Control of working memory: Effects of attention training on target recognition and distractor salience in an auditory selection task

BR A IN RE S E A RCH 1 4 30 ( 20 1 2 ) 6 8 –77

Available online at www.sciencedirect.com

www.elsevier.com/locate/brainres

Research Report

Control of working memory: Effects of attention training on target recognition and distractor salience in an auditory selection task Robert D. Melaraa,⁎, Yunxia Tongb , Aparna Raoc a

Department of Psychology, City College, City University of New York, 138th Street and Convent Avenue, NAC 7/120, New York, NY 10031, USA b Gene, Cognition, and Psychosis Program, National Institute of Mental Health, National Institutes of Health, 9000 Rockville Pike, Building 10, Room 3C101, Bethesda, MD 20892, USA c Department of Speech–Language–Hearing Sciences, University of Minnesota, 115 Shevlin Hall, 164 Pillsbury Drive SE, Minneapolis, MN 55455, USA

A R T I C LE I N FO

AB S T R A C T

Article history:

Behavioral and electrophysiological measures of target and distractor processing were

Accepted 20 October 2011

examined in an auditory selective attention task before and after three weeks of

Available online 26 October 2011

distractor suppression training. Behaviorally, training improved target recognition and led to less conservative and more rapid responding. Training also effectively shortened the

Keywords:

temporal distance between distractors and targets needed to achieve a fixed level of

Event related potential

target sensitivity. The effects of training on event-related potentials were restricted to the

Working memory

distracting stimulus: earlier N1 latency, enhanced P2 amplitude, and weakened P3 ampli-

Attention training

tude. Nevertheless, as distractor P2 amplitude increased, so too did target P3 amplitude,

Inhibitory control

connecting experience-dependent changes in distractor processing with greater distinctive-

Auditory selective attention

ness of targets in working memory. We consider the effects of attention training on the processing priorities, representational noise, and inhibitory processes operating in working memory. © 2011 Elsevier B.V. All rights reserved.

1.

Introduction

1.1.

Background

Attention is essential for implementing flexible, goal-directed behavior. Yet attentional processing is by no means isolated from other cognitive processes, and can be influenced by perceptual, language, memory, and response mechanisms. Here, we consider the role of working memory in directing attention. Information held in working memory can increase or decrease

the efficiency of attentional processing (Awh et al., 2006; Cowan, 1995; Downing, 2000; Melara and Nairne, 1991; Postle et al., 2004). In fact, when performing attention and working memory tasks, distributed, overlapping brain networks are activated, including the mid-dorsolateral prefrontal cortex (PFC), the ventrolateral PFC, the parietal cortex, the anterior cingulate cortex, and the temporal cortex (Banich et al., 2000a, 2000b, 2001; Bledowski et al., 2004; Fan et al., 2003; Liu et al., 2004; MacDonald et al., 2000; Milham et al., 2001, 2003). Not surprisingly, then, recent theoretical models have considered explicitly

⁎ Corresponding author. Fax: + 1 212 650 5659. E-mail addresses: [email protected] (R.D. Melara), [email protected] (Y. Tong), [email protected] (A. Rao). 0006-8993/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2011.10.036

BR A I N R ES E A RCH 1 4 30 ( 20 1 2 ) 6 8 –77

how working memory modulates attentional processing (Awh et al., 2000; Cabeza et al., 2003; Lavie, 2005; Melara and Algom, 2003; Melara et al., 2005). One suggested functional role of working memory is in maintaining an attentional bias (Banich et al., 2000a, 2000b; Desimone and Duncan, 1995) or processing priority (de Fockert et al., 2001) that guides attentional selection. In the Stroop (1935) task, for example, certain PFC activations reliably signal the task requirements (Brass and von Cramon, 2004; Derrfuss et al., 2005; for a review, see Brass et al., 2005) or task-relevant information (Banich et al., 2000a, 2000b, 2001) presumably held in working memory. As the load on working memory increases in an attention task, task-irrelevant information tends to be processed more extensively, manifested as increased activation in stimulus-specific areas of sensory cortex and resulting in larger behavioral interference from distractors (Banich et al., 2001; Lavie et al., 2004). One interpretation is that high memory load obscures processing priorities, allowing task-irrelevant information to undermine target recognition (de Fockert et al., 2001). This is perhaps why individuals with small working memory span are relatively more prone to distractions from irrelevant stimuli (Vogel and Fukuda, 2009). One aim of the current study is to ask whether improved inhibitory control over the task-irrelevant information held in working memory helps reestablish task-relevant priorities, thereby enhancing selective attention performance.

1.2.

Effects of training on inhibitory control of distractors

To manipulate inhibitory control we trained participants to suppress task-irrelevant information in an auditory selective attention task. Studies of auditory discrimination and auditory selective attention indicate that tonal experience yields reliable, long-lasting electrophysiological changes associated with specific auditory functions. Training aimed at improving sensory discrimination, for example, has been shown to boost the amplitude of several distinct event-related potential (ERP) waves, including the mismatch negativity (Atienza and Cantero, 2005; Atienza et al., 2004; Kraus et al., 1995; Näätänen et al., 1993), the N1 wave (Reinke et al., 2003; Tremblay et al., 2001; Tremblay and Kraus, 2002; see also Menning et al., 2000), and the P2 wave (Alain et al., 2007; Atienza et al., 2002; Reinke et al., 2003; Shahin et al., 2005; Sheehan et al., 2005; Tremblay and Kraus, 2002; Tremblay et al., 2001; Tremblay et al., 2010). The current study seeks to identify effects of training aimed at improving selective attention on the auditory N1, P2, and P3 waves. The auditory P2 wave has proven particularly amenable to experiential influence (Salo et al., 2003; Shahin et al., 2005). In discrimination paradigms, P2 is believed to reflect automatic access to perceptual representations (Tong et al., 2009). In selective attention paradigms, automatic access to distractors can be dampened through inhibitory control, and P2 indexes the course and extent of distractor inhibition (Garcia-Larrea et al., 1992; Melara et al., 2002). P2 also may serve as a preattentive alerting mechanism to improve perception (Tremblay and Kraus, 2002) or the fidelity of traces available in short-term memory (Atienza et al., 2002). The present study sought to link experience-dependent changes in the behavior of the P2 wave evoked by distractors – alongside the N1 and P3 waves – with

69

improved inhibitory control of distractor information in working memory (Ceponiené et al., 2005).

