The state of memory-matched distractor in working memory influence the visual attention.

Information in working memory (WM) can guide visual attention towards matched features. While recent work has suggested that cognitive control can act upon WM guidance of visual attention, little is known about how the state of memorized items retaining in WM contribute to its influence over attention. Here, we disentangle the role of inhibition and maintenance on WM-guided attention with a novel delayed match-to-sample dual-task. The results showed that active inhibition facilitated searching by diminishing sensory processing and deterring attentional guidance, indexed by an attenuated P1 amplitude and unaffected N2pc amplitude, respectively. By contrast, active maintenance impaired searching by attentional guidance while sensory processing remained unimpaired, indexed by an enhanced N2pc amplitude and unchanged P1 amplitude, respectively. Furthermore, multivariate pattern analyses could sucessfully decode maintenance and inhibition, suggesting that two states differed in modulating visual attention. We propose that remembered contents may play an anchoring role for attentional guidance, and the state of those contents retaining in WM may directly influence the shifting of attention. The maintenance could guide attention by accessing input information, while the inhibition could deter the shifting of attention by suppressing sensory processing. These findings provide a possible reinterpretation of the influence of WM on attention.

Introduction

In recent years, the interaction between working memory (WM) and attention has become a topic of considerable interest [1-3]. Some researchers have indicated that the maintained contents automatically guide visual attention [4-6], while some other researchers have found that presenting a memory-matched distractor did not capture individuals’ attention [7, 8]. These finding imply that there are different WM states which may have distinct impacts on visual attention. Olivers et al. (2011) proposed that WM representations may be in one of two different states: active and accessory. The active items are within the focus of executive processes and become an attentional template for guiding attention during visual tasks. By contrast, the accessory items are peripheral to the current mental manipulation and have relatively little influence on selective attention.

An intriguing question is whether the state of memorized items retaining in WM (e.g., inhibition and maintenance) contribute to this attentional guidance when the memory-matched items are presented as distractor in the search task. As explained earlier, the different WM representations may have different degrees of maintenance, implying that the state of memorized items retaining in WM play a critical role in the guidance effect. Remembered content and maintenance state (indicating the to-be-maintained content that is task-relevant) influence selective attention in the same direction (guiding attention to memory-matched information). Notably, the other WM state is inhibition, which can lead to rejecting or suppressing specific (e.g., irrelevant) information in WM [9]. In this case, participants may set up a “negative template” and avoid attending to these memory-matched contents, which may in turn improve distractor supression and therefore benefit target detection [7, 10, 11]. Thus, the inhibition state and remembered content may have opposite influences on selective attention: the former may force attention away from an memory-matched content, while the latter will direct attention to that item.

Olivers et al. (2006) reported that inhibited WM content matching a distractor resulted in a trend to facilitate visual attention. In their dual task, two consecutively presented colors were memorized; then, a color was cued to become relevant (to-be-maintained content) while another became irrelevant (to-be-inhibited content) for a subsequent memory test. Thus, it is possible that WM formed two separate memory templates based on the two colors. Because only one object can be actively maintained in WM at a time [3], participants might focus on the template of the to-be-maintained content. In this case, the operation on the template of the to-be-inhibited content should be passive and weak, which may account for the non-significant facilitation.

In the present study, we planned to use a revised dual task to investigate the influences of the state (especially, inhibition) of memorized items retaining in WM on visual attention. Specifically, two colors were first memorized simultaneously, and thereafter respectively cued as to-be-maintained content (cued condition) and to-be-inhibited content (uncued condition) in WM (see Fig 1), ensuring that they were remembered in the same WM template. In this case, participants were supposed to actively conduct both maintenance and inhibition on this template, which may benefit from top-down modulation of sensory processing [12]. Accordingly, the influence of inhibition on selective attention could be enhanced and revealed.

Fig 1

Schematic illustration of the trial sequence, stimulus displays, and response requirements in the present study.

The Search variable (cued, uncued, and neutral colors) was presented contralateral to the search target.

We recorded behavioral and EEG data of visual attention modulated by WM contents with different states, and focused on two ERP components: P1 and N2pc. P1 is a positive deflection peaking at about 100 ms, indexing selective attention at the sensory gating stage [13, 14]. N2pc is a negative-going voltage deflection that is typically observed around 200 ms after the onset of a visual search stimulus, which is sensitive to the shifts of selective attention [15-17]. Given that multivariate pattern analyses (MVPA) [18] can combine all characteristics of the data to help us understand the difference between inhibition and maintenance from the multivariate level, we used MVPA to maximally discriminate two WM states in modulating attention. We expected that they would result in distinct modulation patterns. The findings of this study not only highlight the role of WM state in the guidance of WM on visual attention, but may also open a new avenue to illuminate the interaction between WM and attention.

Methods

Participants

A priori statistical power analysis (power > 0.95) was performed based on assuming a small-to-medium effect size of 0.25 [19]. The projected minimum sample size needed was 14 participants. Considering the sample sizes of published studies, twenty-four (11 females, age from 18 to 23 years) right-handed college students were recruited and monetarily rewarded for participation. Data from two participants were excluded because of excessive errors (exceeding three standard deviations) in search performance. All participants reported normal or corrected-to-normal visual acuity and normal color vision. The written informed consent was obtained from each participant. This study was approved by the Review Board of Southwest University (China) for Human Participant Research. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Design

The state of WM manifesting in the search task were manipulated. Specifically, the contents remembered in WM appeared in the search task in half of the trials, among which the cued color (asking for maintenance) and the uncued color (asking for inhibition) were presented in the opposite visual field to the search target with equal probability. In the other half of the trials, colors other than the cued or the uncued color were displayed in the search task, which served as the neutral condition. Therefore, three levels of the Search variable (cued, uncued, and neutral conditions) were established in a 1:1:2 proportion, to avoid any bias to the memory contents. In the memory test, the cued and the uncued color appeared equally often and independently of each other, thus generating four combinations according to whether the cued and uncued colors were presented.

