High-resolution behavioral time series of Japanese quail within their social environment

The behavioral dynamics within a social group not only could depend on individual traits and social-experience of each member, but more importantly, emerges from inter-individual interactions over time. Herein, we first present a dataset, as well as the corresponding original video recordings, of the results of 4 behavioral tests associated with fear and aggressive response performed on 106 Japanese quail. In a second stage, birds were housed with conspecifics that performed similarly in the behavioral tests in groups of 2 females and 1 male. By continuously monitoring each bird in these small social groups, we obtained time series of social and reproductive behavior, and high-resolution locomotor time series. This approach provides the opportunity to perform precise quantification of the temporal dynamics of behavior at an individual level within different social scenarios including when an individual showing continued aggressive behaviors is present. These unique datasets and videos are publicly available in Figshare and can be used in further analysis, or for comparison with existing or future data sets or mathematical models across different taxa.

The behavioral dynamics within a social group not only could depend on individual traits and socialexperience of each member, but more importantly, emerges from inter-individual interactions over time. Herein, we first present a dataset, as well as the corresponding original video recordings, of the results of 4 behavioral tests associated with fear and aggressive response performed on 106 Japanese quail. In a second stage, birds were housed with conspecifics that performed similarly in the behavioral tests in groups of 2 females and 1 male. By continuously monitoring each bird in these small social groups, we obtained time series of social and reproductive behavior, and high-resolution locomotor time series. This approach provides the opportunity to perform precise quantification of the temporal dynamics of behavior at an individual level within different social scenarios including when an individual showing continued aggressive behaviors is present. These unique datasets and videos are publicly available in Figshare and can be used in further analysis, or for comparison with existing or future data sets or mathematical models across different taxa.

Background & Summary
The behavioral dynamics within a social group depends on many factors (i.e. individual traits, prior social-experience of its members, environmental context) and emerges from the interactions between its members over time. For example, individuals considered to be more aggressive toward conspecifics, may show more dominant behavior, and be more explorative, bold and active 1,2 . In farm animals it is well known that animals can be selected based on certain behavioral traits and this impacts, at least on a population level, on the overall behavioral performance in a wide variety of contexts. For example, quail selected by their high andrenocortical response to restraint, are more fearful in a wide variety of tests 3 but also more aggressive in social groups, in comparison with those with low responsiveness 4 . Also, quail selected as chicks as highly sociable, are less fearful and less aggressive in social groups as juveniles than less sociable birds. Thus, the use of selection tests can favor a higher proportion of individuals with a desired characteristic 5 .
Although individual traits such as fearfulness and aggressiveness can affect the outcome of social interactions and the establishment of dominance hierarchies [6][7][8] , other factors such as previous fighting experience 9 , group size [10][11][12] and housing conditions (i.e. size of box or cage 10,13 , presence or not of enrichment 14,15 ) as well as the dynamical interaction between them 16 can determine the social dynamics of a group. For instance, in large groups it has been proposed that tolerant social dynamics, that does not require individual recognition per se, emerges as the predominate social strategy 10,11 , while hierarchy formation is predominant in small groups 12 . Housing that is relatively small in relation to group size, can lead to high stocking density hence an increase of frequency of agonistic acts 10,13 .
In poultry, like other birds, within small groups hierarchies are established through a peck-order, according to which the animal that rank highest pecks at conspecifics and it is not pecked in return, and the opposite happens to the animals at the extreme bottom of the rank 17,18 . Hence, when two or more unacquainted adult birds are brought together, fights and pecks usually occur until each bird has established a dominance-subordination relationship with each other 12 . Thus, dominance is an emergent property that springs from the interaction of at least www.nature.com/scientificdata www.nature.com/scientificdata/ the Social interaction test, see below) that were submitted to a short photoperiod light cycle (06 h light: 18 h dark) beginning at 4 weeks of age until testing ended 58 .
At 28 days of age, test animals were sexed by plumage coloration, marked with a numbered wing band and randomly housed in pairs of 1 male and 1 female in cages of 20 × 40 × 20 cm (width × length × height respectively).