1.3.

P3 as an index of salience in working memory

As perceivers gain control over the distractors contained in working memory they are better able to recognize and classify task-defined targets. The P3 ERP wave provides an electrophysiological gauge of the salience and ease of classification of stimuli held in working memory (Donchin, 1981; Donchin and Coles, 1988; Karis et al., 1984). Distraction weakens P3 amplitude to targets (Melara et al., 2002), a result consistent with the established role of equivocation (Ruchkin and Sutton, 1978) or task complexity on P3 (e.g., Kramer et al., 1986; Okita et al., 1985). In the context of processing priorities, the presence of distracting stimuli may blur the perceived distinctiveness of targets from surrounding stimuli. To the extent that participants learn to inhibit distractors, as indexed by P2, one would expect both an increase in target salience in working memory (Alvarez and Cavanagh, 2004), measured by greater P3 amplitude to targets, and a decrease in distractor salience in working memory (Lu and Dosher, 1998; Wilken and Ma, 2004), measured by weakened P3 amplitude to distractors.

1.4.

Theoretical predictions of working memory on attention

The present study considers how inhibition training affects auditory selection in the context of three current theories of attention: biased competition, signal detection, and tectonic theory. Each theory implicates working memory in modulating task-irrelevant information. From the perspective of biased competition (Desimone and Duncan, 1995), an attentional template held in working memory tilts processing in favor of taskrelevant representations. The theory predicts that repeated stimulus exposure during training gradually highlights those features of distractor representations that are distinct from the template, leading to improved target detection in the face of distraction (Duncan and Humphreys, 1989). A signaldetection theoretic approach attributes decreased target sensitivity from distraction to overlap between target (signal) and distractor (noise) distributions (Wilken and Ma, 2004); as participants learn to ignore irrelevant signals, the representational activity of distractors in working memory would be expected to decrease (e.g., Lu and Dosher, 1998), thus reducing distributional overlap as a function of training. The tectonic theory of Melara and Algom (2003) holds that prefrontal control of working memory boosts activation to task-relevant information and suppresses activation of task-irrelevant information; training in distractor suppression would enhance the precision of control processes acting on distractor representations, thereby improving target recognition. Later we explore how accurately these theories capture the effects of inhibitory training on a set of behavioral and ERP indices of auditory selection.

1.5.

Changes in selection efficiency with temporal distance

A final aim of the current study was to examine the effects of training as a function of the temporal separation between distractors and targets in the stimulus stream. Distractors and targets were separated by one, two, three, or four intervening

70

BR A IN RE S E A RCH 1 4 30 ( 20 1 2 ) 6 8 –77

standard stimuli. Targets that follow soon after a distractor are apt to suffer most in detectability because of lingering distractor activation in working memory (Roeber et al., 2003). Conversely, the strength of distractor inhibition should build as the temporal distance between distractors and targets lengthens. We expected inhibition training to enhance target detection across the range of temporal distances between distractors and targets by boosting suppression of recent distractor memories. Again, decreased distractor activity from training, as associated here with temporal distance, should result in an enhanced salience of targets in working memory, as measured by P3 amplitude. We asked whether such inhibition training endures by testing participants one week and again one month after training.

2.

Results

1.6.

2.1.

Behavioral analyses

2.1.1.

Perceptual sensitivity

To investigate the mechanisms of distractor processing operating in working memory during selective attention, a single experiment was conducted that examined the effects of attention training during performance of an auditory selection task. Fig. 1 depicts the dual-channel task (Näätänen, 1992) employed: Participants were engaged in detecting targets (T) among a stream of standards (S) in one ear as they practiced suppressing a distractor (D) in the other ear, which was gradually increased in intensity across the three weeks of training. We manipulated the number of standards between a distractor in the ignored ear and a subsequent target in the attended ear; four distractor-target distances were created – 1, 2, 3, and 4 – with each value designating the number of intervening standards. Behavioral and electrophysiological measurements were made before training, on the last day of training, one week after training, and one month after training to assess the effects of selection experience to the distractor, which was physically unchanged across testing sessions. The study had two primary goals. First, we sought to examine the effects of training on distractor suppression and its possible consequences to the distinctiveness of targets in working memory. Second, we sought to

Dual Channel Task ATTEND

IGNORE

TARGET

DISTRACTOR

A summary of behavioral performance appears in Panels A and B of Fig. 2. The correlation between speed and accuracy was .90. No participant falsely responded to the distractor as target. Separate ANOVAs were performed on the three behavioral measures to targets, sensitivity (d′), response criterion (c), and RT, with Training (4 levels: pretest, last training day, posttest 1, posttest 2) and Distance (4 levels) as within-subject factors. For d′, there was a main effect of Training, F(3,24)=5.74, p<.01, MSe =.59. Post-hoc analysis (Newman–Keuls, .05 criterion) revealed that performance at pretest (2.36) was significantly worse than performance at the end of training (2.74) or in either Posttest 1 (3.11) or Posttest 2 (2.79). Fig. 2A suggests that the effects of training dissipated over time: Performance was best one week after training; by one month after training, performance retreated to levels seen immediately after training. Fig. 2A also reveals a main effect of Distance

A Sensitivity (d')

The present study

examine the effects of training on distractor-target distance in working memory. These goals were achieved using behavioral analysis of target sensitivity, response criterion, and RT, and electrophysiological analysis focused on the auditory N1, P2 and P3 waves to distractors and the P3 wave to targets. Correlational analyses were used to connect behavioral and electrophysiological outcomes. To anticipate, we found that both the training and distance manipulations significantly affected selective attention performance, but that the influences of these two factors were largely distinct and independent.

3.75 3.5 3.25 3 2.75 2.5 2.25 2 1.75 1.5

Pretest Training Posttest1 Posttest2

1

2

3

4

Temporal Distance

B STANDARD

Response Bias (c)

1.4

S S

T S S S

D

1.2

Pretest Training

1

Posttest1

0.8

Posttest2

0.6 0.4 1

2

3

4

Temporal Distance

Fig. 1 – Schematic of the dual-channel task used in this study.

Fig. 2 – Summary of sensitivity (Panel A) and response bias (Panel B) for the four distractor-target distances at pretest, training, and posttests. Error bars indicate 1 SD.