The participants in the practice session completed 24 trials with feedback on the memory test, which were excleded from the data analysis. Then, the experiment with 24 blocks of 24 trials was performed. Trials were pseudo-randomly assigned. Each block took 2–3 minutes, and participants had self-determined rest breaks.

Apparatus, stimuli, and procedure

E-Prime 1.1 (Psychology Software Tools Inc. Pittsburg, USA) was used to control the stimuli and collect responses. The stimuli were displayed on a 17-inch monitor with a resolution of 1,024 × 768 pixels and a refresh rate of 85 Hz. The distance between the monitor and the participants was approximately 70 cm. Stimuli were presented on a gray background (RGB values: 128, 128, 128). The color patch was a 1.6° × 1.6° visual angle, in the middle of which was a black line (0.8° × 0.1° visual angle) vertically presented (distractor) or tilted 15° (target). Ten colors were chosen from Munsell’s color system [20] with five principal hues (red, yellow, green, blue, and purple). Table 1 listed the hue, brightness, and chroma of these colors. The hue and value were nearly constant. The chroma of each color varied between 8 and 12.

Table 1

Details of the colors used in the experiment.

Number		Munsell (as chosen)			CIE (as measured)
		Hue	Value	Chroma	x	y	Y
Yellow	1	8Y	6	8	0.4297	0.4738	30.05
Yellow-red	2	8YR	6	11	0.5009	0.4316	30.05
Red	3	8R	6	12	0.4999	0.3476	30.05
Red-purple	4	8RP	6	12	0.4172	0.2814	30.05
Purple	5	8P	6	12	0.3136	0.2243	30.05
Purple-blue	6	8PB	6	12	0.2281	0.1980	30.05
Blue	7	8B	6	9.5	0.1996	0.2400	30.05
Blue-green	8	8BG	6.6	8.2	0.2212	0.3125	37.52
Green	9	8G	6.3	9	0.2438	0.3852	33.66
Green-yellow	0	8GY	6	12	0.3398	0.5548	30.05

Note. The Munsell values (red/green/yellow/purple/blue) were converted to red, green, and blue (RGB) values using the Munsell Conversion computer program [21]. The current colors and the original Munsell colors were not an exact match because of limited color space on the computer screens. CIE refers to Commission Internationale de L’Eclairage values.

The procedure scheme was displayed in Fig 1. Each trial began with a black cross presented for 500 ms. After that, the memory display with two different colors was presented for 1500 ms, and participants were required to remember both colors. The color patches were presented 1.5° up and down from the center fixation. The color patches were replaced by Mondrian patches for 500 ms, constituting the mask display. In the last 200 ms of the mask display, an upward or a downward arrow appeared randomly with equal probability on the fixation. The arrow pointed to either of the Mondrian patches where the masked color was to be tested in the later memory test, introducing the maintenance state on the cued color and the inhibition state on the uncued color. An inter-stimulus interval randomized between 1000 ms and 1200 ms was added before the search task. The search task contained four color patches separately in four quadrants on an imaginary clock face with a radius of 6.5°. Two different colors were chosen and assigned to the left and the right visual field, respectively, in which the color in the same visual field was identical. The cued color was located equally in the left and the right visual fields, as was the uncued color. The search target was presented in the four quadrants evenly. Participants discriminated the location of targets where the tilted line appeared by pressing keyboard (“F” for left side of the screen, “J” for right side of the screen). Notably, the search targets were always presented contralateral to the cued or uncued colors. After an interval of 500 ms, two different color patches were presented horizontally on the memory test, for which the participants were required to report whether the cued color was presented: “F”/“J” for the presence; “J”/“F” for its absence. It must be noted that the button responses were counterbalanced across participants in the memory test. Both the search task and the memory test were presented until a response was obtained or for 3000 ms. The inter-trial interval was 500 ms.

The colors presented in the dual task were controlled in certain respects. In the memory display, colors were chosen from two adjacent hues (red, yellow, green, blue, and purple) to ensure visual similarity, which was supposed to ensure visual representation [5]. In the search task, when the cued or the uncued color appeared as one color, the other color was the most dissimilar one to it. Otherwise, two different colors dissimilar to both the cued or the uncued color were displayed. Additionally, the colors were approximately equated with luminance contrast and color distance to avoid differences in physical attributes.

EEG recordings and pre-processing

The EEG data were recorded using a Brain Products system (band pass 0.01–100 Hz, sampling rate: 500 Hz, notch off) that was connected to a 64 scalp Ag-AgCl electrodes placed according to the international 10–20 system. All inter-electrode impedances were kept below 5 kΩ with the references on FCz and a ground electrode on AFz. The EOGs were simultaneously recorded from four surface electrodes, which were placed over the upper and lower eyelids, and then by the outer canthus of the left and right eyes to monitor ocular movements and eye blinks.

Offline EEG data were re-referenced to the bilateral mastoid electrodes and filtered with a band-pass filter of 0.05–40 Hz by using EEGLAB [22]. Continuous EEG data were segmented from -200 to 1000 ms relative to all stimulus markers. The baseline was corrected using the pre-stimulus time interval (-200 ms). Independent component analysis (ICA) decomposition was run on this dataset to remove ocular artifacts and additional noise [23]. In order to improve ICA decomposition, an initial rejection of noisy EEG data was performed by using visual inspection. Trials contaminated by gross movements were rejected. Automated detection and removal of artifactual independent components (e.g., blinks, eye-movements) were accomplished using the ADJUST toolbox [24]. ERP amplitudes were averaged across trials for each Search variable.

Data analyses

For all analyses, Greenhouse-Geisser correction for degrees of freedom was used whenever the assumption of sphericity was violated (p < .05). For multiple comparisons, the Holm-Bonferroni correction was used to adjusted p value.

