Pitch–Luminance Crossmodal Correspondence in the Baby Chick: An Investigation on Predisposed and Learned Processes
Round 1
Reviewer 1 Report
The authors present three experiments designed to test whether there is some form of pitch-luminance crossmodal correspondence in the baby chick. Their starting motivation is to explore whether the association between high luminance and high pitch found in humans and chimpanzees can be found outside the group of mammals. The study does not find any evidence of such crossmodal correspondence in baby chicks and discusses possible reasons for this negative output.
The manuscript is clearly written, its story engaging, the case well explained, well-motivated and well referenced. The three experiments are well-motivated and (mostly) well described. The statistical analysis and the mixed modelling used are sound. I explain below some important points that should be tackled before publication. None of them requires adding any experiment.
Main points:
Replication: as stated above the experiments are clearly described. However, several points are missing to allow full replication. The authors should clearly state what the low and high luminance levels were. Ideally, experimenters would need the reflectance of the white and the black panels, and the reflectance of the grey plastic of the experimental triangular arena. They would also need to know better the lighting of the arena (how many lamps, power, brand etc.) and what the resulting luminance of the experimental panels. The paper should also give the volume of the high and low pitch sounds, both in dB SPL. All these data are fundamental for replication.
What was the rationale for choosing the frequencies of 175 and 1047 Hz? A priori, the avian auditory system is much more sensitive to frequencies in the range of 600-2500 Hz than to a frequency of 175 Hz. How could that affect the outcome of the experiment? Did the authors compensate for the greater sensitivity to the higher frequency by lowering the volume of the high pitch sound? Is that a point we must consider when doing such experiments?
Ethics: as I understand it, the experiments were run at the University of Padova. I cannot find any statement giving the approval of the ethic approval for doing the experiment from the university, or from any ruling organ. Is that an oversight? Don’t such experiments need ethic approval? This point should be clear to allow publication.
A related point is that of the absence of buddy chick in the test arena. Most ethic committees wouldn’t accept such experimental conditions as they would be very stressful for the chicks. Could the stress induced by the absence of buddies influence the output of the experiments? Why didn’t the authors consider introducing buddies in/close to the experimental arena?
In the first experiment, did the authors explore a possible history effect for the choices of the chicks? That is to say, did the authors check whether a chick’s first choice of colour (first trial) was then mainly chosen for the other trials? That can be easily tested using the authors’ statistical modelling by only considering the trials after the first one and the choice for the first trial as an independent variable.
P-values: the authors have chosen the route of assessing their effect using p-values and significant levels. For coherence, I think they should remove any instance of claims such as ‘on merely descriptive level [there was a higher… but not statistically significant]’ or similar. That’s how it should go with p-values. If a p-value is not below the chosen threshold, the right statement is ‘we did not find any effect of …’.
Other points:
L61: please spell out SNARC for the profane.
All figures showing the wavelengths: showing wavelengths to illustrate the low and high frequencies is very useful. The low frequency is shown with a lower amplitude, which is misleading as, a priori, both the low and high pitch sounds were displayed with the same volume (but had probably different loudness, see above). Please make all the amplitudes equal in Figs. 1, 2 and 5 to avoid this confusion.
First use of congruent in the manuscript: the word ‘congruence’ here is based on a phenomenon observed in humans and chimpanzees; I think that should be mentioned when the word congruence is first used to avoid any confusion (the authors could use ‘putative congruence’ at first or similar).
L212: 3.4 Discussion -> 3.4 Interim discussion, or 3.4 Interim discussion, Experiment 1
L289: 4.4 Discussion -> 4.4 Interim discussion, Experiment 2
L308-317: all the descriptions and explanations of the experiments are excellent in the rest of the manuscript, but this motivation is slightly confusing; I think explaining from the very beginning that there are 4 possible configurations would help.
L452: 5.6 Discussion -> 5.6 Interim discussion, Experiment 3
Gallus gallus -> Gallus gallus domesticus ?
Fig. 4 : confidence intervals -> 95% confidence intervals?
As a last point, while the writing of the paper is great, clear, well-structured and engaging, there are some minor languages mistakes that could be easily solved by showing the manuscript to a native.
