Continuous Video Monitoring of Zoo Cheetahs (Acinonyx jubatus jubatus) Reveals Differential Engagement Patterns for Environmental Enrichment Items Based on Sensory Category
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsDear Authors,
Thank you for submitting this manuscript that explores the behaviour of cheetahs when provided with enrichment types. This is an interesting study and there could be some practical applications for those who keep felids in zoo environments.
At current however, there seem to be some large revisions required in the manuscript to ensure the work is scientifically robust. I have attached the PDF version of the manuscript with specific comments. Additionally, please consider the following points:
1. Methods. The methods need to be clearly explained. What behaviour analysis methods was used? Continuous or one zero? Scan or focal? What times were the observations undertaken? Why are only three behaviours recorded? This seems rather incomplete?
2. Interobserver reliability. Please clearly explain how reliability was assessed. If multiple observers are used in a study, it is common to assess interobserver reliability (e.g. using Kappa). Please explain what was done here to ensure reliability.
3. Statistical test choice. It appears that several parametric tests have been selected (ANOVA, Pearson) and so there are concerns that the assumption of normality is violated. Please test data for normality and reassess using appropriate tests (e.g. Spearman correlation) where the tests were inappropriate.
4. Volunteer credit and COI statement. Have the students been given the opportunity for coauthorship? Technically they have engaged in data collection / curation. The conflict of interest statement is also a concern. If the work is not necessarily reflective of the views of the authors, then why have the authors agreed to publication?
Ultimately, this is a small scale study, and the discussion has a tendency to oversell the findings - including reference to anecdotal information (e.g. urination). There needs to be a much more realistic explanation of the results, which takes into account the study's limitations.
Comments for author File: 
Comments.pdf
English is fine - occasional grammar errors are present.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThis study used continuous video monitoring to assess engagement patterns for several enrichment items in zoo-housed cheetahs. The topic explored is definitely of interest to animal caretakers and has potential for further development, particularly the use of continuous video monitoring as a tool to monitor behaviour. I would also like to commend the authors on the diversity of enrichment items offered. However, I have several concerns with the work as presented. Firstly, the experimental design lacks sufficient replication to support the statistical analyses conducted, and the results presented. Based on the methodology presented, each enrichment was only offered twice per experimental treatment (i.e. day or night), which is not a sufficient sample size to conduct analyses or make statements like those presented in line 16. This could have been reduced by offering a single enrichment item per category (i.e. one and not two), and doubling the number of replicates. Similarly, a lack of any baseline data means interpretation of results is difficult. Furthermore, the limited behvavioural observations recorded (i.e. just grooming and enrichment engagement), does not reflect the full behavioural repertoire of the animals and does not utilise continuous visitor monitoring to its full potential. Further behavioural observations, such as stereotypy, resting, sleeping, etc. could have provided a more holistic view of the enrichment responses, and would have also allowed for the assessment of any potential negative responses to enrichment, which was not assessed in the work presented. I also found the description of the statistical analyses slightly confusing, and would recommend revision of this section for clarity. Overall, the work is valuable, and I would like to encourage the authors to revise the methodology to better capture the required information, and support the necessary statistical analyses.
Individual comments can also be found below:
Line 16 – This statement is not supported by the methodology, I do not think it is appropriate to say that visual enrichment significantly increased allogrooming when this was only assessed on two occasions. I would also question whether or not allogrooming could also be viewed as a stress response (as has been shown in other species), and think this could be further explored in the discussion.
Line 63 – The introduction would benefit from further exploration of how felids, and cheetahs in particular, react to different stimuli, and what has been done before in the field.
Line 151 – the relatedness of all cheetahs, the lack of acknowledgement of potential age-specific differences, and the grouping of results across the three animals all need to be better elaborated upon. Why was it not feasible to assess the reaction of each animal individually?
Line 154 – Section 2.5 is hard to follow at times. Further explanation of what exactly was assessed each time would be beneficial.
Line 156 – Does this mean results were grouped for all individuals?
Line 161 – “each behavioral event was analyzed by the mixed procedure of SAS” what statistical test is this referring to?
