Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

I have read the manuscript entitled: “Sensory and Cognitive Modulation of Saltiness Processing in Everyday Food: Evidence from two fNIRS Experiments”. The manuscript deals with the saltiness perception in two soup types (egg and corn), using functional near infrared spectroscopy. The authors conducted two experiments. In the first experiment, five different samples of each soup were used: regular, low-salt,  mono-potassium glutamate enhancer, bacon odor and soy sauce odor. In the second experiment, the soups were presented with packaging labels: regular, low-salt and no label. Results showed that the rating for saltiness in the regular egg soup was similar to the MPG enhanced. Meanwhile, in the corn soup the saltiness perception in the regular soup was higher than the other samples. In the second experiment, results showed that no changes in the perception were observed, although the labels were presented. The study presents results that are novel and interesting in the field. Nevertheless, some issues regarding the methodological process are present. The authors should address these issues before considering for publication in the journal. 

In section 2.1.1 and 2.2.1 the authors should include the bioethics protocol code.

In section 2.1.2 is not clear how did the authors choose the amount of enhancers (MPG and IMP) and the bacon and soy sauce odors. If this choice was made by previous works, the appropriate reference should be cited. 

In section 2.1.3, the procedure present some issues. What was the order of sample presentation for the participants? In the first experiment, how was response anticipation bias avoided? Was a random, balanced, and counterbalanced order used? If the samples were always presented in the same order, how was the saturation of saltiness perception avoided? Was purified water used between samples to break the stimulus from the previous samples? On the other hand, what was the environment like  where the tests were conducted to ensure that stress was avoided in the participants and therefore, prevent bias in the responses? In this type of tests, it is generally preferable to taste a low number of samples to avoid fatigue and stimulus saturation in the participants. How do the authors guarantee that after tasting four or five samples, the participants responded without bias? Did the authors defined previously palatability to the participants? In complex foods, the definition of a sensory attribute or descriptor is important to avoid bias in the responses. Why did the authors choose a scale without labels at the ends? Generally, when tests are conducted in participant without prior training, the scales indicate each of the categories used to rate the sample. 

In section 2.1.4 the authors mentions the wavelengths used, however, the authors do not specify why those particular wavelengths were chosen. 

Please include the appropriate reference for each package used in the R software. 

Author Response

Dear Reviewer 1, we appreciate your kind comments on our manuscript. We corrected our manuscript with your guidance. Please see attached file, where point-to point responses are described.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript addresses a highly relevant public health challenge: reducing dietary sodium intake without compromising food palatability. The authors successfully transition from simplified model solutions to everyday foods (egg and corn soups) using functional near-infrared spectroscopy (fNIRS). The choice of functional near-infrared spectroscopy (fNIRS) is highly appropriate, as it allows for natural eating behaviors compared to motion-sensitive techniques like fMRI or EEG.

The study is well-grounded, the literature review is up-to-date, and the neuroimaging approach under naturalistic eating conditions provides a highly realistic and reliable experimental framework.

The manuscript is well-written, and the English language is acceptable.

Comments and Corrections:

• Line 316: The text incorrectly refers to “Experiment 4 (Figure 10)”. Since this manuscript only describes two experiments, this appears to be a typographical artifact. Please correct this to “Experiment 2 (Figure 4)” and thoroughly audit the entire text to ensure cross-referencing consistency.
• Reference 11 and Reference 48 are completely identical (Liem et al., 2012). Please remove the duplicate and re-index the subsequent citations.

Author Response

Dear Reviewer 2, we appreciate your kind comments on our manuscript. We corrected our manuscript with your guidance. Please see attached file, where point-to point responses are described.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This manuscript addresses an interesting topic by investigating how sensory and cognitive factors modulate saltiness perception in relatively realistic food matrices using behavioral ratings and fNIRS measurements.

However, several aspects of the experimental design, statistical analysis, and interpretation require revision before the manuscript can be considered for publication. In particular, some of the mechanistic conclusions regarding flavor congruency, packaging-induced expectations, and the functional interpretation of individual fNIRS channels appear stronger than what can be directly supported by the experimental design.

Major comments

The manuscript repeatedly attributes the different effects observed in egg and corn soups to differences in flavor congruency. However, congruency was neither directly measured nor experimentally manipulated. Egg and corn soups differ in several other characteristics, including flavor profile, familiarity, sensory properties, sweetness, and potentially saltiness. Moreover, the statistical analyses were conducted separately for the two soup categories. Therefore, differences in statistical significance between egg and corn soups do not by themselves demonstrate that the magnitude of the enhancement effect was significantly different between the two food matrices. The authors should either provide an appropriate statistical comparison/interaction testing the food-dependent effect, if possible, or substantially moderate the causal interpretation of congruency and clearly describe it as a possible explanation rather than a demonstrated mechanism.

