How Do Brief Social Interactions Change Over Time in Pair-Bonded Zebra

Home , Courtship display, Human bonding, Pair bond

bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

1 Monogamy in a moment: how do brief social interactions change over time in pair-bonded zebra

2 finches (Taeniopygia guttata)?

3 Nora H. Prior1, *, Edward Smith1, Robert J. Dooling1, Gregory F. Ball1

5 1 Department of Psychology, University of Maryland, College Park, USA

9 *Corresponding author 10 Biology-Psychology Bldg. 11 University of Maryland 12 4094 Campus Dr. 13 College Park, MD 20742 14 Email: [email protected] 15

1 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

16 Abstract

17 Research on monogamy has largely focused on marked behaviors that are unique to pair bonded 18 partners. However, these marked behaviors represent only a subset of the pair-directed behaviors that 19 partners engage in; the influence of pair bonding on mundane or subtle social interactions among 20 partners remains largely unknown. In the current study, we describe the changes that occur during brief 21 social reunions (or greets) over the course of pair bonding in zebra finches. We quantified pair-directed 22 behavior during five-minute reunions from three stages of pair bonding: initial pairing (between 4-72 23 hrs), early pairing (1-2 weeks) and late pairing (>1 month). These social interactions were 24 operationalized in multiples ways. First, we quantified the overall activity levels (call and movement 25 rates) for both the male and female. Overall, females were more active than males, but for both males 26 and females calling activity was highest during the initial timepoint (between 4-72 hrs post-pairing). We 27 quantified behavioral coordination between partners in two ways, 1) similarity in call and movement 28 rates between partners, and 2) temporal synchrony between calls and movements (via sliding 29 correlation coefficients of time-stamped calls and movements). Overall there were no effects of pairing 30 on behavioral coordination. Finally, we used principal component analyses to disentangle behavioral 31 coordination from the activity levels of the male and female. These results contribute to a growing line 32 of evidence that male and female zebra finches differentially contribute to social dynamics and highlight 33 the influence of pair bonding on the development of social dynamics. Behavioral coordination is clearly 34 important for marked interactions (e.g. duetting, courtship displays and biparental care). Our results 35 raise the question of what the roles of such mundane social interactions are in monogamous 36 partnerships. 37 38 Keywords: Behavioral synchrony, interactional synchrony, sex differences, stack call

2 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

39 Introduction

40 In monogamous species, the formation and maintenance of a pair bond is necessary for the

41 successful rearing of offspring (Lack, 1968; Reichard & Boesch, 2003). The majority of research on the

42 mechanisms underlying pair bonding has focused on marked, pair-specific behaviors and interactions of

43 partners (Aragona et al., 2006; Soma & Garamszegi, 2015; Wachtmeister, 2001; Young et al., 2011;

44 Young & Wang, 2004). This approach has revealed remarkable conservation in the mechanisms of pair

45 bonding across taxa (Bales et al., 2007; Donaldson & Young, 2016; O’Connell & Hofmann, 2012; Walum

46 & Young, 2018).

47 Highly-marked behaviors and interactions are particularly evident during courtship and initial

48 pair-bond formation. For example, in prairie voles (Microtus ochrogaster), the establishment of a pair

49 bond is clearly marked by the development of selective preference/affiliation for a mate (Resendez et

50 al., 2016; Williams et al., 1992; Young et al., 2011). This marked selective attachment has been used to

51 identify the mechanisms underlying pair formation. However, these marked behaviors represent only a

52 subset of those that partners engage in, and more detailed understandings of partner interactions are

53 needed in order to further elucidate mechanisms of pair bonding. During partner preferences tests, the

54 pattern of approach and proximity time between partners continues even after initial establishment of

55 the pair bond, with the focal individual spending more time with their partners at later stages of pair

56 bonding (Scribner et al., 2020). Furthermore, important long-term impacts of pair bonding on the brain

57 and behavior of voles are revealed by differences in approach of an individual to its mate versus a

58 stranger (Scribner et al., 2020). Even in the most commonly studied model systems, remarkably little is

59 known about the consequences of pair bonding on mundane or subtle features of partner interactions.

60 Whereas monogamy is rare in mammals (<4% of species), the vast majority of birds form some

61 type of monogamous partnership (~90% of species) (Reichard & Boesch, 2003). Across bird species there

62 is considerable variation in the phenotype of monogamous bonds. Monogamous bonds vary in how long

3 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

63 they are maintained; they may be transient, lasting only a season, or life-long (Black & Hulme, 1996). As

64 with prairie voles, the majority of studies focus on marked features of pair bonding (e.g. partner

65 preference, proximity time, allopreening and clumping) (Kenny et al., 2017; Prior et al., 2013; Smiley et

66 al., 2012; Tomaszycki & Adkins-Regan, 2005). However, the importance of brief social interactions has

67 also been described in numerous species: at the nest, brief social interactions appear to be essential for

68 the active coordination of parental duties between partners (I. C. A. Boucaud et al., 2016; Mariette &

69 Griffith, 2012, 2015; van Rooij & Griffith, 2013). This raises the question more broadly of how subtle

70 features of social interactions are shaped by pair bonding.

71 Here we describe the effect of pair bonding on brief social interactions in monogamous zebra

72 finch pairs. Zebra finches maintain sexually monogamous life-long pair bonds, are non-territorial, and

73 are highly gregarious. Interestingly, traditional partner preference paradigms may fail to demonstrate

74 selective preference for the partner (Prior et al., 2013), although other behavioral assays clearly show

75 that many aspects of pair-directed behavior are reserved for or more common between partners than

76 familiar conspecifics (Fernandez et al., 2017; Gill et al., 2015). Additionally, intra-pair calling dynamics,

77 across multiple contexts, appear to be important behavioral components of the zebra finch pair bond (I.

78 Boucaud et al., 2016; Pietro Bruno D’Amelio et al., 2017a; Elie et al., 2010; Fernandez et al., 2017; Gill et

79 al., 2015).

80 In the current study, we operationalize brief social reunions (both calls and movements) of pairs

81 over the course of pair bonding: initial pairing (between 4-72 hrs), early pairing (1-2 weeks) and late

82 pairing (>1 month). First, we quantified the overall activity levels (call and movement rates) for both

83 males and females. Second, we used two approaches to estimate the coordination of the activity

84 between partners, including quantifying the similarity in call and movement rates between partners as

85 well as quantifying the sliding correlation coefficients for time-stamped calls and movements (a measure

4 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

86 of temporal synchrony). Finally, we used principal component analyses to disentangle behavioral

87 coordination from the activity levels of the male and female.