Author Response
Reviewer 1
The authors present three experiments designed to test whether there is some form of pitch-luminance crossmodal correspondence in the baby chick. Their starting motivation is to explore whether the association between high luminance and high pitch found in humans and chimpanzees can be found outside the group of mammals. The study does not find any evidence of such crossmodal correspondence in baby chicks and discusses possible reasons for this negative output.
The manuscript is clearly written, its story engaging, the case well explained, well-motivated and well referenced. The three experiments are well-motivated and (mostly) well described. The statistical analysis and the mixed modelling used are sound. I explain below some important points that should be tackled before publication. None of them requires adding any experiment.
We wish to thank the reviewer for their comments, and thoughtful suggestions. We believe the reviewer’s feedback helped us to significantly improve the clarity of our manuscript.
Main points:
Replication: as stated above the experiments are clearly described. However, several points are missing to allow full replication. The authors should clearly state what the low and high luminance levels were. Ideally, experimenters would need the reflectance of the white and the black panels, and the reflectance of the grey plastic of the experimental triangular arena. They would also need to know better the lighting of the arena (how many lamps, power, brand etc.) and what the resulting luminance of the experimental panels. The paper should also give the volume of the high and low pitch sounds, both in dB SPL. All these data are fundamental for replication.
We thank the reviewer for pointing this out. We have included all the required information in the revised manuscript. Please see lines: 134-137, 147-150, and 167-168.
What was the rationale for choosing the frequencies of 175 and 1047 Hz? A priori, the avian auditory system is much more sensitive to frequencies in the range of 600-2500 Hz than to a frequency of 175 Hz. How could that affect the outcome of the experiment? Did the authors compensate for the greater sensitivity to the higher frequency by lowering the volume of the high pitch sound? Is that a point we must consider when doing such experiments?
As this work was aiming at replicating the original effect found in humans and chimpanzees, we decided to maintain the same stimuli previously used by Ludwig and colleagues. These frequencies still fall in the range of frequencies that chicks can perceive, although, as the reviewer correctly says, there might be an effect related to an enhanced sensitivity of chicks’ ear in the range 600-2500 Hz. Yet, chicks could hear both sounds, and we believe that the difference in pitch could still have resulted in a crossmodal association, would this effect be present. Consistently with this idea, chicks can process and imprint on auditory sounds below the 600-2500 Hz window (Bolhuis and Van Kampen, 1991). We do agree with the reviewer that this aspect requires further clarification, and we included this consideration in the General Discussion. Please see lines: 134-138 and 576-585.
With respect to the second point raised by the reviewer, i.e., adjusting the volume between the two sounds, we decided not to control for this variable when designing the experiment. In fact, volume and frequency are two dimensions of sounds that co-varies in nature. Because of our work being a first attempt to study visual-auditory correspondences in chicks, we decided to maintain the ecological characteristics of the auditory stimuli, planning to add further controls to disentangle the different roles of either volume, or frequency, in case of finding evidence of crossmodal correspondences.
Ethics: as I understand it, the experiments were run at the University of Padova. I cannot find any statement giving the approval of the ethic approval for doing the experiment from the university, or from any ruling organ. Is that an oversight? Don’t such experiments need ethic approval? This point should be clear to allow publication.
The reviewer is totally right, of course this study had obtained the appropriate ethical permissions, we apologies for missing out this information. We have added the ethical statement in the subject section, please see lines: 124-126.
A related point is that of the absence of buddy chick in the test arena. Most ethic committees wouldn’t accept such experimental conditions as they would be very stressful for the chicks. Could the stress induced by the absence of buddies influence the output of the experiments? Why didn’t the authors consider introducing buddies in/close to the experimental arena?