The use of Benjamini–Hochberg correction for the channel-wise fNIRS analyses is appropriate, however the manuscript should provide more detail regarding exactly which family/families of tests were included in the correction. Similarly, the methods used for the hemodynamic response–behavioral correlation analyses should be described more explicitly. Please specify the correlation method, the unit of analysis, how repeated observations were handled, and whether multiple testing was controlled for across channels and behavioral variables. This is particularly important because several reported correlations are weak (e.g., r values around 0.15–0.25), and their statistical significance should not be interpreted as evidence of a strong functional association.

The behavioral results show no significant effect of packaging on saltiness or palatability, whereas neural differences were observed in several channels. The authors appropriately acknowledge this discrepancy, but the discussion should more clearly distinguish between an observed neural modulation and an inferred cognitive mechanism. In particular, packaging-induced expectations were not directly measured before tasting. Therefore, the current data cannot directly establish that packaging altered expectations and that these expectations subsequently modulated taste processing. The relevant conclusions should be phrased more cautiously, for example as being consistent with a possible influence of contextual or expectation-related processing.

The interpretation of individual channels in terms of specific cognitive functions should be moderated. fNIRS provides relatively limited spatial resolution, and several channels appear to overlap multiple anatomical regions. Moreover, some of the discussed regions are not exclusively related to taste processing. Statements assigning specific functions such as semantic integration, expectation monitoring, or higher-order evaluation to individual channels should therefore be presented as hypotheses consistent with previous literature rather than as direct demonstrations of the underlying mechanism. This is particularly relevant to the proposed sequence of processing involving the Rolandic operculum, temporal regions, and prefrontal cortex.

The use of commercially available soups and fNIRS during actual consumption is a strength of the study. Nevertheless, the experimental procedure remained highly controlled, including fixed holding, swallowing, and rating periods. Therefore, expressions such as “everyday eating behavior” or highly naturalistic consumption should be used cautiously. The authors may consider describing the paradigm as more ecologically relevant than studies based on simple taste solutions or fMRI rather than as a direct representation of everyday eating.

The Methods do not clearly state whether the order of presentation of the experimental conditions was randomized or counterbalanced. Because each condition was presented only once, order effects, adaptation, carry-over effects, and sensory fatigue may potentially influence the results. Please clarify how the order of the 10 conditions in Experiment 1 and the 12 conditions in Experiment 2 was determined. If the order was not randomized/counterbalanced, this should be acknowledged as a limitation.

Both experiments included 30 participants, but no rationale or power analysis is reported. Please provide a justification for the selected sample size, particularly given the large number of fNIRS channels and statistical comparisons. If an a priori power analysis was not performed, this should be acknowledged as a limitation.

The preprocessing pipeline is described in considerable detail, which is appreciated. However, please clarify whether short-separation channels or any procedure for correcting systemic physiological fluctuations were available and/or applied. This is particularly relevant in a food-consumption paradigm involving hand and mouth movements and changes in respiration. If such correction was not possible, the potential contribution of systemic signals should be acknowledged as a limitation.

The presentation of the soup formulations should be clarified, particularly in Table 1. The table appears to report 6.5 g and 18.6 g under the “Salt Content” column, whereas the text reports 1.2 versus 0.7 g NaCl per serving for egg soup and 1.0 versus 0.48 g NaCl per serving for corn soup. Please verify the table headings and clearly distinguish soup powder amount from actual sodium/salt content. The concentration/composition of the odor solutions should also be reported more explicitly, including the solvent and concentration where applicable, rather than only the volume applied to the cotton pads.

There appears to be a typographical error in Section 2.2.3, where the procedure is referred to as “Experiment 4 (Figure 10)”. This should presumably refer to Experiment 2 and Figure 4. Please carefully check all figure and experiment cross-references throughout the manuscript.

Some conclusions appear broader than the results support. For example, the statement that external factors alter taste-related processing “rather than salt content itself” should be reconsidered because the present data clearly demonstrate substantial effects of salt content on both saltiness and palatability ratings. A more accurate interpretation would be that sensory and cognitive/contextual factors can modulate taste-related processing in addition to the effects of salt content. Similarly, the statement that packaging cues are “critical” for shaping product acceptance appears too strong given that packaging did not significantly affect explicit palatability ratings in the present study. The conclusions should focus more closely on the demonstrated neural modulation and avoid implying a demonstrated effect on consumer acceptance.

 

Minor comments

• Please carefully revise the English language throughout the manuscript. Several sentences are grammatically awkward or overly complex, and some expressions should be refined for clarity and scientific precision.
• Please use consistent terminology for “saltiness enhancer”, “saltiness enhancement”, “low-salt”, and “reduced-salt” throughout the manuscript.