89 Materials and Methods

90 Subjects and establishment of pairs

91 Twenty adult zebra finches (5-6 months old) were used in this study (10 females and 10 males).

92 Throughout the study, zebra finches were housed with ad libitum seed, water and grit on a 12L:12D light

93 cycle. This same cohort of zebra finches was also used for a subsequent experiment (Prior et al., 2019).

94 Prior to pairing, zebra finches were housed in same-sex flocks. In order to provide the

95 opportunity for pairs to freely form, birds were moved to mixed-sex flocks for 72 hours. Providing

96 individuals with the opportunity to freely form monogamous bonds is important as forced pairing can be

97 associated with lower pair fecundity (Griffith et al., 2017). Thus, birds were housed in mixed-sex flocks

98 with either a male- or female-biased sex ratio (2 females with 3 males or 3 females with 2 males). Pair

99 bonding was assessed visually each day; occurrences of selective affiliative behaviors (i.e. clumping,

100 allopreening, and coordinated preening) between individuals were scored during 5 min behavioral

101 observations. After 72 hours, birds were removed from mixed-sex flocks and housed with their mates

102 for the duration of the study. Note that it is typical for not all birds in mixed-sex flocks to form pair

103 bonds (Scalera & Tomaszycki, 2018; Smiley et al., 2012). Here we identified four clearly-established pairs

104 (paired). Another four pairs were selected from the same mixed-sex flocks of birds that showed little

105 evidence of pairing. The last two pairs were composed of birds unfamiliar with each other. Thus, we

106 established 10 zebra finch pairs with varying extents and patterns of prior experience. We predicted that

107 prior experience would affect behavioral coordination in our reunion paradigm. Thus, despite that there

108 were few birds in each group, we used prior experience as a factor in our later analyses (Paired N=4,

109 Weakly Paired N=4, Force Paired N=2). All pairs, regardless of prior experience, were seen being highly

5 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

110 affiliative in the home cage (and were not seen interacting aggressively). Additionally, after this

111 experiment, all pairs attempted to breed (nest building and/or egg laying) and 8 out of 10 pairs

112 successfully fledged chicks.

113

114 Experimental Design

115 A timeline of the behavioral recordings is presented in Figure 1. We recorded brief social

116 interactions from each pair nine times over the course of the first month of pairing. These nine

117 recordings were not evenly distributed over the course of the month, but were instead situated within

118 periods of the pair bonding process commonly described in the research. Pairing can be conceptually

119 divided up into three stages: a brief courtship phase, a short pair formation phase, and an indefinite pair

120 maintenance phase. Although these stages are commonly referenced (Prior & Soma, 2015; Resendez et

121 al., 2016; Scribner et al., 2020; Smiley et al., 2012), what distinguishes these stages and how long they

122 last is unclear. In general, pair maintenance encompasses anything that occurs after the establishment

123 of a pair bond. In zebra finches, pair bond formation can take up to two weeks; however, it is typically

124 assumed to occur much more quickly (on the order of hours to days) (Zann, 1996).

125 Here we recorded the first social reunion on the day that pairs were moved from mixed-sex

126 flocks to a new home cage with their partner (“initial pairing”). At this timepoint, all pairs had been

127 given the opportunity to engage in courtship behaviors and copulate; however, depending on how they

128 were paired (described above) this time ranged from 4-72 hours together. Thus, we would predict that a

129 mix of courtship and pair formation could be occurring at this timepoint. Next, we recorded reunion

130 behavior from 3 timepoints during the following two weeks (“early pairing”). By two weeks, we would

131 expect all pair bonds to have become established. The majority of research on pairing does not extend

132 past 2-3 weeks post-pairing (Scalera & Tomaszycki, 2018; Smiley et al., 2012; Tomaszycki & Adkins-

133 Regan, 2005; Tomaszycki & Zatirka, 2014), and anything beyond that is typically considered pair

6 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

134 maintenance (Prior et al., 2013; Scribner et al., 2020; Tomaszycki & Adkins-Regan, 2006). Finally, we

135 recorded reunion behavior at 5 timepoints during a late stage of pairing (>1 moth post-pairing), which

136 would be unambiguously considered pair maintenance.

137

138 Behavioral recordings

139 Recordings were made in a sound-attenuated test room, separate from the colony room.

140 Partners were transported, one at a time, in a small covered cage to separate, individual cages each

141 equipped with a tie-clip microphone and a piezo electric sensor attached to the perch (Figure 1). The

142 transport of each partner took 2-3 min, thus partners were separated for 4-6 minutes during

143 transportation. This brief separation period is sufficient to elicit a social reunion (greeting behavior).

144 Partners were not always transported in the same order (some days the male was transported first and

145 other days the female was transported first). Upon placement of the second partner in the cage, reunion

146 behavior was recorded. All four channels (one movement and one acoustic channel for each partner)

147 were recorded using a multi-channel Zoom recorder (F8). Thus, we were able to make single recordings

148 with temporally aligned, individually identifiable acoustic and movement behavior from each partner.

149 We have previously demonstrated that the majority of activity during this behavioral assay

150 occurs within the first few minutes (<5 min) (Prior et al., 2019). First, we minimized the stress of

151 handling by habituating birds to the behavioral and transport procedure prior to the start of the

152 experiment. Additionally, once in the testing room, birds were not handled and were allowed to enter

153 the testing cage/exit the transport cage on their own, with the researcher present. Second, we

154 habituated birds to the procedure prior to the start of experimentation to ensure bird that any changes

155 in behavior would not be due to habituation to the paradigm. Habituation included 10 consecutive days

156 of transport to the testing room (transport only = 4 days; testing room alone (individual recording) = 3

157 days; and social reunion with a same-sex flock mate = 3 days). The last day of habituation was the day

7 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

158 before birds were moved to mixed-sex flocks. Third, we quantified behavior in our assay multiple times

159 during each pairing stage (after the initial pairing). Despite that our behavioral assay clearly elicits

160 socially-directed greeting/reunion behavior, we were concerned that such behavior during these brief

161 behavioral assays would be easily influenced by other factors such as and individuals experience just

162 prior to the assay (such as what was happening in the home cage prior to testing, and aspects of the

163 transport). Finally, we included prior experience at the start of pairing as a factor in our analyses, as it

164 may influence pairing dynamics.

