I would like to thank the authors for their response letter. They have further improved their report significantly and addressed all my major concerns, with only two probably minor points remaining: 

Generality of improvements:
I appreciate that the authors are going to implement the reviewers’ suggestions to use Bayesian statistics to quantify the evidence for null effects. However, why did they only specifically include the test in analysis plan A2.1? In the analysis plans A1, A2.2, A3 and A4, the outcome interpretations O1.2, O2.2.2, O3.1.2, O3.2.2, O4.1.2 and O4.2.2 also concern null effects, so they would also require Bayesian tests.
Similarly, the new precise definition of the IOI for the MA analysis based on 1.5 SD around the mean RT is good. However, why is it specifically stated in A3.1 and A4.1 but not in A3.2 and A4.2?
I think it should be clarified that these tests/criteria will be applied in all cases.

Effect sizes: 
Now that the authors have computed effect sizes using EROS data from Opt-3d, why does the study design table still say that “Effect size for EROS data is not computable”?

I wish the authors all the best for this interesting study and look forward to their results.

I would like to thank the authors again for thoroughly addressing my concerns. I have no further comments and think the planned study will make a valuable contribution to the field.

I would like to thank the authors for their detailed and informative response letter. They have carefully addressed most of the reviewers’ concerns and further improved the report significantly. On the other hand, some issues remain unresolved even though they could be addressed. 

Major points

Interval of interest: 
The IOI for the LOC analysis is now clearly defined. However, how exactly is the IOI for the MA analysis “based on mean RTs”? Is it based on each participant’s mean RT, or on the grand average? Moreover, how is the interval defined? For example, does it include a specific interval around, before or after the peak?

Negative results:
Concerning evidence of absence and Bayesian statistics, the authors responded that “EROS statistical analyses are “limited” to the use of […] Opt3d, which does not allow computing Bayesian statistics.” However, the Opt3d manual (https://www.nitrc.org/docman/view.php/868/2031/opt3d%20Manual) says on p. 1 and 27 that the data can be exported into standard text files (File > Save), including the option “ROI Analysis: Provides data from the average across all voxels defined in the ROI”. In the case of time courses, one would only need to average the samples across time. Bayes factors could easily be computed using, for example, the GUIs of jamovi (which the authors have used for their behavioral data) or JASP (https://jasp-stats.org/) and default priors without previous data. For details and tutorials, see Keysers et al. (2020, Nat Neurosci). Since the PCI RR guide for authors (https://rr.peercommunityin.org/PCIRegisteredReports/help/guide_for_authors) specifically encourages Bayesian tests for evidence of absence, they might want to consider this option.

Sample size estimation:
Concerning the computation of effect sizes, the authors responded that “exporting EROS data would mean extracting a very huge amount of data (i.e., a matrix of data for each voxel and for each time point) which would be too difficult to be managed without a dedicated software. For this reason, EROS analyses are constrained to the use of Opt3d software.” However, as mentioned above, Opt3d allows the export of time courses averaged within the voxels of a ROI. Thus, the authors would only get one time course per condition and ROI. Then, they could easily average the samples and compute the effect sizes in, e.g., jamovi or JASP. 

ROIS:
I agree that basing ROIs on the previous study (Colombari et al., 2024) is perfectly fine. However, is circularity / double-dipping (Kriegeskorte et al., 2009) also avoided when investigating the “novel ROIs over areas responsible for visual processing and motor execution” selected based on “visual inspection of functional data” (l. 368)? That is, are these novel ROIs selected a priori or in a data-driven (potentially non-orthogonal) manner?

Minor points

The authors adequately introduced previous no-report NCC studies using EEG and fMRI (l. 54-75), but only mentioned results focusing on the temporal domain (i.e., VAN and LP in EEG studies) while neglecting spatial results (e.g., LOC effects in fMRI studies). In their introduction of spatial aspects (l. 76-107), the authors highlight that EROS can provide “accurate information […] both from the temporal and spatial point of view” (l. 106), but later mention the „relatively low signal-to-noise ratio of EROS” (l. 182). Thus, it would be worth mentioning that previous no-report studies using both EEG and fMRI (e.g., Dellert et al., 2021; Kronemer et al., 2022) have also found awareness effects in the lateral occipital complex (LOC) and linked it to the VAN, which is the focus of the present study. 