250

300

350

400

450

100

150

200

250

250

300

350

400

450

12 F4 10 8 6 4 2 0 -50 -2 0 50 -4 -6 -8 -10 -12

100

Time (in ms)

300

350

400

450

Voltage (in microvolts)

Voltage (in microvolts)

100

200

N1

Time (in ms) 12 C3 10 8 6 4 2 0 -50 -2 0 50 -4 -6 -8 -10 -12

150

Voltage (in microvolts)

200

P2

12 Cz 10 8 6 4 2 0 -50 -2 0 50 -4 -6 -8 -10 -12

100

Time (in ms)

150

200

250

300

350

400

450

Time (in ms) Pretest

Training

150

200

250

300

350

400

450

300

350

400

450

Time (in ms)

Voltage (in microvolts)

150

Voltage (in microvolts)

Voltage (in microvolts)

100

12 10 Fz 8 6 4 2 0 -50 -2 0 50 -4 -6 -8 -10 -12

Posttest 1

12 C4 10 8 6 4 2 0 -50 -2 0 50 -4 -6 -8 -10 -12

100

150

200

250

BR A I N R ES E A RCH 1 4 30 ( 20 1 2 ) 6 8 –77

12 F3 10 8 6 4 2 0 -50 -2 0 50 -4 -6 -8 -10 -12

Time (in ms) Posttest 2

Fig. 3 – Grand-averaged ERP waveforms at Fz, F3, F4, Cz, C3, and C4 during pretest, training, and posttests.

71

72

BR A IN RE S E A RCH 1 4 30 ( 20 1 2 ) 6 8 –77

A

2.1.2.

Distractor P2 Peak (in microvolts)

on perceptual sensitivity, F(3,24)=3.64, p<.05, MSe =.37. At each point in training sensitivity at the closest proximity (2.48) was significantly worse than at distances 3 (2.93) or 4 (2.82). There was no interaction between Test and Distance, F(9,72)=.93, ns, MSe =.12.

Response criterion

A similar pattern was found for response bias. There was a main effect of Training, F(3,24) = 5.43, p < .01, MSe = .12, with participants responding less conservatively to targets after training than before training (see Fig. 2B). There also was a main effect of Distance, F(3,24) = 8.95, p < .001, MSe = .08, with participants less conservative at the longer distractor-target distances than at the shorter distances (see Fig. 2B). Again, there was no interaction, F(9,72) = .86, ns, MSe = .03.

2.1.3.

2.2.1.

Distractor N1 wave

2.2.2.

Distractor P2 wave

Training also affected the amplitude of the P2 wave, an effect restricted to frontocentral electrode sites, F(30,240) = 1.81, p < .01, MSe = 5.20. As can be seen in Fig. 4A, the amplitude of P2 was significantly greater (Newman–Keuls post-hoc test, .05 criterion) during the first and second posttests than during pretest or immediately after training (see also Fig. 3). Across training sessions there was a significant association (r = .87) between the amplitude of the P2 wave to the distractor and the amplitude of the P3 wave to the target.

2.2.3.

Distractor P3 wave

Training weakened the magnitude of the P3 wave to the distractor, F(3,24) = 9.49, p <.001, MSe =34.37, particularly over frontal electrode locations, F(30,240) =2.91, p <.001, MSe =3.72. As shown in Fig. 4B, post-hoc analyses revealed distractor P3 to be significantly larger at pretest (6.33 μV) than immediately after training (2.32 μV) or during the first (3.32 μV) or second (2.74 μV) posttests.

5 4 3 2 1

Training

Posttest 1

Posttest 2

Testing Session

B 8

Distractor P3 peak (in microvolts)

ERP analyses

An ANOVA of the latency of the distractor N1 wave showed that it peaked earlier in the first and second posttests than either before or immediately after training, F(3,24)=4.71, p<.05, MSe =399.27. Distractor N1 latency correlated significantly (r=.78) with the response criterion to the target: The earlier the peak latency to the distractor, the less biased the responding.

6

Pretest

Reaction time

Grand-averaged ERP waveforms at frontocentral electrode sites in each testing session appear in Fig. 3. There was no interaction between distance and training in any of the ERP analyses. Moreover, training had no effect on ERP waves to the target or to the standards. Training affected electrophysiological activity to the distractor in three different ERP waves, namely, N1, P2, and P3.

7

0

The effect of training on RT was only marginal, F(3,24) = 2.85, p = .06, MSe = 274.63, with RTs slowest on average at pretest. There was, however, a significant effect of distance on RT, F (3,24) = 9.79, p < .001, MSe = 525.71. Participants responded more slowly at the shortest distractor-target distance (524 ms) than at any of the longer distances (distance 2 = 505 ms; distance 3 = 498 ms; distance 4 = 499 ms). There was no interaction between Test and Distance, F(9,72) = .40, ns, MSe = 253.41.

2.2.

8

7 6 5 4 3 2 1 0

Pretest

Training

Posttest 1

Posttest 2

Testing Session Fig. 4 – Average amplitudes of P2 component to distractors (Panel A) and P3 component to distractors (Panel B) during pretest, training, and posttests. Error bars indicate 1 SD. Distractor P3 amplitude correlated significantly (r= −.67) with target sensitivity: The smaller the P3 wave to the distractor, the more sensitive participants were at detecting the target.

2.2.4.

Target P3 wave

The distance between distractor and target affected the P3 wave to targets: The amplitude of target P3 increased significantly with distance, F(3,24)=3.67, p<.05, MSe =6.15 (see Fig. 5). Interestingly, the amplitudes of target P3 and distractor P3 did not covary (r=.03). Yet target P3 amplitude correlated strongly (r=−.96) across distances with the participants' response bias: The larger the magnitudes of the P3 wave to targets, the more lax the criterion.

3.

Discussion

Our experiment revealed unique effects of attention training and temporal separation on behavioral and electrophysiological measures of target and distractor processing. Behaviorally, the three weeks of training led to long-lasting improvements in observers' ability to detect targets in the face of intermittent distractors. Improvements in target sensitivity were accompanied by less conservative and more rapid responding to targets, in line with previous studies of attention training (Melara et al., 2002). Performance was worst when distractor-target proximity

BR A I N R ES E A RCH 1 4 30 ( 20 1 2 ) 6 8 –77

Target P3 peak (in microvolts)

5

73

contrast, behavioral gains from training were greatest by the first posttest. These results fit with previous studies showing that behavioral and electrophysiological measures follow different time courses in response to training (Atienza et al., 2002; Atienza et al., 2005; Tremblay et al., 1998), thus perhaps reflecting different learning mechanisms (Ahissar and Hochstein, 2004). The slow changes we observed in P2 amplitude may gauge long-term consolidation of the learning experience (Alain et al., 2007; Karni and Sagi, 1993) or the transfer of training to attention situations outside the laboratory.

