New Evidence for Retrospectively Cued Perception

Round 1

Reviewer 1 Report

This study by Szaszkó et al. aimed to validate and substantiate the proposition of a perception-memory continuum. Their findings demonstrated an increased reliance on post-target cues for target detection only when target visibility decreased. Notably, they observed the impact of post-target cueing on target perception even without a postcue response prompt, emphasizing a true continuity between perception and memory rather than a task-specific effect of memorization prompted by a response cue. Their results extend prior research conducted with liminal targets and substantiate the perspective of a perception-memory continuum. The manuscript exhibits clear and effective writing. While the work presented holds the potential to make a valuable contribution to the existing data on this topic, I have a few points that require clarification from the authors. I acknowledge that some of my concerns may stem from my confusion, but I trust that addressing the following points could enhance the overall quality of the manuscript.

Following the postcue, participants were given 1 second to specify the orientation of the Gabor patch. It is important to clarify whether, during this duration after the postcue, the screen displays a blank screen or the fixation screen. Additionally, it needs clarification whether the 1-second timeframe is measured from the onset of the target. Ensuring clarity on these points is essential.

For both cued and uncued trials, is the target consistently presented in the cue display? This implies that for uncued trials, the cue display mirrors the target display exactly, potentially introducing a perception enhancement effect. In my interpretation, if the target always appears on the cue display, the dimming of the placeholder circles may not cause significant interference as it might not capture much attention. This interpretation aligns with the results showing no validity effects for the lowest eccentricities condition concerning accuracy and reaction time. Additionally, the absence of differences between invalid and uncued trials in accuracy suggests that the dimming did not have a notable impact. While the study found that invalid cues led to significantly slower responses compared to uncued trials, there remains the possibility that this is attributed to the perceptual enhancement observed in uncued trials where the displays exactly matched the target display. The authors also discussed the absence of a cueing benefit under valid conditions, suggesting ceiling performance as a potential explanation. It would be beneficial for the authors to address whether they can exclude the possibility that the cue does not work at all.

Building upon the previous inquiry, in Experiment 2, the validity effect emerged with the dimming of the placeholder circle after reducing target visibility. Is this cue's effectiveness attributed to contrast compatibility? Would employing a more salient cue, rather than a dimmed one, result in a more pronounced postcue effect?

Author Response

Reviewer: This study by Szaszkó et al. aimed to validate and substantiate the proposition of a perception-memory continuum. Their findings demonstrated an increased reliance on post-target cues for target detection only when target visibility decreased. Notably, they observed the impact of post-target cueing on target perception even without a postcue response prompt, emphasizing a true continuity between perception and memory rather than a task-specific effect of memorization prompted by a response cue. Their results extend prior research conducted with liminal targets and substantiate the perspective of a perception-memory continuum. The manuscript exhibits clear and effective writing. While the work presented holds the potential to make a valuable contribution to the existing data on this topic, I have a few points that require clarification from the authors. I acknowledge that some of my concerns may stem from my confusion, but I trust that addressing the following points could enhance the overall quality of the manuscript.

Following the postcue, participants were given 1 second to specify the orientation of the Gabor patch. It is important to clarify whether, during this duration after the postcue, the screen displays a blank screen or the fixation screen. Additionally, it needs clarification whether the 1-second timeframe is measured from the onset of the target. Ensuring clarity on these points is essential.

BS: We would like to thank the reviewer for their valuable comment. We inserted according information in line 250 about the starting point of the measurement. We also inserted the following passage in Line 251 for clarification: “during that time and after target offset, the fixation display was shown again (note that the fixation cross and the placeholder circles were always present during the entire experiment, except during breaks)”

Reviewer: For both cued and uncued trials, is the target consistently presented in the cue display? This implies that for uncued trials, the cue display mirrors the target display exactly, potentially introducing a perception enhancement effect. In my interpretation, if the target always appears on the cue display, the dimming of the placeholder circles may not cause significant interference as it might not capture much attention. This interpretation aligns with the results showing no validity effects for the lowest eccentricities condition concerning accuracy and reaction time. Additionally, the absence of differences between invalid and uncued trials in accuracy suggests that the dimming did not have a notable impact. While the study found that invalid cues led to significantly slower responses compared to uncued trials, there remains the possibility that this is attributed to the perceptual enhancement observed in uncued trials where the displays exactly matched the target display.

BS: Thanks to the reviewer for pointing out our mistake. The target was never present in the cueing display – this was simply a mistake in the figure and we apologize for the oversight; we have updated Figure 1. We suppose that by doing so, the problem is solved because uncued trials would not lead to a perceptual enhancement without a target present.