Behavioral analyses

In the memory test, to exclude the influence of other confounders, only data from the neutral condition in search task were entered in the analysis of accuracy and reaction time (RT). All data were analyzed with a 2 (Presence of cued: Cued_presence, Cued_absence) × 2 (Presence of uncued: Uncued_presence, Uncued_absence) repeated-measures ANOVA.

In the search task, data from trials in which the search task and memory test were correct were entered in the analyses of RT and ERP. Trials with RT exceeding three standard errors for each participant were removed, leaving at least 104 trials for each condition for the analysis of RT. One-way repeated-measures ANOVA was carried out on the Search variable.

Univariate analyses

The P1 and N2pc components were analyzed in the search task. The P1 component peaked at around 100 ms poststimulus, which was analyzed in 60 to 120 ms interval to examine whether the states of remembered items retaining in WM influence the early visual procedding [13, 25]. The P1 component was analyzed at the lateral posterior electrode sites (PO7, PO8). A 2 (Lateral: ipsilateral, contralateral) × 3 (Search variable: cued, uncued and neutral) repeated-measures ANOVA was executed on P1 amplitudes.

The N2pc component was sourced at the pooled occipital electrodes (PO7/PO8) and obtained by subtracting the ipsilateral waveforms from the contralateral waveforms of the target, which could explore the attentional guidance for different WM states [15, 17]. Thus, the N2pc amplitude was analyzed in the Search variable with a one-way ANOVA. Additionally, the mean amplitudes averaged across 170–250 ms were compared between the contralateral and ipsilateral electrodes with paired-sample t tests for each Search variable, to estimate the guidance effect of the WM.

Multivariate pattern analyses

The MVPA was performed to maximally discriminate between inhibition and maintenance. A cross-subject approach was used in the present study, in which a classifier was trained on the data of 21 subjects and then applied to the remaining subject by using the leave-one-out cross-validation method. We applied an MVPA combining feature selection techniques with a support-vector machine (SVM) by using the LIBSVM toolbox [26]. The exceptionally high dimensionality of the data was reduced by selecting the most relevant features. Considering the main purpose of the current study, we chose attention-related components (P1, N2pc) to train the SVM classifier, as these components were closely related to visual attention [4, 14].

We employed attention-related features for which we combined data across corresponding electrodes (P1 amplitude: averaged across electrodes PO7 and PO8; N2pc amplitude: contralateral—ipsilateral waveforms) and time intervals of two ERP components. Each participant contained two WM states, and the averaged ERP waveforms at given time windows and corresponding electrodes were fed into the SVM classifier. To improve the signal-to-noise ratio, the data from a given state were averaged across trials, and decoding was performed on these averages rather than on single-trial data. Thus, two averaged ERP difference waves (maintenance = cued—neutral; inhibition = uncued—neutral) that can eliminate baseline interference served as classification features to decode maintenance and inhibition. To investigate whether decoding accuracy would be higher than chance level (50%), a permutation test was performed for the classification of WM state. To this end, each dataset was randomly labeled as maintenance or inhibition. The classifier was trained on these features. The randomization process was repeated 1000 times, and then the distribution of classifications was compared with the classification based on the real data.

Furthermore, we compared the predictive value of the averaged ERP difference wave approach with additional approaches that do not consider the influence of the baseline or both ERP amplitudes. The decoding procedure for additional approaches was the same as the above approach.

Results

Behavioral results

Memory test

Analyses of the accuracy showed a significant Presence of cued × Presence of uncued interaction (Fig 2a; F(1,21) = 6.53, p = .018, η2 = 0.24). Follow-up tests demonstrated that participants could obtain a high accuracy rate when the cued color was present in the memory test, regardless of whether the uncued color was present (90.3% vs. 92.7%; F(1,21) = 3.47, p = 0.076). However, when the cued color was absent, participants showed worse performance for the presence of the uncued color than its absence (73.7% vs. 81.1%; F(1,21) = 30.73, p < .001, η2 = 0.59). Also, the main effect of Presence of cued was significant (F(1,21) = 24.71, p < 0.001, η2 = 0.54), with higher accuracy in the presence (91.5%) than in the absence of the cued color (77.4%). And the main effect of Presence of uncued were significant (F(1,21) = 29.64, p < 0.001, η2 = 0.58), with lower accuracy in the presence (82.0%) than in the absence (86.9%) of the uncued color. Furthermore, when only the cued or uncued color was presented in the memory test, participants showed better performance for the presence of the cued color relative to the presence of the uncued color (t(21) = 6.44, p < 0.001).

Fig 2

Results of reaction time and accuracy in the search task and memory test, respectively.

Error bars denote the standard error of the mean (SEM). n.s. = non-significant, * p < .05, *** p < .001.

Only the correct trials were entered in the analysis of RT. There was a significant interaction between Presence of cued and Presence of uncued (Fig 2b; F(1,21) = 10.73, p = .004, η = 0.34). When the cued color was presented in the memory test, participants tended to respond faster in the presence than in the absence of the uncued color (793 vs. 821 ms; F(1,21) = 7.75, p = .011, η = 0.27). In contrast, when the cued color was absent, the difference between the presence and the absence of the uncued color was not significant (889 vs 864 ms; F(1,21) = 2.64, p = .119). The main effect of Presence of cued was significant (F(1,21) = 10.57, p = 0.004, η2 = 0.34), with faster response in the presence (807 ms) than in the absence (876 ms) of the cued color. The main effect of Presence of uncued was not significant (F(1,21) = 0.05, p = 0.831). Additionally, if we merely presented the cued or uncued color in the memory test, participants showed faster responses in the presence of the cued color than in the presence of the uncued color (t(21) = -2.86, p = 0.009).

Search task

Participants obtained high accuracy (over 99%) in all Search variables (Fig 2c; F(2,42) = 0.23, p = .799). The mean RT for three conditions was 764 ms. There was a significant main effect of the Search variable on the measures of RT (Fig 2d; F(1,31) = 33.94, p < .001, η = 0.62). The RT in the cued condition (803 ms) was slower than those in the neutral (752 ms; p < .001) and uncued conditions (738 ms; p < .001). Importantly, the RT in the uncued condition was significantly better than those in the neutral condition (p = .024). These results indicated that the appearance of the cued color opposite to the search target significantly hindered selection of the target, while the presence of the uncued colors significantly facilitated visual search.