The rationale for testing chicks individually was to avoid possible confounding caused by the presence of another animal. This includes a reduced motivation to engage the task or reduced attentional resources to the experimental setting. In this specific case, moreover, a conspecific would have added both luminance and auditory stimulations that could have interfered with the crossmodal association that we aimed to explore (see General Discussion, lines: 569-673). Yet, as the reviewer rightfully says, there is still the possibility of chicks becoming stressed or anxious when separated from their cage companion. This possibility was carefully considered during training and testing procedures. Chicks could spend a first phase of acquaintance to the experimental arena, in which they could freely explore the environment, and receive some food (i.e., the mealworm hidden behind the central panel at training). This helped in creating a positive association with the experimental setting. Moreover, as the paradigm did not involve any form of punishment or negative reinforce, chicks were never subject to any unpleasant experience that could lead to distress or frustration with respect to the experimental procedures. In addition, chicks were monitored for the entire duration of the training and test by the experimenter, and in case any sign of stress (e.g., sustained distress calls, feather pecking, loss of appetite) was detected, the experiment had to be immediately ceased and the chick was returned to the home cage. This situation never happened during our study.
We really appreciate the reviewer’s concern for animal wellbeing, and we agree that when designing an experiment, it is crucial to account for factors such as paradigm-related stress. We agree that this is a sensitive issue, and we clarified the rationale behind our choice in the manuscript. Please see lines: 116-123
In the first experiment, did the authors explore a possible history effect for the choices of the chicks? That is to say, did the authors check whether a chick’s first choice of colour (first trial) was then mainly chosen for the other trials? That can be easily tested using the authors’ statistical modelling by only considering the trials after the first one and the choice for the first trial as an independent variable.
We controlled for an effect of the first choice on both the second choice, and the total performance. In both cases we did not find any significant effect. We have not included this information in the manuscript, as we believe it might overwhelm the reader without conveying any insightful information. However, we are ready to amend the manuscript and include these analyses if the reviewer deems it is worthy to mention.
Additionally, when designing the experiment, we had considered the possibility of the first choice at test having an effect on chicks’ behaviour. For this reason, we added four last trials in the training condition, in which chicks could circumnavigate a single central black panel, and a single central white panel, for two times each. This was aimed at allowing the chick to associate the black and white panel with obtaining a reward, reducing the possibility of them creating a single luminance-reward association at the very first test trial that could bias the following choices. We have clarified this passage in the text. Please see lines: 167-172.
P-values: the authors have chosen the route of assessing their effect using p-values and significant levels. For coherence, I think they should remove any instance of claims such as ‘on merely descriptive level [there was a higher… but not statistically significant]’ or similar. That’s how it should go with p-values. If a p-value is not below the chosen threshold, the right statement is ‘we did not find any effect of …’.
We thank the reviewer for pointing this out. We have amended this kind of sentences from the result sections. We still believe that in terms of discussion it might be useful to include considerations based on qualitative observations. Thus, we maintained such considerations in the discussion, while avoiding the use of syntactic structures as those highlighted by the reviewer. See for example lines: 272-275, and 328-330.
Other points:
L61: please spell out SNARC for the profane.
Thank you for pointing this out, we have added the explanation for the acronym. Please see line: 65.
All figures showing the wavelengths: showing wavelengths to illustrate the low and high frequencies is very useful. The low frequency is shown with a lower amplitude, which is misleading as, a priori, both the low and high pitch sounds were displayed with the same volume (but had probably different loudness, see above). Please make all the amplitudes equal in Figs. 1, 2 and 5 to avoid this confusion.
We agree with the reviewer that the figures might be misleading, and we adjusted them so that they now all display the same amplitude. Please see the revised Fig. 1, Fig. 2, Fig. 5.
First use of congruent in the manuscript: the word ‘congruence’ here is based on a phenomenon observed in humans and chimpanzees; I think that should be mentioned when the word congruence is first used to avoid any confusion (the authors could use ‘putative congruence’ at first or similar).
We have added a sentence in the manuscript to explain the fact that the terms “congruent” and “incongruent” are used with reference to the results from humans and chimps. See lines: 139-142.
L212: 3.4 Discussion -> 3.4 Interim discussion, or 3.4 Interim discussion, Experiment 1
We have modified the paragraph title indicating it is as interim discussion.
L289: 4.4 Discussion -> 4.4 Interim discussion, Experiment 2
We have modified the paragraph title indicating it is as interim discussion.
L308-317: all the descriptions and explanations of the experiments are excellent in the rest of the manuscript, but this motivation is slightly confusing; I think explaining from the very beginning that there are 4 possible configurations would help.
We have revised this part, trying to improve clarity as suggested. Please see lines: 341-353.