L. 108: In line with the abstract, „peculiar“ should also be changed to “distinctive” here.

L. 188: I appreciate the clear new exclusion criteria. On a minor note, “will not also be included in the analyses” should be fixed.

L. 211: “the left portion of the primary visual cortex, which is known to be anatomically closer to the skull compared to the right one”. Could the authors provide references for this?

L. 340: The authors have clarified in their response that the baseline correction will not be conducted using the whole 486 ms prestimulus interval of the segment (which the current report would suggest) but only 204 ms. This is good due to the reasons stated in my previous review, but the actual 204 ms baseline correction interval should also be stated in l. 340.

The authors clearly improved the manuscript. All my comments have been addressed, and passages of the manuscript that were lacking certain details have been supplemented with information.

I believe that if the study is carried out as planned, it will be a valuable contribution to the field.

I would like to thank the authors for thouroughly addressing my concerns. However, I have two important points where I think the registered report could still improve:

(1)    Use of Bayesian statistics. I understand that the EROS data might not be easily tested using Bayesian statistics given constraints of the Opt3D software. However, as written in the Opt3D Manual (https://www.nitrc.org/docman/view.php/868/2031/opt3d) data can be exported. Given this option and the fact that the behavioral data is independent from Opt3d constrains I strongly recommend including Bayesian statistics in the analysis plan (for the behavioral data as well as for the EROS data)  for testing null hypothesis. If no prior knowledge exists default priors can be used. One option would be e.g. JASP (https://jasp-stats.org/) as it is similar to Jamovi or any other statistics program like R.

(2)    Interpretation of non-significant t-tests concerning response times: I still think the description of the behavioral pilot data is misleading. The results (e.g. “Paired sample (two-tailed) t-test performed with Jamovi (version 2.3.28) highlighted that there was no significant difference between the two conditions (t(4) = 1.88, p = .134, Cohen’s d =  .839), suggesting that they are comparable.”) do not indicate that the values are comparable. While it is true that there is no significant difference with N = 5, only Bayesian statistics could offer the conclusion of evidence for the absence of an effect. One option would be to delete the ‘suggesting that they are comparable’ and ‘This indicated that there was no difference in the responsiveness between the two conditions.’ or to include Bayesian evidence (which given the t-values would likely be inconclusive, thus leading to a more fitting bottom line).

Summary

In the present registered report, Colombari and colleagues describe a study planned to investigate neural correlates of conscious visual perception and to isolate them from neural activity related to motor responses. To this goal, they plan to use a GO/NOGO detection task and the event-related optical signal (EROS) technique.

General impression

In my opinion, the registered report is very well written and satisfies most of the PCI RR Stage 1 criteria. First, I will shortly address each criterion, followed by a more detailed description of major and minor issues. 1A: The research questions are generally valid and convincingly derived from the literature. However, I would suggest incorporating some directly relevant previous studies. 1B: The proposed hypotheses are logical, coherent and plausible. On the other hand, the regions and especially intervals of interest are not stated clearly enough. Moreover, absence of evidence should not be interpreted as evidence of absence, and interaction effects should be tested directly. 1C: The methodology and analysis pipeline seem generally sound and feasible, and the experimental design is elegant. Concerning the sampling plan and statistical power analysis, some aspects should be clarified. 1D: The methodology is mostly described very clearly and in sufficient detail for close replication, with the exception of specific critical tests, exclusion criteria and the temporal aspects of the EROS data analysis. 1E: In my opinion, the authors have considered sufficient outcome-neutral conditions and quality checks. A more objective exclusion criterion for the EROS data quality would reduce flexibility. In the following, the abovementioned issues are described in more detail.

