Aggressive dominance can decrease behavioral complexity on subordinates through synchronization of locomotor activities

Social environments are known to influence behavior. Moreover, within small social groups, dominant/subordinate relationships frequently emerge. Dominants can display aggressive behaviors towards subordinates and sustain priority access to resources. Herein, Japanese quail (Coturnix japonica) were used, given that they establish hierarchies through frequent aggressive interactions. We apply a combination of different mathematical tools to provide a precise quantification of the effect of social environments and the consequence of dominance at an individual level on the temporal dynamics of behavior. Main results show that subordinates performed locomotion dynamics with stronger long-range positive correlations in comparison to birds that receive few or no aggressions from conspecifics (more random dynamics). Dominant birds and their subordinates also showed a high level of synchronization in the locomotor pattern, likely emerging from the lack of environmental opportunities to engage in independent behavior. Findings suggest that dominance can potentially modulate behavioral dynamics through synchronization of locomotor activities.

2 -Line 133: "Diversity in the establishment of dominance relationships within social groups" is unclear; you must state clearly -diversity of what? Diversity in the process of dominance establishment? In the relationships? 3 -Lines 169-171: "...with a lower α-value (Fig. 2, both dark gray in "a" and red line in "b") represent a more complex fractal dynamic, a higher fractal dimensions and a "rougher" appearance than the time series with a higher α-value (Fig. 2, blue line in "b") which can be associated with faster switching between ambulation and immobility in the time series" -this part needs rephrasing because, as it is, the final part of the sentence seems to refer to the "higher α-value". Besides, I suggest changing the term "rougher" for a more technical one; 4 -Legends of figures 1 and 4: I suggest removing the explanation of the technique used to define dominants, subordinates and neutrals. On the other hand, information on the circumstance data was collected (after 48 hours of group-forming) is missing in these and other legends; 5 -Try to reduce the number of figures. I suggest removing figures 5 and 8; 6 -Lines 253-255: "The social groups with a dominant individual showed a significantly higher (P = 0.03) mean interindividual distance between group members than the groups with only neutral members" -This result was not discussed in the Discussion section; 7 -Legend of Figure 6 -Review for clarity; 8 -Lines 382-385: "Considering that in our experimental setup food and water resources were provided ad libitum, no actual immediate benefit is obtained through the establishment of dominance, thus a neutral group conformation would be energetically favorable for all members." This argument is not valid, since there were groups, kept under the same conditions, where hierarchy was established. 9 -Some paragraphs of the manuscript are too long; split them into two or three for clarity (e.g., lines 355-399; 400-450); 10 -Lines 510-511: State clearly how many birds made part of the experiment -how many dominants, how many subordinates, how many neutrals. This is important information for reproducing the experiment; 11 - Table 2 -Please, align text in the left-hand column cells with the first line of the corresponding cell in the right-hand column; 12 -Supplementary Tables 1 and 2 -Please, revise the Table legends; they inform the recorded interactions took place "between each female" (1) or "between each male" (2), but both Tables include data from interactions with males and females. Besides, the information included below the Tables need revising as well ("...pecking towards and time receiving pecks"; make it cleartowards who?); 13 -Supplementary Figure 1 -The legend and the title of the X axis contains the word "peaks", instead of "pecks"; 14 -Supplementary Figures 3 and 4 -The figures have several parts, and no identification was made of each part; 15 -Supplementary material -Lines 129-130 -The last sentence of the paragraph is incomplete.