L452: 5.6 Discussion -> 5.6 Interim discussion, Experiment 3
We have modified the paragraph title indicating it is as interim discussion.
Gallus gallus -> Gallus gallus domesticus ?
We have corrected the manuscript, indicating the full scientific name of the species.
Fig. 4 : confidence intervals -> 95% confidence intervals?
Thank you for noticing the missing information. We have corrected the figure legend indicating 95% confidence intervals.
As a last point, while the writing of the paper is great, clear, well-structured and engaging, there are some minor languages mistakes that could be easily solved by showing the manuscript to a native.
We really thank the Reviewer for their kind comments. We have carefully checked the manuscript for language mistakes and typos, and modified some passages to improve readability (not highlighted).
Reviewer 2 Report
My main concern regards the authors’ choice to test baby chicks. Is the space luminescence crossmodal corrispondence described in young individuals of other species (like humans or chimpanzees reported in the study)?
Moreover, regarding the data analysis and results, I think that further considerations are needed, with particular reference to the influence of laterality on chicks’ responses.
Did the initial position of the chick influence the side chosen? (as represented in the figure 2, the body is already oriented toward the black panel while the sight direction is not clear)
Potential effects of learning on the animal responses should also be investigated, for instance those related to the rewarded side. And in particular: were there differences in side preference in each individual (regardless the visual stimulus presented on that side)? Did you find similar preference also in the detour behaviour during training (for instance chicks that preferred to turn around the panel on the left during training showed a strong preference toward the left-placed panel during test?). It has been found both in chicks and in other species (e.g. dogs) that visuo-spatial abilities are lateralized and could be related to motor laterality (e.g. already reported in Diekamp, B., Regolin, L., Güntürkün, O., & Vallortigara, G. (2005). A left-sided visuospatial bias in birds. Current Biology, 15(10), R372-R373; //// Siniscalchi, M., d’Ingeo, S., Fornelli, S., & Quaranta, A. (2016). Relationship between visuospatial attention and paw preference in dogs. Scientific reports, 6(1), 1-8.), and that lateralization of multisensory processing affects auditory learning in chicks (Harshaw, C., Ford, C. B., & Lickliter, R. (2021). Hearing better with the right eye? The lateralization of multisensory processing affects auditory learning in Northern bobwhite quail (Colinus Virginianus) chicks. Applied animal behaviour science, 236, 105274.)
Since a preference for the white panel is reported (even though not statistically significant), are there any differences in other parameters like the animal speed in reaching the reward (to test the "performance" related to the stimuli association?)
Regarding the subject tested, it is not clear to me how testing only males could have help minimizing the sample size (line 91). Since only males were tested, I think that results should highlight this limit, since sex differences have been found in learning of visual and acoustic stimuli in chicks (see Versace, E., Spierings, M. J., Caffini, M., Ten Cate, C., & Vallortigara, G. (2017). Spontaneous generalization of abstract multimodal patterns in young domestic chicks. Animal cognition, 20(3), 521-529.)
Please specify how many individuals were tested in each experiments and their age (if changed).
As for the methodology employed, it would be necessary to provide further details about the “passing criterion” of Experiment 3. How did the correct trials rated? Were they considered "correct" when the chick went directly to visit the rewarded box? Did the animals have a time limit for peaking the correct box?
Again, as for laterality effect, war the position of the "correct stimulus" constant in the acquisition and reversal phase or was it randomized?
Further reference to cross-modal learning in animals should be provided (e.g. -Cloke, J. M., Jacklin, D. L., & Winters, B. D. (2015). The neural bases of crossmodal object recognition in non-human primates and rodents: a review. Behavioural brain research, 285, 118-130.;
-Quaranta, A., d’Ingeo, S., & Siniscalchi, M. (2020). Odour-Evoked Memory in Dogs: Do Odours Help to Retrieve Memories of Food Location?. Animals, 10(8), 1249;
-Korzeniowska, A. T., Simner, J., Root-Gutteridge, H., & Reby, D. (2022). High-pitch sounds small for domestic dogs: abstract crossmodal correspondences between auditory pitch and visual size. Royal Society Open Science, 9(2), 211647.)
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Round 2
Reviewer 2 Report
All the issues have been successfully addressed by the authors.