Univariate analyses: P1 and N2pc

The analysis of P1 amplitude showed significant main effects of the Search variable (F(2,42) = 4.37, p = .019, η = 0.172). Pairwise comparisons of the Search variable (see Fig 3) showed smaller P1 amplitudes in the uncued condition (0.466 μV) than in the neutral condition (0.78 μV; p = .028), but no significant difference between the cued (0.77 μV) and neutral conditions (p = .92). The main effect of Lateral was not significant (F(1,21) = 0.18, p = .674). P1 amplitudes induced by three conditions were not lateralized, and equally present at electrodes ipsilateral and contralateral to the target. The interaction between Lateral and the Search variable was not significant (F(2,42) = 0.007, p = .993).

Fig 3

The P1 amplitudes for the search variable at the posterior electrode sites (averaged across electrodes PO7 and PO8) are presented.

Topography represents the P1 component measured as the mean amplitude between 60 and 120 ms. The left and right sides of the outline head represent left and right hemisphere locations, respectively. n.s. = non-significant, * p < .05.

As indicated by Fig 4a, 4b and 4c, the N2pc amplitude was strongly affected by the cued colors presented as distractors, with large amplitudes for contralateral waveforms than ipsilateral waveforms in the cued condition (t(21) = 5.00, p < .001). However, no difference emerged in the uncued (t(21) = 1.13, p = .273) and neutral conditions (t(21) = -0.57, p = .574). These observations were confirmed by the ANOVAs on the mean difference waves in the N2pc time window (170–250 ms). This analysis showed a remarkable main effect of the Search variable (Fig 4d; F(2,42) = 20.08, p < .001, η2 = 0.49), indicating that the N2pc amplitudes were larger for the cued condition (-0.881 μV) than for the uncued (-0.131 μV; p < .001) and neutral conditions (0.052 μV; p < .001); no difference emerged between the uncued and neutral conditions (p = .17).

Fig 4

The N2pc waveforms (a, b, c) for contralateral and ipsilateral electrical activity and mean difference wave (d) are presented.

Topography represents the N2pc component (170–250 ms). The left and right sides of the outline head represent ipsilateral and contralateral electrode sites of the target, respectively. *** p < .001.

MVPA analyses

MVPA were used to predict the WM states (maintenance or inhibition) based on attention-related ERP responses. First, the SVM classifier was trained on both ERP difference waves, resulting in a maximum accuracy of 68.18%. Permutation tests confirmed that this accuracy was significantly greater than chance (50%; p = 0.028).

Next, we compared our approach with one that did not consider the influence of neutral condition or both attention-related ERP components. We therefore averaged the ERP amplitude for each Search variable instead of calculating the difference wave between the cued/uncued and neutral conditions. This approach resulted in a maximum accuracy of 68.18% (p = 0.032). Then, we conducted a classification based on the single ERP amplitude or single difference wave. Classification based on single P1 or N2pc difference waves showed a maximum accuracy of 61.36% and 70.45%, respectively. Besides, single P1 or N2pc amplitude served as classification features, resulting in a maximum accuracy of 50% and 65.91%, respectively.

Overall, multivariate assessment of the ERP patterns allowed for significant predictions of the modulating pattern of WM states from brain activity, suggesting inhibition and maintenance are significantly different with each other in modulating visual attention.

Discussion

We used a new delayed match-to-sample procedure, in which both maintenance and inhibition of WM were actively executed, and recorded EEG data to investigate the influences of the WM states on visual search. The behavioral results showed that the time of visual search significantly increased in the cued condition but significantly speeded up in the uncued condition, compared to the neutral condition. Furthermore, the ERP results indicated that P1 amplitude was significantly attenuated in the uncued condition relative to the neutral condition, but was not different between the cued and neutral conditions. Interestingly, N2pc was significantly enhanced in the cued condition compared with the neutral and uncued conditions, but did not differ between the latter two conditions. Besides, the two WM states could be decoded using the SVM classifier, suggesting they were significantly different in modulating visual attention.

In the uncued condition, one color of the search display matching the uncued color (to-be-inhibited color in WM) was always located on the side contralateral to the target. To maximally reduce interference from the to-be-inhibited color in WM, the uncued color served as a template for rejection (negative template) that biased attention away from the contralateral side during visual search [10, 27–29], which in turn benefited attention to the ipsilateral side. Attenuated P1 amplitude in the uncued condition relative to the neutral condition suggested that the inhibition of WM diminished sensory processing [30, 31], which may have eliminated the guidance effect in the uncued condition; owing to this, the N2pc amplitude was comparable between the uncued and neutral conditions [15]. It is possible that diminished sensory processing helped participants effectively avoid noticing the contralateral color; in this case, their attention would not be guided to the contralateral side. Moreover, relative to the neutral condition, the uncued condition facilitated visual search behaviorally. Because the interference signal coming from uncued color matched location would obscure the signal coming from the location of search target, and the active suppression of interference signal would lead to improved behavioral accuracy (facilitation effect). In this case, attenuated P1 may reflect the mechanism of active suppression deriving from the inhibition [25].

