Mechanisms and neuronal networks involved in reactive and proactive cognitive control of interference in working memory.

G Model

ARTICLE IN PRESS

NBR 1973 1–13

Neuroscience and Biobehavioral Reviews xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Neuroscience and Biobehavioral Reviews journal homepage: www.elsevier.com/locate/neubiorev

Review

1

Mechanisms and neuronal networks involved in reactive and proactive cognitive control of interference in working memory

2

3

4

Q1

Kerstin Irlbacher ∗ , Antje Kraft, Stefanie Kehrer, Stephan A. Brandt Department of Neurology, Universitätsmedizin Charité, Berlin, Campus Charité Mitte, Charitéplatz 1, 10117 Berlin, Germany

5 6

a r t i c l e

7 20

i n f o

a b s t r a c t

8

Article history: Received 2 September 2013 Received in revised form 22 June 2014 Accepted 27 June 2014 Available online xxx

9 10 11 12 13 14

19

Keywords: Decision making Interference in working memory Proactive cognitive control Reactive cognitive control

21

Contents

15 16 17 18

1. 2. 3.

22 23 24 25 26 27 28 29 30 31 32 33 34

4.

35 36 37 38 39 40

Q2

Cognitive control can be reactive or proactive in nature. Reactive control mechanisms, which support the resolution of interference, start after its onset. Conversely, proactive control involves the anticipation and prevention of interference prior to its occurrence. The interrelation of both types of cognitive control is currently under debate: Are they mediated by different neuronal networks? Or are there neuronal structures that have the potential to act in a proactive as well as in a reactive manner? This review illustrates the way in which integrating knowledge gathered from behavioral studies, functional imaging, and human electroencephalography proves useful in answering these questions. We focus on studies that investigate interference resolution at the level of working memory representations. In summary, different mechanisms are instrumental in supporting reactive and proactive control. Distinct neuronal networks are involved, though some brain regions, especially pre-SMA, possess functions that are relevant to both control modes. Therefore, activation of these brain areas could be observed in reactive, as well as proactive control, but at different times during information processing. © 2014 Published by Elsevier Ltd.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The concept of interference in working memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Reactive control of interference in working memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Brain regions and mechanisms involved in reactive control of interference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.1. Familiarity-inhibition hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.2. Context-retrieval hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1.3. Summary and interim conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Temporal dynamics of reactive interference control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1. Analysis of event-related potentials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2. Interference-based approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.3. Summary and interim conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Functional role of the ACC in reactive interference control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Summary and interim conclusions from studies investigating reactive control of interference in working memory . . . . . . . . . . . . . . . . . . . . . . Proactive control of interference in working memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Mechanisms and brain regions involved in proactive control of interference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.1. DLPFC and proactive control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.2. Median and inferior frontal cortex and proactive control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.3. Effect of the level of information processing at which conflict occurs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.4. Summary and interim conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

∗ Corresponding author. Tel.: +49 30 450560248; fax: +49 30450560912. E-mail address: [email protected] (K. Irlbacher). http://dx.doi.org/10.1016/j.neubiorev.2014.06.014 0149-7634/© 2014 Published by Elsevier Ltd.

Please cite this article in press as: Irlbacher, K., et al., Mechanisms and neuronal networks involved in reactive and proactive cognitive control of interference in working memory. Neurosci. Biobehav. Rev. (2014), http://dx.doi.org/10.1016/j.neubiorev.2014.06.014

G Model NBR 1973 1–13

K. Irlbacher et al. / Neuroscience and Biobehavioral Reviews xxx (2014) xxx–xxx

2

5.

41

42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97

ARTICLE IN PRESS

4.2. Temporal dynamics of proactive interference control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. Summary and interim conclusions from studies investigating proactive control of interference in working memory . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction Decision making is dependent on the ability to exert cognitive control over reflexive and habitual responses (Miller, 2000). Cognitive control is assumed to consist of multiple components (e.g. Badre and Wagner, 2007; Banich et al., 2000; Parks and Madden, 2013). One influential theory, the dual mechanisms of control (DMC) theory (Braver et al., 2007, 2009; Braver, 2012; De Pisapia and Braver, 2006), differentiates between a reactive and a proactive control mode with distinct temporal dynamics. In the reactive control mode, control processes are recruited as late correction mechanisms, for instance after detection of interference between automatic and controlled responses. In the proactive control mode, goal-relevant information is actively maintained in an anticipatory manner over a period of time, even before the occurrence of a cognitively demanding event or before the registration of a conflict between reflexive and controlled behavior. Thus, proactive control aims to minimize interference from internal or external sources on decision making, whereas reactive control aims to reduce the effect of interference on the decision making process after its detection. Both modes also differ with regard to the effort and the attentional commitment required, which explains the benefits and disadvantages of either strategy, depending on the frequency and expectancy of the cognitively demanding event. Proactive control consumes resources and implements a form of sustained mental set that reduces sensitivity to unexpected but potentially relevant sources of information. Reactive control requires a retrieval or activation of goal representations only at the time at which they are needed, and is therefore computationally efficient. On the other hand, it is late acting and stimulus-dependent, and efficiency depends on the saliency of the stimulus and on the strength of associated cues that enable the retrieval of stored goals. The brain is able to shift flexibly between both modes according to task demands (Braver, 2012). In this framework, intra-individual variability regarding the preferred cognitive control strategy results from a change in situational factors, like interference expectancy (Burgess and Braver, 2010). Inter-individual differences are explained by factors like working memory capacity, fluid intelligence (Burgess and Braver, 2010), aging (Paxton et al., 2008), and personality factors, such as reward sensitivity. These factors influence the value estimates of the relative benefits and disadvantages of the preferred mode of control (Jimura et al., 2010). The DMC theory is applicable in interpreting findings from different paradigms and domains in the research of cognitive control, such as behavioral switching, task-switching, interference in working memory, Stroop task, n-back task and Go/Nogo task (e.g. Czernochowski et al., 2010; Grandjean et al., 2012; Marklund and Persson, 2012; Ullsperger and King, 2010; West and Bailey, 2012). The neuronal interrelation of both types of cognitive control is currently under debate: Are they mediated by different neuronal networks? Or are there neuronal structures that have the potential to act in a proactive as well as in a reactive manner? The DMC account predicts that proactive control should be associated with sustained and/or anticipatory activation of the lateral prefrontal cortex (PFC), whereas reactive control transiently involves the lateral PFC, and activates, either via detection of interference (through

00 00 00 00 00

engagement of conflict monitoring regions such as anterior cingulate cortex (ACC)), or via associative and episodic associations, the posterior or medial temporal lobe regions (Braver, 2012). This review illustrates the way in which the integration of knowledge gathered from behavioral studies, functional imaging with its high spatial resolution and human electroencephalography with its high temporal resolution proves useful in answering the following questions: What is known about the mechanisms underlying proactive and reactive control, and is it conceivable that they are independent from one another? Are they mediated by different neuronal networks? Or are there neuronal structures that have the potential to act in a proactive as well as in a reactive manner? The DMC theory proposes that the ACC plays a role in the proactive as well as the reactive control networks. Others, like Ullsperger and King (2010) suggest a dichotomy of medial frontal regions (pre-supplementary motor area (pre-SMA) and ACC) in the sense that ACC is involved in reactive control, whereas pre-SMA is activated in proactive cognitive control, regardless of the level of information processing at which conflict occurs. The medial frontal cortex is thought to play an important role in regulating cognitive control and different theories postulate its involvement in conflict monitoring (e.g. Botvinick et al., 2001, 2004; Carter and van Veen, 2007), prediction of task difficulty or error likelihood (Brown and Braver, 2005; Carter et al., 1998; Nieuwenhuis et al., 2007) or reward-based decision-making (see for instance Hayden and Platt, 2010). Is there evidence supporting the hypothesis of a dichotomy of medial frontal cortex regions, as suggested by Ullsperger and King (2010)? To answer these questions, we focus mainly on studies investigating interference resolution at the level of verbal working memory representations, because neuronal networks and mechanisms involved in cognitive control of conflict or interference seem to be, at least partially, dependent on the material (e.g. Badre and Wagner, 2005; Leung and Zhang, 2004; Mecklinger et al., 2003) and the level of information processing at which conflict occurs (see for instance Bisset et al., 2009; Friedman and Miyake, 2004; Nee and Jonides, 2008; Nelson et al., 2003). It has to be taken into account Q3 that the general question of a dichotomy of medial frontal cortex regions in cognitive control is addressed here with only one specific requirement as verbal working memory. Therefore, it allows no statement regarding cognitive control in other task domains, such as inhibitory motor control or task switching. On the other hand, working memory plays an integral role in most forms of intelligent behavior (Nee et al., 2007), and capacity differences are related to differences in intelligence, reasoning, reading comprehension and problem-solving (Cowan et al., 2005; Daneman and Merikle, 1996; Just and Carpenter, 1999). A major factor in determining the capacity of short-term memory is the ability to protect it against interference from previously relevant information. Therefore, interference resolution at the level of working memory representations was intensively studied in recent years and a wealth of neuropsychological, functional imaging, electrophysiological, and lesion studies were performed to identify the mechanisms and neuronal networks involved. In the following section, the concept of interference in working memory is introduced. Thereafter, mechanism and networks involved in reactive and proactive control are presented and discussed separately.