Major points

Introduction

Previous research:
The introduction (l. 48- 71) would profit from incorporating directly relevant previous studies which also specifically addressed the role of early posterior versus late centroparietal activity (i.e., VAN and LP) in conscious visual perception versus post-perceptual processes (e.g., Pitts et al., 2014; Shafto & Pitts, 2015; Dellert et al., 2021, 2022; Kronemer et al., 2022). In line with the authors’ goals (l. 75-79), some of these studies also reported brain activity with both high spatial and temporal resolution by means of EEG and fMRI (e.g., Dellert et al., 2021; Kronemer et al., 2022). Moreover, there are additional relevant studies that specifically investigated effects of attention and response requirements on the VAN (e.g., l. 55-67 and l. 111-114), both with positive (e.g., Bola & Doradzińska, 2021; Dellert et al., 2021; Doradzińska & Bola, 2024) and negative (e.g., Koivisto et al., 2006; Dellert et al., 2022; Ciupińska et al., 2024) results.

Methods

Intervals of interest:
The temporal aspect of the EROS data analysis should be described more clearly, e.g., which samples exactly are compared in the critical tests (l. 307-334). In all hypotheses and analysis plans (table in 3. Study design), the term “early temporal/time window” is too vague. All hypotheses should define specific intervals of interest or describe exactly how the intervals will be selected based on the data (e.g., in the case of collapsed localizers). Moreover, how will the statistical analyses control for multiple comparisons in the temporal domain? How were the intervals in the preliminary results (l. 405-414) selected?

Negative results:
All interpretations of negative results (column “Interpretation given different outcomes” of the study design table) are based on the absence of significant effects in frequentist null hypothesis testing. Especially given the limited statistical power of the planned sample size (n = 24), the absence of effects should be substantiated by Bayesian hypothesis testing. This also concerns the preliminary results (l. 378-382), where effects with d = .839 or -.791 are interpreted as “no difference” with n = 5.

Interaction effects:
In hypothesis 2, the authors “expect to find the same activation” in LOC in GO and NOGO trials. How will this equality be tested? Do the authors refer to a significant awareness effect in each condition (GO, NOGO) or actually the absence of a difference between the two effects, i.e., of an interaction effect (i.e., (Aware-NOGO – Unaware-NOGO) – Aware-GO – Unaware-GO))? In this context, it should also be considered that the presence of an effect (e.g., Aware-Unaware) in one condition (e.g., GO) and its absence in another (e.g., NOGO) should not be interpreted as a difference between the conditions – instead, a direct interaction test should be performed (see Nieuwenhuis et al., 2011).

Results

Figure 4 (l. 415): What exactly do the statistical maps present? What are the underlying data, units and scales? 

Minor points

Introduction

L. 40: “However, which one [VAN or LP] represents the true signature of conscious vision is still under debate”. This creates a false dichotomy because it could also be both or neither of them.

Methods

Sample size estimation:
L. 134-136 and 3. Study design / sampling plan): Concerning the “technical constraints of the employed dedicated software” (Opt-3d) which does not allow to calculate effect sizes, could the authors export the data in order to calculate them (at least in their own upcoming data)?
L. 142/151/152/346: Since 12.944 is much closer to 13 than 12, I would suggest to revise these lines. 
L. 145/153 At this point, it is unclear what the research questions (Qx) are, so perhaps they could be paraphrased here.
L. 146: An alpha level of 2% seems unusual to me (see also “3. Study design”). Could the authors explain why they chose it?

Perceptual threshold assessment: 
Concerning the perceptual threshold assessment (l. 155-163, 181-189), it is unlikely that one of the nine stimuli is perceived in exactly 50% of the cases (“acknowledged/identified as perceived the 50% of the times”). Perhaps it would be more appropriate to frame the criterion for the final stimulus as, e.g., “the one perceived a minimum of 25%, a maximum of 75%, and closest to 50% of the times”? 