165

166 Scoring behavior during social reunions

167 Operational definition of calls and movements

168 Overall movement and calls rates were quantified for each partner during the first five minutes

169 of the social reunion. Using an in-house written MATLAB program (written by E.S.) we automatically

170 identified time-stamped movements for each partner (Prior et al., 2019). Following initial observations

171 and pilot experiments we decided to group all movements together for three main reasons. First,

172 individuals produced a small repertoire of behaviors. All of the behaviors we observed appeared to be

173 related to the social interactions and included larger body movements (perch hops) and small

174 movements (including head tilts, fluff ups, and wing movements). Note that our goal was to reliably

175 elicit social interactions, so the cages were small and contained no additional stimulation or distraction.

176 Second, it was difficult to disentangle discrete movements (both during observations and from the

177 recording on the timewaveform) because large and small movements often occurred simultaneously or

178 happened in rapid succession. Finally, we had no a priori reason to differentiate between large and small

179 body movements.

180 Calls were semi-automatically classified to identify time-stamped calls for each partner (in-

181 house written MATLAB program written by E.S.). An initial automatic classification identified all noises,

8 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

182 including vocalizations and non-vocalizations. One researcher (N.H.P) manually classified all auditory

183 events as non-vocalizations or calls from either Bird 1 or Bird 2 (left or right channel, respectively).

184 Auditory events were manually classified based on visual assessment of the spectrogram and

185 timewaveform and assigned to the appropriate channel based on the amplitude of the signal on the

186 timewaveform. While call types were not distinguished in the current dataset, the large majority of calls

187 produced were stack calls (with some distance calls). Stack calls are the most common call type used in

188 this behavioral assay (Prior et al., 2019), are commonly used between mates outside of a breeding

189 context (D'Amelio, 2018; Pietro Bruno D’Amelio et al., 2017b; Gill et al., 2015), and encode information

190 on sex and individual identity (Pietro B. D’Amelio et al., 2017; Prior et al., 2018). Stack calls appear to be

191 important for communicating information about movement of partners when they are separated by

192 short distances (D'Amelio, 2018; Pietro Bruno D’Amelio et al., 2017b).

193

194 Operational definition of behavioral coordination

195 The coordination of movements and calling was quantified separately (Prior et al., 2019). First,

196 the similarity in activity rates within a dyad were calculated (e.g. Similarity of Calling = Call Rate of Bird 1

197 - Call Rate of Bird 2 / Call Rate of Bird 1). Second, the temporal synchronization of social dynamics was

198 estimated using sliding correlation coefficients (based on Pearson correlations) of the time-stamped list

199 of movements and events that was generated for each recording. More specifically, the sliding

200 correlation coefficients were calculated separately within the vocalizations and movements using

201 MATLAB “corrcoef” function. The step size for the sliding correlations was chosen based on the natural

202 temporal dynamics of the movements and calls which we assessed during preliminary observations and

203 development of the programs. For calculation of calling synchrony, a 1 ms sliding correlation timestep

204 was used. For calculation of movement synchrony, a 40 ms sliding correlation timestep was used. Inputs

205 to the sliding correlation computations were two vectors of ones and zeros, with a one indicating

9 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

206 presence of a movement or call during the sliding window (1 ms for calls and 40 ms for movements) and

207 a zero indicating absence of a movement or call. The sensor signal power in each time step was

208 computed. For statistical analyses we used the maximum Pearson’s correlation coefficient value

209 (“corrcoef”) based on all possible temporal offsets.

210

211 Disentangling activity and coordination for calls and movements

212 The above approaches allow us to quantify two crucial components of the social reunion:

213 activity and coordination. However, both mathematically and conceptually, the amount of activity and

214 the coordination of activity may be related. Furthermore, the above approaches operationalize

215 movement and calling separately but we would expect these two behaviors to be related. Therefore we

216 used principal components analyses (PCAs) to determine whether we could disentangle activity from

217 behavioral coordination, as well as to describe multi-modal behavioral profiles of these social

218 interactions. We conducted two separate PCAs, one for females and one for males (function ‘prcomp’)

219 because there was a significant effect of sex on activity (see results) . Four dependent variables (call rate,

220 movement rate, sliding correlation of time-stamped calls, sliding correlation of time-stamped

221 movements) were loaded into each PCA.

222 For females, the first two components explained 67% of behavioral variation (PC1 explained

223 36% and PC2 explained 31%). PC1 describes overall activity, with call rate and movement rate positively

224 related (Loadings: Call Rate = -0.625; Movement Rate = -0.732, Sliding Correlation Coefficient of Calls =

225 0.185, Sliding Correlation Coefficient of Movements = 0.227), whereas PC2 describes coordination of

226 activities with a positive relationship between call rate, the coordination of calling and the coordination

227 of movements (Loadings: Call Rate = 0.444; Movement Rate = -0.018, Sliding Correlation Coefficient of

228 Calls = 0.661, Sliding Correlation Coefficient of Movements = 0.605).

10 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

229 For males, the first two components explained 74% of the behavioral variation (PC1 explained

230 46% and PC2 explained 28%). PC1 describes overall activity and the coordination of calls with call rate

231 and movement rate positively related (Loadings: Call Rate = -0.663; Movement Rate = -0.642, Sliding

232 Correlation Coefficient of Calls = -0.372, Sliding Correlation Coefficient of Movements = 0.104), whereas

233 PC2 describes coordination of activities with call rate and movement rate positively related (Loadings:

234 Call Rate = 0.190; Movement Rate = 0.227, Sliding Correlation Coefficient of Calls = -0.502, Sliding

235 Correlation Coefficient of Movements = -0.812).

236

237 Statistical analysis

238 All statistical analyses were carried out in R (v.3.2.3, R Foundation for Statistical Computing). We

239 used linear-mixed models (function lmer from the lme4 Package). For each model, prior to

240 interpretation, we transformed data as necessary based on a visual assessment of the distribution of the

241 residuals. All data presented in graphs is non-transformed.

242 The effect of pair bonding on activity levels (call rate and movement rate) of males and female

243 partners during social reunions was assessed using linear-mixed models with Pairing Stage and Sex as

244 fixed factors and Individual identity as a random factor (CallRate~Sex*Pairing Stage+(1|BirdID)). The

245 effect of pair bonding on the coordination of activities was assessed using linear-mixed models with

246 Pairing Stage as a fixed factor and pair ID as a random factor (Sliding correlation coefficient calling~

247 Pairing Stage+(1|PairID)). Similarly, the effect of pair bonding on multimodal principal components was

248 assessed separately for males and females with Pairing Stage as a fixed factor and Individual ID as a

249 random factor.