98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155


ARTICLE IN PRESS K. Irlbacher et al. / Neuroscience and Biobehavioral Reviews xxx (2014) xxx–xxx

156

3

2. The concept of interference in working memory

A typical example of everyday experience, which illustrates how memory of past events influences processing in the present is the frustrating experience of forgetting where a car is parked in a regu159 larly used lot, which is at least partially attributable to interference 160 from memories established during prior occasions of parking the 161 car in the lot (Badre and Wagner, 2005). It has long been rec162 ognized that this interference from memory of past events is a 163 fundamental processing constraint not only in long-term memory 164 Q4 (McGeoch, 1942; Underwood, 1950, 1957), but also in cognition 165 (Hasher and Zacks, 1988), and in short-term memory (Brown, 1958; 166 Peterson and Peterson, 1959; Keppel and Underwood, 1962). In typ167 ical experimental situations, subjects are presented a target set, 168 comprising letters, words or numbers. After a delay, a probe is pre169 sented, and subjects have to decide if it was a part of the actual 170 target set or not. If in this task, called item recognition task, items 171 were not shown in the current target set, but were shown in the 172 previous target set, they have a high residual familiarity which is 173 inappropriate in the actual context. This illustrates how previously 174 relevant content in working memory can interfere with currently 175 relevant memory representations (e.g. Badre and Wagner, 2005; 176 Jonides and Nee, 2006). The phenomenon is called “proactive inter177 ference”. To prevent confusion with the term proactive control, 178 the expression “interference in working memory” is used in the 179 following. 180 In recognition memory, two separate processes are thought to 181 be involved, namely familiarity processing and recollection (e.g. 182 Yonelinas, 2002). Whereas familiarity processing is assumed to be a 183 fast assessment of memory trace strength, recollection is supposed 184 to be a slower process in which source and context information 185 about a prior study event is retrieved (e.g. Yonelinas, 2002; Goethe 186 and Oberauer, 2008). Theoretical models and empirical evidence 187 suggest that both processes take place not only in long-term mem188 ory, but also in short-term memory recognition (e.g. Oberauer, 189 2008; Oeztekin and McElree, 2007). Interference between the out190 comes of familiarity- and recollection-based processes of item 191 recognition accounts for the slow reaction time and high error 192 percentages on recent negative trials (Feredoes and Postle, 2010). 193 The task typically used to investigate interference in work194 ing memory is called the “recent probes task” (Jonides et al., 195 1998; Monsell, 1978; Sternberg, 1966). Here, subjects must decide 196 157 158

Fig. 1. The experimental procedure of the recent negative task with the inter-trial interference effect of recent negative probes is illustrated. A target set, comprising four letters is displayed. After a delay, a probe is presented, and subjects have to decide if it was a part of the actual target set or not. If the probe was not shown in the current target set but was a member of the previous target set, is has a high residual familiarity which is inappropriate in the actual context. This induces interference in working memory.

whether or not a probe letter, presented after a delay, matches a set of four target letters (positive trials match, negative trials do not). In a subset of negative trials, the probe of the current trial is not part of the actual target set, but was presented in the target set of the previous trial (recent negative trials) (see Fig. 1). In recent negative trials, interference occurs after probe presentation and induces transient reactive control. Sustained proactive control could be induced by a high proportion of recent negative trials in an experimental session, i.e. when greater interference could be expected. It occurs before probe presentation, and might be measured during the delay phase (see Figs. 1 and 2a). As explained earlier, proactive control is assumed to be realized via active maintenance of goal-related representations and aims to anticipate and prevent interference prior to its occurrence. One advantage of the recent probes task therefore is that it allows to study the temporal dynamics of interference control by measuring brain activity in functional magnetic resonance imaging (fMRI) or electroencephalography (EEG) related to the delay phase or to the probe (Burgess & Braver, 2010). Besides the recent probes task other tasks were employed to evaluate and advance the DMC theory in verbal working memory which should be shortly introduced in the following. These

Fig. 2. Schematic drawing illustrating the time course and proposed mechanisms of proactive (a) and early reactive (b) as well as late reactive (c) control of interference in working memory. Brain regions known to be involved in the different sub-processes are highlighted in different colors, which indicate the assumed sequential activation. Proactive as well as early and late reactive control mechanisms involve the left IFG at a different time during processing, but sub-regions do not overlap. In late reactive control, left inferior frontal regions are involved in different sub-processes. The left IFG (Brodmann area 45), which is also called mid-VLPFC, is employed in the registration of conflict and post-retrieval selection, whereas the anterior VLPFC (Brodmann area 47) is involved in an increased retrieval of episodic details. Overlapping activity in proactive and reactive control of interference was reported only in the pre-SMA.


197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218

G Model NBR 1973 1–13 4


Fig. 3. The experimental procedures of a 2-back task (a) and an item recognition task with a directed forgetting manipulation (b) are illustrated. In the 2-back task, subjects are required to determine whether each item (e.g. letter) in a list matches the item that was presented 2 positions before. Interference occurs on so-called lure trials, where a no-longer-relevant item in the neighborhood of the 2-back item corresponds to the actually presented item. In the item recognition task with a directed forgetting manipulation, a target set (e.g. four letters) is followed by an instruction cue to forget a part of the target set. After a delay, a probe is presented. If it is one of the to-be-forgotten items, it is called a lure, because interference exists between the required negative response and the higher amount of familiarity.

comprise for instance the n-back task (e.g. Smith and Jonides, 1997; 220 Q5 Gray et al., 2003). In this item recognition task, subjects are required 221 to determine whether each item (e.g. letter) in a list matches the 222 item that was presented n positions before (see Fig. 3a). There223 fore, subjects have to remember a specified number (n) of the 224 most recently presented items in serial order (n-back). The sub225 jects have to continuously update the memorized string of n most 226 recent items. Interference occurs on so-called lure trials, where a 227 no-longer-relevant item in the neighborhood of the n-back item 228 corresponds to the actually presented item (e.g. Gray et al., 2003; 229 Szmalec et al., 2011). 230 Another example is an item recognition task combined with a 231 directed forgetting procedure (see Fig. 3b). In this task, a target set 232 (e.g. four letters) is presented, which is followed by an instruction 233 cue to forget a part of the target set (e.g. two of the four letters). 234 After a delay, a probe is presented. If it is one of the to-be-forgotten 235 items, it is called lure. It requires a negative response, but carries a 236 higher amount of familiarity. This interference accounts for longer 237 reaction times and higher error rates to lures (Zhang et al., 2003). 219

238

239 240

241 242

3. Reactive control of interference in working memory 3.1. Brain regions and mechanisms involved in reactive control of interference Functional imaging studies that investigated reactive control of interference in working memory using the recent negative task,

consistently identified increased probe-related activation of the left inferior frontal gyrus (IFG) associated with recent negative probes (D’Esposito et al., 1999; Jonides et al., 1998; Nelson et al., 2003; for review see Jonides and Nee, 2006). Hence, an important functional role of the left IFG in reactive interference resolution is assumed (Badre and Wagner, 2005; Jimura et al., 2009; Jonides and Nee, 2006; Oeztekin et al., 2008). Congruently, patients with left IFG damage demonstrate a higher susceptibility to recent negative probes compared to patients with other frontal lobe lesions and healthy controls (e.g. Hamilton and Martin, 2005; Thompson-Schill et al., 2002). In recent years, two main classes of hypotheses were established regarding the mechanisms underlying reactive control of interference: familiarity-inhibition and context-retrieval models (Badre and Wagner, 2005; Jonides and Nee, 2006). Both models propose similar cognitive control processes to overcome interference, namely representational selection and monitoring, and both recognize that interference is a result of residual familiarity of the actual negative probe. However, the models differ with respect to the nature of the representations that give rise to interference and that are consequently selected and monitored: familiarity-inhibition theories assign interference to a conflict between the familiar nature of the probe and the required negative response due to its status as a non-member of the current target set. The resolution of this conflict is thought to proceed through inhibition of the familiar representation, the inappropriate response or the attribution of familiarity (Jonides et al., 1998; D’Esposito et al., 1999; Nelson et al., 2003). In contrast context retrieval models posit that interference exists between target familiarity and criterial context (here, membership in the current target set), which engages context retrieval to resolve interference (Badre and Wagner, 2005). We think that different task demands, like emphasis on speed versus accuracy, might favor the preferential use of the different mechanisms to resolve interference. If subjects should respond as fast as possible, they will rely more on the fast assessment of familiarity, which early activates an inappropriate response to recent negative probes. Consequently, an inhibition of familiarity or an inhibition of the inappropriate stimulus-response mapping might be the appropriate mechanism. Emphasis on accuracy allows employing the slower process of recollection and context retrieval. Additionally, working memory tasks with a higher demand on contextual recollection (like the n-back task, see Fig. 3a), will strongly favor context-retrieval models (Badre and Wagner, 2005). Both classes of models are supported by a large amount of literature, which should be reviewed in the following. 3.1.1. Familiarity-inhibition hypothesis Familiarity-inhibition models state that selection between multiple competing representations arises through the use of an attentional template. This contains goal-relevant information pertinent to the current trial. If the sources of information, as for instance temporal information, familiarity, as well as episodic features of other members of the target set are incompatible, they can compete in the assignment of the probe item as either a match or a non-match (Desimone and Duncan, 1995; Kan and Thompson-Schill, 2004; Jonides and Nee, 2006; Prabhakaran and Thompson-Schill, 2011). Reactive interference control operates through the assignment of a bias to representations of probe items based on the extent to which they match the contextual features of the current trial (biased competition model). In the context of the familiarity-inhibition model Mecklinger et al. (2003) proposed that interference resolution could be realized through a re-selection of the appropriate (negative) response to a recent negative probe which elicited a positive response (Mecklinger et al., 2003). In accordance with their assumption, they reported activity in the left


243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288

289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307



308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328

329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372

IFG and in the ACC related to recent negative probes. Remarkably, due to the experimental procedure introduced by Mecklinger et al. (2003), the recent negative probe was not only a part of the target set but also a positive probe in the previous trial, and therefore carried a larger amount of response conflict. Consequently, Nelson and co-workers (2003) tested if the inappropriate familiarity of the recent negative probe per se activates the ACC, or if the additional response conflict of recent negative probes which were positive probes in the previous trial leads to ACC activation. To this end, Nelson and co-workers (2003) directly compared probe related activity to recent negative probes with additional response conflict (negative probes which were positive probes in the previous trial) and without additional response conflict (negative probes which were part of the previous target set). They clearly demonstrated a double-dissociation between response conflict, activating the ACC, and interference in working memory due to inappropriate familiarity of the negative probe, activating the left IFG. Although others, like Bunge et al. (2001) reported some ACC activation in the recent negative task, the activation in this region was not correlated with the size of the interference effect, but was correlated with difficulty in the working memory task overall. 3.1.2. Context-retrieval hypothesis A number of functional imaging data are available in support of the context-retrieval hypothesis (Badre and Wagner, 2005). In the reduced experimental situation of the recent negative task, different context across trials means for instance different examples in the particular target sets, and especially, according to Badre and Wagner (2005) the contextual information of being a part of the actual or recent target set or not. The hypothesis here is that a negative probe with high residual familiarity due to a membership in the previous target set activates associated details from the previous context. These activations compete with retrieved details from the current context, i.e. that the probe is not part of the actual target set. Interference resolution is achieved through an increased selection of relevant episodic details in order to assign the probe to a task-relevant context. A monitoring process is required to integrate the retrieved information with decision criteria to obtain a decision for action. Thus, a multi-component cognitive control system is assumed, including the IFG (Brodmann area 45), which is also called the left mid-ventrolateral prefrontal cortex (VLPFC), as well as the frontopolar, right prefrontal, and parietal areas (Badre and Wagner, 2005, 2007; Nee et al., 2007). Activity in anterior portions of the left VLPFC (Brodmann area 47), which is concerned with controlled retrieval of long-term semantic representations, co-varies with that in the lateral temporal cortex (Badre and Wagner, 2007). The role of the left mid-VLPFC in interference resolution is considered to be post-retrieval selection, affecting simultaneously active representations (Badre and Wagner, 2007). Badre and Wagner (2005) revealed, that an increased activation of the left IFG was associated with greater behavioral indices of interference, whereas increased activation in the left frontopolar cortex was associated with decreased behavioral indices of interference. These findings were corroborated by a study of Nee and co-workers (2007), who reported a positive correlation of activity in the left IFG, and a negative correlation of activity in the left frontopolar cortex with a behavioral index of interference (Nee et al., 2007). These different correlations with behavioral measures of interference suggest a different functional role for the left IFG and the left frontopolar cortex in interference resolution. Whereas the left IFG is engaged in interference detection and in processes of interference resolution, the left frontopolar cortex evaluates and integrates the outcome of these processes to bias response processes. In support of the multi-component hypothesis of interference resolution, no functional connectivity was observed between both regions, and connectivity analyses