4

3

2

1

0 1

2

3

4

3.2.

Temporal distance and the attentional blink

Temporal Distance Fig. 5 – Average amplitude of P3 component to target at each of four distractor-target distances.

was near, but training effectively shortened the temporal distance needed to achieve a fixed level of target sensitivity. In the current study the contributions of distance and training to performance were largely distinct and independent, with a noticeable absence of statistical interaction between the two factors in the behavioral analyses and, perhaps not surprisingly, no ERP effects of distractor suppression training observed in analyses of target stimuli. By contrast, ERP effects of training were widely observed in analyses of the distracting stimuli. Training enhanced the P2 wave to distractors, an outcome previously interpreted as increased distractor inhibition (Melara et al., 2002, 2005). Training also reduced the P3 amplitude to distractors, suggesting a weakening of distractor distinctiveness in working memory. The effects of training on target salience, measured as the amplitude of the P3 wave to targets, were only indirect: As distractor P2 amplitude increased, so too did target P3 amplitude (see also Gjini et al., 2010). Nevertheless, our manipulation of distractortarget distance yielded marked electrophysiological effects on target processing: Target P3 amplitude increased as the temporal distance from distractors lengthened.

3.1.

Time course of inhibition training

The current results replicate and extend previous effects of training on discrimination and attention uncovered in our and other laboratories (Lu and Dosher, 1998; Melara et al., 2002; Reinke et al., 2003; Reisberg et al., 1980; Sturm et al., 1997; Tong et al., 2009). We found that a training regimen in which distractor intensity was progressively increased was associated with both improved behavioral performance and enhanced P2 amplitude. Although the present study did not include a separate group withheld from training (see also Banai et al., 2010; Wright and Sabin, 2007), we found previously (Melara et al., 2002) that control participants given the same stimulus exposure and discrimination experience as experimental participants, but no inhibitory training, showed little change from pretest to posttest in either selective attention performance or P2 amplitude (see also Reinke et al., 2003; but see Sheehan et al., 2005). The current study further revealed that the effects of training on P2 were not immediate, and did not peak until the last posttest, which occurred several months after the last training session. By

Training yielded constant improvement across the four temporal distances, with progressively worse performance as fewer standards intervened between a distractor and a subsequent target. The role of temporal distance in our dual-task paradigm can be compared with the role of lag in the attentional blink paradigm, in which detection of a second target during rapid serial presentation is poor if relatively few items (short lag) intervene after presentation of the first target (Broadbent and Broadbent, 1987; Raymond et al., 1992; Shapiro et al., 1997). Although most investigations of attentional blink have been limited to visual modality, an auditory analog exists (Shen and Mondor, 2006). According to at least one model (Olivers, 2007; Olivers and Meeter, 2008; Olivers et al., 2011), the attentional blink emerges from inhibitory processing of the intervening items. Applying this view to findings of the current study, the effects of temporal distance can be explained by a progressive dissipation of distractor inhibition as successive standards lead up to the next target. Yet, to the extent that our training regimen enhanced distractor inhibition, this account inaccurately predicts that training should lengthen rather than shorten the temporal distance needed to maintain target sensitivity. The results of the current study are more consistent instead with the view that distractors deplete the cognitive resources needed to process targets (Chun and Potter, 1995; Dell'acqua et al., 2009; Jolicoeur and Dell'Acqua, 1998), which perhaps are restored as a byproduct of inhibitory training. Future research is needed to determine whether the same mechanisms are at work in generating the auditory attentional blink and the effects of temporal distance found here.

3.3.

Three working memory accounts of selective attention

Several different approaches are possible in explaining the pattern of behavioral and electrophysiological results from the current study. We consider three models: biased competition, signal-detection theory, and tectonic theory. The processes involved in each model are not mutually exclusive. Nevertheless, the models represent three distinct perspectives for explaining the effects of attention training and temporal proximity on performance. The biased competition model of Desimone and Duncan (1995) suggests that an attentional template held in working memory shifts processing in favor of target representations during selective attention tasks (see also Näätänen, 1982). The fewer features of non-target (distractor) stimuli in common with the template, the stronger the bias toward target features and hence the greater their competitive advantage

74

BR A IN RE S E A RCH 1 4 30 ( 20 1 2 ) 6 8 –77

in processing. On this view, repeated stimulus exposure during training would gradually highlight those features of distractor representations that are distinct from the template, indexed perhaps as earlier distractor N1 latency or greater distractor P2 amplitude (Atienza et al., 2002), thereby effectively reducing the perceived similarity between targets and distractors (Duncan and Humphreys, 1989). One consequence might be an experience-dependent renewal of the task-relevant priorities in working memory, leading to improved target detection in the face of distraction. One could further hypothesize that greater temporal distance from distractor stimuli facilitates efficient comparison of targets with the template in working memory, indexed by increased target P3 amplitude. From the perspective of signal-detection theory (Green and Swets, 1966; MacMillan and Creelman, 1991), decreased target sensitivity in the face of distraction results from overlap, and the consequent uncertainty in choosing, between the distribution of target representations (signal) and the distribution of distractor representations (noise) (Wilken and Ma, 2004). The noise of distractor activation might also reconfigure attentional filters to non-optimal settings (Di Lollo et al., 2005). From this perspective, the perceptual context established by the presence of distractors in the stimulus environment influences the distinctiveness of target stimuli in memory (Durlach and Braida, 1969). By reducing distractor-target separation, the average temporal distance between standards and targets lengthens, hence boosting context variance in memory, which effectively lowers target sensitivity (Berliner and Durlach, 1973; Tanner, 1961) and perhaps also target P3 amplitude (Tong and Melara, 2007). Conversely, one would expect participants to be more confident in their decision-making as target salience grows, in line with the correlation we found between response bias and target P3 amplitude. Moreover, as participants learn to ignore irrelevant signals, noise accounts suggest that the representational activity of distractors in working memory would decrease (e.g., Lu and Dosher, 1998), thus reducing distractor salience (as measured by P3 amplitude) and decisional uncertainty (as measured by response bias) as a function of training. The tectonic theory of Melara and Algom (2003) suggests that prefrontal control processes operate continuously during selective attention tasks on representations in perception and working memory, boosting activation to task-relevant information and suppressing activation of task-irrelevant information until evidence favoring a target decision has reached a criterion. Such excitatory and inhibitory processes are functionally separate, and therefore separately modifiable through experience. Training in distractor suppression would therefore be expected to (1) enhance the precision of control processes acting on distractor representations, reflected in enhancements in distractor P2 amplitude (Melara et al., 2002), (2) decrease distractor salience in memory, measured by distractor P3 amplitude, and (3) link distractor inhibition to improved target recognition, seen here in the strong correlation between distractor P2 amplitude and target P3 amplitude and between distractor P3 amplitude and target sensitivity. Inhibitory processing triggered by distractor onset requires time to reach optimal levels, suggesting that both target sensitivity (d′) and target salience (P3) would vary with distractor-target proximity. It was not the intent of the present study to formally evaluate these three accounts. Thus, at present, it is unclear whether