If an animal showed any indication of illness or escaped from their cage during rearing, they and their companion cagemate were completed excluded from the experiment. preselection of quail. One-hundred six quail were first evaluated in 4 preselection tests, separated between each other roughly by 30 days in order to favor independence between tests. These tests were used as a preselection criterion for social group testing. All data registered during the preselection tests are available in file "PreselectionTestsQuail.xls" and stored in the public repository Figshare 52 . Original video recordings are also available 53 . A schematically representation of the experimental design, tests and variables registered is shown in Fig. 1. In each test the order of testing of cages was randomized avoiding evaluation of adjacent cages consecutively. Both birds of the cage were always evaluated simultaneously. Moreover, the experimenters were always blinded regarding the prior history of the animals allocated in each group. With the exception of the tonic immobility test, all tests were recorded onto a computer and video recordings were analyzed the following days after testing by one previously trained experimenter.
Partial mechanical restraint. This test has been proposed as a method to measure fear in quail 6,59,60 . Moreover, in juveniles, subsedative 57 anxiolytic doses of Propofol have shown to reduce struggling (see below) to durations bellow 60 s (See pilot study results in Fig. 2). This test was performed at 40 days of age. This test consists in restricting the movement of the animal between two walls of a melamine box of 20 × 10 cm (height per width, respectively) with the characteristic that the front wall was made of glass (it allows the visualization of the animal and video recording of its behavior), and the back wall was adjustable to induce immobility in such a way that the animal cannot open the wings, but can move the head and legs 9 . The experimenters retreated out of the birds' sight, and the test was during 5 minutes recorded with a video camera place in front of the box. All the birds were tested in 31 batches of 4 animals each, where the birds had no visual or physical contact between each other. The video was analyzed manually, and the following variables were recorded: the latency to struggle considered as time between the initiation of restraint until the first struggling episode (defined as the birds making fast movement with their legs when aiming to escape from the test apparatus) and the number of struggles events during the observation period. The struggle events were considered different if they were separated by 5 s or more. The immobility of the animal during the test has been widely considered in the literature as an indicator of fear 6,59,60 . Struggling during such restraint is known to be more pronounced in genetic lines of quail showing low rather than high levels of underlying fearfulness 6,40 . Those whose latency of struggle was >60 s were considered as fearful (Fig. 2) 61 . Caliva et al. (2017) and measures levels of aggressiveness towards a non-aggressive photocastrated male opponent in a novel environment 58   www.nature.com/scientificdata www.nature.com/scientificdata/ allowed constant monitoring and recording during the test while out of the sight of the birds. Using commercially available behavioral tracking software (ANY-maze ™ , 2015) the number of events (i.e. continuous time performing the behavior separated from the following event by at least 5 s), duration of behavior (i.e. seconds performing behavior) and latency to initiate behavior (i.e. time from the start of the test until bird shows the first event) of the following aggressive behaviors were recorded:

Social interaction (SI) test. This test is described in detail in
Pecks: when one bird raises its head and vigorously pecks the other bird's body (usually on the head). Grabs: when a male catches ("grabs") with their beak the neck or head region of the female. Mounts: while performing a grab, the male approaches a female from behind, and places both feet on the dorsal surface of its torso, stepping over the females' tail (adapted from 62 ).
Cloacal contacts: during mounting, the male lifts his tail and tilts his pelvis underneath the other bird and briefly presses its cloaca against the female (adapted from 62 ).
Threats: one bird stands with its neck and head raised in front of the other bird that usually has its' head at a lower level than the first (adapted from 62 ).
Chase: a bird runs after another that is escaping (adapted from 63 ). Herein, when grabs, mounts or cloacal contacts were performed by one male towards another male, they were considered as aggressive behaviors 64 . Birds that performed more than 5 aggressiveness behaviors were considered aggressive, and birds that did not perform any aggressive behavior towards the photocastrated opponent were considered non-aggressive 58 . 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 damage, the interaction was immediately interrupted 65 . Caliva et al. (2017) showed that in the SI test only 8% of the photostimulated females showed clear signs of aggression towards photocastrated opponent. The authors proposed that this is most likely due to the short duration of the test (up to 5 min), and that longer test durations are needed in females to observe significant aggressiveness (i.e. 3 h tests are performed in hens to establish dominance 16,48,66 ). Thus, expression of aggressive behavior in females was very low.