5

revealed distinct networks connected with them (Nee et al., 2007). Left VLPFC showed greater functional connectivity with left premotor cortex, right medial temporal cortex, right ACC, left inferior temporal pole, right posterior cingulate cortex, and left caudate during interference trials. Left frontopolar cortex, on the other hand, showed greater functional connectivity with left ACC. In accordance to Badre and Wagner (2007), Nee et al. (2007) assume that left IFG function is to select among memorial representations or contextual features in order to identify whether a probe is a member of the target set or not (Badre and Wagner, 2005; Nee et al., 2007). The connectivity with the right medial temporal cortex may be caused by an increased selection of episodic details, whereas the connectivity with the left premotor cortex was interpreted to reflect the use of those details to bias decision processes. Earlier findings of a similar network of right posterior medial cortex, left IFG and left premotor cortex, activated during encoding of items whose contexts were subsequently recollected (Ranganath et al., 2003) suggest that this same network might be re-activated during retrieval, when contextual information is needed to resolve interference (Nee et al., 2007). The frontopolar cortex has an important role in evaluation, monitoring, and integration processes in the domain of working memory and episodic retrieval. This is supported by a large amount of functional imaging and electrophysiological studies (for review see Badre and Wagner, 2005). That the outcome of these processes is used to bias response processes is underpinned by the increased functional connectivity between this cortical region and the ACC (Nee et al., 2007). A correlated pattern of activity between frontopolar cortex and ACC was demonstrated already by Badre and Wagner (2004), who also demonstrated that the ACC was not solely sensitive to response conflict, but showed a general sensitivity to increased cognitive control demands, irrespective of whether they arose from response conflict or from demands to engage in subgoaling/integration. Wolf et al. (2010) performed a behavioral and fMRI study, in which they investigated the impact of increased context processing on brain activity related to interference resolution. The authors used a variant of the item recognition task (Wolf and Walter, 2005): stimulus presentation of three capital letters was held constant, regardless of load and trial type to minimize load-dependent activation effects. One, two or three of the initially presented letters were briefly highlighted at the end of the stimulus period. Starting from these letters, subjects were instructed to memorize only those letters that directly followed in the alphabet. After a delay, the probe (a lower-case letter) occurred, and subjects had to indicate whether or not the probe was part of the manipulated set. Two types of negative probes were included in addition to positive probes: in non-interference trials, the negative probe did not match the manipulated set and were not included in the context set or stimulus set of the previous trial, whereas in interference trials, probes were taken from the highlighted context (highlighted letters prior to manipulation). Thus, interference from stimuli presented within the same trial, not inter-trial interference was induced. No directed-forgetting procedure was used (Wolf et al., 2010). As hypothesized by the context-processing hypothesis, they found that a network of VLPFC, dorsolateral prefrontal cortex (DLPFC) and frontopolar cortex was associated with interference resolution, and that frontopolar cortex activation increased with context demands, but had no direct impact on interference susceptibility. What are the findings of Braver and colleagues, regarding the cortical network involved in reactive interference control in working memory? Burgess and Braver (2010) used an a priori Region-Of-Interest (ROI)-based approach, and analysis was restricted to twenty-five regions, identified in prior studies to be related to working memory and executive functions. Six ROIs showed a pattern of reactive interference-related probe activity.


373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438

G Model NBR 1973 1–13 6 439 440 441 442


They were located within the left DLPFC and lateral parietal lobe, right pre-SMA, as well as bilateral IFG. There was no positive correlation between activity in one of the brain regions and effectiveness of the reactive control function (Burgess and Braver, 2010).

457

3.1.3. Summary and interim conclusions A network of bilateral IFG, left DLPFC, lateral parietal lobe and right pre-SMA is involved in reactive control of interference in working memory. Evidence was presented for early and late reactive control mechanisms, namely familiarity inhibition and biased competition as well as increased context retrieval. ACC and frontopolar cortex are employed in evaluation, monitoring, and integration processes. Functional imaging data are in accordance with the assumption, that both mechanisms of reactive control of interference in working memory might be used (see Fig. 2b and c). In the following section, which discusses the results of electrophysiological studies, further evidence is given for that hypothesis. Furthermore, the electrophysiological studies provide insight in the temporal dynamics of the different reactive control mechanisms.

458

3.2. Temporal dynamics of reactive interference control

443 444 445 446 447 448 449 450 451 452 453 454 455 456

3.2.1. Analysis of event-related potentials ERP studies are less diagnostic to the spatial localization of 460 effects, but allow a fine temporal resolution and are therefore a 461 valuable extension to the functional imaging data. What is actu462 ally known about electrophysiological correlates and the temporal 463 dynamics of reactive interference control? 464 Tays et al. (2008) used event relates potentials (ERP) to exam465 ine the time course of cognitive control of interference in working 466 memory in a recent negative task. Because the authors aimed to 467 evaluate the effect of aging, they compared a group of young sub468 jects with a group of older subjects. Young subjects showed an 469 increased correlate of early target categorization, called P3a, to 470 positive as compared to negative probes (mean peak latency for 471 positive probes: 274 ms), whereas this effect was not evident in 472 older subjects. At this processing stage, older adults thus seem to 473 fail to discriminate targets from non-targets and allocate atten474 tional resources equally across condition. Nonetheless, both groups 475 produced similar patterns of differentiation between positive and 476 neutral probes in their P3b response (mean peak latency around 477 500 ms), with greater amplitudes associated with positive probes. 478 Tays and co-workers suggest that the P3b might reflect the ongo479 ing application of attentional resources. This allocation of attention 480 may allow for controlled evaluation of salient/familiar stimuli and 481 automatic activation of prepotent stimulus–response mappings 482 Q6 (e.g. Nieuwenhuis et al., 2005). 483 An ERP component which was present in young subjects only, 484 was a medial frontal negativity (MFN) at around 450 ms after probe 485 presentation in recent negative trials. This was interpreted as an 486 index of (reactive) interference control. In accordance with this 487 assumption, a trend for a negative correlation between amplitudes 488 of the P3a and of the MFN was reported, suggesting that young 489 adults who showed greater neural discrimination between posi490 tive and negative probes tended to produce smaller MFNs to recent 491 negative probes, indicating that they experienced less interference. 492 Informed dipole source analyses revealed a temporal co-activation 493 of the inferior frontal and anterior cingulate cortex, suggesting that 494 these regions may interact during interference resolution. Tays 495 and co-workers (2008) suggest that the IFG may play a role in 496 biasing posterior brain systems to support stimulus-response map497 pings. This hypothesis is in accordance with familiarity-inhibition 498 theories, which suppose that interference results from differ499 ent stimulus attributes or stimulus-response mappings, whereas 500 context-retrieval models posit interference as competition among 501 459

specific episodic details that can assign a probe to a particular temporal context. No difference was reported between amplitudes and latencies of the P3b for interference and non-interference negative trials. This result could support the early familiarity-inhibition hypothesis. Because fMRI data (Nelson et al., 2003) has shown that response-based interference elicits activations in the ACC while familiarity-based interference activates IFG, Tays and co-workers (2009) argue that the MFN, in response to interference in working memory, may be more reflective of activations from IFG than ACC and therefore more sensitive to familiarity-based interference. However, general interference from frequent stimulus repetition can undercut familiarity-based interference effects and the activation of brain regions contributing to these responses may also be underspecified (Tays et al., 2009). The MFN was identified earlier in interference trials in the Stroop task, where it has been shown to occur in trials involving interference from both response and non-response conflict (West et al., 2004). Generators were assumed to be located in ACC and other anterior prefrontal regions (West et al., 2004). Two further ERP studies revealed seemingly contradictory results regarding the time point at which interference is resolved. Du et al. (2008) used an item recognition task with a directed forgetting procedure as depicted in Fig. 3b. In short, a target set of four letters was presented. After a delay, two of the items of the target set were shown again and the subjects were prompted to forget these items. After a second delay, a positive probe, which belonged to the actual to-be-remembered target set, or a negative probe, which was not shown in the actual target set, or a lure was presented. The lure was part of the actual target set but was shown again in the delay phase as a cue to indicate to subjects that they should forget this part of the target set. Therefore, the lure carries a high familiarity, but subjects are required to respond negatively. Du et al. (2008) reported a reduced fronto-central N2 around 300 ms after presentation of the lure, when compared to the negative probes. Positive probes elicited the smallest amplitudes. Because no differences between negative probes and lures were obtained in the P3 waveform, results were interpreted in the framework of familiarity-inhibition models as indicative of an early mechanism of reactive interference control (Du et al., 2008). The fronto-central N2 component is generally regarded to reflect increased cognitive demand. Source localization (e.g. van Veen and Carter, 2002; Nieuwenhuis et al., 2003; Yeung et al., 2004), in convergence with Q7 fMRI findings (e.g. Carter et al., 1998), and intracranial studies (e.g. Wang et al., 2005) identify the ACC as the source of the frontocentral N2. Typically, the fronto-central N2 amplitude is enhanced in interference trials (e.g. Yeung and Nieuwenhuis, 2009). We suppose that because the fronto-central N2 amplitude was not enhanced after presentation of the lure but was reduced, the results do not favor the hypothesis, that the N2 component reflects interference control, but rather reflects identification of increased familiarity of the lure. Nonetheless, the latency and amplitude of the P3 differed not between interference trials and non-interference trials, suggesting an early mechanism of reactive control of interference in that study. A later study by the same group (Zhang et al., 2010) using the recent negative probe task (as described in Section 2, see Fig. 1), revealed differences between recent and non-recent negative trials at around 550 ms. Authors therefore claimed that interference in that task is resolved later in the trial. Here, no differences were found in amplitudes of the N2 for recent negative and non-recent negative probes. Instead, the late parietal positive component (LPC), which is maximal over central parietal regions, showed reduced amplitudes for recent negative probes. Positive and nonrecent negative probes evoked a larger LPC with non-significant greater amplitudes in positive than in non-recent negative trials.


