The Effect of Prior Viewing Position and Spatial Scale on the Viewing of Paintings

Round 1

Reviewer 1 Report

SUMMARY

The current study investigates how people visually attend to different aspects of paintings—specifically, faces, theme, and context (ROIs)—while trying to remember the painting. The researchers compared whether the initial view was centered or varied, and manipulated whether the participants were focused on global or local features of a figure prior to viewing the painting. The main dependent variables were recognition memory and eye movement patterns (total time viewing the painting, and likelihood, mean duration, and number of fixations for each ROI. The researchers concluded that both initial viewing location and prior spatial scale impacted memory for paintings and eye movement patterns while viewing the paintings.

EVALUATION

The researchers are essentially interested in whether viewing paintings is similar to viewing natural scenes. This is an important step in developing a comprehensive literature on how we understand our (visual) world. The methods are precise, and the data analyses are sophisticated. However, I am not convinced that the authors’ conclusions follow their results, primarily because it was difficult to understand how the analyses addressed the central questions set forth in the introduction. Below I offer recommendations for improving the manuscript.

1. The introduction needs to be expanded to include more context.

a. Currently, the introduction sets up the main question of whether the central bias is due to the starting location of the viewer or to the importance of objects in the center of the painting. However, more specific hypotheses are needed. For example, if the central bias is due to initial viewing location, then we should see x, y, and z. On the other hand, if the bias is due to the importance of the central objects, then we should see a, b, and c. This would help situate the results in a larger context.

b. In contrast to the first question, there is scant evidence regarding the second question, which addresses whether the spatial scale at which one is focusing on with a prior object affects viewing patterns of subsequent objects (lines 88-101). Is there no prior literature on this? If so, then it needs to be included. If not, then explain in more detail how the current study addresses this gap. Further, specific hypotheses should be included. For example, if attending to the global aspect of a previous object, then we should see x, y, and z, whereas attending to the local aspect should result in a, b, c. Again, this would help put the results in a larger context.

2. The methods need to be clearer.

a. The explanation of the main task (Stage 2) was difficult to follow. It would be easier to understand if it were illustrated in a figure. The text should follow the same order as in the figure.

b. The other measures should be explained in a little more detail. For example, what was in the art knowledge questionnaire? Perhaps provide a couple of examples. Also include how many questions there were; the scale used to answer the questions; how the overall score was calculated; and what the minimum and maximus scores were. In addition, the art knowledge statistics should be included in Table 1. The other measures should be described similarly, but briefly.

c. I realize that the race/ethnicity of participants is not as important across the globe as it is in the U.S., but it would be helpful to include this information in the Participants section. We know far too little about who we are generalizing our results to. (For background on this comment, see Qu, Y., Jorgensen, N. A., & Telzer, E. H. (2021). A call for greater attention to culture in the study of brain and development. /Perspectives on Psychological Science, 16/(2), 275–293.)

3. The results need to be grounded in the questions and hypotheses being addressed.

a. I suggest that the results section have section headings for each specific question, rather than including both in all the analyses. In the first section, provide only the basic analyses that address the question of whether the starting location of attention (central, varied) affects memory and eye movements, using the five dependent variables (recognition memory, total fixation time, likelihood of making a fixation to each of the ROIs, mean duration of fixations for each ROI, and number of fixations for each DOI. Relate the results to the hypotheses.

b. The second section should focus on the question of whether a local or global scale of attention affects memory and eye movements. Present the accuracy results for the NAVON task first to assure the reader that participants’ focus was on the correct scale. Then follow the same analyses as in the first section except for global vs. local.

c. Based on eye movement data from natural scenes, you should also present analyses for data from the first 5 seconds only, to test whether the influences are relatively short-lived.

d. The painting recognition accuracy results should be presented in terms of hits (correctly identifying paintings) and false alarms (incorrectly saying yes to a foil) to highlight any biases for saying yes that may differ among the conditions. Otherwise, the main comparisons should be on the accuracy of recognizing paintings only. Hits and false alarms should be analyzed separately unless there were specific hypotheses that require the inclusion of both in a single analysis.

e. I did not understand why the linear models compared face vs. context and face vs. theme, especially for the local/global analyses. I believe that the comparison should be between context and theme because that’s where one would see the effects of local vs. global spatial scale. This comment ties back to the question of what the hypotheses are.

f. Any further analyses (e.g., Figures 6 and 7) need to be explained more clearly, including any expectations for the results.