attention training mainly sharpens or restores processing priorities (biased competition), reduces distractor noise (signal detection theory), or boosts inhibitory control (tectonic theory). Still, certain considerations suggest that the three views differ in explanatory power. The representational noise model, for example, is satisfying as a descriptive account of the current results, but not as a mechanistic one. An important unanswered question of the model is: What neural or psychological processes would cause a decline in distractor noise after training? Moreover, technically, the decision space in our task pitted targets (signal) against standards (noise), not distractors, leaving open the question of how distractor noise actually affects target decisions, particularly since our distractors did not elicit false alarms, either before or after training. The explanatory power of the biased competition account hinges on how closely the similarity between targets and distractors determines performance. Yet we intentionally chose a distractor stimulus perceptually distinct in auditory frequency from both target and standard. Indeed, in a previous study designed explicitly to compare model predictions (Melara et al., 2005), we found that distractor-target similarity (biased competition) fares worse than distractor salience (tectonic theory) in accounting for instances and magnitudes of selection failure. Moreover, if training in the current study improved selection success by enhancing the distinctiveness of distractor features from the attentional template, why would the primary electrophysiological measure of distractor salience (distractor P3 amplitude) shrink rather than grow with training? One limitation of the biased competition account, then, is its theoretical emphasis on mechanisms of target processing to the relative neglect of distractor processing. In our view, a mechanism of increased inhibitory control from training (tectonic theory) holds the greatest explanatory promise among the three accounts, while raising the fewest unanswered questions. The tectonic account formally links inhibitory activity to decreased activation of distractor representations in working memory and increased rates of evidence accumulation in reaching target decisions. Hence, the model handles well the range of behavioral, electrophysiological, and correlational results uncovered in the present study, including effects of training on distractor P2 and P3 amplitudes and the association between distractor P3 amplitude and target sensitivity. Nevertheless, in its present form, the model's only formal predictions concern behavioral data. The model also lacks an explicit learning mechanism to account for training effects. Thus, future efforts at modeling must be aimed at a clearer understanding of the precise electrophysiological consequences of distractor suppression and other forms of attention training.

4.

Experimental procedures

4.1.

Participants

Nine right-handed volunteers from Purdue University (4 men, 5 women) participated for course credit or payment. All participants had normal hearing sensitivity, defined as pure tone sensitivity better than or equal to 20 dB HL bilaterally for octave frequencies of 0.5, 1, 2, and 4 kHz. The nature of the

BR A I N R ES E A RCH 1 4 30 ( 20 1 2 ) 6 8 –77

procedures was explained fully, and informed consent was obtained from each participant. The Institutional Review Board of Purdue University approved the protocol.

keypad. Behavioral and electrophysiological data were collected during each session. 4.3.

4.2.

75

Behavioral analysis

Stimuli, apparatus, and procedure

Pure tones in the 1000 Hz range served as stimuli. Sinusoids were generated by an analog signal processor, digitized to 8 bits at a sampling rate of 22 kHz, and stored on a Macintosh Quadra 650 computer. Each tone was 100 ms in duration, including 10 ms rise/fall linear ramps presented through Nova 40 headphones. Sound intensity was calibrated on the A scale of a General Radio 1565-B sound-level meter fitted to a 2.4 cm diameter coupler. The stimulus set comprised a standard stimulus (1000 Hz), an infrequent distractor stimulus (1200 Hz), and an individually determined target deviant (ranging from 1002 Hz to 1014 Hz). The target stimulus was identified in a series of preexperimental baseline tasks, performed separately for each ear of each participant until a deviation from the standard was found that the participant could detect reliably at a d′ of approximately 2.5. The experiment proper employed a dual-channel oddball task (Näätänen, 1992; see Fig. 1). The task contained 336 trials, including 240 presentations of the standard, 48 presentations of the distractor, and 48 presentations of the target. The number of trials (standards) between a distractor in the ignored ear and a subsequent target in the attended ear was manipulated. Four distractor-target distances were created – 1, 2, 3, and 4 – with each value designating the number of intervening standards. We expected target detection to worsen as fewer standards intervened between the target and the preceding distractor. The onset-to-onset interval between any two stimuli ranged randomly between 703 ms and 870 ms in rectangular distribution. The study lasted eight weeks. Each participant was tested in a pretest session (1st week) and two posttests sessions (5th week [Posttest 1] and 8th week [Posttest 2]) occurring at the same time of day each week. Each test session contained three consecutive blocks of trials, with tones in the attended channel presented at 62 dB SPL and tones in the ignored channel at 82 dB SPL. The − 20 dB ratio between signal (target) and noise (distractor) was employed as a means of enhancing exogenous orienting to the distractor. Participants were instructed to detect appearances of the target in the attended channel by pressing with the dominant hand the zero key on a keypad as quickly and accurately as possible, disregarding all tones in the ignored channel. Three training sessions, each involving three blocks of 336 trials, were held during the second, third, and fourth weeks. Across the training sessions the signal-to-noise ratio was progressively reduced by increasing the intensity of infrequent distractors, from 42 dB in the first week of training, to 62 dB in the second week, to 82 dB in the final week, similar to Melara et al. (2002). The purpose of the training was to systematically expose the participants to progressively stronger intensities of the irrelevant signal to improve their ability to suppress distracting events (cf. Lu and Dosher, 1998; Reisberg et al., 1980; Sturm et al., 1997). Participants were tested individually in a dimly lit, sound attenuated, and electrically shielded Industrial Acoustics Company (New York) chamber. Each participant was seated in a comfortable recliner, arms on rests, within reach of the