502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567



568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612

613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632

The LPC is regarded as an index of familiarity and episodic retrieval in long-term memory recognition (Voss and Paller, 2009), as well as in short-term memory (Danker et al., 2008). From that perspective, one would expect a larger amplitude in recent as compared to non-recent trials because of greater familiarity of recent negative probes. A plausible explanation of the findings, discussed by Zhang et al. (2010) might be that reduced confidence during recognition due to performance decrements in recent negative trials led to the reported effect. This assumption is indeed supported by findings of different studies which investigate the consequences of reduced confidence during recognition on the late LPC (e.g. Curran, 2004; Woodruff et al., 2006). Thus EEG studies of reactive interference control which aim to differentiate between an early (familiarity inhibition) or late (context retrieval) reactive control mechanism remained inconclusive. The differences between the studies of Du et al. (2008) and Zhang et al. (2010) may be interpreted at least partially as a consequence of the different manipulations to induce interference in working memory (intra-trial interference in a directed forgetting manipulation versus recent negative probes). In contrast, the studies of Tays and co-workers (2008) as well as Zhang and co-workers (2010) both used the recent negative task. Nonetheless, Tays et al. (2008) favored the familiarity-inhibition mechanism, whereas Zhang et al. (2010) proposed the late acting context retrieval mechanism. The main difference between both studies is an additional manipulation in the study of Tays and co-workers (2008), with trials comprising both, interference in working memory as well as response conflict. This was achieved by using a probe which was not only part of the previous target set but was also the positive probe of the previous trial. This additional condition was introduced according to Nelson et al. (2003). The MFN was larger in interference trials with additional response conflict as compared to trials which comprise only interference in working memory, but no additional sources were obtained. We think that the additional response conflict might have favored the early familiarity-inhibition mechanism of reactive control, but this remains speculative. Therefore, both familiarity inhibition as well as controlled retrieval seems to be a possible mechanism of reactive interference control (see Fig. 2b and c). The mechanisms used might depend on different, currently unknown factors, which could comprise study design or task instructions. An emphasis on speed could for instance prompt the subjects to favor an early familiarity inhibition to reactively control interference whereas emphasis on accuracy might facilitate the context retrieval mechanism of reactive control of interference. 3.2.2. Interference-based approach Support for an early involvement of the left IFG in reactive control of interference was provided from a study applying repetitive transcranial magnetic stimulation (rTMS) over the left IFG early (0–250 ms) and late (500–750 ms) after probe presentation (Feredoes and Postle, 2010). The rationale was to evaluate whether the left IFG is involved in early processes of familiarity assignment to the probe or in later processes of context-retrieval to resolve interference. Only early rTMS modulated the false alarm rate of recent negative probes. One offered interpretation is that the left IFG supports the identification or tagging of memory-related signals regarding familiarity, such that downstream decision mechanisms, probably in cooperation with the frontopolar cortex, would be able to weight each source of recognition-related information. The results of this study gave no support for a post-retrieval selection function of the left IFG, because later intervention had no effect on interference resolution. On the other hand, accrual of recollection-based information might start as early as 100–200 ms after probe onset (Feredoes and Postle, 2010). In other words, the onset of the subsequent rTMS train could have been too late.

7

Therefore, a more precise temporal resolution probably would have revealed an involvement of the left IFG in different processes at different time points. And indeed, there is a large body of evidence that the left IFG is involved in familiarity processing as well as in encoding and retrieval processes (e.g. Yonelinas et al., 2005; Duarte et al., 2005; Eichenbaum et al., 2007). For example, fMRI studies showed a significant correlation of activity in prefrontal regions, including left IFG with variation in probe familiarity (Yonelinas et al., 2005). Furthermore, lesions of the left hemisphere, including left IFG impaired selectively familiarity-based recognition from long-term memory (Duarte et al., 2005). Familiarity is a fast signal, processed earlier than recollection (Diana et al., 2006; Yonelinas, 2002). A response deadline procedure applied to a recent probe task revealed evidence for an early influence of familiarity at around 200–300 ms, and a late influence of trial-related contextual information at around 500 ms (Feredoes and Postle, 2010). It seems therefore, that left inferior frontal regions are involved in different processes at different time points during information processing. As stated by Badre and Wagner (2005), and Braver and colleagues (2009), distinct and independent neural processes may occupy colocalized voxels even at the resolution of fMRI (for another example see Badre and Wagner, 2004). 3.2.3. Summary and interim conclusions The different electrophysiological studies support the notion that reactive control of interference in working memory could take place either early (300–450 ms after probe presentation) or later (at around 550 ms) after probe presentation. Early signatures of interference control were interpreted in the context of familiarity inhibition models, whereas late signatures were regarded to indicate increased context-retrieval. Application of rTMS over the left IFG as an interference-based approach revealed an early involvement of the left IFG in reactive control between 0 and 250 ms after probe presentation. Nonetheless, because accrual of recollectionbased information might start as early as 100–200 ms after probe onset (Feredoes and Postle, 2010), the duration of the rTMS intervention was too long to clearly separate familiarity processing from context retrieval. 3.3. Functional role of the ACC in reactive interference control Are there arguments from electrophysiological or functional imaging studies supporting a role of the ACC in reactive interference control? The interference-related MFN, reported by Tays et al. (2008), and the fronto-central N2 effect reported by Du et al. (2008) suggest an involvement of the ACC, because for both potentials a source in the ACC was reported. In contrast, fMRI studies with their high spatial resolution did not suggest a direct involvement of the ACC in interference detection or resolution. Rather they suggest a recruitment of the ACC with increased cognitive control demands (e.g. Badre and Wagner, 2005; Nee et al., 2007). Functional connectivity was revealed between ACC and the left frontopolar cortex. Accordingly, reactive control activation in the medial frontal cortex was reported in the pre-SMA rather than in the ACC (Burgess and Braver, 2010). This is different in tasks with a larger amount of response conflict, like Stroop task, n-back task or Stop-signal task. In the Stop-signal task, the right IFG and pre-SMA are involved in the inhibition of hand responses, and the ACC is supposed to monitor stopping performance (Verbruggen and Logan, 2008). Transient right IFG activation was also found in reactive control in the n-back task, and was supposed to reflect executive control processes corresponding to response inhibition rather than interference resolution per se (e.g. Marklund and Persson, 2012). In that study, analysis was restricted to specific regions, and ACC activation was not studied, but earlier studies underline the role of the ACC in reactive


633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654

655 656 657 658 659 660 661 662 663 664 665 666 667 668 669

670

671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695

G Model NBR 1973 1–13 8 696 697 698 699 700 701 702 703 704

705 706

707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730

731

732 733

734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755


control of response conflict (e.g. Braver et al., 2001; Pastötter et al., 2010). The right IFG was associated earlier with both preparation for response conflict resolution (e.g. in cued stop-signal paradigms), and the actual inhibition of prepotent response tendencies once they threaten to interfere with successful task performance (e.g. Chikazoe et al., 2009; Jahfari et al., 2010). This illustrates, that the level of information processing, at which conflict occurs, determines reactive control mechanisms and involved brain regions. 3.4. Summary and interim conclusions from studies investigating reactive control of interference in working memory Reactive interference control could be regarded as a multicomponent process involving the left IFG, functionally connected with temporal, parietal and premotor brain regions as well as the frontopolar cortex with its functional connections to the ACC (see Fig. 2). Because evidence from functional imaging, patients and behavioral studies support that the left IFG is, at different time points, involved in familiarity processing (200–300 ms) as well as in controlled retrieval (∼500 ms), it is plausible that it is involved in reactive control of interference either early after probe presentation by familiarity inhibition and biased competition, or later by an increased retrieval of contextual information. Accordingly, functional imaging and electrophysiological studies of reactive control of interference in working memory suggest that different mechanisms (familiarity inhibition and context retrieval) might operate in different time windows, which partly involve identical brain regions, like the left IFG (Fig. 2b and c). It would be interesting to study, if sub-regions of the left IFG could be identified with high resolution fMRI engaged in familiarity processing and context retrieval, and if they could be experimentally dissociated. Regarding the question, which region of the medial frontal cortex is involved in reactive control of interference in working memory, the actual literature does not prove an involvement of the ACC, but instead of the pre-SMA, whereas ACC and the left frontopolar cortex are generally recruited with increased cognitive control demands. 4. Proactive control of interference in working memory 4.1. Mechanisms and brain regions involved in proactive control of interference In the framework of the DMC model, the underlying mechanism of the proactive control mode is a sustained maintenance of goal relevant context information (Braver, 2012). This takes place either in the delay period or across entire task blocks (see Fig. 2a). Burgess and Braver (2010) conducted an fMRI study to compare brain networks involved in proactive versus reactive control in a recent negative task with a five item memory set using English words as stimuli. Individual differences were studied by contrasting subjects with high and low fluid intelligence. Interference expectancy was manipulated across blocks by varying the frequency of the recent-negative probes. In the high expectancy condition, where proactive control is induced, especially in subjects with a high fluid intelligence, blocks contained 40% recent negative trials, 10% recent positive trials, 10% were novel negatives and 40% were novel positives. In the low expectancy condition, where reactive control is the preferred mode of cognitive control, 10% of trials were recent negatives, 40% were recent positives, 40% were novel negatives, and 10% were novel positives. Brain regions involved in proactive control were measured in the delay period of a recent negative task. In contrast, the brain network involved in reactive control is active after probe presentation. In the high expectancy condition, Burgess and Braver (2010) identified increased