Reviewer #3 (Remarks to the Author): The study entitled "Aggressive dominance can decrease behavioral complexity on subordinates through synchronization of locomotor activities" analyze, under the perspective of the influence of the social environment on individual behavior from the perspective of a complex system in birds, the temporal dynamics of behavior in small social groups of birds exhibiting divergent characteristics. The study evaluate the temporal dynamics of the behavior and spatial use of the individual within their social environment, by using mathematical tools from the field of study of complex systems: scale-based analysis and DFA. This approach permit to obtain a precise quantification of patterns that are often limited to average durations of a behavior or the use of traditional patterns, which are often limited to average durations of a behavior. By continuously tracking each bird authors were able to quantitatively assess individual behavioral dynamics, and monitor possible behavioral synchronization between animals. Authors hypothesized that the presence of a dominant bird has an effect not only on the type of behaviors displayed by its subordinate but also on the temporal organization of behavior. The study address questions about the influence of the social environment on the individual behavior, highlighting if there a biological rhythms and if this rhythms emerge from social dynamics. These questions are aimed at deepening knowledge of social dynamics using powerful tools of analysis, and study the problem of aggression in farm birds that threatens animal welfare. It is therefore a basic study of social behavior using very innovative techniques with results that, although expected, can give arguments to respond to the management challenges of social groups in captivity and their impact on the final efficiency of the system. Therefore, I consider that this work should be published with the certainty that it will be very useful both for the applied ethology research groups and for the analysts and managers of the captive bird production systems. Line 68. Please cite the following article in this context: Marıa, G. A., Escós, J., & Alados, C. L. (2004). Complexity of behavioural sequences and their relation to stress conditions in chickens (Gallus gallus domesticus): a non-invasive technique to evaluate animal welfare. Applied Animal Behaviour Science, 86(1-2), 93-104. Line 68 to line 82. Why is it avoided the most representative social groups of social production groups (ie laying hens, free range chicken) where there are groups only of females or only of males that although it is not a "natural" social group is the most frequent in management of these production systems? Line 103. Temporal organization of the group? Line 134. Endpoint criteria provided in cases of extreme aggression? Line 140-142 The definition of the group called neutral is not clear, it seems that these animals can eventually move to well-defined categories as dominant and subordinate? Line 355-365 How do you think that subordinate birds can be affected their productive efficiency? ... Was growth and conversion of food measured? ... Is it likely that these birds that spend more time walking develop a strategy to obtain resources with an avoidant personality of the dominant ones and which are finally more efficient? ... Can the energy cost of the dominant ones be higher than the energy cost of developing a strategy of avoidance and a better use of resources? Line 366-374 Maybe it is more appropriate to talk about different "personalities" or adaptive strategies or coping styles of individuals and maybe it is good to review the literature in this regard, I recommend citing the following work: Miranda-de la Lama, G. C., Pascual-Alonso, M., Aguayo-Ulloa, L., Sepúlveda, W. S., Villarroel, M., & María, G. A. (2019). Social personality in sheep: Can social strategies predict individual differences in cognitive abilities, morphology features and reproductive success?. Journal of Veterinary Behavior. Page 18. Density and lighting are two of the aspects that affect the social groups of bird production. In your case, the density is not inconvenient and the lighting follows a natural regime. Taking into account the high densities (ie broiler) used in production and the 24-hour light rhythm during the first and last stage of fattening, the results obtained here can be useful in intensive bird production systems. ? Line 479. Explain the "broad implication for farm animal welfare"… Line 485. Give more information about the feeding regime of the animals and performance traits (consumption, efficiency, diet etc.) Density and lighting are two of the aspects that affect the social groups of bird production. In your case, the density is not inconvenient and the lighting follows a natural regime. Taking into account the high densities (ie broiler) used in production and the 24-hour light rhythm during the first and last stage of fattening, the results obtained here can be useful in intensive bird production systems. ?