Sensitivity, measured as d′ (i.e., z-transformed hits minus ztransformed false alarms), response criterion (c = −[z(hits) + z (false alarms)] / 2) (Green and Swets, 1966; MacMillan and Creelman, 1991), and RT to correct detections, were examined in separate ANOVAs. The hit rate was defined as the proportion of trials in which a participant pressed a key to the presentation of the target tone in the attended ear. The false alarm rate was defined as the proportion of trials in which a participant pressed a key to the standard tone in the attended ear. Any key press that occurred between 100 ms after stimulus onset and before the next stimulus onset was treated as valid. Within-subject factors for the ANOVA were Training (pretest, last training day, posttest1 and posttest2) and Distance (1, 2, 3, or 4 intervening standards). 4.4.

Electrophysiological analysis

In each session, local software (written in C) was used to control Grass Model 12 amplifiers in collecting continuous electrophysiological data from fourteen scalp locations (Fz, F3, F4, Cz, C3, C4, Pz, P3, P4, T3, T4, T5, T6, and Oz, 10–20 International System), with linked mastoids as reference. A cap fitted with tin electrodes (Electro-Cap International) was used to ensure proper placement of electrodes on the scalp. Fpz served as the ground electrode. Inter-electrode impedance below 1.5 kΩ (relative to the reference) was achieved by light scalp abrasion. Eye movements were monitored by bipolar horizontal and vertical EOGs. Vertical EOG was recorded from tin cup electrodes placed at the supra- and infraorbital ridges of the right eye. Horizontal EOG was recorded from cup electrodes placed at the outer canthus of each eye. EEG and EOG signals were amplified 5000 times and analog filtered with a bandpass from 0.1 to 100 Hz (−3 dB cutoffs), digitized at 256 Hz. Trials contaminated by blinks, eye movements, or motor movements (in excess of 100 μV) were rejected and re-presented later in the task. The sweep time was 500 ms (including a 50 ms pre-stimulus baseline), sufficient to allow the longest latency ERP wave we measured (P3) to return to baseline levels in each participant. Prior to statistical analysis, local software was used offline to average the time-locked waveforms of each stimulus type (standard, target, and distractor), separately for each day of testing (pretest, last training day, posttest 1, and posttest 2). Although EEGs were recording during each of the three training days, ERPs were examined only on the last training day to equate stimulus properties of the distractor across the testing sessions (i.e., −20 dB signal-to-noise ratio). ERPs were filtered digitally with a low pass of 40 Hz using a Blackman window (61 dB/octave). Peak amplitude and peak latency to each stimulus were measured at each scalp location for five ERP waves — P1 (43–82 ms post-onset search epoch), N1 (70–141 ms), P2 (125–227 ms), N2 (164–301 ms), and P3 (250–450 ms; targets and distractors only). ERPs to target stimuli were restricted to trials involving a correct behavioral response. The ERP waves were evaluated in separate ANOVAs, separately for each stimulus, with Training (pretest, last training day, posttest 1, posttest 2)

76

BR A IN RE S E A RCH 1 4 30 ( 20 1 2 ) 6 8 –77

and Distance (1, 2, 3, 4) as within-subject factors. All main effects and interactions reported as significant were reliable after Greenhouse–Geisser correction (Greenhouse and Geisser, 1959).

5.

Conclusions

The primary goals of the current study were to examine the effects of training on the distinctiveness of targets in working memory and on distractor-target proximity in working memory. We found in both cases that training had, at best, an indirect effect on target processing. In particular, training primarily affected the timing and amplitude of distractor processing which, in turn, modulated the salience of targets in working memory at each distractor-target distance. One possible explanation is enhanced prefrontal inhibitory control of representations in working memory conferred by extensive experience in distractor suppression.

REFERENCES

Ahissar, M., Hochstein, S., 2004. The reverse hierarchy theory of visual perceptual learning. Trends Cogn. Sci. 8, 457–464. Alain, C., Snyder, J.S., He, Y., Reinke, K.S., 2007. Changes in auditory cortex parallel rapid perceptual learning. Cereb. Cortex 17, 1074–1084. Alvarez, G.A., Cavanagh, P., 2004. The capacity of visual short-term memory is set both by visual information load and by number of objects. Psychol. Sci. 15, 106–111. Atienza, M., Cantero, J.L., 2005. Redefining memory consolidation. Behav. Brain Sci. 28, 64–65. Atienza, M., Cantero, J.L., Dominguez-Marin, E., 2002. The time course of neural changes underlying auditory perceptual learning. Learn. Mem. 9, 138–150. Atienza, M., Cantero, J.L., Stickgold, R., 2004. Posttraining sleep enhances automaticity in perceptual discrimination. J. Cogn. Neurosci. 16, 53–64. Atienza, M., Cantero, J.L., Quiroga, R.Q., 2005. Precise timing accounts for posttraining sleep-dependent enhancements of the auditory mismatch negativity. Neuroimage 26, 628–634. Awh, E., Anllo-Vento, L., Hillyard, S.A., 2000. The role of spatial selective attention in working memory for locations: evidence from event-related potentials. J. Cogn. Neurosci. 12, 840–847. Awh, E., Vogel, E.K., Oh, S.H., 2006. Interactions between attention and working memory. Neuroscience 139, 201–208. Banai, K., Ortiz, J.A., Oppenheimer, J.D., Wright, B.A., 2010. Learning two things at once: differential constraints on the acquisition and consolidation of perceptual learning. Neuroscience 165, 436–444. Banich, M.T., Milham, M.P., Atchley, R.A., Cohen, N.J., Webb, A., Wszalek, T., Kramer, A.F., Liang, Z.P., Barad, V., Gullet, D., Shah, C., Brown, C., 2000a. Prefrontal regions play a predominant role in imposing an attentional ‘set’: evidence from fMRI. Cogn. Brain Res. 10, 1–9. 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., 2000b. 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., Jacobson, B.L., Webb, A., Wszalek, T., Cohen, N.J., Kramer, A.F., 2001. Attentional selection and the processing of task-irrelevant information: insights from fMRI examinations of the Stroop task. Prog. Brain Res. 134, 459–470.