Reviewer: The authors also discussed the absence of a cueing benefit under valid conditions, suggesting ceiling performance as a potential explanation. It would be beneficial for the authors to address whether they can exclude the possibility that the cue does not work at all.

BS: We would argue that observed selective interference by invalid cues (compared to neutral cues) cannot be explained in any other way than by (temporary) distraction of attention away from the target through the spatial difference between cue and target position. Response interference as an alternative explanation of this cue-target position incongruity effect can be ruled out because participants pressed buttons to discriminate target orientations and because target orientations and target positions varied orthogonally. What else but spatial attention being distracted away from the target for a time could account for this cue-target position incongruity effect? This is explained in lines 347-359.

Reviewer: Building upon the previous inquiry, in Experiment 2, the validity effect emerged with the dimming of the placeholder circle after reducing target visibility. Is this cue's effectiveness attributed to contrast compatibility? Would employing a more salient cue, rather than a dimmed one, result in a more pronounced postcue effect?

BS: This is a very interesting comment. If we understand the reviewer correctly, he/she believes that participants might be more sensitive to post-target cues that are relatively (“contrast-“)similar to the current target in Experiment 2 than to the less target-similar cues in Experiment 1? This is theoretically possible because (1) cue-background and cue-target contrasts were both more similar in Experiment 2 than in Experiment 1 and (2) commencing target processing in a contrast range more similar to that of the dimmed cue in Experiment 2 could have primed the perception of the cues of a like contrast more than it was the case for the high-contrast targets (that were more contrast-different from the same dimmed cues) in Experiment 1. In turn, a stronger target-cue priming effect would have facilitated attentional post-cueing effects in Experiment 2 compared to Experiment 1. Luckily, we have some data available to answer the open question of the reviewer. In a pilot experiment, we used more salient cues (white crossfades of the same contrast as the target, presented around the placeholder circles, but otherwise in a similar experimental protocol). These high-contrast cues were more similar in terms of the contrast to the high-contrast targets used in Experiment 1. In contrast to the theoretical prediction based on contrast-priming of post-target cues in Experiment 2 above, in the pilot experiment, we observed inverse validity effects for short target-cue stimulus-onset asynchronies, with faster responses and higher accuracy in invalid than valid trials. We believe that this was the case because the high-contrast cues in the pilot experiment were backward-masking the target to some extent, worsening performance in valid compared to invalid conditions. For long SOAs, where masking should have been less relevant, we could not find a validity effect in neither accuracy rates, nor response times. (By the way, in the pilot experiment, just as in Experiment 1 of the manuscript, again, we found interference by invalid cues compared to uncued trials [5 ms], indicated by a significant main effect of validity, F(1,25) = 23.00, p < .001, partial eta2 =.48, BFincl > 100.)

For the sake of brevity and because results in the pilot experiment with the salient cues were otherwise almost identical to the results reported in Experiment 2 of the manuscript (see interference in invalid compared to neutral conditions), we now have not included the pilot experiment in the revision and only reported the results of this pilot experiment in the response letter. However, we are happy to include the pilot experiment in the manuscript if the reviewer thinks this is necessary.

To conclude, we continue to believe that contrast gain is responsible for the results in Experiment 2: Attention and, hence, a post-target cueing effect is more decisive for the visibility of targets of a lower than of a higher target-noise contrast. This also makes sense from an economic point of view, because attention, as an energy-consuming process, fosters target processing where this is really needed and not according to a scattergun approach.

Reviewer 2 Report

The manuscript delves into the connection between perception and memory, particularly in how cues presented after the conclusion of a visual stimulus (post-stimulus) can influence the perception of that stimulus. The research conducted two experiments to test this relationship and confirmed that there's a significant influence of post-stimulus cues on the detection of the visual stimulus, especially when the visibility of the stimulus is reduced. This suggests a higher dependence on post-stimulus cues to detect the visual stimulus, indicating a contrast gain.

The article is well structured, provides an adequate review of the literature, describes the methodological process and development of the experimental study, analyses and discusses the results, and presents the conclusions.

However, I recommend that the authors improve the conclusion, indicating the contributions (academic, scientific and medical) of the study, the limitations and directions for future research.

Congratulations and good luck.

Author Response

Reviewer: The manuscript delves into the connection between perception and memory, particularly in how cues presented after the conclusion of a visual stimulus (post-stimulus) can influence the perception of that stimulus. The research conducted two experiments to test this relationship and confirmed that there's a significant influence of post-stimulus cues on the detection of the visual stimulus, especially when the visibility of the stimulus is reduced. This suggests a higher dependence on post-stimulus cues to detect the visual stimulus, indicating a contrast gain.