Tonic immobility (TI). The TI test was performed at 100 days of age. According to Jones 67 this test induces an unlearned antipredator response that is triggered by a brief period of physical restraint. In the test the individual was place in the left lateral decubitus and hold for 15 seconds (the necessary time required to unleash the muscular immobilized tonic behavior), holding him with both hands against a support base (one hand on the head and another in the body). We recorded the number of inductions to achieve an immobility of at least 10 seconds and the duration of TI once induced. Maximum duration of TI was fixed at 5 minutes. Duration of TI implies both a behavioral and physiological response modulated by frightening situations and is considered as a measure of the level of fearfulness 68,69 . Thus, a long duration of TI and a smaller number of necessary inductions is indicative of a high level of fear as opposed to a short response 67 . If IT was not attained after 5 successive attempts, the bird was considered not to be susceptible and scores of 0 were given for TI duration. Thus "non-fearful" birds were selected based on those that needed 4 or more inductions, while the birds considered "fearful" required a single induction in the test.
Welfare assessment. At 96 and 108 days of age female skin lesions and plumage status were evaluated following a procedure proposed by Pellegrini et al. 46 that is an adapted version of the protocol proposed by Welfare Quality ® consortium 70 . Pellegrini et al. showed in Japanese quail that male aggressions toward a female cagemate can predict aggressiveness toward unknown conspecifics 46 . Head, neck and back skin lesions were determined using a score scale from 0 to 2 where "0" represents no lesions (punctiform damage <0.25 cm diameter) or scratches, www.nature.com/scientificdata www.nature.com/scientificdata/ "1" represents less than 3 lesion or scratches, and "2" reflects 3 or more lesion or scratches. Head, neck and back plumage damage was also determined using a score scale from 0 to 2 as follows: "0" represents individuals with no plumage damage or slight wear (only single feathers lacking), "1" represent individuals with one or more body parts that have moderate wear (i.e. damaged feathers worn or deformed) or one or more featherless areas <1.5 cm in diameter at the larger extent and "2" corresponded to individuals that have at least one featherless area >1.5 cm in diameter at the largest extent. None of the males showed plumage damage or lesions, thus, only male aggression towards females were considered in this analysis. Plumage damage induced by males (score > 0) to their female cagemates were considered as indicative of male aggressiveness 46 . Non-aggressive males were those in which no plumage damage was seen in female cagemates 46 . It should be noted that if at any point in the study a bird showed severe lesions they were separated from their cagemates and thus both cagemates were eliminated from the study in order to protect the welfare of the animals. Due to this systematic standard laboratory procedure, very aggressive birds were excluded from the study even at a young age. performance assessment. Birds were weighed at 28 days of age. The weight of birds transferred to cages ranged between 100-150 g. Thereafter, weight and male cloacal gland width and length, and male foam production were recorded weekly until 9 weeks of age, when all males showed completed gonadal development (Cloacal gonadal volume CGV > 1000 mm 3 ). Cloacal gland volume was estimated as (4/3 × 3.5414 × a × b2), where a = 0.5 × length, and b = 0.5 × width 71 . Foam production was quantified by subjective scaling of the amount of foam ejected upon manual expression (squeezing) of the foam gland, using a scale of 1 (no foam expressed) to 5 (maximum amount of foam expressed). Female quail egg production was monitored throughout the study and all females reached peak egg production. All birds were also weighed after the last behavioral test at 92 days of age, and male cloacal gland size and foam production also assessed. This data is available in file a b c d e f  Fig. 4, and illustrates the relationship between variables, and selection criteria. Statistical independence (R2 = 0.04 and ~90° angles in the PCA biplot (lines in Fig. 4)) was observed between variables of two different tests, namely latency to struggle during partial mechanical restraint and number of inductions for tonic immobility. However, 74% of females that were fearful in the tonic immobility test (i.e. only needed one induction) also showed low level of struggling (i.e. ≤3 struggling bouts), in comparison to 34% (P < 0.05, 2-tailed proportion test) that showed low struggling and were found to be less fearful (i.e. tonic immobility was not induced or only after 5 inductions). This was not evident in males (65% and 68% fearful and non-fearful during tonic immobility, respectively). In all, these results show that, at least in females, highly fearful birds during tonic immobility on average were also more fearful during restraint.