delay-related activity of regions in bilateral DLPFC, IFG and pre-SMA, and within right lateral parietal regions, when compared to the reactive control mode. The right pre-SMA and the right posterior PFC demonstrated greater delay-related activity for subjects with a high fluid intelligence, which should show a greater tendency to utilize proactive control mechanisms compared to subjects with a low fluid intelligence (Braver et al., 2007). Interestingly, the right pre-SMA showed both probe-related and delay-related interference effects. The more anterior portion of this region showed a delay-related interference effect, whereas the more posterior portion showed a probe-related effect. The central portion showed both effects. Consequently, Burgess and Braver (2010) suggested that the pre-SMA is involved in both forms of interference control. The pre-SMA is thought to predict the expected cognitive demand and, in the proactive control mode, to register an increased need for cognitive control. 4.1.1. DLPFC and proactive control Further evidence for a role of the DLPFC in proactive control comes from a different study. Wolf et al. (2010), mentioned earlier in Section 3.1, performed a behavioral and fMRI study to investigate the impact of increased context processing on brain activity related to interference resolution in a recent negative task. Although this study was designed to evaluate the mechanism of reactive control of interference, and especially to prove the context-processing hypothesis, an interesting finding most likely related to proactive control was reported: If subjects were divided into those with high and low susceptibility to interference, those with low interference susceptibility showed an increased activation of a region in the left DLPFC under conditions of high context processing. In other words, lower susceptibility to interference was associated with an increase of activation of the left DLPFC, corresponding to Brodmann area 9. This finding was interpreted as to emphasize the relevance of this region for interference monitoring when cognitive demand is increased. The design of this study did not allow a differentiation between delay-specific or sustained (proactive) and probe-specific or transient (reactive) activation. Therefore it is not possible to decide whether this left DLPFC activation is associated with reactive or proactive control of interference. Because activity in this region was not specifically related to interference resolution but to low interference susceptibility, it seems nonetheless possible, that it might be related to proactive cognitive control. Yi et al. (2009) used a different task design to investigate proactive control. In an item recognition task with a memory set of four digits, a selection cue presented during the delay phase indicated which part of the target set should be kept in the focus of attention. The probe was either positive or negative. In the interference condition the probe was a digit which was presented in the original memory set, but was no longer relevant after the cue. This probe was called intrusion probe and prompted a negative response. Slow reaction times and high error percentage in interference trials are consistent with the notion that the no-longer-relevant information, though removed from the focus of attention in working memory, remains highly familiar and influences delayed recognition. In an fMRI study, brain activity related to informative cues versus neutral cues was measured (Yi et al., 2009). Remarkably, there was a negative correlation between individual differences in interference cost in reaction times and informative cue-related activity (beta weights from the selection-neutral contrast) in the lateral PFC and PPC, including the left inferior parietal lobe (IPL), left middle frontal gyrus (DLPFC), and precuneus. The negative correlation indicated that increases in differences in activity between the selection and neutral condition corresponded to decreases in the intensity of interference. The results suggest that this prefrontal-parietal network may contribute to proactive control, i.e. one’s ability to focus on the task relevant information in working memory. Accordingly,


756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771

772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820



821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863

864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885

the activity in this network influenced later experience of interference caused by the no-longer-relevant information. The results of Yi et al. (2009) are in accordance with the results of Burgess and Braver (2010), who identified increased activity of regions in bilateral DLPFC, IFG and pre-SMA and within right lateral parietal regions in proactive as compared to reactive control. The left IFG activation, identified in the study of Burgess and Braver (2010), was not seen in the study performed by Yi et al. (2009). This could be due to the different material (words versus digits), the different procedure to induce interference, and last but not least it could depend on the fact that proactive control in the study of Yi et al. (2009) was cue-related, which might activate different mechanisms and brain regions as compared to the internally generated proactive control due to high interference expectancy. The reported left DLPFC activation in subjects with lower interference susceptibility as compared with subjects with higher interference susceptibility reported by Wolf et al. (2010) is in line with its proposed role in proactive control. There is a large amount of prior studies regarding the functions of the DLPFC (see for review Badre and Wagner, 2004). From one perspective, DLPFC functions are conceptualized in terms of the biased competition framework of attention-mediated dynamic filtering (e.g. Cohen and Servan-Schreiber, 1992; Dehaene and Changeux, 1995; Desimone and Duncan, 1995). DLPFC computations serve to represent task context and provide a top-down bias that favors task relevant response pathways over competitors. Neuroimaging, neuropsychological and electrophysiological data suggest that DLPFC regions implement a context-dependent biasing or dynamic filtering (e.g. Banich et al., 2001; Miller and Cohen, 2001). The DLPFC shows enhanced activity during maintenance of task relevant information in the face of interference or distraction (e.g. Knight et al., 1999; Miller, 2000). The DLPFC is part of a cognitive control network including dorsal ACC/pre-SMA, dorsal premotor cortex, anterior insular cortex, inferior frontal cortex, and posterior parietal cortex (e.g. Cole and Schneider, 2007). DLPFC and dorsal ACC/pre-SMA show particular highly correlated activation patterns across a large number of studies (Duncan and Owen, 2000; Cole and Schneider, 2007). Cole and Schneider (2007) hypothesized that ACC/pre-SMA activity during working memory delays is found due to its involvement in preparatory processes at a point at which the anticipated need for cognitive control is highest, whereas DLPFC activity results from active maintenance of task goal information. 4.1.2. Median and inferior frontal cortex and proactive control The DMC model postulates that the ACC plays a role in both the proactive as well as the reactive control networks. One hypothesis regarding ACC function is that it integrates conflict over a short time-scale to signal the immediate need for reactive control (Botvinick et al., 2001), and it may also integrate repeated interference over a longer time-scale to signal the need for proactive control in anticipation of conflict (De Pisapia and Braver, 2006). Using the Stroop task, De Pisapia and Braver (2006) performed several computational simulations with a neural-network model, in which ACC-PFC interactions are described by two distinct conflictcontrol loops. Whereas in the Stroop task, ACC activation was reported to be relevant for reactive control, the results reported in Section 3 do not consistently support a relevant role of the ACC in reactive control of interference in working memory. Here, the ACC activation reported in fMRI studies correlated with activation in the frontopolar cortex and seems to reflect general sensitivity to control demands. Fronto-median activity reported by Burgess and Braver (2010) in the proactive control mode was located in the pre-SMA. Therefore, no evidence was found for ACC activation in proactive control of interference in working memory.

9

In the left IFG, the proactive control mode elicited delay-related activity, whereas the reactive control mode elicited a probe-related activity. There was no overlap between activated voxels in the left IFG in both control modes, suggesting that different sub-regions of the left IFG are involved (Burgess and Braver, 2010). To conclude, a cognitive control network comprising pre-SMA, (left) DLPFC, (left) IFG and inferior parietal regions is involved in proactive control of interference in working memory (see Fig. 2a). According to Cole and Schneider (2007), ACC/pre-SMA activity during working memory delays signals the involvement in preparatory processes at a point at which the anticipated need for cognitive control is highest. ACC and pre-SMA are capable of tuning activity in the lateral PFC and thereby, support adjustments of cognitive control over time (e.g. Kerns, 2006). Based on anatomical studies in monkeys (e.g. Luppino et al., 1993) as well as a diffusion tensor imaging study in humans (Johansen-Berg et al., 2004), the pre-SMA was shown to have anatomical connections with the DLPFC. Increased pre-SMA activity results in a stronger activation of (left) DLPFC and parietal regions. This proactive activity in lateral PFC may reflect preparation and maintenance of task goal representations, to facilitate the optimal updating and integration of memory-set information into an attentional bias regarding the upcoming probe (Burgess and Braver, 2010). Because interference is reduced proactively, a reduced probe-related activation of the left IFG would be expected. Although this interaction effect in the left IFG was not observed by Burgess and Braver (2010), there was a tendency for increased delay-related activity in the lateral PFC in the group of subjects with a high fluid intelligence, and of increased probe activity associated with recent negatives in the group of subjects with low fluid intelligence. These activation patterns suggest that the subjects with high fluid intelligence activated proactive control mechanisms to a greater degree and reactive control mechanisms somewhat less, than the subjects with low fluid intelligence (Burgess and Braver, 2010). 4.1.3. Effect of the level of information processing at which conflict occurs Do the neuronal network and underlying mechanisms involved in proactive control depend on the level of information processing, at which conflict occurs? In a verbal n-back task, containing a larger amount of response conflict, Marklund and Persson (2012) investigated sustained and transient activation in proactive and reactive control in the fMRI. Proactive control was induced by embedded cues predictive of high-interference trials. The aim of this study was to test whether proactive control could be utilized to facilitate performance in more complex working memory tasks in which concurrent processing of intervening items and updating is mandatory during contextual cue maintenance. Analysis was restricted to the prefrontal cortex, medial temporal lobe, striatum and midbrain. In blocks where proactive control was induced, the authors obtained an increased sustained activation in the right DLPFC, right IFG and the brainstem. In the cued as compared to the un-cued condition, activation peaked earlier in the right IFG and right caudate nucleus. Marklund and Persson (2012) interpreted this finding as a further argument for early proactive mechanisms served by these brain regions. Regarding the transient activation, significant differences between cued and un-cued trials were obtained in the bilateral IFG, bilateral medial temporal lobe, right DLPFC and bilateral caudate nucleus. In the proactive control mode, shorter reaction times were correlated with less transient activation of the right IFG, suggesting that proactive control reduced the need for interference control at the time of lure presentation. What are the supposed mechanisms underlying proactive and reactive control in that task? The right IFG has been strongly linked with response inhibition (e.g. Aaron et al., 2004; Rubia et al., 2003). Furthermore, an involvement in sustained focussing of attention and


886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919

920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950

G Model NBR 1973 1–13 10 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968


task-set-maintenance was proposed (e.g. Marklund et al., 2007). Different studies underline its role in preparation for response conflict resolution (e.g. in the cued stop-signal paradigm), and for its role in inhibition of prepotent response tendencies when they interfere with successful task performance (e.g. Chikazoe et al., 2009; Jahfari et al., 2010). In a stop signal task, proactive control is mostly related to a regulation of the level of excitability of the motor system. Thereby, the proactive control system sets the threshold for initiating a response. In making these adjustments the proactive system has to negotiate the trade off between speed (reaction time) and accuracy (cancelation likelihood) (Bogacz et al., 2010). It is assumed that the dorsomedial frontal cortex, including the SMA, is the source of the proactive control signal that modulates the baseline motor activity (Stuphorn and Emeric, 2012). Thus a pattern emerges which supports the hypothesis that mechanisms and involved brain regions in proactive control differ, depending on the level of information processing at which interference occurs.