250 As we described above, we distributed the nine social reunion recordings based on key stages or

251 pair bonding, rather than evenly throughout the month. For that reason, we chose to use Pairing Stage

252 rather than Date as our primary variable. However, as a double check, the significant main effects that

11 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

253 we report here for Pairing Stage are also present if we use Date as the primary factor. Additionally,

254 because we had the expectation that prior experience at the time of pairing would influence social

255 interactions, we subsequently ran linear-mixed models for key dependent variables including Prior

256 experience as a fixed factor (Sliding correlation coefficient calling~ Pairing* Prior Experience+(1|PairID).

257

258 Results

259 Activity levels of females and males during social reunions

260 While isolated (recorded individually in the testing room), both males and females were largely

261 inactive (Figure 2). Eighteen out of the twenty individuals had a movement rate of <1/min (9 females

262 and 9 males) and fifteen out of the twenty individuals had a call rate of <1/min (6 females and 9 males).

263 We recorded each partner alone in the testing room as a control and it is not included in our statistical

264 models: however, the low levels of activity during isolation emphasizes the extent to which we are able

265 to elicit a social reunion. During the social reunions both call rate and movement rate were higher for

266 females than males, regardless of the stage of pair bonding (Figure 2. Call Rate 2 (1) = 4.65, P = 0.031;

267 Movement Rate 2 (1) = 5.72, P = 0.017).

268

269 Calling activity is modulated by pairing

270 For both males and females call rates were highest during the initial recording (Figure 2. Pairing

271 Stage 2 (2) = 8.35, P=0.015; Pairing Stage  Sex 2 (2) = 0.677, P=0.713). This main effect was driven by

272 a difference between the initial time point and the later timepoint (summary LMER t value = -1.971 P=

273 0.051. There is a similar pattern of decreased movement rate during later pairing, but this effect is not

274 significant (Figure 2. Pairing Stage 2 (2) = 4.52, P = 0.104; Pairing Stage  Sex 2 (2) = 0.30, P = 0.861).

275 The results from our PCA analyses further support the interpretation that the effect of pairing stage is

276 specific to call, not movement, rate. For both males and females, PC1 represented a composite

12 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

277 multimodal activity score (call rate and movement rate were positively correlated see methods). There

278 was no effect of pairing stage on PC1 for females (2 (2) = 4.21, P=0.122) or males ( 2 (2) = 3.43,

279 P=0.180).

280

281 Pairing stage had no effect on the coordination of activities

282 There was no effect of pairing stage on coordination of activity for either calls or movements

283 (Figure 3). Pairing stage had no effect on the percent difference in call rate (F:M) (2 (2) = 1.47, P =

284 0.481), nor on the sliding correlation coefficient of calling activity (2 (2) = 2.85, P = 0.240). Likewise,

285 there was no main effect of pairing stage on the coordination of movements (percent difference 2 (2) =

286 0.11, P = 0.944; sliding correlation coefficient 2 (2) = 1.00, P=0.605). Again, the results of our PCA are

287 consistent with the raw data. For both males and females PC2 represented a composite multimodal

288 coordination score (the sliding correlation coefficient of calls and movements were positively correlated

289 see methods). There was no effect of pairing stage on PC2 for females (2 (2) = 1.09, P=0.581) or males

290 (2 (2) = 3.35, P=0.187).

291

292 Additional factors that may influence social reunion

293 Pairing stage was the primary factor we investigated. However, we had the opportunity to also

294 ask whether there was the potential for a relationship between two additional features of the

295 partnership and behavior during the social reunion. First, we asked whether prior experience at the time

296 of pairing may influence behavioral coordination. Indeed, there was a significant interaction between

297 Experience and Pairing Stage on activity (Figure 4A-D. Calling Prior Experience 2 (2) = 0.10, P=0.950;

298 Pairing Stage  Prior Experience 2 (4) = 17.44, P = 0.001; Movement: Prior Experience 2 (2) = 0.20,

299 P=0.903; Pairing Stage  Prior Experience 2 (4) = 17.88, P = 0.001). There was also a significant

300 interaction between Experience and Pairing Stage on calling synchrony (sliding correlation coefficient of

13 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

301 calls: Prior Experience 2 (2) = 3.31, P=0.191; Pairing Stage  Prior Experience 2 (4) = 17.93, P = 0.001).

302 However, there was no effect of prior experience on any other measure of the coordination of activities

303 (Percent Difference in Calling: Prior Experience 2 (2) = 2.03, P = 0.363; Pairing Stage  Prior Experience

304 2 (4) = 7.01, P = 0.135; Percent Difference in Movement: Prior Experience 2 (2) = 1.16, P = 0.561;

305 Pairing Stage  Prior Experience 2 (4) = 2.57, P = 0.633; Sliding Correlation of Movements: Prior

306 Experience 2 (2) = 0.30, P = 0.863; Pairing Stage  Prior Experience 2 (4) = 4.49, P = 0.383).

307 Second, we tested whether behavioral coordination during social interactions during early

308 pairing influences breeding success. After the conclusion of this study, pairs had multiple opportunities

309 to breed together. In the first breeding attempt (two months after the end of this experiment) all 10

310 pairs engaged in nest building behavior and laid eggs. Seven out of the 10 pairs went on to successfully

311 fledge chicks (between 2-4 chicks per clutch). One of these only built a nest (a strongly paired dyad). The

312 two other pairs laid eggs that did not hatch (a weakly paired and force paired dyad). There was no

313 difference in calling synchrony or movement synchrony between pairs that successful fledged chicks

314 (Sliding Correlation of Calls: 2 (2) = 0.24, P = 0.621; Sliding Correlation of Movements: 2 (2) <0.01, P =

315 0.978).

316

317 Discussion

318 Here we describe the subtle changes that occur during brief social reunions between

319 monogamous partners over the first month of pairing. The primary effect was elevated calling of both

320 the male and female partner during the initial social reunion timepoint. Interestingly, we saw no change

321 in coordination of calling or movement activity over the course of pairing. This is particularly notable

322 because we have demonstrated previously that social reunions with novel conspecifics elicit less robust

323 behavioral responses and that those interactions are less coordinated (Prior et al., 2019). Together this

14 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

324 suggests that zebra finches are very rapidly able to develop patterns of interactions between familiar

325 individuals.