4. The discussion will need to be modified and I will not comment on it here.

5. Other suggestions.

a. Starting location was not manipulated, as suggested throughout the manuscript. Rather, the results from the current study (varied locations) were compared with results from a prior study (central location).

b. Use “central” and “varied” rather than “current” and “2022” to emphasize the comparison being made.

c. In the linear modeling, the baseline (= 0) seems to change between analyses. For recognition memory, baseline is the central condition (2022). For eye movements, the baseline seems to be the varied condition (current). State the baseline for each analysis.

d. The verbiage in the tables and figures do not always align with that used in the text. For example, Figure 4 uses the terms Fixation Index, and Saccade Amplitudes, which are not used in the text, making it difficult to understand the results.

e. The authors say that the impact of initial location and spatial scale are long-lasting without being clear about how they are using the phrase (e.g., ms, sec, min, hr). I suggest using the phrase “relatively long lasting” and mention that the effects are on the order of seconds. (But see comment 3c)

none

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Line 122: Does mean and sd refer to age? It's not made explicit, and if so is SD = 9.2 correct? Seems high variability for a sample of 31 undergraduate students.

Line 127-129: Rephrase sentence. Is it missing a verb?

Line 135-140: For clarity, please indicate what type of hypothesis test was used to derive the p-values.

Line 217:  "on distribution eye movements" --> "on the the distribution of eye movements"

Line 219-221: Rephrase sentence. Remove "we analyzed" perhaps.

Figure 3: Change font size of "Foils" to same as for "Global" and "Local".

Paragraph 3.2: No mention of which fixation/saccade detection algorithm + parameter settings were applied to the data. Should be stated for clarity and for aiding reproducibility of results.

High quality

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Dear Authors,

I would like to report some aspects that, in my opinion, require revision:

You start by stating that visual attention, expressed through fixations, is mostly located at the center of visual stimuli such as paintings. Do you mean this happens all the time? People observe different parts of paintings during visual exploration. If you mean that they initially center their vision, it is not only limited to paintings; there is evidence of this phenomenon in various visualizations, advertisements, or even maps.

You state that the central fixation bias is of particular interest in art perception studies. Could you explain why it is particularly important in art perception studies? I am not familiar with studies that claim this central bias somehow determines the visual perception of a painting.

The central location fixation is often required in visual studies as an interstimulus in the experimentation process to have a standardized eye position for every participant. Could you elaborate on the impact of different positions on the experimentation process in visual sciences?

For a small sample group, the difference between men and women should be as small as possible. Please provide an explanation for the unequal number of participants from the two sexes.

I am a bit confused because the main hypothesis says something different, and the discussion starts with a sentence that aims at different things than the aim or hypothesis. In the introduction, there is no mention of painting memorization related to the fixation starting position. Please clarify this at the beginning of the paper.

The paintings used in the experiment varied in terms of size. You did not consider this as an important factor in the data analysis. Would you expect differences in fixation duration in relation to the actual size of the displayed painting?

Overall, please review and address these points in order to improve the clarity and coherence of your paper.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

I thank the authors for carefully considering my recommendations. Their adjustments to the introduction highlighted more explicitly the hypotheses to be tested. Changes to the methods and procedures section addressed my concerns. Though my specific suggestions for the results were not included, adjustments were made to highlight the main questions being addressed. The discussion is generally complete.

I have a couple of minor suggestions.

1. Lines 212-220. Thank you for including more details. I suggest adding that the total possible score is 48 ((10 x 3) + (6 x 3)).

2. In the discussion, compare the current results with those from viewing natural scenes (see lines 75-78).

3. In the discussion, bring the story back to the real-life situation (see lines 57-60).

I also have a couple of comments regarding my earlier review.

1. I think it’s great that you are asking questions regarding how culture may influence how we view paintings. My comment regarding race/ethnicity (2c) was not intended to say that we need to compare race/ethnicity in any analyses, however; rather, my suggestion was merely to describe the sample.

2. Thank you for conducting the analyses I suggested and for clarifying why the face ROI should be the basis for comparison.

3. We come from different analytic traditions regarding recognition data. My background is with signal-detection theory, where we commonly distinguish between sensitivity (d-prime) and bias.

4. My suggestion for analyzing the first 5 seconds came from the results from natural scenes (see lines 75-77). The fact that your analyses suggest that the first 4 seconds are important indicates that viewing paintings and viewing natural scenes are similar.

none

Author Response

Please see the attachment.

Author Response File: Author Response.pdf