Berliner, J.E., Durlach, N.I., 1973. Intensity perception. IV. Resolution in roving-level discrimination. J. Acoust. Soc. Am. 53, 1270–1287. Bledowski, C., Prvulovic, D., Goebel, R., Zanella, F.E., Linden, D.E.J., 2004. Attentional systems in target and distractor processing: a combined ERP and fMRI study. Neuroimage 22, 530–540. Brass, M., von Cramon, D.Y., 2004. Selection for cognitive control: a functional magnetic resonance imaging study on the selection of task-relevant information. J. Neurosci. 24, 8847–8852. Brass, M., Derrfuss, J., Forstmann, B., von Cramon, D.Y., 2005. The role of the inferior frontal junction area in cognitive control. Trends Cogn. Sci. 9, 314–316. Broadbent, D.E., Broadbent, M.H., 1987. From detection to identification: response to multiple targets in rapid serial visual presentation. Percept. Psychophys. 42, 105–113. Cabeza, R., Dolcos, F., Prince, S.E., Rice, H.J., Weissman, D.H., Nyberg, L., 2003. Attention-related activity during episodic memory retrieval: a cross-function fMRI study. Neuropsychologia 41, 390–399. Ceponiené, R., Alku, P., Westerfield, M., Torkia, M., Townsend, J., 2005. ERPs differentiate syllable and nonphonetic sound processing in children and adults. Psychophysiology 42, 391. Chun, M.M., Potter, M.C., 1995. A two-stage model for multiple target detection in rapid serial visual presentation. J. Exp. Psychol. Hum. Percept. Perform. 21, 109–127. Cowan, N., 1995. Attention and Memory: An Integrated Framework. Oxford University Press, New York. de Fockert, J.W., Rees, G., Frith, C.D., Lavie, N., 2001. The role of working memory in visual selective attention. Science 291, 1803–1806. Dell'acqua, R., Jolicoeur, P., Luria, R., Pluchino, P., 2009. Reevaluating encoding-capacity limitations as a cause of the attentional blink. J. Exp. Psychol. Hum. Percept. Perform. 35, 338–351. Derrfuss, J., Brass, M., Neumann, J., von Cramon, D.Y., 2005. Involvement of the inferior frontal junction in cognitive control: meta-analyses of switching and Stroop studies. Hum. Brain Mapp. 25, 22–34. Desimone, R., Duncan, J., 1995. Neural mechanisms of selective visual attention. Annu. Rev. Neurosci. 18, 193–222. Di Lollo, V., Kawahara, J., Ghorashi, S.M.S., Enns, J.T., 2005. The attentional blink: resource depletion or temporary loss of control? Psychol. Res. 69, 191–200. Donchin, E., 1981. Surprise … surprise. Psychophysiology 18, 493–513. Donchin, E., Coles, M.G.H., 1988. Is the P300 component a manifestation of context updating. Behav. Brain Sci. 11, 357–374. Downing, P.E., 2000. Interactions between visual working memory and selective attention. Psychol. Sci. 11, 467–473. Duncan, J., Humphreys, G.W., 1989. Visual search and stimulus similarity. Psychol. Rev. 96, 433–458. Durlach, N.I., Braida, L.D., 1969. Intensity perception. I. Preliminary theory of intensity resolution. J. Acoust. Soc. Am. 46, 372–383. Fan, J., Flombaum, J.I., McCandliss, B.D., Thomas, K.M., Posner, M.I., 2003. Cognitive and brain consequences of conflict. Neuroimage 18, 42–57. Garcia-Larrea, L., Lukaszewicz, A.C., Mauguiere, F., 1992. Revisiting the oddball paradigm. Non-target vs neutral stimuli and the evaluation of ERP attentional effects. Neuropsychologia 30, 723–741. Gjini, K., Arfken, C., Boutros, N.N., 2010. Relationships between sensory “gating out” and sensory “gating in” of auditory evoked potentials in schizophrenia: a pilot study. Schizophr. Res. 121, 139–145. Green, D.M., Swets, J.A., 1966. Signal Detection Theory and Psychophysics. Wiley, New York.

BR A I N R ES E A RCH 1 4 30 ( 20 1 2 ) 6 8 –77

Greenhouse, S.W., Geisser, S., 1959. On methods in the analysis of profile data. Psychometrika 24, 95–112. Jolicoeur, P., Dell'Acqua, R., 1998. The demonstration of short-term consolidation. Cogn. Psychol. 36, 138–202. Karis, D., Fabiani, M., Donchin, E., 1984. P300 and memory — individual differences in the Von Restorff effect. Cogn. Psychol. 16, 177–216. Karni, A., Sagi, D., 1993. The time course of learning a visual skill. Nature 365, 250–252. Kramer, A., Schneider, W., Fisk, A., Donchin, E., 1986. The effects of practice and task structure on components of the event-related brain potential. Psychophysiology 23, 33–47. Kraus, N., Mcgee, T., Carrell, T.D., King, C., Tremblay, K., Nicol, T., 1995. Central aAuditory-system plasticity associated with speech-discrimination training. J. Cogn. Neurosci. 7, 25–32. Lavie, N., 2005. Distracted and confused?: selective attention under load. Trends Cogn. Sci. 9, 75–82. Lavie, N., Hirst, A., de Fockert, J.W., Viding, E., 2004. Load theory of selective attention and cognitive control. J. Exp. Psychol. Gen. 133, 339–354. Liu, X., Banich, M.T., Jacobson, B.L., Tanabe, J.L., 2004. Common and distinct neural substrates in an integrated Simon and spatial Stroop task as assessed by event-related fMRI. Neuroimage 22, 1097–1106. Lu, Z.L., Dosher, B.A., 1998. External noise distinguishes attention mechanisms. Vis. Res. 38, 1183–1198. MacDonald, A.W., Cohen, J.D., Stenger, V.A., Carter, C.S., 2000. Dissociating the role of the dorsolateral prefrontal and anterior cingulate cortex in cognitive control. Science 288, 1835–1838. MacMillan, N.A., Creelman, C.D., 1991. Detection theory: A user's guide. Cambridge University Press, New York. Melara, R.D., Algom, D., 2003. Driven by information: a tectonic theory of Stroop effects. Psychol. Rev. 110, 422–471. Melara, R.D., Nairne, J.S., 1991. On the nature of interactions between the past and the present. J. Exp. Psychol. Learn. Mem. Cogn. 17, 1124–1135. Melara, R.D., Rao, A., Tong, Y.X., 2002. The duality of selection: excitatory and inhibitory processes in auditory selective attention. J. Exp. Psychol. Hum. Percept. Perform. 28, 279–306. Melara, R.D., Chen, S.F., Wang, H.J., 2005. Inhibiting change: effects of memory on auditory selective attention. Cogn. Brain Res. 25, 431–442. Menning, M., Roberts, L.E., Pantev, C., 2000. Plastic changes in the auditory cortex induced by intensive frequency discrimination training. Neuroreport 11, 817–822. Milham, M.P., Banich, M.T., Webb, A., Barad, V., Cohen, N.J., Wszalek, T., Kramer, A.F., 2001. The relative involvement of anterior cingulate and prefrontal cortex in attentional control depends on nature of conflict. Cogn. Brain Res. 12, 467–473. Milham, M.P., Banich, M.T., Barada, V., 2003. Competition for priority in processing increases prefrontal cortex's involvement in top-down control: an event-related fMRI study of the Stroop task. Cogn. Brain Res. 17, 212–222. Näätänen, R., 1982. Processing negativity: an evoked-potential reflection of selective attention. Psychol. Bull. 92, 605–640. Näätänen, R., 1992. Attention and Brain Function. Lawrence Erlbaum, Hillsdale, NJ. Näätänen, R., Schroger, E., Karakas, S., Tervaniemi, M., Paavilainen, P., 1993. Development of a memory trace for a complex sound in the human brain. Neuroreport 4, 503–506. Okita, T., Wijers, A.A., Mulder, G., Mulder, L.J., 1985. Memory search and visual spatial attention: an event-related brain potential analysis. Acta Psychol. 60, 263–292. Olivers, C.N.L., 2007. The time course of attention — it is better than we thought. Curr. Dir. Psychol. Sci. 16, 11–15. Olivers, C.N.L., Meeter, M., 2008. Feature priming in visual search does not depend on the dimensional context. Vis. Cogn. 16, 785–803. Olivers, C.O., C. N. L., Hulleman, J., Spalek, T., Kawahara, J., Di Lollo, V., 2011. The Sparing is far from spurious: reevaluating