The facilitation effect demonstrated that the to-be-inhibited color in the uncued condition was still represented in WM during the search task [16, 32], which was also supported by the finding that when the to-be-inhibited color appeared in the subsequent memory test, the test for the to-be-maintained color was significantly affected. Traditionally, content remembered in WM is thought to enable guidance of visual attention; if this is the case, then visual search would be impaired in the uncued condition. The overall pattern of facilitation in the uncued condition suggested that the facilitation effect induced by the inhibition of WM was stronger than the impairment effect induced by the remembered color; although, it is possible that the state of memorized items retaining in WM rather than WM content influenced visual attention. Regardless of which explanation will be verified, the current results clearly showed that WM representation is essential for guidance, in which it may play an anchoring role in attentional guidance. That is, only when one color was remembered in WM could the WM states (inhibition and maintenance) influence visual attention. Moreover, the current results highlighted the necessity of separating the two aspects of one WM representation (the remembered content and the state of those items retaining in WM) when studying how WM guides attention, because the evidence indicated that they make different but complementary contributions to guidance. Besides, although both colors were stored in WM representation, to-be-maintained color had larger bias signal with it under the model of Bundesen’s theory of visual attention (TVA) [33], which might generate attentional guidance; but to-be-inhibited color had a smaller attentional weight compared to neutral and to-be-maintained colors, which may facilate the search performance shown in Fig 2d. It is worth noting that this weight difference might be caused by maintenance and inhibition, suggesting the importance of WM states in guiding visual attention.

In accordance with previous studies [1, 4, 5, 8], our behavioral results showed that the to-be-maintained color in the cued condition impaired the visual search. In the present study, the memory-matched color was a distractor that was always presented contralateral to the search target. The to-be-maintained items might be viewed as a template for rejection in the search task [10], while these items should be maintained in the working memory task. In this case, the requirement of working memory and search demand engender different response, evoking a conflict [34-36]. Therefore, this conflict hindered the effect of rejection template, leading to slower response time in the cued condition than in the neutral condition. In our opinion, the maintenance state in the cued condition allowed participants’ attention to access input from a color matching the remembered color in WM. Since the color matching, attention would be directed to the contralateral side, indexed by the enhanced N2pc amplitude in the cued condition relative to the neutral condition, which therefore impaired the visual search. In the present study, the cued items induced larger N2pc amplitudes in contralateral side than in ipsilateral side; but no difference was found for the uncued/neutral items between the two sides. Thus, different from the cued condition, the uncued and neutral conditions would not capture attention at all. Here, the remembered color in WM could still play an anchoring role for the shifting of attention. Although shifting attention to the contralateral side is disruptive for the search task, access from the remembered color to the input color kept by the maintenance state favored the normal sensitivity of sensory processing in the cued condition, indexed by unchanged P1 amplitude between the cued and neutral conditions. Thus, despite the detrimental influence on the visual search, a shift of attention did occur in the cued condition, which also indicated the automaticity feature of guidance induced by WM representation [37]. From the current results, we further propose that the automaticity may be due to the maintenance state of WM, and therefore reflects an effect of active executive processing.

To decode the two states of WM from attention-related ERP response, the SVM classifier was trained and resulted in an accuracy of 68.18%. An MVPA approach can extract an identifying pattern of brain activity by constructing a classifier that maximally discriminates between mental states [18, 38, 39]. We applied multivariate patterns of ERP components (P1 and N2pc) to predict the WM states in modulating visual attention, as those components were commonly used to index selective attention [8, 14]. Thus, both P1 and N2pc components are important for identifying the WM states, which may help to build a reliable objective neuronal marker to explore the interaction between the WM states (maintenance and inhibition) and visual attention.

In addition, although the proportion between the presence and absence of the cued color was same in the memory test, participants showed better performance in the prior condition. Remembered contents would be stored in WM representation [5, 32], especially when those cued contents were related to the subsequent task. The representation of the cued items might be in an active state [3]. As indicated by previous study, visual WM capacity is a limited resource that can be dynamic shifted between objects [40]. The precision of WM is determined by the proportion of resource allocation to each object. In the present study, to-be-maintained items would have higher precision of WM than to-be-inhibited items. Therefore, whether the cued items appeared would induce unbalanced performance in the memory test. Besides, we also found that the uncued items might be mistaken for the cued items when the cued items were absent, suggesting that both the cued and uncued items existed in WM. However, they had distinct influence on visual search, which might be caused by the different states of remembered items retaining in WM.

These findings suggest a possible reinterpretation of the influence of WM on selective attention that is often described as top-down modulation [12]. We propose that the two aspects of WM representation should play different roles in the guidance effect: remembered content may be an anchor for the shift of attention, while the state of memorized items retaining in WM may maintain access from WM content to input information. The current findings highlight the roles of inhibition and maintenance in the guidance of WM on visual attention, in which to-be-maintained items function as an actively attentional template which may directly capture the attention and further impair the search for target. By contrast, to-be-inhibited items may be viewed as a “negative template” and can facilitate the search for target. Given that the other negative template studies involved an actively-maintained negative template that is being actively utilized during search [28, 41], future research should explore the role of WM inhibition in the active condition. Finally, as compared to univariate methods, the multivariate approach built a reliable instrument to explore the complex patterns of working memory in modulating visual attention.

In summary, by assembling to-be-maintained and to-be-inhibited colors in a common WM template, and especially utilizing the opposite influences of the inhibition and remembered content in WM on visual attention, we investigated the roles of the WM states in the guidance of WM during visual search. As the color matching the remembered content in WM which was always located on the side contralateral to the target, to-be-maintained colors impaired the visual search by guiding attention to the contralateral side but not diminishing early sensory processing. In contrast, to-be-inhibited colors facilitated visual search by diminishing sensory processing but not guiding attention to the contralateral side.

26 Aug 2020

PONE-D-20-19909

The Roles of Executive Operations in the Guidance of Working Memory on Visual Attention

PLOS ONE

Dear Dr. Chen,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Both reviewers agree that the study were well conducted in general. On the other hand, there are some places to be refined in terms of experimental settings and result explanation. I think these comments are direct and I will not repeat them here.