The article is well structured, provides an adequate review of the literature, describes the methodological process and development of the experimental study, analyses and discusses the results, and presents the conclusions.

However, I recommend that the authors improve the conclusion, indicating the contributions (academic, scientific and medical) of the study, the limitations and directions for future research.

BS: We appreciate the reviewer's insightful feedback and are pleased that our manuscript aligns well with their perspective. In response, we have revised the conclusions to address the differences in findings between Sergent et al.’s study and ours, emphasizing the unique contributions and limitations of our work. Additionally, we have expanded the discussion on future research directions, specifically highlighting the potential for extending our findings to other sensory modalities and exploring cross-modal integration in the context of post-cueing.

Reviewer 3 Report

The current paper demonstrates through two experiments the effects of post-cues on response time and response accuracy for targets at varying levels of visibility, extending previous research working strictly with liminal targets. Results indicate that there are influences of post-cues, but that they are not always as simple as increased accuracy and decreased response time for validly cued targets, depending on the condition. The studies appear well designed and results are logically interpreted. Please see below for some suggestions and needed changes for the next draft of the manuscript:

 1. Abstract: Perhaps some brief mention that results in these experiments were not entirely the same as previous studies would be beneficial.

2. Page 5-6, Lines 231-249: A description of when targets appeared and for how long is missing from the description of the trial sequence here. While this is illustrated/described in the figure, it should be clarified in the main text as well.

3. Page 10, Figure 4: Target display should be 100 ms.

4. Page 11, Figure 5: The figure is unclear/does not appear to match the experiment/figure caption. Namely, what is meant to be displayed on the x-axis? I would assume that there would need to be five points on the x-axis for the noise orientation, and possibly two graphs side by side for the two SOAs.

Page 7, Line 301, Page 9, Line 377, Page 11, Line 462: The word “Data” is plural, “Datum” being the singular. Thus the phrase “Data were…” should be used, as opposed to “Data was…”

Author Response

Reviewer: The current paper demonstrates through two experiments the effects of post-cues on response time and response accuracy for targets at varying levels of visibility, extending previous research working strictly with liminal targets. Results indicate that there are influences of post-cues, but that they are not always as simple as increased accuracy and decreased response time for validly cued targets, depending on the condition. The studies appear well designed and results are logically interpreted. Please see below for some suggestions and needed changes for the next draft of the manuscript:

1. Abstract: Perhaps some brief mention that results in these experiments were not entirely the same as previous studies would be beneficial.

BS: We would like to thank the reviewer fot pointing out this important source of possible confusion. We changed the abstract accordingly to show the main discrepancy, namely the demonstration of the postcueing effect through RTs in invalid vs. uncued trials in Experiment 1, compared to the performance difference between valid and invalid conditions in past studies like the one of Sergent et al. (2013).

2. Page 5-6, Lines 231-249: A description of when targets appeared and for how long is missing from the description of the trial sequence here. While this is illustrated/described in the figure, it should be clarified in the main text as well.

We apologize for the oversight and have now included this description in the manuscript.

3. Page 10, Figure 4: Target display should be 100 ms.

Thanks to the reviewer for pointing this out to us. We have changed this now accordingly.

4. Page 11, Figure 5: The figure is unclear/does not appear to match the experiment/figure caption. Namely, what is meant to be displayed on the x-axis? I would assume that there would need to be five points on the x-axis for the noise orientation, and possibly two graphs side by side for the two SOAs.

This is a valid remark. We have corrected the glitch in the figure.

Comments on the Quality of English Language

Page 7, Line 301, Page 9, Line 377, Page 11, Line 462: The word “Data” is plural, “Datum” being the singular. Thus the phrase “Data were…” should be used, as opposed to “Data was…”

Thanks to the reviewer. We have changed these phrasings.

Reviewer 4 Report

see attached file

Comments for author File: Comments.pdf

Author Response

Reviewer: This study should be of interest to readers of Vision. It addresses a fundamental problem in the transition from perception to memory. The author adopted a post-cue procedure due to Sergent et al. (2013) but with above threshold targets, to test a more general percep?on-memory continuum hypothesis. All very good. However their cuing procedure is confusing, and I cannot evaluate the work without understanding it better.

