suggests that not only response onflits due to irrelevant stimulus features but also variations in attentional demands in the primary task produe subsequent proessing adjustments. At first glane our results seem to be at odds with aounts suggesting that the proessing of task-irrelevant stimulus features is bloked after onflit (e.g., Stürmer et al., 2002). Stritly speaking, if the onflit adaptation mehanism ats merely on the task-irrelevant dimensions, speifi aspets of subsequent taskrelevant proessing should not be affeted (e.g., ertain qualities about number size judgments). However, our results show that a Simon onflit in trial N 1 not only redued irrelevant loation interferene in trial N but also affeted the size judgment in trial N. In sum, our findings onfirm the idea that flexible adaptation of ognitive ontrol serves to support the ognitive system in fulfilling the primary task and in shielding the system against distration from irrelevant information. In the present study, we found that inreased proessing requirements in the primary task required resoures that draw from the pool needed for the suffiient ativation of ontrol mehanisms in Simon-onflit trials. Along the same lines, an experiened (Simon N 1 ) response-onflit triggered adjustments of ognitive ontrol for interferene redution in the subsequent primary task that altered the subsequent numberomparison proess. Consequently, our study points to the neessity of haraterizing not only the influene of task-irrelevant stimulus features leading to onflit and onflit adaptation but moreover, to integrating and regarding the speifi requirements of proessing task-relevant stimulus features that might also determine adjustments of ognitive ontrol. 4 In this respet a further oneivable interpretation would propose a general flueny monitoring system that might monitor the flueny of proessing or the flueny of responding. Therefore, this system might be partiularly sensitive to proessing alterations that slow down responses, suh as response onflit (Simon), diffiult proessing ( numbers), ommitted errors, or generally outomes that are less than expeted. We thank an anonymous reviewer for direting our attention to this possibility. Referenes Botvinik, M. M., Carter, C. S., Braver, T. S., Barh, D. M., & Cohen, J. D. (2001). Conflit monitoring and ognitive ontrol. Psyhologial Review, 108, Botvinik, M. M., Cohen, J. D., & Carter, C. S. (2004). Conflit monitoring and anterior ingulate ortex: An update. Trends in Cognitive Sienes, 8(12), Botvinik, M. M., Nystrom, L. E., Fissell, K., Carter, C. S., & Cohen, J. D. (1999). Conflit monitoring versus seletion-for-ation in anterior ingulated ortex. Nature, 402(6758), Dehaene, S., Dupoux, E., & Mehler, J. (1990). Is numerial omparison digital? Analogial and symboli effets in two-digit number omparison. Journal of Experimental Psyhology: Human Pereption and Performane, 16, Dreisbah, G., & Haider, H. (2006). Preparatory adjustment of ognitive ontrol in the task swithing paradigm. Psyhonomi Bulletin & Review, 13, Egner, T., & Hirsh, J. (2005). Cognitive ontrol mehanisms resolve onflit through ortial amplifiation of task-relevant information. Nature Neurosiene, 8, Freitas, A. L., Bahar, M., Yang, S., & Banai, R. (2007). Contextual adjustments of ognitive ontrol aross tasks. Psyhologial Siene, 18, Goshke, T., & Dreisbah, G. (2008). Conflit-triggered goal shielding: Response onflits attenuate bakground-monitoring for onurrent prospetive memory ues. Psyhologial Siene, 19, Gratton, G., Coles, M. G. H., & Donhin, E. (1992). Optimizing the use of information: Strategi ontrol of ativation of responses. Journal of Experimental Psyhology: General, 121, 506. Hommel, B., Protor, R. W., & Vu, K.-P. L. (2004). A feature-integration aount of sequential effets in the Simon task. Psyhologial Researh, 68, Kerns, J. G., Cohen, J. D., MaDonald, A. W. I., Cho, R. Y., Stenger, V. A., & Carter, C. S. (2004, February 13). Anterior ingulate onflit monitoring and adjustments in ontrol. Siene, 303, Mayr, U., Awh, E., & Laurey, P. (2003). Conflit adaptation effets in the absene of exeutive ontrol. Nature Neurosiene, 6, Miller, E. K., & Cohen, J. D. (2001). An integrative theory of prefrontal ortex funtion. Annual Review of Neurosiene, 24, Moyer, R. S., & Landauer, T. K. (1967). Time required for judgments of numerial inequalities. Nature, 215, Notebaert, W., Soetens, E., & Melis, A. (2001). Sequential analysis of a Simon task: Evidene for an attention-shift aount. Psyhologial Researh, 65,

7 718 BRIEF REPORTS Oriet, C., Tombu, M., & Jolioeur, P. (2005). Symboli distane affets two proessing loi in the number omparison task. Memory & Cognition, 33, Simon, J. R. (1990). The effets of an irrelevant diretional ue on human information proessing. In R. W. Protor & T. G. Reeve (Eds.), Stimulus-response ompatibility (pp ). Amsterdam: North- Holland. Stürmer, B., Leuthold, H., Soetens, E., Shröter, H., & Sommer, W. (2002). Control over loation-based response ativation in the Simon task: Behavioral and eletrophysiologial evidene. Journal of Experimental Psyhology: Human Pereption and Performane, 28, Wendt, M., Kluwe, R. H., & Peters, A. (2006). Sequential modulations of interferene evoked by proessing task-irrelevant stimulus features. Journal of Experimental Psyhology: Human Pereption and Performane, 32, Wühr, P. (2005). Evidene for gating of diret response ativation in the Simon task. Psyhonomi Bulletin & Review, 12, Wühr, P., & Ansorge, U. (2005). Exploring trial-by-trial modulations of the Simon effet. Quarterly Journal of Experimental Psyhology, 58(A), Appendix Further Disussion of Adjustments of Cognitive Control Table A1 Response Times in Milliseonds (Perent Error) for Simon Compatibility (Compatible [C] vs. Inompatible [IC]) and Numerial Distane (ND, Far vs. Close) in Previous (N 1) and Current (N) Trials ND N 1 ND N Simon C N 1 Simon IC N 1 C N IC N C N IC N Close 454 (1.5) 478 (6.3) (4.4) 482 (3.9) Close 426 (0.1) 467 (3.2) 446 (1.9) 452 (1.1) Far 446 (2.1) 493 (11.3) 491 (5.3) 482 (6.5) Far 422 (1.5) 455 (3.5) 432 (0.8) 442 (1.1) With respet to the ND N 1 Simon-modulation N interation, we argued that the ativation of ontrol adjustments is less suffiient when the proessing mode binds resoures in N 1 (ND ). Inspetion of these data in Table A1 might suggest that this is the ase only when ND is in trials N 1 and N (see row 1). However, it does not appear plausible that attentional limitations in trial N are responsible for this insuffiient ontrol adjustment beause ontrol adjustments also lead to strong sequential modulations in the onditions ND N 1 and ND N (see row 3). Reeived August 4, 2007 Revision reeived January 28, 2008 Aepted January 31, 2008