Please submit your revised manuscript by Oct 10 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Zaifeng Gao

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For more information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially sensitive information, data are owned by a third-party organization, etc.) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

3. PLOS requires an ORCID iD for the corresponding author in Editorial Manager on papers submitted after December 6th, 2016. Please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field. This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager. Please see the following video for instructions on linking an ORCID iD to your Editorial Manager account: https://www.youtube.com/watch?v=_xcclfuvtxQ

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Title: The Roles of Executive Operations in the Guidance of Working Memory on Visual Attention

The current study explored whether state of memorized items retaining in working memory (WM) influence attentional guidance when these items are presented as distractors in visual search task. To do so, the author employed a WM change-detection task and inserted a visual search task in maintenance phase, with EEG records. Participants were first required to memorize two colors and then maintenance the cued one as well as inhibit the uncued one. Then they were required to do a visual search task in which either the cued or uncued color could be distractors. Results showed that in visual search task, uncued condition had shorter search time and attenuated P1 amplitude than neutral condition, while cued condition had longer search time and enhanced N2pc amplitude than neutral condition. Furthermore, multivariate pattern analyses could classify cued and uncued conditions accurately. Together the authors suggest that the active item in WM could guide attention by accessing input information, while the inhibit item in WM could deter the shifting of attention by suppressing sensory processing.

Overall, the current study addressed an interesting question, the whole manuscript (ms) was elegantly organized, the data is solid. However, I have a few concerns about the current ms which discourage my enthusiasm. I will list them below regardless of importance.

Major Points:

1) The ms described current question as “whether the operations of WM contribute to attentional guidance when the contents remembered in WM are presented as distractor in the search task”. In terms of experimental procedure, I think the more accurate describe of current question is about state of memorized items retaining in WM instead of operations of WM. On the other hand, the current study focuses on the influence of distractors in a visual search task when the distractors are items retaining in a related WM task. So I would suggest the author added the information of distractors in title to make it more focal.

2) In line 200, “for multiple comparisons, p values were adjusted using the Least Significant Difference (LSD) correction.” The more appropriate way for multiple comparisons is Bonferroni correction.

3) In line 230, “a classifier was trained on 43 datasets and 231 then applied to the remaining dataset”. What the 44 datasets were? In context, the 44 datasets seems from 22 participants × 2 conditions (maintenance vs. inhibition). If so, I would suggest the author combining data from the same participant and treat it as one dataset, then train on 21 datasets from 21 participants and test the remaining dataset using the leave one-out cross-validation method. Because data from the same participants may have common characteristics related to the participants, which may lead to unexpected feature extraction in SVM.

4) WM task in ms is to judge whether the cued item is presented in memory test. In this task, proportion of presence and proportion of absence are the same, but task performance in cued presence condition and cued absence condition are quite different. I would suggest the author to discuss this unbalanced performance.

Minor Points:

5) In line 158, dose “hemisphere” means hemisphere of brain? Or does it means left side or right side of the screen?

6) Behavioral results in ms only report interaction now. I would suggest the author to report intact ANOVA results with main effects as well as interaction.

7) It will be a better exhibition of data in Figure 4 if add a brain electrical activity mapping like in Figure 3.

Reviewer #2: The present study presents an intriguing finding. I don't find any serious flaw in experimental design and data analyses.

However, I have several concerns regarding the interpretation of results and discussion.

First, not only the uncued item, but the cued item should also play a role in formulating the template for rejection. In the experiment, working memory items, maintained or inhibited, were always presented contralaterally to the target. Hence, any memory-matching stimulus is a distractor. In this case, the cued item should also be regarded as a template for rejection in visual search. Certainly, the cued item should be maintained in working memory, evoking a conflict between search demand and working memory task. This conflict might have hindered the effect of rejection template. For the uncued item, due the inhibitory tagging imposed on the stimulus, the rejection template could be established more effectively.

Second, it is nice to see that the amplitudes of the Pd and N2pc well match behavioral results. Contrary to this, I don't see any merit in the MVPA analyses. What does it add to the behavioral results and univariate ERP results? The cued item should be maintained in working memory and the uncured item is not in working memory. It is not surprising that this difference would yield huge difference in behavior and neural responses.

As a minor issue, English in the manuscript is not too bad, but it has to be more scholarly. Some of expressions are too colloquial and not completed sentences.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

8 Oct 2020

Authors’ Response to the Review Comments

Response to Comments from Reviewer 1

Reviewer #1: Title: The Roles of Executive Operations in the Guidance of Working Memory on Visual Attention

The current study explored whether state of memorized items retaining in working memory (WM) influence attentional guidance when these items are presented as distractors in visual search task. To do so, the author employed a WM change-detection task and inserted a visual search task in maintenance phase, with EEG records. Participants were first required to memorize two colors and then maintenance the cued one as well as inhibit the uncued one. Then they were required to do a visual search task in which either the cued or uncued color could be distractors. Results showed that in visual search task, uncued condition had shorter search time and attenuated P1 amplitude than neutral condition, while cued condition had longer search time and enhanced N2pc amplitude than neutral condition. Furthermore, multivariate pattern analyses could classify cued and uncued conditions accurately. Together the authors suggest that the active item in WM could guide attention by accessing input information, while the inhibit item in WM could deter the shifting of attention by suppressing sensory processing.

Overall, the current study addressed an interesting question, the whole manuscript (ms) was elegantly organized, the data is solid. However, I have a few concerns about the current ms which discourage my enthusiasm. I will list them below regardless of importance.

Major Points:

1) The ms described current question as “whether the operations of WM contribute to attentional guidance when the contents remembered in WM are presented as distractor in the search task”. In terms of experimental procedure, I think the more accurate describe of current question is about state of memorized items retaining in WM instead of operations of WM. On the other hand, the current study focuses on the influence of distractors in a visual search task when the distractors are items retaining in a related WM task. So I would suggest the author added the information of distractors in title to make it more focal.

Response: We greatly appreciated the reminder about this important issue. We totally agree that using “the state of rememorized items retaining in working memory” to describe the current question is more accurate. According to your advice, we revised the title of the manuscript and relevant content in the manuscript.

The new title is that “The state of memory-matched distractor in working memory influence the visual attention”.