Line 50: Sergent et al. did not use a ‘visual search task’. This brings to mind variation in the number of distractors, with the case of one distractor being of no special consequence, with target spatial location randomized. Instead, Sergent’s task was to identify the orientation of a Gabor patch that could be in one of two known locations, as described correctly in line 51.

Thanks to the reviewer for pointing this out. As it might help to prevent the wrong associations, we no longer call the task a “visual search task”. The decisive point that we wanted to emphasize is that spatial uncertainty about target positions is of essence in the investigation of a cue’s spatial attention effects on target processing. However, we agree with the reviewer that it is not necessary to use the label “visual search” and that the label could be dropped to prevent misleading associations.

Reviewer: Line 52: Sergent’s post-cue was a brief dimming of a ring surrounding the target or non-target location. This should be stated here, for readers unfamiliar with that paper. Sargent’s cue was typical of most research papers on attention in which any ‘cue’ is clearly distinct from the target and ideally does not affect it visually. (Sergent et al.’s ring was only 0.4 deg larger than the target and, as they do state, at 4 deg eccentricity it could have interfered with their target.)

We are grateful for this important comment that will increase the comprehensibility of the manuscript. We changed the information accordingly, pointing it out (now) in line 57.

Reviewer: Line 70: “where it is impossible to derive any firm conclusions about this impact as the authors did not detail the exact luminance and contrast values of their Gabor targets at threshold [5]. “

There is no need to specify the exact luminance or contrast, because the decrease of RT as contrast increases above the incremental threshold has been known for a century or more, and occurs at all photopic luminance levels. I suggest deleting this remark.

Thanks the reviewer for the suggestions. We have deleted this information in the manuscript.

Reviewer: Line 95. Xia et al. claimed the visual interactions affected the degree of cuing, even if the cue occurred 200 ms after the target. This may be, as the gap from the target to the ring cue was only 0.4 deg wide (as also in Sergent’s study). However metacontrast is optimal at an ISI of 80 msec, whereas the benefit of the cue in Xia et al. diminished smoothly until an ISI of 200 msec. An ‘attentional’ explanation seems more likely to me, as accuracy with the cue was overall better than with no cue. Knowing the target timing and location in advance (‘endogenous attention’), which was Xia’s control for attention, may not be as useful as also having a physical cue to help one (‘exogenous attention’), especially if the cue occurs close in time to the target (i.e., within 160ms -ring- Expt. 1, Xia’s fig. 2, left panel; or <100 ms -dots - Expt 2).

This is a great remark. We agree that we oversimplified matters. We have corrected the corresponding passage to: “Another non-attentional explanation was implied by the study of Xia et al. (2016) [33]. These authors found that cues, as were used by Sergent et al. [5], effectively improved target perception even under conditions of certainty about an unchanging target position, meaning that visual line-orientation interactions between target and postcues at target position alone could account for the postcueing effect and that spatial orienting of attention is not decisive. In contrast to a masking account, however, Xia et al. [33] argued for an enhancing effect of postcues on target-orientation perception.”

Reviewer: Line 91. Dubious. Sergent dimmed the valid cue for only 50 ms. Any visual spatial interactions between the orientations in the ring and those in the target would have not only been weak, given that a ring does not alter orientation thresholds for straight lines, but also extend over the full duration of presentation, and so be the same for both valid and invalid cues.

We revised the paragraph in question, now between lines 99 and 106, as follows: "

Another non-attentional explanation was implied by the study of Xia et al. (2016) [33]. These authors found that cues, as were used by Sergent et al. [5], effectively improved target perception even under conditions of certainty about an unchanging target position, meaning that visual line-orientation interactions between target and postcues at target position alone could account for the postcueing effect and that spatial orienting of attention is not decisive. In contrast to a masking account, however, Xia et al. [33] argued for an enhancing effect of postcues on target-orientation perception. "

Reviewer: Fig. 1 was almost entirely black in my copy of the paper, although I was able to see it in power point. I suggest re-drawing it to make the stimuli more visible.

Thanks to the reviewer for mentioning this. We improved the figure by thickening of the rings and the fixation dot. We also inserted a corresponding sentence into to the Figure caption: “For improved legibility, stimuli are not drawn to scale.”

Reviewer: Fig. 1 shows that the post-cue included Gabors, which does not agree with the text, in which the post- cues are described as briefly dimming the ring.

Perhaps the authors showed the target Gabor twice, with a ‘blank display’ between the two presentations? The pot-cue was then solely a dimmed circle, either around the target (‘valid’) or opposite to it (‘invalid’), which occurred during the second presentation. This makes sense of Fig. 1, but it does not agree with the text.