983

4.1.4. Summary and interim conclusions According to the actual literature we suppose that a cognitive control network comprising pre-SMA, (left) DLPFC, (left) IFG and inferior parietal regions is involved in proactive control of interference in working memory (see Fig. 2a). Pre-SMA neurons provide a continuously updated prediction of the expected cognitive demands. Activity in the pre-SMA determines the activity in the lateral PFC and supports adjustments of cognitive control over time. An increased pre-SMA activity might result in a stronger activation of the left DLPFC and parietal regions which exert an increased top-down bias that favors task relevant representations, which are actively maintained. Active maintenance of task-relevant representations in working memory activates the left IFG (see Fig. 2a). This would explain the increased delay-related activity in the left IFG.

984

4.2. Temporal dynamics of proactive interference control

969 970 971 972 973 974 975 976 977 978 979 980 981 982

985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013

Yi and Friedman (2011) explicitly aimed to investigate the timing and ERP correlates of proactive control of interference in working memory. As in the fMRI study of Yi et al. (2009) described in Section 4.1, they used a cued item recognition task. The authors investigated cue-related and probe-related activity in event-related potentials. Regarding the cue-related activity, larger amplitudes of evoked potentials after informative cues when compared to non-informative cues were observed between 260 and 420 ms over parietal sites in a positive potential, between 420 and 560 ms over left frontal regions in a negative potential, and between 660 and 700 ms again over parietal regions in a positive component. It was hypothesized that the first positive activity reflects aspects of the selection process. For the negative component over left frontal regions, an inhibitory mechanism was assumed (to keep the noncued items out of the focus of working memory), whereas the later positivity over parietal regions was assumed to reflect the continued rehearsal of task-relevant information within the focus of attention (Yi and Friedman, 2011). Analysis of probe-related activity revealed that intrusion-probes elicited a negative-going component with a left fronto-central scalp focus during the time period of 500–540 ms after probe presentation. The authors interpreted their results within the framework of familiarity-inhibition theories of interference resolution, as explained in Section 3.2. They suggest that the selection cue, assumed to induce proactive control, requires an inhibitory process to remove irrelevant items out of the focus of attention in working memory to select the appropriate response. After an intrusion probe (reactive), cognitive control is needed to inhibit the tendency for an incorrect “yes”-response to the intrusion probe, which requires a negative response. Thus the

inhibitory processes differ with regard to the level of information processing, at which they occur: Whereas cue-related inhibition acts on internally-generated representations, probe-related inhibition is assumed to act at the response level (Yi and Friedman, 2011). Proactive control induced by a cue as defined in that study is different from the sustained proactive control induced by expectation changes due to a variance in the probability of conflict. Whereas the former might indeed involve a cue-related inhibition of internal representations of a part of the target set, which should be kept out of the focus of attention, a higher proportion of interference trials might be registered in the pre-SMA and induce a different mechanism, namely an increased top-down bias that favors task-relevant representations, which are actively maintained. To the best of our knowledge, no study has been published until now that investigated proactive control induced by probability changes in the recent probe task. As in the previous sections, we shortly mention studies which examine correlates of proactive control of interference at a different level of information processing. Investigating ERPs in the counting Stroop task, West and Bailey (2012) and West et al. (2012) reported a (transient) MFN at around 300–450 ms after stimulus presentation and a sustained lateral frontal activity, which were associated with adjustments of control across trials. The authors related them to proactive control, although in the context of the DMC theory, trial-to-trial adjustments should be regarded as a reactive control mechanism. Other studies have failed to reveal an association between proactive control and the MFN (Larson et al., 2009, 2012); instead demonstrating an effect of the congruency of the current stimulus on the MFN. We think that to study delay-related changes in neuronal activity in the EEG activity as a correlate of proactive control, time-frequency analyses and inter-regional connectivity analyses via cross-frequency coupling of different frequencies would be applicable. To summarize, up to now, there was no study investigating electrophysiological correlates of proactive control in the recent negative task with a manipulation of interference expectancy. Therefore, no information regarding the exact time course and electrophysiological correlates of proactive control induced by expectancy changes is available.

4.3. Summary and interim conclusions from studies investigating proactive control of interference in working memory A cognitive control network comprising pre-SMA, (left) DLPFC, (left) IFG and inferior parietal regions seems to be involved in proactive control of interference in working memory. Pre-SMA neurons provide a continuously updated prediction of expected cognitive demands. Activity in the pre-SMA determines the activity in the lateral PFC and supports adjustments of cognitive control over time. An increased pre-SMA activity might result in a stronger activation of the left DLPFC and parietal regions which exert an increased top-down bias that favors task-relevant representations, which are actively maintained. Active maintenance of task-relevant representations in working memory activates the left IFG. This explains the increased delay-related activity in the left IFG. Electrophysiological correlates of proactive control in working memory induced by a cue suggest an initial inhibition of internal representations of a part of the target set. Interpreted in the framework of familiarity-inhibition hypotheses of interference resolution, an effective cue-related inhibition would result in less reactive control, i.e. less inhibition to suppress familiarity-induced activation of the inaccurate “yes” response to recent negative trials.


1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055

1056 1057

1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076



1077

1078 1079 1080 1081 1082 1083 1084

1085 1086 1087 1088 1089 1090 1091 1092 1093 1094

1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139

5. Conclusions The DMC framework postulates that proactive and reactive control may involve potentially independent mechanisms with distinct temporal dynamics (Braver, 2012). Furthermore, the DMC model suggested that proactive and reactive control could be mediated by the same PFC regions, which might be able to switch their activity depending on task demands (Braver et al., 2009). In this review, we aimed to answer the following questions: i) What are the potential mechanisms of proactive and reactive control of interference in working memory? ii) It is conceivable that they are independent from one another, and are they mediated by different neuronal networks? iii) Are there neuronal structures that have the potential to act in a proactive as well as in a reactive manner? iv) Is the ACC involved in reactive and/or proactive control of interference in working memory or is there evidence of a dichotomy of medial frontal cortex function with an ACC activation in reactive control and a pre-SMA activation in proactive control? The presented findings gathered mainly from fMRI studies and electrophysiological studies of interference in working memory do indeed support the assumption that different mechanisms are responsible for both control modes. Reactive control of interference in working memory could be realized either by an early reactive control mechanism of familiarity-inhibition and biased competition or a late context-retrieval mechanism. Evidence suggests that both mechanisms might be used, probably depending on task conditions, manipulation to experimentally induce interference in working memory, and emphasis on speed or accuracy. Proactive control, on the other hand, seems to rely on an increased top-down bias of task-relevant representations. This leads either to an updating or increased maintenance of goal-relevant information. All presented studies confirm the distinct temporal dynamics of both control modes, as conceptualized by the DMC model. Different neuronal networks, involved in proactive and reactive control of interference in working memory were identified (see Fig. 2). Both activate the left IFG, but the activated subregions do not overlap (Burgess and Braver, 2010). Because it was demonstrated that the left IFG is involved in different functions, like familiarity processing, encoding and active maintenance of representations in working memory and also in the selection of representations from working memory, it is reasonable that different parts of this brain region are involved in processes supporting both control modes. However, one region outside of the left IFG was identified which displayed delay-related activity as well as probe-related activity, suggesting a role in proactive as well as reactive control of interference in working memory: the pre-SMA showed both a probe-related and a delay-related interference activity. Generally, the pre-SMA is supposed to predict the expected cognitive demand. As displayed in Fig. 2, this explains its increased activation in reactive control of interference in working memory. In blocks with a high proportion of interference trials, the pre-SMA registers the increased need for cognitive control and is regarded to initiate the sustained control processes involved in proactive control. Therefore, not the ACC, but the pre-SMA seems to be the medial frontal cortex region involved in both control modes. Studies investigating cognitive control in other tasks, like Stroop task, n-back task or stop-signal provided evidence that mechanisms and neuronal networks involved in proactive and reactive control depend on the level of information processing, at which conflict occurs. Regarding the hypothesis of Ullsperger and King (2010), that results from behavioral switching could be generalized to cognitive control of interference at different levels of information

11

processing, it is plausible to assume that cognitive control can vary “generally along a reactive-proactive continuum” (Ullsperger and King, 2010). Nonetheless, it seems questionable from the reported studies, that conflict at different levels of information processing uniformly activates the pre-SMA in proactive control and the ACC in reactive control. A more differentiated picture emerges, that shows distinct networks involved in tasks with conflict at working memory representations versus tasks with a larger amount of response conflict. Whether or not proactive and reactive control mechanisms are independent from another could not be determined from the available literature. In a future study using an rTMS-based interference approach, this question could be addressed by selectively impairing the function of areas related either to proactive or reactive cognitive control. An interference-based approach would also be applicable to prove the hypotheses regarding the functional role of the left IFG and the pre-SMA in proactive and reactive control of interference in working memory. Other outstanding questions regard the factors that favor either early or late reactive control mechanisms. Furthermore, detailed studies of the evolution and time-course, and the identification of neuronal signatures of proactive control are required. Delay-related changes in neuronal activity in the EEG that could be a measure of proactive control could be studied for instance by employing time-frequency analyses. To study inter-regional connectivity, cross-frequency coupling of different frequencies would be applicable. Further theoretical challenges, as already outlined in a recent review article by Braver (2012), comprise for instance the need of a formal mechanistic model, explaining the mechanisms by which different factors can lead to shifts or fluctuations in control state and the establishment of robust behavioral markers of proactive and reactive control. Uncited reference