“An intriguing question is whether the state of memorized items retaining in WM (e.g., inhibition and maintenance) contribute to this attentional guidance when the memory-matched items are presented as distractor in the search task.” (Introduction. Line 56-58)

“In the present study, we planned to use a revised dual task to investigate the influences of the state (especially, inhibition) of memorized items retaining in WM on visual attention.” (Introduction. Line 80-82)

2) In line 200, “for multiple comparisons, p values were adjusted using the Least Significant Difference (LSD) correction.” The more appropriate way for multiple comparisons is Bonferroni correction.

Response: Thanks for your reminder. We conducted the new data analyses, in which we adjusted p values by using Holm-Bonferroni correction instead of using LSD in the revised manuscript. And the new results of behavior and ERP were similar to previous results. The detailed results can be found in Results.

According to your advice, we revised the manuscript as below:

“For multiple comparisons, the Holm-Bonferroni correction was used to adjusted p value.” (Data analyses. Line 205-206)

3) In line 230, “a classifier was trained on 43 datasets and 231 then applied to the remaining dataset”. What the 44 datasets were? In context, the 44 datasets seems from 22 participants × 2 conditions (maintenance vs. inhibition). If so, I would suggest the author combining data from the same participant and treat it as one dataset, then train on 21 datasets from 21 participants and test the remaining dataset using the leave one-out cross-validation method. Because data from the same participants may have common characteristics related to the participants, which may lead to unexpected feature extraction in SVM.

Response: Thank you very much. In order to avoid unexpected feature extraction in SVM, we used a cross-subject approach to maximally discriminate between inhibition and maintenance in the new data analyses. To this end, we combined data from the same participant in the new analyses. A classifier was trained on the data of 21 subjects and then applied to the remaining subject by using the leave-one-out cross-validation method. The accuracy of classification was 68.18% (p = 0.028), suggesting that inhibition and maintenance are significantly different with each other in modulating visual attention.

According to your suggestion, we revised the corresponding contents as below:

“The MVPA was performed to maximally discriminate between inhibition and maintenance. A cross-subject approach was used in the present study, in which a classifier was trained on the data of 21 subjects and then applied to the remaining subject by using the leave-one-out cross-validation method. We applied an MVPA combining feature selection techniques with a support-vector machine (SVM) by using the LIBSVM toolbox [26].” (Methods. Data analyses. Line 233-238)

“First, the SVM classifier was trained on both ERP difference waves, resulting in a maximum accuracy of 68.18%. Permutation tests confirmed that this accuracy was significantly greater than chance (50%; p = 0.028).” (Results. MVPA analyses. Line 338-340)

4) WM task in ms is to judge whether the cued item is presented in memory test. In this task, proportion of presence and proportion of absence are the same, but task performance in cued presence condition and cued absence condition are quite different. I would suggest the author to discuss this unbalanced performance.

Response: Thank you for the insightful suggestion. We think that the unbalanced performance between the cued and the uncued conditions might be caused by different precision of working memory in the memory test. The visual working memory capacity is limited, and the proportion of resource allocation to each object determines the precision of working memory (Bays & Husain, 2008). In the revised manuscript, we discussed this important point.

According to your advice, we added the content to the Discussion.

“In addition, although the proportion between the presence and absence of the cued color was same in the memory test, participants showed better performance in the prior condition. Remembered contents would be stored in WM representation [5,32], especially when those cued contents were related to the subsequent task. The representation of the cued items might be in an active state [3]. As indicated by previous study, visual WM capacity is a limited resource that can be dynamic shifted between objects [40]. The precision of WM is determined by the proportion of resource allocation to each object. In the present study, to-be-maintained items would have higher precision of WM than to-be-inhibited items. Therefore, whether the cued items appeared would induce unbalanced performance in the memory test. Besides, we also found that the uncued items might be mistaken for the cued items when the cued items were absent, suggesting that both the cued and uncued items existed in WM. However, they had distinct influence on visual search, which might be caused by the different states of remembered items retaining in WM.” (Discussion. Line 443-456)

Reference:

3. Olivers CN, Peters J, Houtkamp R, Roelfsema PR (2011) Different states in visual working memory: when it guides attention and when it does not. Trends in Cognitive Sciences 15: 327-334.

5. Olivers CN, Meijer F, Theeuwes J (2006) Feature-based memory-driven attentional capture: visual working memory content affects visual attention. Journal of Experimental Psychology: Human Perception and Performance 32: 1243.

32. Kornblith S, Quiroga RQ, Koch C, Fried I, Mormann F (2017) Persistent single-neuron activity during working memory in the human medial temporal lobe. Current Biology 27: 1026-1032.

40. Bays PM, Husain M (2008) Dynamic shifts of limited working memory resources in human vision. Science 321: 851-854.

Minor Points:

5) In line 158, dose “hemisphere” means hemisphere of brain? Or does it means left side or right side of the screen?

Response: Thank you for reminding us about this important point. Participants were introduced to discriminate the locations of targets. If the targets were presented on the left or right side of the screen, they pressed the “F” or “J” key. Below is the revised content:

“Participants discriminated the location of targets where the tilted line appeared by pressing keyboard (“F” for left side of the screen, “J” for right side of the screen).” (Apparatus, stimuli, and procedure. Line 162-164)

6) Behavioral results in ms only report interaction now. I would suggest the author to report intact ANOVA results with main effects as well as interaction.

Response: Thanks for your suggestion. In the revised manuscript, we reported the main effect and interaction in the behavioral results.

According to your advice, we reported the main effect in the analyses of the accuracy as below:

“Also, the main effect of Presence of cued was significant (F(1,21) = 24.71, p < 0.001, ηp2 = 0.54), with higher accuracy in the presence (91.5%) than in the absence of the cued color (77.4%). And the main effect of Presence of uncued were significant (F(1,21) = 29.64, p < 0.001, ηp2 = 0.58), with lower accuracy in the presence (82.0%) than in the absence (86.9%) of the uncued color.” (Results. Behavioral results. Memory test. Line 270-275)

Besides, we also reported the main effect in the analyses of the response time as below:

“The main effect of Presence of cued was significant (F(1,21) = 10.57, p = 0.004, ηp2 = 0.34), with faster response in the presence (807 ms) than in the absence (876 ms) of the cued color. The main effect of Presence of uncued was not significant (F(1,21) = 0.05, p = 0.831).”(Results. Behavioral results. Memory test. Line 284-287)

7) It will be a better exhibition of data in Figure 4 if add a brain electrical activity mapping like in Figure 3.