77

within-trial contingency effects in the attentional blink. J. Exp. Psychol. Hum. Percept. Perform. 37, 396–408. Postle, B.R., Awh, E., Jonides, J., Smith, E.E., D'Esposito, M., 2004. The where and how of attention-based rehearsal in spatial working memory. Cogn. Brain Res. 20, 194–205. Raymond, J.E., Shapiro, K.L., Arnell, K.M., 1992. Temporary suppression of visual processing in an RSVP task — an attentional blink. J. Exp. Psychol. Hum. Percept. Perform. 18, 849–860. Reinke, K.S., He, Y., Wang, C., Alain, C., 2003. Perceptual learning modulates sensory evoked response during vowel segregation. Cogn. Brain Res. 17, 781–791. Reisberg, D., Baron, J., Kemler, D.G., 1980. Overcoming Stroop interference — effects of practice on distractor potency. J. Exp. Psychol. Hum. Percept. Perform. 6, 140–150. Roeber, U., Widmann, A., Schroger, E., 2003. Auditory distraction by duration and location deviants: a behavioral and event-related potential study. Cogn. Brain Res. 17, 347–357. Ruchkin, D.S., Sutton, S., 1978. Emitted P300 potentials and temporal uncertainty. Electroencephalogr. Clin. Neurophysiol. 45, 268–277. Salo, S.K., Heikki Lang, A., Salmivalli, A.J., Johansson, R.K., Peltola, M.S., 2003. Contralateral white noise masking affects auditory N1 and P2 waves differently. J. Psychophysiol. 17, 189–194. Shahin, A., Roberts, L.E., Pantev, C., Trainor, L.J., Ross, B., 2005. Modulation of P2 auditory-evoked responses by the spectral complexity of musical sounds. Neuroreport 16, 1781–1785. Shapiro, K.L., Arnell, K.M., Raymond, J.E., 1997. The attentional blink. Trends Cogn. Sci. 1, 291–296. Sheehan, K.A., McArthur, G.M., Bishop, D.V., 2005. Is discrimination training necessary to cause changes in the P2 auditory event-related brain potential to speech sounds? Cogn. Brain Res. 25, 547–553. Shen, D., Mondor, T.A., 2006. Effects of distractor sounds on the auditory attentional blink. Percept. Psychophys. 68, 228–243. Stroop, J., 1935. Studies of inference in serial verbal reactions. J. Exp. Psychol. 18, 643–662. Sturm, W., Willmes, K., Orgass, B., Hartje, W., 1997. Do specific attention deficits need specific training? Neuropsychol. Rehabil. 7, 81–103. Tanner Jr., W.P., 1961. Physiological implications of psychophysical data. Ann. N. Y. Acad. Sci. 89, 752–765. Tong, Y.X., Melara, R.D., 2007. Behavioral and electrophysiological effects of distractor variation on auditory selective attention. Brain Res. 1166, 110–123. Tong, Y., Melara, R.D., Rao, A., 2009. P2 enhancement from auditory discrimination training is associated with improved reaction times. Brain Res. 1297, 80–88. Tremblay, K.L., Kraus, N., 2002. Auditory training induces asymmetrical changes in cortical neural activity. J. Speech Lang. Hear. Res. 45, 564–572. Tremblay, K.L., Kraus, N., McGee, T., 1998. The time course of auditory perceptual learning: neurophysiological changes during speech-sound training. Neuroreport 9, 3557–3560. Tremblay, K.L., Kraus, N., McGee, T., Ponton, C., Otis, B., 2001. Central auditory plasticity: changes in the N1–P2 complex after speech sound training. Ear Hear. 22, 79–90. Tremblay, K.L., Inoue, K., McClannahan, K., Ross, B., 2010. Repeated stimulus exposure alters the way sound is encoded in the human brain. PLoS ONE 5 (4), e10283. Vogel, E.K., Fukuda, K., 2009. In mind and out of phase. Proc. Natl. Acad. Sci. U. S. A. 106, 21017–21018. Wilken, P., Ma, W.J., 2004. A detection theory account of change detection. J. Vis. 4, 1120–1135. Wright, B.A., Sabin, A.T., 2007. Perceptual learning: how much daily training is enough? Exp. Brain Res. 180, 727–736.