Exclusion criteria (l. 166): The rule that “participants whose EROS signals could not be detected properly during the experiment will not be included in the analyses” and will be replaced (l. 170) is rather vague and leaves some room for arbitrary exclusions. A more objective criterion for signal quality would be beneficial. 

Fatigue: The very high number of trials (3120 per participant) is commendable. However, the duration of 3 hours per session (l. 197), further prolonged by the threshold assessment (20 min) on the first day, seems very long. How can adequate task performance be ensured despite potential fatigue?

Stimulus position: Why will the stimulus always be presented “in the lower right quadrant of the screen” rather than in the center (l. 212)? Does it have to do with the left-lateralized ROI? If so, this could be explicated. 

Tests: Are the t-tests going to be one- or two-tailed (e.g., l. 287, l. 296)?

Epoch: Why does each epoch already begin at 486 ms (l. 306) specifically, and will the whole pre-stimulus segment be used for baseline correction (l. 302)? Since pre-stimulus brain activity can influence conscious perception, this baseline correction of pre-stimulus effects could potentially induce artificial post-stimulus effects. 

ROIs: “Critical ROIs will be selected on the basis of the results obtained in the above-mentioned experiment (Colombari et al., under review) and by visual inspection of functional data.” (l. 330). Does “functional data” refer to the new data to be collected here? If so, how is circularity / double-dipping (Kriegeskorte et al., 2009) avoided when basing ROIs on effects?

References

Bola, M., & Doradzińska, Ł. (2021). Perceptual awareness negativity—Does it reflect awareness or attention? Frontiers in Human Neuroscience, 15, 601. https://doi.org/10.3389/fnhum.2021.742513

Ciupińska, K., Orłowska, W., Zębrowski, A., Łępa, L., Koculak, M., Bola, M., & Wierzchoń, M. (2024). The influence of spatial and temporal attention on visual awareness—A behavioral and ERP study. Cerebral Cortex, 34(6), bhae241. https://doi.org/10.1093/cercor/bhae241

Dellert, T., Krebs, S., Bruchmann, M., Schindler, S., Peters, A., & Straube, T. (2022). Neural correlates of consciousness in an attentional blink paradigm with uncertain target relevance. NeuroImage, 264C, 119679. https://doi.org/10.1016/j.neuroimage.2022.119679

Dellert, T., Müller-Bardorff, M., Schlossmacher, I., Pitts, M., Hofmann, D., Bruchmann, M., & Straube, T. (2021). Dissociating the neural correlates of consciousness and task relevance in face perception using simultaneous EEG-fMRI. Journal of Neuroscience, 41(37), 7864–7875. https://doi.org/10.1523/JNEUROSCI.2799-20.2021

Doradzińska, Ł., & Bola, M. (2024). Early electrophysiological correlates of perceptual consciousness are affected by both exogenous and endogenous attention. Journal of Cognitive Neuroscience, 1–28. https://doi.org/10.1162/jocna02156

Koivisto, M., Revonsuo, A., & Lehtonen, M. (2006). Independence of visual awareness from the scope of attention: An electrophysiological study. Cerebral Cortex, 16(3), 415–424. https://doi.org/10.1093/cercor/bhi121

Kriegeskorte, N., Simmons, W. K., Bellgowan, P. S. F., & Baker, C. I. (2009). Circular analysis in systems neuroscience: The dangers of double dipping. Nature Neuroscience, 12(5), 535–540. https://doi.org/10.1038/nn.2303

Kronemer, S. I., Aksen, M., Ding, J. Z., Ryu, J. H., Xin, Q., Ding, Z., Prince, J. S., Kwon, H., Khalaf, A., Forman, S., Jin, D. S., Wang, K., Chen, K., Hu, C., Agarwal, A., Saberski, E., Wafa, S. M. A., Morgan, O. P., Wu, J., … Blumenfeld, H. (2022). Human visual consciousness involves large scale cortical and subcortical networks independent of task report and eye movement activity. Nature Communications, 13(1), Article 1. https://doi.org/10.1038/s41467-022-35117-4