Line 163 – 165 - Further explanation is needed to show how time of day was incorporated into the analyses. Was this frequency per hour as presented in the tables?
Line 171 – Why were “autogrooming, allogrooming, and total grooming” the only behaviours looked at?
Line 187 – I am unsure which values in the table the chi-square results are referring to.
Line 203 – I am unsure which values in the table the chi-square results are referring to.
Figure 3 - I am unsure which statistical analysis was used for this figure.
Figure 4 - I am unsure which statistical analysis was used for this figure.
Line 239 – “fog” spelling
Line 245 – 246 – the statement made here is not represented in Figure 6.
Figure 6 – The text size on the graph appears too small
Line 258 – Figure 7 is missing.
Line 282 – This sentence is inappropriate as you did not look at or assess all “desired naturalistic behaviours”.
Line 286 – 287 – This statement needs to be referenced, as the situation is more nuanced than what is presented, especially as many zoos invest significant time in extending periods of time animals engage with food-related enrichment.
Line 355 - 356 – The idea that cheetahs had less capacity to maintain engagement seems doubtful and I do not believe it is evinced by the results presented.
Line 358 – was the sound on when the movie was being played? If so this is also auditory enrichment and needs to be considered separately.
Line 362 – 363 – What was the former enrichment regime for these animals and were they exposed to similar enrichment in the past? This will impact the novelty of enrichment.
Line 391 – 392 – I think the idea that an ipad mimicked moonlight is questionable and I would like to see this elaborated upon if retained.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsMajor comments:
I think the introduction needs more background information. Some questions to consider addressing within the introduction include: What is currently known about felid enrichment? What are the key differences between tactile, visual, and olfactory enrichment and why are each valuable to felids? What research has been done with each of these different enrichment types? What is the distinction between positive versus negative valence and arousal? What are the challenges and opportunities associated with enriching groups versus meeting the needs of single animals? What is known about the relationship between enrichment and group dynamics / social behavior – especially in carnivores or more specifically, felids?
Additionally, as the study is currently described, there are no clear hypotheses or predictions. This is a major weakness of the study that should be addressed. For example, are you expecting engagement with different enrichment types to differ? Why and in what direction? Are you expecting results to differ between daytime and nighttime? Why and in what direction? Why are you assessing the impact enrichment has on autogrooming and allogrooming? What is the expected relationship between enrichment and grooming?
Ln 100: Are there any previous studies that examined felid interactions with mirrors? I wonder if this might be considered a stressor to see “additional” cheetahs in their environment. Alternatively, is there any previous research to suggest felids can recognize themselves in mirror tests? If there are studies on this, they should be cited in the paper. Minimally, it might be beneficial to address this within the introduction.
Ln 108: Please add a sentence or two describing why cinnamon and cologne were chosen as opposed to more natural scents the cheetahs might encounter in their natural environments. Earlier in the text, the firehouse is described as resembling a prey animal and the Jolly Egg is similarly described, so the biological relevance of those enrichment items is clear. Please provide similar justifications for the cinnamon and cologne.
Ln 122: Please add information about the cheetahs’ previous experience with the various enrichment items being tested. If the cheetahs had not previously encountered these items or if their experience varies with the different items, then it is possible neophobia could be confounding the results of the study.
I think the section on “Potential Study Limitations” would fit better in the Discussion rather than in the Methods. Additionally, in my comments above, I bring up a few other potential limitations that should be addressed.
Minor comments:
Ln 22: Remove “s” from “highlights”
Ln 42: Change “its environment” to “their environment”
Ln 45: It’s unclear what is meant by “Differences in behavioral flexibility…” Perhaps a follow up sentence with a specific example would help illustrate the point being made here. If there are studies to cite here, that would be even better.
Ln 52: I think it’s important to clarify here the distinction between positive vs. negative valence and arousal.
Ln 58. Missing period after “time-consuming”
Ln 81: Please specify whether the cheetahs had access to indoor and outdoor areas 24/7 or whether access was limited.