326 Beyond the primary effect of pairing on calling activity, this study raises additional questions

327 about what factors influence the social reunion of partners. First, across pairing stages there were

328 differences between females and males in their respective contributions to the interaction. Overall,

329 females were more active than males (having higher call rates and movement rates). Second, pairs

330 differed in how long they had been paired and the nature of the courtship experience (prior experience).

331 Our data suggests this prior experience influenced patterns of behavior (both activity and the

332 coordination of activities) during the initial, but not subsequent reunions.

333

334 Changes in pair-directed affiliative behavior over time

335 For species that maintain long-term social bonds, including humans, the ability to maintain such

336 bonds is as important as the ability to form them. Importantly, there is evidence that the neurobiological

337 mechanisms underlying the formation and maintenance of bonds are different (Aragona et al., 2006;

338 Prior & Soma, 2015; Resendez et al., 2016; Smiley et al., 2012). There are many highly-marked affiliative

339 behaviors that are associated with pair bonding, which can be useful in some species for identifying the

340 point when pair bonds become established. Research investigating the changes in pairing over time has

341 tended to parse these stages of pair bonding into discrete stages. In species such as zebra finches, where

342 a mating event may not be needed for pair bond formation and individuals do not form a strong partner

343 preference, it is particularly challenging to distinguish discrete changes.

344 Whereas the presence of highly-marked affiliative behaviors (e.g. clumping (close proximity),

345 side-by-side perching facing the same direction, and allopreening (Black & Hulme, 1996; Reichard &

346 Boesch, 2003; Zann, 1996)) are associated with the establishment of a bond, it is less clear how these

347 behaviors change over time following initial pair bond formation. There is some evidence from both

15 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

348 prairie voles and zebra finches that even after pair bond establishment, pairs continue to increase the

349 amount of time they spend in close proximity (Pietro Bruno D’Amelio et al., 2017b; Scribner et al., 2020).

350 In zebra finches, D’Amelio and colleagues (2017) described the changes in social dynamics of

351 new vs established zebra finch pairs in their home cage over a week. On the first day, newly paired zebra

352 finches spent significantly less time in physical proximity (clumping) compared to the established pairs

353 but groups were already more similar by the third day of pairing. This timeline is consistent with when

354 we see most significant changes in social reunion behavior in the current study (within the first few days

355 of pairing). However, the trend for a difference between new and established pairs in D’Amelio et al.,

356 2017 was present throughout the week of recording. Similarly, in prairie voles time spent in close

357 proximity between partners appears to increase over the first month of pairing (Scribner et al., 2020).

358 D’Amelio and colleagues (2017) also carefully described the effect of pair bonding on vocal

359 interactions of new and established zebra finch pairs using continuous vocal recordings. On the first day

360 of recording, newly paired individuals called less than predicted. Additionally, over the first week of

361 pairing, calling dynamics between partners became more symmetrical: shifting from a scenario where

362 one bird calls more and the other answers to a scenario where each partner calls and answers similarly.

363 Importantly, even newly paired birds were clearly motivated to engage in call-response. It is striking that

364 we see similar patterns emerging during the brief social reunion assay. The types of calling exchanges

365 we elicited here are very similar to those elicited by D’Amelio et al., 2017, and are predominately made

366 up of stack-stack calls and responses. This highlights the ethological relevance of our current social

367 reunion test. Our current study combined with recent research highlights the need for a more

368 comprehensive description of the subtle ways that social dynamics of partners change over time

369 (Scribner et al., 2020).

370 Monogamy and moment-to-moment behavioral coordination

16 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

371 Broadly, behavioral coordination across timescales is associated with gregariousness (Conradt &

372 Roper, 2005; Focardi & Pecchioli, 2005) and is thought to increase social cohesion (King & Cowlishaw,

373 2009; Pays et al., 2007) and affiliative behavior (Sakai et al., 2010). Furthermore, behavioral

374 coordination between two individuals has been shown to promote prosocial behavior (Ashton–James et

375 al., 2007; Gueguen et al., 2009; Van Baaren et al., 2004) (reviewed in (Duranton & Gaunet, 2016)), which

376 we would expect to reflect the presence of affiliative bonds. Even on extremely acute timescale, such as

377 what we are focused on here, there is evidence in humans that social or interactional synchrony

378 promotes the formation and reinforcement of affiliative bonds (Feldman, 2007; Feldman & Eidelman,

379 2007; Feldman et al., 2011). In songbirds specifically, there has been extensive research investigating the

380 function of behavioral coordination in monogamous pairs particularly as it relates to breeding success

381 and the coordination of biparental care (I. C. A. Boucaud et al., 2016; Mariette & Griffith, 2012, 2015;

382 van Rooij & Griffith, 2013). Interestingly, brief social interactions at the nest, only a few minutes long

383 appear to be essential for the coordination of parental duties across many species (review Prior 2020).

384 In zebra finches specifically, there are several lines of evidence demonstrating that parental duties are

385 actively coordinated during interactive calling exchanges (I. Boucaud et al., 2016; Boucaud et al., 2017;

386 Elie et al., 2010; Villain et al., 2016). During incubation, female calling behavior predicts whether or not

387 the male will relieve her and begin incubation (Boucaud et al., 2017). Experimentally preventing the

388 male from returning to the nest to relieve the female, causes her to modify calling behavior which

389 subsequently predicted the timing of the females time off the nest (I. C. A. Boucaud et al., 2016).

390 Experimental evidence from other avian species suggests that the coordination of parental behavior is

391 an emergent consequence of the behavioral interactions between partners, not simply a summation of

392 both partners contributions (Ball & Silver, 1983). How partners develop such patterns of communication

393 and what makes partners good communicators remains an open question.

17 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

394 It is possible that the patterns of a partner’s communication during mundane social interactions

395 lays the foundation for these marked-moments. After the end of our current experiment, all the pairs

396 were given the opportunity to breed. Seven out of 10 pairs eventually went on to successfully fledge

397 chicks, although there were no differences in behavioral coordination of the pairs that successfully

398 fledged chicks and those that did not. The fact that we saw no relationship between behavioral

399 coordination and breeding success is not surprising given our very low sample size (3 vs 7) as well as the

400 amount of time after the last social reunion recording and breeding. Furthermore, we would expect that

401 there are nuances beyond temporal synchrony alone which are important for the development of these

402 social interactions. Future research will directly test the relationship between moment-to-moment

403 coordination of activities and the coordination of parental behavior in order to further elucidate the

404 function of behavioral variation in these brief interactions.