Nieuwenhuis, S., Forstmann, B. U., & Wagenmakers, E.-J. (2011). Erroneous analyses of interactions in neuroscience: A problem of significance. Nature Neuroscience, 14(9), 1105–1107. https://doi.org/10.1038/nn.2886

Pitts, M. A., Padwal, J., Fennelly, D., Martínez, A., & Hillyard, S. A. (2014). Gamma band activity and the P3 reflect post-perceptual processes, not visual awareness. NeuroImage, 101(Supplement C), 337–350. https://doi.org/10.1016/j.neuroimage.2014.07.024

Shafto, J. P., & Pitts, M. A. (2015). Neural signatures of conscious face perception in an inattentional blindness paradigm. Journal of Neuroscience, 35(31), 10940–10948. https://doi.org/10.1523/jneurosci.0145-15.2015

The Stage 1 Registered Report The role of extra-striate areas in conscious motor behavior: a registered report with Fast-Optical Imaging by Elisabetta Colombari, Giorgia Parisi, Sonia Mele, Chiara Mazzi and Silvia Savazzi addresses the question of the spatiotemporal resolution of neural correlates of consciousness (NCC) and the distinction between true correlates of consciousness and motor or task-specific processes. The authors chose a contrastive design realized by a Go/NoGo-Task and plan to measure brain activity using the EROS technique. Neural data is planned to be analyzed using the Granger causality analysis.

The report is well-written and clear, and it has the potential to make an interesting and valuable contribution to the field. In the following, I will raise some issues and questions to be thought about before realizing the study:

Abstract

The authors should reconsider the term, as it does not seem appropriate. Perhaps they should use 'distinctive' or 'special' instead.

Introduction

      i.        In the introduction section, the authors briefly summarize the state of research on NCCs without confounding concurrent neural dynamics. An overview of the field of no-report paradigms and the spatiotemporal characteristics of conscious processing is given. I would suggest considering the inclusion of Hense A, Peters A, Bruchmann M, Dellert T, Straube T. Electrophysiological correlates of sustained conscious perception. Sci Rep. 2024;14(1):10593. in the list of referenced works on sustained conscious perception. I believe it would complement the existing literature the authors have discussed.

     ii.        The authors aim to separate true consciousness effects and task effects by their paradigm. The Go/NoGo-Paradigm includes task effects. The authors should further elaborate on this issue.

Methods

      i.        I wonder why the analysis of the Catch Trials was not included in the exclusion criteria. The authors should provide some insights into this decision.

     ii.        Is there any control of eye movement effect planned? This issue should be addressed, as the critical stimuli are presented offside the center of the screen.

    iii.        The authors chose nine different stimuli to be shown before the experiment to determine the stimulus of the main experiment. Please discuss this choice. Doesn't the restriction to nine stimuli lead to a high drop-out rate if none of the stimuli leads to the desired behavior? Wouldn't the choice of a continuous parameter be more favorable?

   iv.        The selection of the ROIs used for analysis based on a posteriori inspection of the data seems a bit arbitrary. The authors should explain in more detail what criteria will be used for ROI selection. Please explain why the full left motor cortex and posterior visual areas are not selected as ROIs. 

     v.        The authors should explain how a correction for multiple comparisons is realized. 

Pilot study

      i.        The authors should comment on how the results were obtained. Are the effects based on an analysis across all available data or do the results refer to the mentioned ROIs? 

     ii.        Again, how do the authors account for errors due to multiple comparisons?

Review

This registered report proposes a study on the role of extra-striate areas in conscious motor behavior. The authors propose to conduct a study where a cartwheel stimulus with a bar at perception threshold is presented and participants have to report whether they perceive an increase in bar thickness or not. Response requirement will be counterbalanced in different blocks, in which sometimes participants press a button when aware and vice versa. Neural data will be recorded via fast optical imaging. I think the study addresses an interesting question and would expand other studies that addressed similar questions with EEG.