Author Response

Dear Reviewer 3, we appreciate your kind comments on our manuscript. We corrected our manuscript with your guidance. Please see attached file, where point-to point responses are described.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The manuscript describes a sensory and cognitive modulation of saltiness processing in everyday food.

The abstract describes the work and the principal aim of the research adequately

For the introduction section.

Lines 67-79, it is recommended to add some references.

Lines 98-122, it is recommended to edit and separate the information included in the introduction from the description of experiments that could be included in the methodology

Lines 115-188, it is recommended to use the third person.

For the methodology section

Line 136 it is recommended to use the third person.

For the results section

In Figures 11 and 19, given the low correlation between responses in both experiments, what can you discuss and argue about the impact on consumer responses? This part is not clear in the discussion. As mentioned, it is difficult to identify the factors influencing neural responses.

For discussion

Line 713 it is recommended to use the third person.

Line 777 it is recommended to use the third person.

Line 806 it is recommended to use the third person.

Line 809 it is recommended to use the third person.

It is recommended to homogenize references because, in some parts of the section, it uses author, [number], and in others, [number]

Author Response

Dear Reviewer 4, we appreciate your kind comments on our manuscript. We corrected our manuscript with your guidance. Please see attached file, where point-to point responses are described.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have made the appropriate changes to the manuscript. This new version has increased the clarity, quality and scientific rigor of the study. With these changes, the manuscript can be considered for publication in the journal. 

Author Response

We really appreciate for your important comments and discussion. 

Reviewer 3 Report

Comments and Suggestions for Authors

The revision has improved the manuscript. However, several issues concerning the experimental design and the statistical support for the main conclusions remain unresolved, and some of them affect the central claims of the paper.

Major concerns
The Methods now state that, within each block, the low-salt soup was always presented first and the regular soup always fifth, with only the three intermediate conditions balanced. The two reference conditions, against which all enhancement effects are evaluated, are therefore fully confounded with serial position. This affects both the behavioral and the neural data. For fNIRS, the first trial in a block is expected to elicit a larger response (novelty, orienting, absence of habituation). Consistent with this, in corn soup the responses at CH19 to both the low-salt (position 1) and the regular soup (position 5) exceeded the response to the bacon-odor condition (intermediate positions), and in egg soup CH19 responded more strongly to the low-salt soup than to the other conditions. The Discussion nevertheless interprets the greater CH19 response to low-salt soup as reflecting unfamiliarity or greater cognitive demand. Acknowledging this as a limitation (Section 4.4) is not sufficient. I ask the authors to (a) analyse the effect of serial position where possible (e.g., within the balanced intermediate conditions, or with mixed models including position as a covariate), (b) remove or substantially qualify the interpretations of the neural contrasts involving the low-salt and regular conditions, and (c) clarify how “planned, balanced orders” were implemented across 30 participants and 15 participants per block order (six possible orders of the three intermediate conditions cannot be equally distributed).

In egg soup, the MPG + IMP condition shares a letter with both the low-salt and the regular soup (Figure 6a). It therefore does not differ significantly from the low-salt soup, and it also does not differ significantly from the regular soup. The conclusion that umami “enhances” perceived saltiness, and that it does so more effectively than the other approaches, rests on the absence of a significant difference from the regular soup. This is the same problem that Reviewer 3 raised regarding egg versus corn soup. In corn soup, the mean ratings for the low-salt and the MPG + IMP conditions are virtually identical. Statements such as “Umami substances enhance the perceived saltiness of foods with complex flavors” (Section 4.3) and the related passages in the Abstract and Conclusions should be revised. Planned contrasts of each enhancer versus the low-salt soup, with effect sizes and confidence intervals (and equivalence tests where the authors wish to claim similarity to the regular soup), are needed.

The families of tests are now described, but the number of channels entering each family after quality pruning is not reported, and it is unclear whether the reported p-values are raw or BH-adjusted. Based on the F statistics and degrees of freedom reported, several effects would be unlikely to survive BH correction across the stated number of channels (e.g., CH17 and CH19 in corn soup in Experiment 1, with p ≈ 0.01 across 20 channels; CH32 in Experiment 2, with p ≈ 0.002 across 34 channels), unless the family size was smaller than stated. Please report the number of retained channels, raw p-values, and adjusted q-values for every reported effect, and explain why all tests have 29 denominator degrees of freedom despite the channel-pruning procedure (were no participant–channel data excluded?). Furthermore:

• • The “channels of interest” in Experiment 2 are reported with p < 0.05 uncorrected, described as “marginally significant”, and then used in the Discussion (including the argument on opposite DLPFC patterns across experiments) and in the correlation analyses. These should be explicitly labeled as exploratory.
  • Four channels of interest were defined in Section 2.2.6 (CH19, CH10, CH20, CH3), but results are reported only for CH19 and CH20. The results for CH10 and CH3 must be reported, including null findings.
  • Figure 11 includes channels that, according to the text, did not show a significant effect in the respective soup category (e.g., CH16 in corn soup, CH17 in egg soup). Please make the channel selection consistent with the text.
  • In Experiment 2, the family for the correlation correction (ten tests) is restricted to channels already selected for showing significant effects, which does not account for the selection step.