Reviewers' comments:
Reviewer #1 (Remarks to the Author): The manuscript presents an investigation of how social interactions of individuals who are members of a social group affect the behavior of individual members of the group, depending on their social position compared to other group members -dominant, neutral or subordinate. The authors study small social groups of Japanese quail, and rise a novel hypothesis that social interactions among group members would be reflected in the various types of behavior displayed by the individuals as well as in the temporal patterns and organization embedded in their ambulatory activity. Modern concepts and approaches derived from fractal theory and stochastic process are utilized to provide quantitative assessment of the behavioral dynamics of individual group members, and to identify differences in dynamic characteristics based on the role and status of individuals in the social group. Further, through sophisticated wavelet-based analysis, the authors provide first evidence of coordinatedbehavior and synchronized ambulation patterns between group individuals with different social roles. Remarkably, the authors demonstrate how behavioral types and dynamical patterns involve in time as a function of increased familiarity among individuals. The validity of the reported observations is supported by several complimentary methods of analysis, measures of complexity and statistical tests. The manuscript is well written and presents an elegant study of how social interactions impact behavioral dynamics of the individual. I recommend publication.

Re: Thank you very much for your positive comments, detailed analysis of our work and constructive comments. We have addressed all the points mentioned, incorporated all recommended papers as references and modified Figure 2 in order to include the ambulation time series and the corresponding DFA scaling function F(n).
There are several technical points related to the details of analysis and interpretation of results that the authors should address before publication: 1. On page 4, line 68, the authors discuss temporal organization, long-term correlations and fractal dynamics and self-similarity in locomotion. This should be put in context of earlier works which have pioneered these concepts and demonstrated scale-invariant organization in animal and human locomotion - Re: We agree that these are very important papers and have now cited them accordingly.
2. On page 6, line 113-115, the authors connect the value of the self-similarity parameter α with higher level of fluctuation between ambulation and immobility events, and more complex dynamics with lower level of long-range correlations (lower α-values). This is not quite accurate. The DFA fluctuation function F(n) depends on the size of fluctuations (standard deviation of the ambulation time series), however, the scaling exponent α does not depend on the size of fluctuations. Also, the α value reflects only the strength of long-range power-law correlations (i.e. degree of persistent behavior), and does not impart on the complexity of the dynamics -the DFA α is a linear measure of self-similarity (fractality) in the temporal organization of a time series. The authors should correct this in their interpretation.
Similar statements are made on page 9, line 169-171, and also on page 17, line 344-345, which should also be revised. Lower α value does not mean more complex fractal dynamics, but simply transition from long-range correlated scale-invariant behavior to a trivial uncorrelated (white noise) scale-invariance for α=0.5.
Re: These statements have now been rewritten to improve accuracy on this very important concept throughout the manuscript. In particular, the phrase on page 9 now reads: In this context, the higher α-value (closer to 1) indicates that activity fluctuations are characterized by strong long-range positive correlations, and thus are not dominated by random factors 28 . In contrast, the lower α-value found in dominant and neutral birds (Fig. 2, dark gray in "a" and red line in "b") indicate a more uncorrelated scale-invariance (more random activity fluctuations). In addition, these time series with lower α-value, have a higher fractal dimensions 37 than the time series with a higher α-value (Fig. 2, blue line in "b"). Table 1, page 6,pos one can notice that although there is large increase in eating time in both female and male in day 3, the ratio of time spent for eating between female and male during day 1 is similar to this ratio for day 3. Same can be observed for the ratio of foraging time between female and malesimilar ratios for both day 1 and 3. Would that mean that there is certain constancy in social interaction between female and male independent of familiarity?

In
Re: Yes, seemingly females eat and forage approximately twice as much as males both on day 1 and 3. It is an interesting point, and we speculate that it is most likely associated with the high energetic requirements of egg production in females. It is important to note that in quail, as in other poultry species, females place roughly one egg per day along several months (Mills et al.,1997, Neurosci Behav 21: 261-281). Also, females are larger (heavier) (Mills et al.,1997, Neurosci Behav 21: 261-281) than males in this species. Unfortunately, our experimental setup does not allow us to accurately estimate individual feed consumption nor feed conversion in order to further explore this assumption. Nevertheless, a sentence informing the readers about quail male/female characteristic is now added in the results section as follows: Also, in general, females spent significantly more time performing pecks, and foraging, and tended to spend more time eating than males (Table 1). These last two variables could reflect higher energy requirements needed in females for daily egg laying and larger size in comparison to males 36 .  Table 1) or d) a group with a dominant (red) and subordinate (blue) female (box 2, see Supplementary Table 1). Notice the similar pattern of activity and inactivity between dominant and subordinate birds. Grey boxes in panels b and d also represent the region of time series amplified in the inset. c,e) Detrended Fluctuation Analysis (DFA) of the locomotor time series, corresponding to the same time series shown integrated in "b" and "d" respectively. Fluctuation functions were offset by 0.5 in order to improve visibility. Lines show the actual fitting region used (n, scales between 7.5 and 322 s) and the numbers represent the selfsimilarity parameter (α-value) obtained for each of the locomotor time series. Note that the time series with the lowest α-value such as those in panel b, present high level of switching between immobility and mobility events, thus shorter events (see also Fig. 5) as can be observed in insets. Higher α-value (blue line in panel e) shows longer periods of continuous immobility or ambulation. 5. On page8, figure 1, line 161, for birds that are neither dominant nor subordinate: perhaps the inequality should be with inversed sign "4.9 > aggressive score > -4.9"?
Re: This typo was corrected.

Re: These typos have been corrected.
7. Data were recorded and analyzed for one hour between 9-10 am. Would the results differ if same analysis is performed in a different hour for the day, having in mind circadian effects?
Re: They definitely could be different. We have previously documented both circadian as well as ultradian rhythms in locomotion (Guzman et al., 2017, Sci Rep. 7(1):684). Moreover, the time of day was specifically selected given that in the morning both locomotor and reproductive activity is maximum (see Delville et al., 1986, Horm Behav 20: 13-22  These potential effects on the results should be discussed with reference to prior works.