1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169

Q8

Sheth et al. (2012). Acknowledgements Funding was obtained from the German Research Foundation Q9 (FOR 778, IR 48/2-2). We wish to thank two anonymous referees for useful comments and suggestions on an earlier version of this manuscript. We thank J. England for corrections to the English text. References Aaron, A.R., Robbins, T.W., Poldrack, R.A., 2004. Inhibition and the right inferior frontal cortex. Trends Cogn. Sci. 8, 170–177. Badre, D., Wagner, A.D., 2004. Selection, integration, and conflict monitoring: assessing the nature and generality of prefrontal cognitive control mechanisms. Neuron 41, 473–487. Badre, D., Wagner, A.D., 2005. Frontal lobe mechanisms that resolve proactive interference. Cereb. Cortex 15, 2003–2012. Badre, D., Wagner, A.D., 2007. Left ventrolateral prefrontal cortex and the cognitive control of memory. Neuropsychologia 45, 2883–2901. Banich, M.T., Milham, M.P., Atchley, R., Cohen, N.J., Webb, A., Wszalek, T., Kramer, A.F., Liang, Z.P., Wright, A., Shenker, J., Magin, R., 2000. fMRI studies of Stroop tasks reveal unique roles of anterior and posterior brain systems in attentional selection. J. Cogn. Neurosci. 12, 988–1000. Banich, M.T., Milham, M.P., Atchley, R., Cohen, N.J., Webb, A., Wszalek, T., Kramer, A.F., Liang, Z.P., Wright, A., Shenker, J., Magin, R., 2001. FMRI studies of the Stroop tasks reveal unique roles of anterior and posterior brain systems in attentional selection. J. Cogn. Neurosci. 12, 988–1000. Bogacz, R., Wagenmakers, E.J., Forstmann, B.U., Nieuwenhuis, S., 2010. The neural basis of the speed-accuracy tradeoff. Trends Neurosci. 33, 10–16. Botvinick, M.M., Braver, T.S., Barch, D.M., Carter, C.S., Cohen, J.D., 2001. Conflict monitoring and cognitive control. Psychol. Rev. 108, 624–652. Botvinick, M.M., Cohen, J.D., Carter, C.S., 2004. Conflict monitoring and anterior cingulate cortex: an update. Trends Cogn. Sci. 8, 539–546. Braver, T.S., Barch, D.M., Gray, J.R., Molfese, D.L., Snyder, A., 2001. Anterior cingulate cortex and response conflict: effects of frequency, inhibition and errors. Cereb. Cortex 11, 825–836, http://dx.doi.org/10.1093/cercor/11.9.825.


1170

1171

1172

1173 1174 1175 1176 1177

1178

1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204

G Model NBR 1973 1–13 12

1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290


Braver, T.S., Gray, J.R., Burgess, G.C., 2007. Explaining the many varieties of working memory variation: dual mechanisms of cognitive control. In: Conway, A.R.A., Jarrold, C., Kane, M.J., Miyake, A., Towse, J.N. (Eds.), Variation in Working Memory. Oxford University Press, New York, NY, pp. 76–106. Braver, T.S., Paxton, J.L., Locke, H.S., Barch, D.M., 2009. Flexible neural mechanisms of cognitive control within human prefrontal cortex. Proc. Natl. Acad. Sci. U. S. A. 106, 7351–7356. Braver, T.S., 2012. The variable nature of cognitive control: a dual mechanisms framework. Trends Cogn. Sci. 16, 106–113. Brown, J.W., Braver, T.S., 2005. Learned predictions of error likelihood in the anterior cingulate cortex. Science 307, 1118–1121. Bunge, S.A., Ochsner, K.N., Desmond, J.E., Glover, G.H., Gabrieli, J.D.E., 2001. Prefrontal regions involved in keeping information in and out of mind. Brain 124, 2074–2086. Burgess, G.C., Braver, T.S., 2010. Neural mechanisms of interference control in working memory: effects of interference expectancy and fluid intelligence. PLoS ONE 5 (9), e12861. Carter, C.S., Braver, T.S., Barch, D.M., Botvinick, M.M., Noll, D., Cohen, J.D., 1998. Anterior cingulate cortex, error detection, and the online monitoring of performance. Science 280, 747–749. Carter, C.S., van Veen, V., 2007. Anterior cingulate cortex and conflict detection: an update of the theory and data. Cogn. Affect. Behav. Neurosci. 7, 367–379. Chikazoe, J., Jimura, K., Hirose, S., Yamashita, K., Miyashita, Y., Konishi, S., 2009. Preparation to inhibit a response complements response inhibition during performance of a stop-signal task. J. Neurosci. 29, 15870–15877. Cohen, J.D., Servan-Schreiber, D., 1992. Context, cortex, and dopamine: a connectionist approach to behavior and biology in schizophrenia. Psychol. Rev. 99, 45–77. Cole, M.W., Schneider, W., 2007. The cognitive control network: integrated cortical regions with dissociable functions. Neuroimage 37, 343–360. Cowan, N., Elliott, E.M., Scott Saults, J., Morey, C.C., Mattox, S., Hismjatullina, A., Conway, A.R., 2005. On the capacity of attention: its estimation and its role in working memory and cognitive aptitudes. Cogn. Psychol. 51, 42–100. Curran, T., 2004. Effects of attention and confidence on the hypothesized ERP correlates of recollection and familiarity. Neuropsychologia 42, 1088–1106. Czernochowski, D., Nessler, D., Friedman, D., 2010. On why not to rush older adults – relying on reactive cognitive control can effectively reduce errors at the expense of slowed responses. Psychophysiology 47, 637–646. Daneman, M., Merikle, P.M., 1996. Working memory and language comprehension: a meta-analysis. Psychon. Bull. Rev. 3, 422–433. Danker, J.F., Hwang, G.M., Gauthier, L., Geller, A., Kahana, M.J., Sekuler, R., 2008. Characterizing the ERP Old–New effect in a short-term memory task. Psychophysiology 45, 784–793. D’Esposito, M., Postle, B.R., Jonides, J., Smith, E.E., Lease, J., 1999. The neural substrate and temporal dynamics of interference effects in working memory as revealed by event-related fMRI. Proc. Natl. Acad. Sci. U. S. A. 96, 7514–7519. Dehaene, S., Changeux, J.P., 1995. Neuronal models of pre-frontal cortex functions. In: Grafman, J., Holyoak, K.J., Boller, F. (Eds.), Structure and Functions of the Human Prefrontal Cortex. New York Academy of Sciences, New York, pp. 305–319. De Pisapia, N., Braver, T.S., 2006. A model of dual control mechanisms through anterior cingulate and prefrontal cortex interactions. Neurocomputing 69, 1322–1326. Desimone, R., Duncan, J., 1995. Neural mechanisms of selective visual attention. Annu. Rev. Neurosci. 18, 193–222. Diana, R.A., Reder, L.M., Arndt, J., Park, H.Y., 2006. Models of recognition: a review of arguments in favour of a dual-process account. Psychon. Bull. Rev. 13, 1–21. Du, Y., Xiao, Z., Song, Y., Fan, S., Wu, R., Zhang, J.X., 2008. An electrophysiological signature for proactive interference resolution in working memory. Int. J. Psychophysiol. 69, 107–111. Duarte, A., Ranganath, C., Knight, R.T., 2005. Effects of unilateral prefrontal lesions on familiarity, recollection, and source memory. J. Neurosci. 25, 8333–8337. Duncan, J., Owen, A.M., 2000. Common regions of the human frontal lobe recruited by diverse cognitive demands. Trends Neurosci. 23, 475–483. Eichenbaum, H., Yonelinas, A.P., Ranganath, C., 2007. The medial temporal lobe and recognition memory. Annu. Rev. Neurosci. 30, 123–152. Feredoes, E., Postle, B.R., 2010. Prefrontal control of familiarity and recollection in working memory. J. Cogn. Neurosci. 22, 323–330. Friedman, N.P., Miyake, A., 2004. The relations among inhibition and interference control functions: a latent variable analysis. J. Exp. Psychol.: Gen. 133, 101–135. Goethe, K., Oberauer, K., 2008. The integration of familiarity and recollection information in short-term recognition: modeling speed-accuracy trade-off functions. Psychol. Res. 72, 289–303. Grandjean, J., D’Ostilio, K., Phillips, C., Balteau, E., Degueldre, C., Luxen, A., Maquet, P., Salmon, E., Collette, F., 2012. Modulation of brain activity during a Stroop inhibitory task by the kind of cognitive control required. PLoS ONE 7, e41513. Hamilton, A.C., Martin, R.C., 2005. Dissociations among tasks involving inhibition: a single-case study. Cogn. Affect. Behav. Neurosci. 5, 1–13. Hayden, B.Y., Platt, M.L., 2010. Neurons in anterior cingulate cortex multiplex information about reward and action. J. Neurosci. 30, 3339–3346. Jahfari, S., Stinear, C.M., Claffey, M., Verbruggen, F., Aron, A.R., 2010. Responding with restraint: what are the neurocognitive mechanisms? J. Cogn. Neurosci. 22, 1479–1492. Jimura, K., Yamashita, K., Chikazoe, J., Hirose, S., Miyashita, Y., Konishi, S., 2009. A critical component that activates the left inferior prefrontal cortex during interference resolution. Eur. J. Neurosci. 29, 1915–1920.