405

406 Effect of familiarity and prior experience on social reunions

407 An intriguing potential confound exists when examining the long-term effects of pair bonds on

408 social dynamics: to what extent can the nature of the social relationship (formation of a monogamous

409 pair bond) be disentangled from familiarity and shared social experience. Using the same social reunion

410 behavioral assay, we recently demonstrated that familiarity itself, not pair bonding per se, influences

411 social reunion behavior (Prior et al., 2019). When we compared reunion behavior between different

412 social dyads (monogamous partners, familiar same-sex dyads, familiar opposite-sex dyads, novel same-

413 sex dyads, and novel opposite-sex dyads), we found that both activity levels and the coordination of

414 activity was higher in familiar social dyads (Prior et al., 2019). In our current study, only the two force

415 paired dyads (paired for ~4 hours), appeared similar to the novel dyads described previously. Thus, we

416 can assume that pairs were able to familiarize and stabilize their degree of coordination very quickly,

417 between 4 – 72 hours. Combined, we interpret the results of these two studies as demonstrating an

18 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

418 effect of prior social experience on behavioral coordination and patterns of behavior during brief social

419 interactions.

420 Over the course of pair bonding, the fact that coordination is maintained but the activity

421 decreases could be evidence that the strength of established bonds is positively correlated with the

422 length of social exchanges required during. Perhaps well-established partners require ever-briefer, and

423 less intense, social exchanges. Perhaps it is less the extent of coordination, but rather the extent of the

424 effort it takes to coordinate that reflects pair bond strength (although prior experience had no lasting

425 effect on behavioral profiles beyond behavior during the initial courtship timepoint). Despite behavioral

426 coordination/ interactional synchrony being influenced by social bonding; the processes by which this

427 happens may be a shared biological foundation of social alignment, rather than a pair bonding process

428 specifically.

429

430 Conclusion

431 Here we show one way that mundane social interactions is affected by pair bonding, likely via shared

432 prior social experience. Broadly speaking, this is consistent with the idea that social relationships are a

433 culmination of repeated social interactions between familiar individuals. Understanding how brief social

434 exchanges are modulated by experience and social bonding may provide an entry point to describe the

435 wide diversity in the phenotypes of social relationships.

436

437 Acknowledgements

438 We would like to thank the entire Ball/Dooling lab for help with animal husbandry, data collection, and

439 discussion of results. An earlier version of these results was presented at the Animal Behavior Society

440 meeting in Chicago, IL in July 2019. We are grateful for all of the feedback we received there. For

441 feedback on this manuscript, we thank Dr. Matthew D. Taves and Dr. Benjamin A. Sandkam.

19 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

442

443 Ethics Statement

444 This work was conducted in accordance with Association for the Study of Animal Behaviour (ASAB)

445 guidelines and was approved by the Institutional Animal Care and Use Committee (IACUC) (R-15-09),

446 University of Maryland, College Park.

447

448 Funding

449 A T32 training grant to N.H.P (NIDCD T-32 DC00046).

20 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

450 References 451 Aragona, B. J., Liu, Y., Yu, Y. J., Curtis, J. T., Detwiler, J. M., Insel, T. R., & Wang, Z. (2006). 452 Nucleus accumbens dopamine differentially mediates the formation and maintenance of 453 monogamous pair bonds. Nature Neuroscience, 9(1), 133 454 Ashton–James, C., Van Baaren, R. B., Chartrand, T. L., Decety, J., & Karremans, J. (2007). 455 Mimicry and me: The impact of mimicry on self–construal. Social cognition, 25(4), 518- 456 535 457 Bales, K. L., Mason, W. A., Catana, C., Cherry, S. R., & Mendoza, S. P. (2007). Neural 458 correlates of pair-bonding in a monogamous primate. Brain research, 1184, 245-253 459 Ball, G. F., & Silver, R. (1983). Timing of incubation bouts by ring doves (Streptopelia risoria). 460 Journal of Comparative Psychology, 97(3), 213 461 Black, J. M., & Hulme, M. (1996). Partnerships in Birds: The Study of Monogamy. Oxford 462 University Press, UK. 463 Boucaud, I., Mariette, M., Villain, A., & Vignal, C. (2016). Vocal negotiation over parental care? 464 Partners adjust their time spent incubating based on their acoustic communication at the 465 nest. Biol. J. Linnean Soc, 117, 322-336. 466 Boucaud, I. C., Perez, E. C., Ramos, L. S., Griffith, S. C., & Vignal, C. (2017). Acoustic 467 communication in zebra finches signals when mates will take turns with parental duties. 468 Behavioral Ecology, 28(3), 645-656. 469 Boucaud, I. C. A., Smith, M. L. N. A., Valère, P. A., & Vignal, C. (2016). Incubating females 470 signal their needs during intrapair vocal communication at the nest: a feeding experiment 471 in great tits. Animal behaviour, 122, 77-86. 472 Conradt, L., & Roper, T. J. (2005). Consensus decision making in animals. Trends in ecology & 473 evolution, 20(8), 449-456 474 D'Amelio, P. B. (2018). Vocal communication in zebra finches: a focused description of pair 475 vocal activity lmu. 476 D’Amelio, P. B., Klumb, M., Adreani, M. N., Gahr, M. L., & Ter Maat, A. (2017). Individual 477 recognition of opposite sex vocalizations in the zebra finch. Scientific Reports, 7(1), 1-10. 478 D’Amelio, P. B., Trost, L., & Ter Maat, A. (2017a). Vocal exchanges during pair formation and 479 maintenance in the zebra finch (Taeniopygia guttata). Frontiers in Zoology, 14(1), 13 480 D’Amelio, P. B., Trost, L., & ter Maat, A. (2017b). Vocal exchanges during pair formation and 481 maintenance in the zebra finch (Taeniopygia guttata). Frontiers in Zoology, 14(1), 13. 482 Donaldson, Z. R., & Young, L. J. (2016). The neurobiology and genetics of affiliation and social 483 bonding in animal models. In Animal models of behavior genetics (pp. 101-134). 484 Springer. 485 Duranton, C., & Gaunet, F. (2016). Behavioural synchronization from an ethological 486 perspective: overview of its adaptive value. Adaptive Behavior, 24(3), 181-191 487 Elie, J. E., Mariette, M. M., Soula, H. A., Griffith, S. C., Mathevon, N., & Vignal, C. (2010). 488 Vocal communication at the nest between mates in wild zebra finches: a private vocal 489 duet? Animal behaviour, 80(4), 597-605. 490 Feldman, R. (2007). Parent–infant synchrony and the construction of shared timing; 491 physiological precursors, developmental outcomes, and risk conditions. Journal of child 492 psychology and psychiatry, 48(3‐4), 329-354 493 Feldman, R., & Eidelman, A. I. (2007). Maternal postpartum behavior and the emergence of 494 infant–mother and infant–father synchrony in preterm and full‐term infants: The role 495 of neonatal vagal tone. Developmental psychobiology, 49(3), 290-302