Considering results obtained in preselection test birds were classified in 2 types: 1) Type A (Fig. 4 black circles) had birds that were fearful in both the Tonic immobility and the partial mechanical restraint test. These males also were aggressive in the Social interaction test or their female cagemate showed higher scores during Welfare assessment. 2) Type B (Fig. 4 gray circles) had females and males that were not fearful in the Tonic immobility test, and males that tested non-aggressive in both the Social interaction test and in their home cage. No differences between body weight, cloacal gland volume or foam production was observed between types. It should be noted that there are males that were not used for the social group tests (Fig. 4 open circles) because they were used in a separate neurobiology experiment. Social group testing. Novel social groups (2 females: 1 male) of animals (156-171 days old) that behaved coherently in the 4 preselection tests described in the previous section, thus half of the 12 social groups evaluated had type A birds and the other half type B birds (see previous subsection). Social groups were housed in a white wooden apparatus measuring 80 × 40 × 40 cm (width × length × height, respectively) with wood-shavings on the floor. A feeder and an automatic nipple drinker were positioned in opposite corners of the apparatus (Fig. 1, left and right bottom corner of box in the photograph, respectively). Nylon monofilament line was extended over the top of the boxes with a 1 cm separation in order to prevent the birds from escaping without interfering with their visualization. A video camera was suspended 1.5 m above the box. Since only 4 social groups could be tested simultaneously, the setup was repeated 3 consecutive times. For convenience, boxes in which each social group were placed were numbered from 1 to 12. Boxes 1-4 were tested simultaneously first, 5-8 second and 9-12 last. Video recordings are publicly available on Figshare 53 .
We used IdTracker 22 in MATLAB R2017a to register x, y coordinates of each animal within the social group during a 1 h period immediately after being placed in the test apparatus between 9 and 10 am, and 48 hours after testing began. Locomotion was than estimated at 0.5 s intervals (7200 time intervals) using customized code www.nature.com/scientificdata www.nature.com/scientificdata/ Locomotion.m that calculates the distance moved by the animal, converts distance expressed in pixel to centimeters, and if the distance moved is above a threshold of 1 cm that the animal is considered mobile 72 . Thus, the recorded behavioral data is expressed in the form of a time series of mutually exclusive states. At any given time, if the bird was moving a number one was recorded or a zero if immobile. These locomotor time series are publicly available and stored in the public repository Figshare 51 .
Time series of non-locomotor behaviors were obtained through visual observation of video recordings using as an interface ANY-MAZE@ to register behavior. For each bird, when the specific behavior was performed the corresponding key was pressed until the bird finished performing the behavior, thus a binary time series, x i, sampled at up to 2 data points per second was constructed for each behavior.
x 0(not performing the behavior) 1(performing behavior) i Only one observer recorded all data in order to avoid inter-individual variability. Prior to video analysis observer performed training sessions than consisted in analyzing the same three behavioral video at least two times. Then, reliability was estimated to be >95%. (formula: number of agreements/number of agreements + number of disagreements).
The following behaviors were recorded: Pecks, Grabs, Mounts, Cloacal contacts, Threats, Chase, as described previously for the SI test, and additionally, Foraging: pecking at the ground or actively moving litter with beak, Feeding: peaking at food in the feeding trough, and Dust bathing: vertical wing shakes in a lying position 73 . From the behavioral time series both frequency and durations of behaviors can be easily estimated. These behavioral time series are also publicly available and stored in the public repository Figshare 50 .

Data Records
Original video recordings 53 of Partial Mechanical Restraint, Social Interaction test and Social groups are provided in avi or mod format. File names include the abbreviation, PMR, SI or Box, respectively. For video file of the Partial Mechanical Restraint and Social Interaction tests file names also include the ID of the animals tested. In the case of social groups box number (1-12) and the day of testing (day1 or day3) is provided in the file name.

CGV (mm 3 )
Cloacal gland volume estimated as (4/3 × 3.5414 × a × b) 71 , where a = 0.5 × length, and b = 0.5 × width of cloacal gland Foam product. Subjective scaling of the amount of foam ejected upon manual expression of the foam gland, using a scale of 1 (no foam expressed) to 5 (maximum).