Ln 135: Was inter-observer reliability quantified? Please add this information. Also, a period is missing after “engagement”
Ln 169: Missing period after “(daytime, nighttime, overall)”
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for AuthorsThe following study examined six different environmental enrichments in three female group-housed cheetahs. The authors found that the cheetahs engaged more with visual enrichment for frequent and longer duration. Although the paper provides some useful information, it would require significant writing and statistical reporting revisions. Although enrichment has been studied across various species in captivity, the intro does not reflect this or give hypothesis-driven predictions. There are also a few methodological flaws, including a lack of a control condition, a lack of inter-rater reliability measures, and unclear or missing statistical information that must be addressed.
The greatest weakness is that there is only 1 group of cheetahs, which limits some of the statistics and interpretations. Another weakness is that there was no control condition in which no enrichment items were presented, and allogrooming or other behavior was measured. If there are recordings from days without enrichment, I would encourage the authors to examine those days to compare differences in allogrooming. Also, since the authors point out that enrichment increases welfare by allowing animals to engage in naturalistic behaviors, other behaviors could be measured other than allogrooming, such as activity, resting, pacing, scent marking, etc. The authors mention scent marking and proximity in the mirror condition (lines 316-324), but no data was presented (an ethogram of more naturalistic behaviors measured would strengthen this paper). In addition, there were 9 individuals that coded the videos. Although the authors mention training, information on what that training entails is missing. Did more than one individual score the same video, and if so, what is the inter-rater reliability? Were raters randomly assigned videos, or was it balanced that all raters watched videos with different enrichment? If not, this could impact the measures if only certain individuals watched one condition (e.g. visual). This information would be necessary before determining if this paper could be accepted.
Additional specific comments that would need to be addressed:
· Please provide the rationale for selecting the enrichment. These sensory categories have been presented in previous research in both zoo and shelter populations of animals; the introduction should reflect this previous work. Also, from all the previous work done on different enrichment categories, please provide more specific hypotheses regarding these different enrichments and why you were chosen at the end of the introduction, where you mention the objectives of this study. As is, it does not seem like this study was hypothesis-driven. Also, were all the enrichments novel to the cheetahs prior to this study? Why choose cinnamon and cologne rather than prey scent? More rationale was provided on what was chosen and how it was compared to previous research in other species.
· Enrichment was provided at two time points: daytime (between 0900 and 1100) and night (1500 to 1700). It also appears that feedings occurred during those times with the evening feeding happening in the middle of the recordings (lines 84-85; 0900 and 1600). Indicate how this interruption was handled with the data analysis. Were all videos, day and night, recorded and analyzed for two hours after feedings? If less because of keeper interruption, please explain how that was addressed in the frequency/hr measures. If behavior in the evening was recorded before feedings (1500) and after feedings (1600-1700), was the behavior different before and after feeding?
· Potential study limitations should be in the discussion, not the methods section, unless it explains how the data was handled for statistical analysis.
· In your statistical analysis (lines 155-58), the authors state, “Enrichment replication was treated as an experimental unit.” I am unclear what this implies. I am assuming that all 3 cheetahs’ behaviors were grouped together to provide a total frequency of engagement with enrichment items each day. Was that total then combined across the 4 days that enrichment type was given?
o Tables 2-4. The Chi-Square should also be reported in the text, and given the frequency data is already in the Table, repeating the info in the text is probably redundant.
o Please change “P” to lowercase (p) throughout.
o Given the sample size (and my comments above on how a unit was determined), I don’t believe an ANOVA can be used without violating assumptions. Please provide evidence if it is appropriate; otherwise, a non-parametric test (e.g., Kruskal-Wallis, Friedman, etc.) and post hoc tests (e.g., Dunn’s, Wilcoxon) would be best.
o Figures 3-5 seem a bit fuzzy. Increase quality for sharpness. Also, in the legend, be clear on what a-d indicates.
o For all statistics including the interaction on lines 244-246, please provide the full stats (F, df, p).
· The discussion is quite lengthy and much of it would be better placed in the Introduction as a prior prediction for the present study rather than ad hoc explanations unless you had no specific predictions before you conducted this study.
· Lines 313-319 refers to data (scent-marking) that was not presented earlier. As I stated earlier, this study would be stronger if it did examine other naturalist behaviors such as scent-marking but if you do not, you cannot refer to it here.