Interpretation of the Experiment 2 neural results remains over-flexible and vulnerable to artefacts. The stronger CH19 response to regular soup under regular packaging (congruent > incongruent) in egg soup and the stronger CH17 response under mismatched conditions (incongruent > congruent) in corn soup are both interpreted as “higher-order integrative processing”. An interpretation that accommodates effects in opposite directions is not falsifiable. A direct contrast of matched versus mismatched conditions should be tested, and the interpretation of the direction of effects should be stated a priori or explicitly flagged as post hoc.

The authors acknowledge that the packaging manipulation was probably weak (similar visual designs), that expectations were not measured, and that no behavioral effect was found. There is no manipulation check, and the null behavioral effects of packaging are interpreted without equivalence testing or Bayes factors. Attributing the neural differences to packaging-related contextual processing therefore remains speculative; this should be reflected in the Abstract and Conclusions.

The channel assigned to the temporal pole (CH17 in Experiment 2) lies at the lower edge of the probe set, close to the temporalis muscle, and is likely to be particularly sensitive to jaw movement, swallowing, and extracerebral hemodynamics. The depth sensitivity of fNIRS for the temporal pole and the Rolandic operculum should be discussed. Since no short-separation regression or physiological recordings were available, I strongly recommend reporting HbR (or HbT) results as a robustness check, and providing MNI coordinates and probability of anatomical assignment for all discussed channels in a table or the Supplementary Material.

Familiarity, expectation, and cognitive load were not measured. Interpretations of prefrontal and frontopolar activity in these terms are reverse inferences and should be clearly framed as such.

The same channel numbers denote different regions in the two experiments (e.g., CH17 is frontopolar in Experiment 1 and temporal pole in Experiment 2; CH19 is DLPFC in Experiment 1 and frontopolar in Experiment 2). Please use experiment-specific labels (e.g., E1-CH19), and justify the cross-experiment correspondences (probe geometries and systems differ), since the argument about “opposite DLPFC patterns” depends on them.

The correlations between saltiness and palatability (r = 0.32 and 0.25 in Experiment 1; r = 0.50 and 0.61 in Experiment 2) are not described. The associated p-values suggest that all observations were treated as independent. Please use repeated-measures correlation or a mixed model, consistent with the neural analysis.

Experiment 1 reports no significant palatability difference between low-salt and regular soups, whereas Experiment 2, with the same stimuli, reports a significant difference (p < 0.01) for both soups. This discrepancy is not discussed and may reflect the order confound (point 1) or other design differences.

Experiment 2 is described as “identical to Experiment 1” except for the cue, yet Figure 4 and the GLM description (T = 10 s) indicate different timing (1-s swallowing instruction and no cup-replacement step). Please correct the description.

 

The Methods state that MPG was used to “exclude the potential influence of sodium ions”, but IMP was used as the disodium salt (and MPG adds potassium). The sodium contribution of IMP and the potential taste contribution of potassium should be stated and discussed.

For the odorants, the response states that commercial flavorings were used undiluted, but the composition (including solvent), the intensity matching, and whether control conditions used untreated cotton pads and identical lids (with holes) are not reported. Soy-sauce odor significantly reduced palatability, which suggests that its intensity may not have been appropriate.

Minor comments

1. Citation numbering appears to be inconsistent after revision: “McClure et al. [29]” in Section 4.2 should presumably be [32]; “[50,51]” in the CH6 paragraph points to Hu et al. and Skaczkowski et al. rather than to the gustatory cortex references [53,54]; in Section 4.4, “[15, 26, 27, 36]” includes [36] (NIRS-SPM) and [27] (a molecular study on umami synergy). Please check all citations.
2. Typographical and language errors remain: “ow-salt soup” (Section 3.2.2); “Another limitation in Experiment is…” (Section 4.4, missing experiment number); “Following the produce of Onuma et al.” (Introduction); “Overall, this findings suggest…” (Abstract); “Lz” in the abbreviation list versus “Iz” in the text; “SD” is defined as source-detector but also used for standard deviation.

Author Response

All responses to the comments from reviewer 3 are described in the attached file. We really appreciate your kind comments for improving our manuscript.

Author Response File: Author Response.pdf