Re: These papers are fundamental and were taken into account in the analysis process. Citations have now added to the Materials and Methods accordingly, as follows (changes are underlined).
The presence of nonstationarities in the signal 72 , such as those associated with polynomial and sinusoidal trends 35 , as well as the coarse-graining method 73 used to obtain the locomotor time series, can lead to crossovers in the scaling curve. Thus, the potential presence of crossovers were systematically studied for detrending orders 1 through 5 71 . A DFA of third order (DFA-3) was the lowest detrending order that eliminated trends in all series and therefore it was applied to all series for estimating α.
In addition, the optimal range of scales n 35 between 7.5 and 322 s (see linear fits in in Fig. 2c,e), was determined using the following criteria: stable values of local slopes, maximum coefficient of variation and minimum sum of squared residuals 24,70,74 . Once again, we really appreciate the comments received and if the Reviewer and or Editor think that further changes should be made, please do not hesitate to contact us again.

Reviewer #2 (Remarks to the Author):
This study investigated the effects of social environment on birds' behavioural complexity. The authors managed to quantitatively evaluate changes in the animals' behavioural repertoire and, particularly, in the temporal dynamics of their behaviour as hierarchy established. They found out that subordinates performed less complex temporal dynamics of locomotion, and that their locomotor pattern was synchronised with locomotion pattern of dominants. These are, as far as I know, novel information, which are relevant not only as basic research -e.g., studies on social behaviour, but also from an applied perspective -from the point of view of animal welfare.
The manuscript was well-written and, from my point of view, has no great flaws that could prevent it from being published. Regarding data analysis, the choice of GLMM was appropriate; however, evaluation of the use of the "Detrended Fluctuation" and "Wavelet" analysis is outside of the scope of my expertise. The authors' interpretation and conclusions are also adequate.
Re: We would like to thank the reviewer for the positive feedback. We greatly appreciate the remarks aimed at improving the organization and clarity of our work. We have addressed all comments in this revised version. If the Reviewer and /or Editor think that further changes need to be done, please do not hesitate to let us know.
Following I list some points in which the manuscript deserves improvement, mostly for clarity: 1 - Table 1: the time unity is missing (I assumed the unity is "seconds", but this must be explicit in the legend); Re: The following statement has been added to the legend: "The time unit seconds (s) was used for all behavioral variables, except ambulation which is expressed as a percentage of time (%)". 2 -Line 133: "Diversity in the establishment of dominance relationships within social groups" is unclear; you must state clearly -diversity of what? Diversity in the process of dominance establishment? In the relationships?
Re: In order to improve clarity, the subtitle was modified to "Variability in social group dynamics".
3 -Lines 169-171: "...with a lower α-value (Fig. 2, both dark gray in "a" and red line in "b") represent a more complex fractal dynamic, a higher fractal dimensions and a "rougher" appearance than the time series with a higher αvalue (Fig. 2, blue line in "b") which can be associated with faster switching between ambulation and immobility in the time series" -this part needs rephrasing because, as it is, the final part of the sentence seems to refer to the "higher α-value". Besides, I suggest changing the term "rougher" for a more technical one; Re: These sentences have been completely rewritten in order to improve accuracy, and also considering reviewer #1 comments, it now reads: In contrast, the lower α-value found in dominant and neutral birds (Fig. 2, dark gray in "a" and red line in "b") indicate a more uncorrelated scale-invariance (more random activity fluctuations). In addition, these time series with lower α-value, have a higher fractal dimensions 37 than the time series with a higher α-value (Fig. 2, blue line in "b").
4 -Legends of figures 1 and 4: I suggest removing the explanation of the technique used to define dominants, subordinates and neutrals. On the other hand, information on the circumstance data was collected (after 48 hours of group-forming) is missing in these and other legends; Re: Figure legends 1 and 4 were rewritten accordingly (changes underlined), in the rest of the figures in the titles the phrase "48 hours after being placed in the novel social environment" was added. Fig. 1. (…) b-f) Schematic representation of the direction of the aggressive social interactions (dark red arrow) present between individuals in the triad of 2 females (grey background) and 1 male (white background) 48 hours after being placed in the novel social environment. Red indicates dominant birds (aggressive score ≥ 4.9), blue, subordinate birds (aggressive score ≤ -4.9) that receive pecks from dominant birds, and black those birds that are neither dominant nor subordinate (4.9 > aggressive score > -4.9). Panel "f" represents neutral groups (i.e. the group did not include any dominant member).  Figure 8 in the main text, given it shows the actual mean group values estimated (a highly relevant information). However, we do not want to preclude publication because of this required change and we are willing to move it to the Supplementary material if the Reviewer and/or Editor feels strongly that is necessary.
6 -Lines 253-255: "The social groups with a dominant individual showed a significantly higher (P = 0.03) mean interindividual distance between group members than the groups with only neutral members" -This result was not discussed in the Discussion section;