Partial mechanical restraint Lat. struggle (s) Time in seconds between the initiation of restraint until the first struggling episode. If struggling was not observed, 300 was recorded.
N. of struggles Number of struggles during the observation period.
Social interaction test N. of pecks Number of events when one bird raises its head and vigorously pecks the other bird's body.
Dur. of pecks (s) Seconds spent pecking at opponent.
N. of grabs Number of events when a bird catches ("grabs") with their beak the neck or head region of the other bird.
Dur. of grabs (s) Seconds spent performing grabs towards opponent.
N. of mounts Number of events while performing a grab, the bird approaches the other bird from behind, and places both feet on the dorsal surface of its torso, stepping over the other birds' tail.
Dur. of mounts (s) Seconds spent performing mounts towards opponent.
N. of C.C. Number of events during mounting, the bird lifts his tail and tilts his pelvis underneath the other bird and briefly presses its cloaca against the other bird.
Dur. of C. C. (s) Seconds spent performing cloacal contacts.
N. of threats Number of events when one of the birds raises its head and neck rapidly, moves forward and backward vigorously in the direction of the opponent without making physical contact. www.nature.com/scientificdata www.nature.com/scientificdata/ The results of the preselection tests are presented in the excel file "PreselectionTestsQuail.xls" stored in the public repository Figshare 52 with the headers of the columns representing the variable analyzed in each test for each animal (rows). Table 1 presents all the column headers as well as a brief definition of the variable.
All time series from this study are stored in Figshare as text files (.txt). For practical purposes, locomotor data obtained from IdTracker are in separate files 51 from the behavioral data time series obtained from AnyMaze 52 . Locomotor data consists of a single column of data sampled at 0.5 s intervals (as explained previously). For the behavioral data the first column refers to the time while the following columns refer to the behavioral data. Considering that the 36 animals were evaluated in 12 mixed-sex groups of 3 birds in individual experimental boxes on the first hour (Day1) and 48 hours (Day3) after test initiation, each subject quail was identified by their experimental group number (Box), ID number of wing band, and sex (femaleA, femaleB or male). In the case of females, an indication A or B is used to discriminate between the two. In the file name, an indication of the corresponding bird is also provided as "BoxN°_IDN°_sex_DayN°" ( Table 2) for recorded obtained.

Technical Validation
All data analysis and technical validation was performed by one observer both in Any-Maze as well as in IdTracker. In both cases the observer was blinded regarding the prior history of the animals allocated in each group. One of the advantages of IdTracker is that the researcher can perform visual observation of the tracking performed on each frame analyzed using the complementary software IdPlayer. A number (see Fig. 1) in the center of the of the animal a number indicates the identity of the bird. In the case of identification errors, they Table 2. Overview of the data files uploaded to Figshare 50,51 grouped in file sets according to time series type (locomotor, and behavioral) recorded from the 2 females and 1 male quail studied during a one-hour period in each of the 12 social groups housed in boxes after 48 hours of habituation to the social environment (Day3). (2019) 6:300 | https://doi.org/10.1038/s41597-019-0299-8 www.nature.com/scientificdata www.nature.com/scientificdata/ were corrected manually using this software. In order to validate the correct tracking and identification of the animal, visual observations of tracking were performed for all birds. The high contrast between the white, well illuminated, box and the dark brown quail feathers facilitated an accurate tracking of the animal. Also animals had small white markings of their backs that allowed identification from video recordings.
Behavioral data sets were collected using the commercially available ANY-maze ™ Video Tracking System software that can be downloaded at www.anymaze.com. Since in this software keystrokes allow observer to register manually behaviors from video recording, a validation period to guarantee reproducibility was first performed. Observer performed validation sessions than consisted in analyzing the same three behavioral video at least two times. Then, reliability was estimated to be >95%. (formula: number of agreements/number of agreements + number of disagreements).

code availability
IdTracker 22 is a videotracking software that keeps the correct identity of each individual during video behavioral analysis and is publicly available at http://www.idtracker.es/. ANY-MAZE@ is a licensed video tracking program, that can be downloaded from http://www.anymaze.co.uk/. The customized Matlab code customized code Locomotion.m in publicly available on Figshare 72 .