21 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

496 Feldman, R., Gordon, I., & Zagoory‐Sharon, O. (2011). Maternal and paternal plasma, salivary, 497 and urinary oxytocin and parent–infant synchrony: considering stress and affiliation 498 components of human bonding. Developmental Science, 14(4), 752-761 499 Fernandez, M. S. A., Vignal, C., & Soula, H. A. (2017). Impact of group size and social 500 composition on group vocal activity and acoustic network in a social songbird. Animal 501 behaviour, 127, 163-178. 502 Focardi, S., & Pecchioli, E. (2005). Social cohesion and foraging decrease with group size in 503 fallow deer (Dama dama). Behavioral ecology and sociobiology, 59(1), 84-91 504 Gill, L. F., Goymann, W., Ter Maat, A., & Gahr, M. L. (2015). Patterns of call communication 505 between group-housed zebra finches change during the breeding cycle. elife, 4. 506 Griffith, S. C., Crino, O. L., Andrew, S. C., Nomano, F. Y., Adkins‐Regan, E., Alonso‐ 507 Alvarez, C., Bailey, I. E., Bittner, S. S., Bolton, P. E., & Boner, W. (2017). Variation in 508 reproductive success across captive populations: methodological differences, potential 509 biases and opportunities. Ethology, 123(1), 1-29 510 Gueguen, N., Jacob, C., & Martin, A. (2009). Mimicry in social interaction: Its effect on human 511 judgment and behavior. European Journal of Social Sciences, 8(2), 253-259. 512 Kenny, E., Birkhead, T. R., & Green, J. P. (2017). Allopreening in birds is associated with 513 parental cooperation and stable pair bonds across years. 514 King, A. J., & Cowlishaw, G. (2009). All together now: behavioural synchrony in baboons. 515 Animal behaviour, 78(6), 1381-1387 516 Lack, D. L. (1968). Ecological adaptations for breeding in birds. 517 Mariette, M. M., & Griffith, S. C. (2012). Nest visit synchrony is high and correlates with 518 reproductive success in the wild zebra finch Taeniopygia guttata. Journal of Avian 519 Biology, 43(2), 131-140. 520 Mariette, M. M., & Griffith, S. C. (2015). The adaptive significance of provisioning and foraging 521 coordination between breeding partners. The American Naturalist, 185(2), 270-280. 522 O’Connell, L. A., & Hofmann, H. A. (2012). Evolution of a vertebrate social decision-making 523 network. Science, 336(6085), 1154-1157 524 Pays, O., Jarman, P. J., Loisel, P., & Gerard, J.-F. (2007). Coordination, independence or 525 synchronization of individual vigilance in the eastern grey kangaroo? Animal behaviour, 526 73(4), 595-604 527 Prior, N. H., Fernandez, M. S. A., Soula, H. A., & Vignal, C. (2019). Water restriction influences 528 intra-pair vocal behavior and the acoustic structure of vocalisations in the 529 opportunistically breeding zebra finch (Taeniopygia guttata). Behavioural processes, 162, 530 147-156. 531 Prior, N. H., Heimovics, S. A., & Soma, K. K. (2013). Effects of water restriction on 532 reproductive physiology and affiliative behavior in an opportunistically-breeding and 533 monogamous songbird, the zebra finch. Hormones and Behavior, 63(3), 462-474. 534 Prior, N. H., Smith, E., Lawson, S., Ball, G. F., & Dooling, R. J. (2018). Acoustic fine structure 535 may encode biologically relevant information for zebra finches. Scientific Reports, 8(1), 536 6212. 537 Prior, N. H., & Soma, K. K. (2015). Neuroendocrine regulation of long-term pair maintenance in 538 the monogamous zebra finch. Hormones and Behavior, 76, 11-22. 539 Reichard, U. H., & Boesch, C. (2003). Monogamy: mating strategies and partnerships in birds, 540 humans and other mammals. Cambridge University Press.

22 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

541 Resendez, S. L., Keyes, P. C., Day, J. J., Hambro, C., Austin, C. J., Maina, F. K., Eidson, L. N., 542 Porter-Stransky, K. A., Nevárez, N., & McLean, J. W. (2016). Dopamine and opioid 543 systems interact within the nucleus accumbens to maintain monogamous pair bonds. 544 elife, 5. 545 Sakai, M., Morisaka, T., Kogi, K., Hishii, T., & Kohshima, S. (2010). Fine-scale analysis of 546 synchronous breathing in wild Indo-Pacific bottlenose dolphins (Tursiops aduncus). 547 Behavioural processes, 83(1), 48-53 548 Scalera, A., & Tomaszycki, M. L. (2018). Acute exogenous corticosterone treatments have few 549 effects on courtship and pair bonding in zebra finches. General and Comparative 550 Endocrinology, 268, 121-127 551 Scribner, J. L., Vance, E. A., Protter, D. S. W., Sheeran, W. M., Saslow, E., Cameron, R. T., 552 Klein, E. M., Jimenez, J. C., Kheirbek, M. A., & Donaldson, Z. R. (2020). A neuronal 553 signature for monogamous reunion. Proceedings of the National Academy of Sciences, 554 117(20), 11076-11084 555 Smiley, K. O., Vahaba, D. M., & Tomaszycki, M. L. (2012). Behavioral effects of progesterone 556 on pair bonding and partner preference in the female zebra finch (Taeniopygia guttata). 557 Behavioural processes, 90(2), 210-216 558 Soma, M., & Garamszegi, L. Z. (2015). Evolution of courtship display in Estrildid finches: dance 559 in relation to female song and plumage ornamentation. Frontiers in Ecology and 560 Evolution, 3, 4 561 Tomaszycki, M. L., & Adkins-Regan, E. (2005). Experimental alteration of male song quality 562 and output affects female mate choice and pair bond formation in zebra finches. Animal 563 behaviour, 70(4), 785-794. 564 Tomaszycki, M. L., & Adkins-Regan, E. (2006). Is male song quality important in maintaining 565 pair bonds? Behaviour, 143(5), 549-567. 566 Tomaszycki, M. L., & Zatirka, B. P. (2014). Same-sex partner preference in zebra finches: 567 Pairing flexibility and choice. Archives of sexual behavior, 43(8), 1469-1475 568 Van Baaren, R. B., Holland, R. W., Kawakami, K., & Van Knippenberg, A. (2004). Mimicry and 569 prosocial behavior. Psychological science, 15(1), 71-74 570 van Rooij, E. P., & Griffith, S. C. (2013). Synchronised provisioning at the nest: parental 571 coordination over care in a socially monogamous species. PeerJ, 1, e232. 572 Villain, A. S., Fernandez, M. S. A., Bouchut, C., Soula, H. A., & Vignal, C. (2016). Songbird 573 mates change their call structure and intrapair communication at the nest in response to 574 environmental noise. Animal behaviour, 116, 113-129 575 Wachtmeister, C.-A. (2001). Display in monogamous pairs: a review of empirical data and 576 evolutionary explanations. Animal behaviour, 61(5), 861-868 577 Walum, H., & Young, L. J. (2018). The neural mechanisms and circuitry of the pair bond. 578 Nature Reviews Neuroscience, 19(11), 643-654 579 Williams, J. R., Catania, K. C., & Carter, C. S. (1992). Development of partner preferences in 580 female prairie voles (Microtus ochrogaster): the role of social and sexual experience. 581 Hormones and Behavior, 26(3), 339-349 582 Young, K. A., Gobrogge, K. L., Liu, Y., & Wang, Z. (2011). The neurobiology of pair bonding: 583 insights from a socially monogamous rodent. Frontiers in neuroendocrinology, 32(1), 53- 584 69 585 Young, L. J., & Wang, Z. (2004). The neurobiology of pair bonding. Nature Neuroscience, 586 7(10), 1048-1054.