Re: A discussion about this result is now included as follows:
Consistently, social proximity, which is commonly associated with underlying sociality (motivation to be near conspecifics) and social cohesion 39 , was also different in both group types. Groups with a dominant individual show lower social proximity (larger interindividual distance), than the groups with only neutral members, which is consistent with shyness and social withdrawal of subordinates. 8 -Lines 382-385: "Considering that in our experimental setup food and water resources were provided ad libitum, no actual immediate benefit is obtained through the establishment of dominance, thus a neutral group conformation would be energetically favorable for all members." This argument is not valid, since there were groups, kept under the same conditions, where hierarchy was established.
Re: The phrase was rewritten to improve accuracy and the statement about "Considering that in our experimental setup food and water resources were provided ad libitum, a neutral group conformation would be energetically favorable for all members." 9 -Some paragraphs of the manuscript are too long; split them into two or three for clarity (e.g., lines 355-399; 400-450); Re: Each of these two paragraphs have now been broken down into 3 paragraphs (please see main manuscript).
10 -Lines 510-511: State clearly how many birds made part of the experimenthow many dominants, how many subordinates, how many neutrals. This is important information for reproducing the experiment; Re: All detailed information about how many birds were dominant, subordinate or neutrals is now included in the Methods Section "Animals and husbandry: Table 2 -Please, align text in the left-hand column cells with the first line of the corresponding cell in the right-hand column;

-
Re: Alignment has now been changed. Tables 1 and 2 -Please, revise the Table legends; they inform the recorded interactions took place "between each female" (1) or "between each male" (2), but both Tables include data from interactions with males and females. Besides, the information included below the Tables need revising as well ("...pecking towards and time receiving pecks"; make it cleartowards who?);

-Supplementary
Re: 13 - Supplementary Figure 1 -The legend and the title of the X axis contains the word "peaks", instead of "pecks"; Re: Typo has been corrected.
14 - Supplementary Figures 3 and 4 -The figures have several parts, and no identification was made of each part; Re: These legends have been rewritten to clarify this aspect: Re: The sentence has been completed as follows: If during the interaction a quail received more than 5 consecutive aggressive pecks, showed a clear and continued escaping (retrieval) behavior, and/or showed any sign of physical injuries testing was finalized immediately.

Reviewer #3 (Remarks to the Author):
The study entitled "Aggressive dominance can decrease behavioral complexity on subordinates through synchronization of locomotor activities" analyze, under the perspective of the influence of the social environment on individual behavior from the perspective of a complex system in birds, the temporal dynamics of behavior in small social groups of birds exhibiting divergent characteristics. The study evaluate the temporal dynamics of the behavior and spatial use of the individual within their social environment, by using mathematical tools from the field of study of complex systems: scale-based analysis and DFA. This approach permits to obtain a precise quantification of patterns that are often limited to average durations of a behavior or the use of traditional patterns, which are often limited to average durations of a behavior. By continuously tracking each bird authors were able to quantitatively assess individual behavioral dynamics, and monitor possible behavioralsynchronization between animals. Authors hypothesized that the presence of a dominant bird has an effect not only on the type of behaviors displayed by its subordinate but also on the temporal organization of behavior. The study address questions about the influence of the social environment on the individual behavior, highlighting if there a biological rhythms and if this rhythms emerge from social dynamics. These questions are aimed at deepening knowledge of social dynamics using powerful tools of analysis, and study the problem of aggression in farm birds that threatens animal welfare. It is therefore a basic study of social behavior using very innovative techniques with results that, although expected, can give arguments to respond to the management challenges of social groups in captivity and their impact on the final efficiency of the system. Therefore, I consider that this work should be published with the certainty that it will be very useful both for the applied ethology research groups and for the analysts and managers of the captive bird production systems.
Re: Thank you very much for the detailed analysis of our work and the positive comments. We have now addressed all the points mentioned. Once again, if the Reviewer and /or Editor think that further changes need to be done, please do not hesitate to let us know.
Re: Yes, it is indeed a very relevant paper in the field that has now been added to our reference list.
Line 68 to line 82. Why is it avoided the most representative social groups of social production groups (ie laying hens, free range chicken) where there are groups only of females or only of males that although it is not a "natural" social group is the most frequent in management of these production systems?