1.       General point in the introduction: I think the motivation for the study comes across clearly and it is true that the manipulation allows dissociating effects of motor responses from consciousness if a go/nogo paradigm is used. However, motor responses are not the only possible confounds in consciousness research, as has been shown by several studies (e.g. Dellert et al., 2021; Pitts et al., 2012; Schelonka et al., 2017; Schlossmacher et al., 2020; Shafto and Pitts, 2015) the sole task relevance  of stimuli (in the absence of a motor response) can also elicit a late positivity. I think a more detailed discussion of these issues would benefit the registered report. I think the phrasing ‘… isolating neural activity strictly related to awareness from response-related mechanisms …’ is a little too strong.

2.       What was the motivation for the use of cartwheel stimuli? Given that the stimuli in itself will always be consciously perceived, the awareness question is specifically for the awareness of radius thickness and not of the stimulus per se. Given the introduction/abstract I would expect e.g. Gabor gratings at the perception threshold (like in Koivisto et al. (2016)). Does this influence the interpretation of the results? Furthermore, will the first radius clockwise always be the radius that is potentially thicker? Could this lead to confounds?

3.       Sample size estimation: I think taking the average sample size from previous EROS studies and increasing it, is a first step to sample size estimation in this particular study. However, using past EROS studies on unrelated topics would only partially be helpful, as not only EROS signal-to-noise ratio but also the question at hand and thus the expected effect size is (even more) important for determining sample size. I would like to suggest to the authors to obtain an effect size estimate from EROS data (e.g. from the previous study of the authors) even if such indices are not computed by the software itself. To my knowledge there are several ways to estimate effect sizes based on z-values/t-tests/F-tests/mean and SD differences. Maybe peak/mean values of such a kind can be extracted from the software and then be converted to estimate an effect size of EROS data for a power analysis in addition to the other arguments for the chosen sample size. 

4.       2.2.2 Exclusion criteria: How will it be quantified whether an EROS signal could not be detected? Is there an objective way to do so? 

5.       Perceptual threshold assessment: If I understood correctly, participants will be excluded if none of the nine predefined radii fits the 50%-threshold. Even if stimuli have to be created beforehand in Matlab, a more fine-grained or wider range of stimuli could probably be obtained to find stimuli that allow inclusion of more participants. 

6.       I fear there is a possibility of perceptual learning that should be addressed. Given that 2 sessions á 3 hours are planned, it seems possible that the stimulus that was initially perceived at the 50% threshold will be perceived more easily due to learning. Behavioral data from the pilot study might also hint at this problem, as three participants were excluded using a slightly higher awareness threshold of 80% compared to the planned experiment (75%). One possibility would be to continuously monitor performance and decrease thickness of the radius if participants cross a predefined performance threshold.

7.       I think the pilot data is very insightful, however, the interpretation of the statistical tests of the behavioral data seems somewhat premature given the small N and rather large differences in observed means, e.g. 50ms difference in reaction times between go aware and go unaware. Given a full sample of 24 participants, it seems possible that significant differences in reported aware trials as well as differences in reaction times will be observed. How would this impact the interpretation of the results? Could differences in motor responsivity measured by EROS (Q3) be attributed to differences in response times? Given that null effects are reported and interpreted, the authors could consider inclusion of Bayesian statistics as they allow a quantification of these effects.

8.       I think it would be interesting to see performance of participants of the pilot study on catch trials in order to see whether the paradigm works as planned. How will catch trials be investigated? If participants also report seeing the thicker bar on a substantial amount of catch trials this could hint at problems in the experimental stimuli/design.

Minor points:

Abstract, l. 16, Introduction l.76: I think peculiar/peculiarity is not the right word to describe the experimental design/methods (see e.g. https://www.merriam-webster.com/thesaurus/peculiar)

l. 67ff This sentence seems incomplete, connector words are missing for it to make sense (like ‘… are consistent with considering…, however the localization …’

Section 2.1: typo of ‘and.’ in the first sentence

Matlab is written inconsistently (sometimes Matlab, sometimes MATLAB)

Different version number of Matlab are reported throughout the manuscript, is this intentional?