· Lines 336-338 you did not define or measure “relaxed” in this study so again, cannot refer to it here.
Overall, although this study is limited with its sample size, I do think some of this data would be valuable if the authors could give more information regarding their methods, data analysis, and rewrite the discussion to reflex what is known about enrichment types in captive animals.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsDear Authors,
Thank you for providing a revised copy of your manuscript and addressing the points.
While the majority of concerns have been addressed, the point pertaining to statistical analysis has not been fully addressed. Please replace the old parametric analysis with the non parametric statistics. This is important for repeatability.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsI thank the authors for addressing all of the concerns raised, and for editing the manuscript accordingly. The work as presented is a valuable piece of work and I have no further concerns or suggestions.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsLn 55-57: There are critiques against the use of behavioral diversity as an indicator of animal welfare. For example, please see Cronin and Ross 2019 "Technical contribution: A cautionary note on the use of behavioural diversity (H-index) in animal welfare science." Therefore, I think it is inaccurate to state that behavioral diversity "directly indicates positive welfare..." This sentence should be toned down. Additionally, within this paragraph it seems that you use behavioral flexibility and behavioral diversity interchangeably, however these mean different things. Please clarify.
Ln 98-100: As currently written, these are not testable hypotheses. "...the olfactory stimuli chosen might produce considerable engagement..." and "tactile items might elicit interest..." are not testable. As currently described, how would you know whether your "hypotheses" are supported from the data?
Ln 102: Please clarify "it..." As currently written, the text is not clear whether "it" refers to rate of engagement or grooming behaviors.
Ln 107: It's still not clear why you're interested in whether patterns vary between daytime and nighttime. What is the background literature on this? Also importantly, what are your hypotheses here?
In Lines 101-102, you hypothesize that enrichment engagement would inversely correlate with grooming behaviors, however this test is not clearly shown in your results section. Perhaps it would help to include a figure that shows the correlation between rates of enrichment engagement and rates of grooming. Related to this, it's odd that Table 5, which has the results of the correlation tests, is not shown until the Discussion section. I realize this has to do with how the journal diagrammed the article, however it makes me wonder if it would be helpful to have fewer figures in the manuscript so that these important results are able to be shown within the Results section where the content is most relevant.
Ln 207: This is misleading. Some reviewers had ratings below 90% according to Supplementary Table 1. You should report the lowest/cutoff inter-observer reliability score that was included in this study.
Ln 289-290: Incomplete sentence.
Ln 318-319: Why did you need to do Pearson's and Spearman's correlations? Please clarify in the manuscript.
Ln 451-471: You should also discuss that neophobia could have influenced your results.
Supplementary Table 2: "replication was" is repeated in the table footnote. Please delete this typo.
Supplementary Table 4: It's unclear why you have the results of the Pearson correlations reported here and in Table 5. Also as previously mentioned, it's unclear why it was necessary to use both Pearson and Spearman tests.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 3
Reviewer 3 Report
Comments and Suggestions for AuthorsLn 89-93: Do you have a citation for this rating system?
Ln 115-116, 118: As I've mentioned in my previous review, "produce considerable engagement" and "illicit interest" are not testable hypotheses. Similarly, Ln 118 states that enrichment items would "impact grooming behaviors" without any clarification on the direction of the prediction.
Ln 232: I think it would be useful for readers if you could describe within the manuscript why parametric and non-parametric tests were performed. In your letter, you stated that this was because another reviewer asked you to. Can you explain in the manuscript why this decision was made? It's unclear to me (and will likely be unclear to readers) why this was necessary. For example, were the data not normally distributed? If that's the reason, then please state that within the manuscript. Even so, this would mean using only non-parametric tests; I don't see why you'd need both parametric and non-parametric tests for the same dataset. Please justify within the manuscript.
Ln 472-474: As I've mentioned in my previous review, I think the discussion would benefit from mention that neophobia to the items used in this study could have impacted your results. I'm not asking you to assess neophobia, rather just acknowledge that neophobia could have influenced the patterns you observed and perhaps cite some research on neophobia. I think this would provide a more nuanced discussion.
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