Or perhaps the second presentation was a different (non-target) Gabor, and the subject had to report on only the first Gabor, the target. This would agree with the text, in which it is stated that the target was presented for only 50 ms. Fig. 1 is too poorly drawn for me to tell by looking, and the text doesn’t appear to say.

We apologize for this mistake, as correctly pointed out by the reviewer. The target was never present in the cueing display – this was simply a mistake in the figure. We have updated Figure 1.

Reviewer: To help the reader, the authors should define the ‘post cue’. Is this just the dimmed circle, or is it an array including both a dimmed circle and a repetition of the target Gabor, or is it an array including both a dimmed circle and a non-target Gabor ? ? I have gone round and round trying to figure this out, and I have not succeeded.

This should be clear after the correction of the figure and our remarks above.

Reviewer: Line 104. I do not follow this comment about the response prompt. Sergent showed a foveal prompt pointing towards the side to be reported on, sometime between 500 and 900 ms after the target. The Gabor stimulus itself was 4 deg peripheral. How could the prompt, so far from the Gabor spatially and temporally, and so distinct from it in terms of shape, ‘corrupt’ a perceptual judgement about the Gabor ? Maybe explicate and add citations, if this is indeed a real possibility.

This is a misunderstanding. Like the reviewer, we do NOT believe that the prompt corrupted target visibility. What we wanted to say is that the prompt corrupted the rationale of the study if the aim is to show that post-cueing works because perception continuously blends into memory under regular, every-day conditions. The reason is that with the prompt, participants are forced to wait with their judgments until the prompt, meaning the perceptual task is artificially transformed into a memory task. The prompt, thus, corrupts the internal validity of the experiment, NOT the target perception. We have rephrased the sentence in question to: “The problem with this procedure is that it is then unclear if the usage of the response prompt corrupted an otherwise perceptual task and changed it into a memory task merely by means of this characteristic of the experimental protocol.”

Reviewer: Line 314-onwards. If the post-cue included both a repeat of the Gabor and a dimming circle, it seems obvious that the subject would pay attention only to the Gabor, to extract its orientation, and ignore the dimming circle entirely. If the post-cue is just the dimming circle, but the second stimulus was a non- target Gabor (vertical, perhaps ?), then the entire experiment shows that cuing a mask (the non- target Gabor) does not change its effectiveness as a mask. Again, as I do not understand Fig. 1, I cannot tell.

See our remarks above regarding Figure 1, as well as the corrected Figure itself.

Reviewer: Note: it could be useful to determine how well the subjects can identify the orientation of the first (50 ms) Gabor. Perhaps they would achieve the same accuracy as in Fig. 2 at each eccentricity, in which case, the experiment would simply indicate ‘parallel processing’ across both hemispheres, with no need for any cuing. A no cue control is normally provided to show that the cue is helping (or, if it is masking, hindering).

For an earlier study of perceptual integra?on with spa?al memory (defined as ‘imagery’), see

Brockmole et al, Journal of Experimental Psychology:
Human Perception and Performance 2002, Vol. 28, No. 2, 315–334 Whose integration effect lasted 800 msec or so.

We hope that this point was clarified as well through the information that there was only one Gabor present in each trial. If not and if we misunderstand this remark, the reviewer should please let us know.

We would especially like to thank the reviewer for his detailed review. His keen eye for detail and his in-depth, logical understanding and expertise have greatly helped us to improve this manuscript.

Round 2

Reviewer 1 Report

The authors addressed my concerns adequately. 

Author Response

We would like to thank the reviewer for their valuable work. Current changes in the manuscript are due to incorporating the feedback of other reviewers.

Reviewer 4 Report

see attach file

Comments for author File: Comments.pdf

Author Response

Please see the PDF attachment for our comments.

Author Response File: Author Response.pdf

Round 3

Reviewer 4 Report

I remain unconvinced by the 5ms effect, but the authors have answered all my criticisms, including showing that it applies also to the medians, and it is their paper, not mine. However I do not want to lend my name to the publication, as readers may assume that I agree with the conclusion. It may be true, but I am not sure.

 I  STRONGLY suggest explaining how the CI's were computed, when they are introduced. The fact that they are so wide, compared to the MSerror, needs to be accounted for. It should not be taken for granted that readers can figure out what is going on here. The explanation in the authors' letter is perfectly adequate, ad should appear in the paper (if it is in there, and I missed it, my apologies).

Author Response

We would like to thank the reviewer for improving the legibility of our paper. We have inserted a remark to the corresponding figures, stating that CIs were calculated based on condition means per participant, and corrected for within-subject designs.