Response: Thanks for your advice. According to your suggestion, we added the Topography of the N2pc component in Fig.4 as below:

Fig. 4. The N2pc waveforms (a, b, c) for contralateral and ipsilateral electrical activity and mean difference wave (d) are presented. Topography represents the N2pc component (170-250 ms). The left and right sides of the outline head represent ipsilateral and contralateral electrode sites of the target, respectively. *** p < .001. (Line 332-335)

Response to Comments from Reviewer 2

Reviewer #2: The present study presents an intriguing finding. I don't find any serious flaw in experimental design and data analyses.

However, I have several concerns regarding the interpretation of results and discussion.

First, not only the uncued item, but the cued item should also play a role in formulating the template for rejection. In the experiment, working memory items, maintained or inhibited, were always presented contralaterally to the target. Hence, any memory-matching stimulus is a distractor. In this case, the cued item should also be regarded as a template for rejection in visual search. Certainly, the cued item should be maintained in working memory, evoking a conflict between search demand and working memory task. This conflict might have hindered the effect of rejection template. For the uncued item, due the inhibitory tagging imposed on the stimulus, the rejection template could be established more effectively.

Response: Thank you for the insightful comment. It is possible that the memory-matched colors (cued and uncued colors) were regarded as a template for rejection in the search task. The search demand and the requirement of working memory might engender the same response in the cued condition; but they might engender the different response in the uncued condition, which would evoke conflict (Kerns et al., 2004; Macleod, 1991; Weissman et al., 2017). Thus, the time of visual search significantly increased in the cued condition but significantly speeded up in the uncued condition, compared to the neutral condition.

According to your advice, we added the content to the Discussion.

“In accordance with previous studies [1,4,5,8], our behavioral results showed that the to-be-maintained color in the cued condition impaired the visual search. In the present study, the memory-matched color was a distractor that was always presented contralateral to the search target. The to-be-maintained items might be viewed as a template for rejection in the search task [10], while these items should be maintained in the working memory task. In this case, the requirement of working memory and search demand engender different response, evoking a conflict [34-36]. Therefore, this conflict hindered the effect of rejection template, leading to slower response time in the cued condition than in the neutral condition.” (Discussion. Line 408-416)

Reference

10. Reeder RR, Olivers CN, Pollmann S (2017) Cortical evidence for negative search templates. Visual Cognition 25: 278-290.

34. Macleod CM (1991) Half a Century of Research on the Stroop Effect: An Integrative Review. Psychological Bulletin 109: 163-203.

35. Weissman DH, Colter KM, Grant LD, Bissett PG (2017) Identifying Stimuli That Cue Multiple Responses Triggers the Congruency Sequence Effect Independent of Response Conflict. Journal of Experimental Psychology-Human Perception and Performance 43: 677-689.

36. Kerns JG, Cohen JD, MacDonald AW, Cho RY, Stenger VA, et al. (2004) Anterior Cingulate conflict monitoring and adjustments in control. Science 303: 1023-1026.

Second, it is nice to see that the amplitudes of the Pd and N2pc well match behavioral results. Contrary to this, I don't see any merit in the MVPA analyses. What does it add to the behavioral results and univariate ERP results? The cued item should be maintained in working memory and the uncured item is not in working memory. It is not surprising that this difference would yield huge difference in behavior and neural responses.

Response: Thank you for comments on this important issue. In the current study, we want to explore whether the state of memorized items contribute to attentional guidance when the contents remembered in WM are presented as distractor in the search task. To this end, we conducted two different methods of data analyses. On the one hand, behavioral and univariate results could describe the difference between cued and uncued items in the level of single component. On the other hand, we combined all ERP components to investigate their difference from the multivariate level by using the MVPA analyses. MVPA analysis may combine all the characteristics of the data to help us understand the differences from a holistic perspective.

Importantly, the results of MVPA showed that the accuracy is maximum when the SVM classifier was trained on both ERP (P1 and N2pc). These findings suggested that the difference of neural responses between maintenance and inhibition dependent on the multivariate characteristics. Overall, these results might provide more evidence to understand the different state of memorized items retaining in working memory.

In order to clarify the role of MVPA analyses, we added some contents in the Introduction about this point.

“Given that multivariate pattern analyses (MVPA) [18] can combine all characteristics of the data to help us understand the difference between inhibition and maintenance from the multivariate level, we used MVPA to maximally discriminate two WM states in modulating attention.” (Introduction. Line 94-97)

As a minor issue, English in the manuscript is not too bad, but it has to be more scholarly. Some of expressions are too colloquial and not completed sentences.

Response: Thanks for your careful check. We feel sorry for the inappropriate description. We have improved the clarity and readability of the manuscript.

Submitted filename: R1_Response to Reviewers.doc

Click here for additional data file.

9 Nov 2020

The state of memory-matched distractor in working memory influence the visual attention

PONE-D-20-19909R1

Dear Dr. Chen,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Zaifeng Gao

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: I think the authors have adequately addressed the comments raised in a previous round of review. I don't find any serious flaw in the manuscript.

Reviewer #2: I think that the authors well addressed all of my issues. I think that the paper is now suitable for publication. I have no further comment.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Xiqian Lu

Reviewer #2: Yes: suk won han

17 Nov 2020

PONE-D-20-19909R1

The state of memory-matched distractor in working memory influence the visual attention

Dear Dr. Chen:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Professor Zaifeng Gao

Academic Editor

PLOS ONE