23 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

587 Zann, R. A. (1996). The zebra finch: a synthesis of field and laboratory studies. Oxford 588 University Press. 589 590

24 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

591 Figure 1

592

593

25 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

594 Figure 2 A 30 Female

25 Male

) n

i Female > Males

20 m

e P < 0.05

a C

c 10

# ( 5

n g g g io in in in t ir ir ir la a a a o P P P s l I a ly te ti r a i a L In E B

) Female > Males n

i 25 P < 0.05

t e

a 20

n n

e 15

e e

v 10

M m

(

n g g g io in in in t ir ir ir la a a a o P P P s l I a ly te ti r a i a L In E 595 596 597

26 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

598 Figure 3 A

0.20

a C

0.15

i t

a 0.10

o C

0.05

i l S 0.00

g g g in in in ir ir ir a a a P P P l a ly te ti r a i a L In E

B s

t 0.05

m e

v 0.04

o 0.03

t a

l 0.02

o C

0.01

d i

l 0.00 S g g g in in in ir ir ir a a a P P P l a ly te ti r a i a L In E 599 600 601 602 603

27 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

604 Figure 4 Females Males 50 B 50 A Force-Paired Weakly-Paired

40 40 )

) Paired

t r

30 t

r 30

l a

l 20 a

20 C

( ( 10 10

0 0 g g g g g g in in in in in in ir ir ir ir ir ir a a a a a a P P P P P P l y e l y e ia rl t ia rl t it a it a a a L n L In E C I E

50 D 50

)

i m

40 m 40

30 t 30

m m

20 m 20

o M

10 M 10

(

0 0 g g g g g g in in in in in in ir ir ir ir ir ir a a a a a a P P P P P P l l a ly te a ly te ti r a ti r a i a L i a L In E In E 605 606

28 bioRxiv preprint doi: https://doi.org/10.1101/2020.06.18.160051; this version posted June 20, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license.

607 Figure Legends 608 Figure 1: Diagram of the timeline of initial pair formation and behavioral recordings during the first 609 month of pairing. All birds were given the opportunity to freely form a pair bond in mixed sex flocks 610 over the course of 72 hours. The first social reunion was recorded on the day that pairs were moved 611 from mixed-sex flocks to a new home cage with their partner (“initial pairing”). At this timepoint, pair 612 had been removed from flock for 4-72 hours. At the time of pairing, four pairs from the mixed-sex flocks 613 clearly engaged in selective pairing behavior, four pairs were formed from birds that had been together 614 in mixed-sex flocks but were not obviously pair bonded, and two pairs were composed of individuals 615 that had not been in the same mixed-sex flocks and had no prior experience. We recorded reunion 616 behavior from 3 timepoints during the following two weeks (“early pairing”), and from 5 timepoints 617 during a late stage of pairing (>1 month post pairing) (“late pairing”). Additionally, as a control, two 618 recordings were made of each individual alone in the testing room (individual recordings). A schematic 619 illustration of social reunion behavior paradigm is shown in the bottom right; redrawn from Prior et al., 620 2019). Briefly, behavior was recorded using four channels of a Zoom recorder (F8): movement was 621 recorded from a piezo sensor attached to the perch of a smaller cage (indicated via oscilloscope), and 622 acoustic behavior was recorded using tie clip microphones (indicated via spectrogram). 623 624 Figure 2: Overall activity rates, call rate (A) and movement rate (B), are shown for females and males 625 tested in isolation and with their partner across the three pairing stages. Overall, females were more 626 active than males. Additionally, male and female call rates were highest during the initial recording 627 (Pairing Stage (2) = 8.35, P=0.015; Pairing Stage  Sex (2) = 0.677, P=0.713). 628 629 Figure 3: Pairing stage had no effect on the coordination of activities between males and females as 630 quantified using the sliding correlation coefficient of time-stamped calls (A), and time-stamped 631 movements (B), across the three pairing stages. As described in the text, there was also no effect of 632 pairing stage on similarity of call rate and movement rate between partners. 633 634 Figure 4: Effect of prior experience on activity (call rate A-B and movement rate C-D) for females (A and 635 C) and males (B and D) across the three stages of pairing. Pairs differed based on the amount of time 636 and prior experience they had when they were initially paired after the 72 hours being housed in mixed- 637 sex flocks. Four pairs had clearly formed (paired). Another four pairs were created from individuals of 638 the same mixed-sex flocks (weakly paired), with little evidence of pairing in the flocks. The final two 639 pairs were composed of individuals that had no prior familiarity with each other (force paired). There 640 was no effect of sex on either calling or movement, however there was a significant interaction between 641 Prior Experience and Pairing Stage (Calling: (4) = 17.44, P = 0.001; Movement: (4) = 17.88, P = 0.001). 642