Jimura, K., Locke, H.S., Braver, T.S., 2010. Prefrontal cortex mediation of cognitive enhancement in rewarding motivational contexts. Proc. Natl. Acad. Sci. U. S. A. 107, 8871–8876. Johansen-Berg, H., Behrens, T.E., Robson, M.D., Drobnjak, I., Rushworth, M.F., Brady, J.M., Smith, S.M., Higham, D.J., Matthews, P.M., 2004. Changes in connectivity profiles define functionally distinct regions in human medial frontal cortex. Proc. Natl. Acad. Sci. U. S. A. 101, 13335–13340. Jonides, J., Smith, E.E., Marshuetz, C., Koeppe, R.A., Reuter-Lorenz, P.A., 1998. Inhibition in verbal working memory revealed by brain activation. Proc. Natl. Acad. Sci. U. S. A. 95, 8410–8413. Jonides, J., Nee, D.E., 2006. Brain mechanisms of proactive interference in working memory. Neuroscience 139, 181–193. Just, M.A., Carpenter, P.A., 1999. A capacity theory for comprehension. Psychol. Rev. 99, 122–149. Kan, I.P., Thompson-Schill, S.L., 2004. Selection from perceptual and conceptual representations. Cogn. Affect. Behav. Neurosci. 4, 466–482. Kerns, J.G., 2006. Anterior cingulate and prefrontal cortex activity in an FMRI study of trial-to-trial adjustments on the Simon task. Neuroimage 33, 399–405. Knight, R.T., Staines, W.R., Swick, D., Chao, L.L., 1999. Prefrontal cortex regulates inhibition and excitation in distributed neural networks. Acta Psychol. (Amst) 101, 159–178. Larson, M.J., Clayson, P.E., Baldwin, S.A., 2012. Performance monitoring following conflict: internal adjustments of cognitive control? Neuropsychologia 50, 426–433. Larson, M.J., Kaufman, D.A., Perlstein, W.M., 2009. Neural time course of conflict adaptation effect on the Stroop task. Neuropsychologia 47, 663–670. Leung, H.C., Zhang, J.X., 2004. Interference resolution in spatial working memory. Neuroimage 23, 1013–1019. Luppino, G., Matelli, M., Camarda, R., Rizzolatti, G., 1993. Corticocortical connections of area F3 (SMA-proper) and area F6 (pre-SMA) in the macaque monkey. J. Comp. Neurol. 338, 114–140. Marklund, P., Fransson, P., Cabeza, R., Petersson, K.M., Ingvar, M., Nyberg, L., 2007. Sustained and transient neural modulations in prefrontal cortex related to declarative long-term memory, working memory, and attention. Cortex 43, 22–37. Marklund, P., Persson, J., 2012. Context-dependent switching between proactive and reactive working memory control mechanisms in the right inferior frontal gyrus. Neuroimage 63, 1552–1560. Mecklinger, A., Weber, K., Gunter, T.C., Engle, R.W., 2003. Dissociable brain mechanisms for inhibitory control: effects of interference content and working memory capacity. Brain Res. 18, 26–38. Miller, E.K., 2000. The prefrontal cortex and cognitive control. Nat. Rev. Neurosci. 1, 59–65. Miller, E.K., Cohen, J.D., 2001. An integrative theory of prefrontal cortex function. Annu. Rev. Neurosci. 24, 167–202. Monsell, S., 1978. Recency, immediate recognition, and reaction time. Cogn. Psychol. 10, 465–501. Nee, D.E., Jonides, J., Berman, M.G., 2007. Neural mechanisms of proactive interference resolution. Neuroimage 4, 740–751. Nee, D.E., Jonides, J., 2008. Neural correlates of access to short-term memory. Proc. Natl. Acad. Sci. U. S. A. 105, 14228–14233. Nelson, J.K., Reuter-Lorenz, P.A., Sylvester, C.Y., Jonides, J., Smith, E.E., 2003. Dissociable neural mechanisms underlying response-based and familiarity-based conflict in working memory. Proc. Natl. Acad. Sci. U. S. A. 100, 11171– 11175. Nieuwenhuis, S., Schweizer, T.S., Mars, R.B., Botvinick, M.M., Hajcak, G., 2007. Errorlikelihood prediction in the medial frontal cortex: a critical evaluation. Cereb. Cortex 17, 1570–1581. Oberauer, K., 2008. How to say no: single- and dual-process theories of short-term recognition tested on negative probes. J. Exp. Psychol. Learn. Mem. Cogn. 34, 439–459. Oeztekin, I., McElree, B., 2007. Retrieval dynamics of proactive interference: PI slows retrieval by eliminating fast assessments of familiarity. J. Mem. Lang. 57, 126–149. Oeztekin, I., Curtis, C.E., McElree, B., 2008. The medial temporal lobe and the left inferior prefrontal cortex jointly support interference resolution in verbal working memory. J. Cogn. Neurosci. 21, 1967–1979. Parks, E.L., Madden, D.J., 2013. Brain connectivity and visual attention. Brain Connect. 3, 317–338. Pastötter, B., Hanslmayr, S., Bäuml, K.H., 2010. Conflict processing in the anterior cingulate cortex constrains response priming. Neuroimage 50, 1599–1605, http://dx.doi.org/10.1016/j.neuroimage.2010.01.095. Paxton, J.L., Barch, D.M., Racine, C.A., Braver, T.S., 2008. Cognitive control, goal maintenance, and prefrontal function in healthy aging. Cereb. Cortex 18, 1010–1028. Prabhakaran, R., Thompson-Schill, S.L., 2011. The price of fame: the impact of stimulus familiarity on proactive interference resolution. J. Cogn. Neurosci. 23, 816–831. Ranganath, C., Yonelinas, A.P., Cohen, M.X., Dy, C.J., Tom, S.M., D’Esposito, M., 2003. Dissociable correlates of recollection and familiarity within the medial temporal lobes. Neuropsychologia 42, 2–13. Rubia, K., Smith, A.B., Brammer, M.J., Taylor, E., 2003. Right inferior prefrontal cortex mediates response inhibition while mesial prefrontal cortex is responsible for error detection. Neuroimage 20 (1), 351–358. Sheth, S.A., Mian, M.K., Patel, S.R., Asaad, W.F., Williams, Z.M., Dougherty, D.D., Bush, G., Eskandar, E.N., 2012. Human dorsal anterior cingulate cortex neurons mediate ongoing behavioral adaptation. Nature 488 (7410), 218–221.


1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376



1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403

Stuphorn, V., Emeric, E.E., 2012. Proactive and reactive control by the medial frontal cortex. Front. Neuroeng. 5, 9, http://dx.doi.org/10.3389/fneng.2012.00009. Szmalec, A., Verbruggen, F., Vandierendonck, A., Kemps, E., 2011. Control of interference during working memory updating. J. Exp. Psychol. Hum. Percept. Perform. 37 (1), 137–151, http://dx.doi.org/10.1037/a0020365. Sternberg, S., 1966. High-speed scanning in human memory. Science 153, 652–654. Tays, W.J., Dywan, J., Mathewson, K.J., Segalowitz, S.J., 2008. Age differences in target detection and interference resolution in working memory: an event-related potential study. J. Cogn. Neurosci. 20, 2250–2262. Tays, W.J., Dywan, J., Segalowitz, S.J., 2009. General proactive interference and the N450 response. Neurosci. Lett. 462, 239–243. Thompson-Schill, S.L., Jonides, J., Marshuetz, C., Smith, E.E., D’Esposito, M., Kan, I.P., Knight, R.T., Swick, D., 2002. Effects of frontal lobe damage on interference effects in working memory. Cogn. Affect. Behav. Neurosci. 2, 109–120. Ullsperger, M., King, J.A., 2010. Proactive and reactive recruitment of cognitive control: comment on Hikosaka and Isoda. Trends Cogn. Sci. 14, 191–192. Verbruggen, F., Logan, G.D., 2008. Response inhibition in the stop-signal paradigm. Trends Cogn. Sci. 12, 418–424. Voss, J.L., Paller, K.A., 2009. An electrophysiological signature of unconscious recognition memory. Nat. Neurosci. 12, 349–355. Wang, C., Ulbert, I., Schomer, D.L., Marinkovic, K., Halgren, E., 2005. Responses of human anterior cingulate cortex microdomains to error detection, conflict monitoring, stimulus-response mapping, familiarity, and orienting. J. Neurosci. 25, 604–613. West, R., Bailey, K., 2012. ERP correlates of dual mechanisms of control in the counting Stroop task. Psychophysiology 49, 1309–1318. West, R., Bailey, K., Tiernan, B.N., Boonsuk, W., Gilbert, S., 2012. The temporal dynamics of medial and lateral frontal neural activity

13

related to proactive cognitive control. Neuropsychologia 50, 3450– 3460. West, R., Bowry, R., McCoville, C., 2004. Sensitivity of medial frontal cortex to response and non-response conflict. Psychophysiology 41, 739–748. Wolf, R.C., Walter, H., 2005. Evaluation of a novel event-related parametric fMRI paradigm investigating prefrontal function. Psychiatry Res. 140, 73–83. Wolf, R.C., Walter, H., Vasic, N., 2010. Increasing contextual load demand modulates anterior and lateral prefrontal brain regions associated with proactive interference. Int. J. Neurosci. 120, 40–50. Woodruff, C.C., Hayama, H.R., Rugg, M.D., 2006. Electrophysiological dissociation of the neural correlates of recollection and familiarity. Brain Res. 1100, 125–135. Yi, Y., Driesen, N.R., Leung, H.-C., 2009. Behavioral and neural correlates of memory selection and proactive interference during a digit working memory task. Cogn. Affect. Behav. Neurosci. 9, 249–259. Yi, Y., Friedman, D., 2011. Event-related potential (ERP) measures reveal the timing of memory selection processes and proactive interference resolution in working memory. Brain Res. 1411, 41–56. Yonelinas, A.P., 2002. The nature of recollection and familiarity: a review of 30 years of research. J. Mem. Lang. 46, 441–517. Yonelinas, A.P., Otten, L.J., Shaw, K.N., Rugg, M.D., 2005. Separating the brain regions involved in recollection and familiarity in recognition memory. J. Neurosci. 25, 3002–3008. Zhang, J.X., Leung, H.C., Johnson, M.K., 2003. Frontal activations associated with accessing and evaluating information in working memory: an fMRI study. Neuroimage 20, 1531–1539. Zhang, J.X., Wu, R., Kong, L., Wenig, X., Du, Y., 2010. Electrophysiological correlates of proactive interference in the recent probes verbal working memory task. Neuropsychologia 48, 2167–2173.


1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431

Cognitive control in context: working memory capacity and proactive control.

Neural mechanisms of proactive and reactive cognitive control in social anxiety.

Frontoparietal networks involved in categorization and item working memory.

Working memory in schizophrenia: behavioral and neural evidence for reduced susceptibility to item-specific proactive interference.

Interference Impacts Working Memory in Mild Cognitive Impairment.

Developmental change in proactive interference across the life span: evidence from two working memory tasks.

Alterations in working memory networks in amnestic mild cognitive impairment.

Neonatal Perirhinal Lesions in Rhesus Macaques Alter Performance on Working Memory Tasks with High Proactive Interference.

Cognitive Impairments in Occupational Burnout - Error Processing and Its Indices of Reactive and Proactive Control.

Proactive control of proactive interference using the method of loci.

Correction: ERP Correlates of Proactive and Reactive Cognitive Control in Treatment-Naïve Adult ADHD.

ERP Correlates of Proactive and Reactive Cognitive Control in Treatment-Naïve Adult ADHD.

The dynamics of proactive and reactive cognitive control processes in the human brain.

Does verbal labeling influence age differences in proactive and reactive cognitive control?

Abnormal proactive and reactive cognitive control during conflict processing in major depression.

Item-nonspecific proactive interference in monkeys' auditory short-term memory.

Aging, working memory capacity and the proactive control of recollection: An event-related potential study.

Neural and behavioral mechanisms of proactive and reactive inhibition.

Challenging Cognitive Control by Mirrored Stimuli in Working Memory Matching.

A Cognitive Paradigm to Investigate Interference in Working Memory by Distractions and Interruptions.

Spatial-simultaneous working memory and selective interference in Down syndrome.

Frontoparietal and Cingulo-opercular Networks Play Dissociable Roles in Control of Working Memory.

Sleep and memory consolidation: motor performance and proactive interference effects in sequence learning.

Mechanisms of Working Memory Impairment in Schizophrenia.