Communication governs the formation and maintenance of social relationships. The interpretation of communication signals depends not only on the signal’s content, but also on a receiver’s individual experience. Experiences throughout life may interact to affect behavioral plasticity, such that a lack of developmental sensory exposure could constrain adult learning, while salient adult social experiences could remedy developmental deficits. We investigated how experiences impact the formation and direction of female auditory preferences in the zebra finch. Zebra finches form long-lasting pair bonds and females learn preferences for their mate’s vocalizations. We found that after two weeks of cohabitation with a male, females formed pair bonds and learned to prefer their partner’s song regardless of whether they were reared with (“normally-reared”) or without (“song-naïve”) developmental exposure to song. In contrast, females that heard but did not physically interact with a male did not prefer his song. In addition, previous work has found that song-naive females do not show species-typical preferences for courtship song. We found that cohabitation with a male ameliorated this difference in preference. Thus, courtship and pair bonding, but not acoustic-only interactions, strongly influence preference learning regardless of rearing experience, and may dynamically drive auditory plasticity for recognition and preference.

Animals. All zebra finches (n=86 females and 24 males, average 9 months old, range 3-26 months) were maintained on a 14:10 light:dark cycle, given ad libitum access to seed, grit, and water, and weekly supplements (e.g. lettuce, egg). We raised females in one of two conditions. One set of females (‘normally-reared’) was raised to 60 days post hatch in a cage with both parents and siblings. A second set of females (‘song-naïve’) were raised in sound-attenuating chambers (‘soundboxes’; TRA Acoustics, Cornwall, Ontario) with only the mother and siblings. Specifically, fathers were removed within five to seven days post-hatch, prior to the period when females memorize the father’s song [1] and male siblings were removed at 30-40 days post-hatch, prior to producing stereotyped song [2]. After 60 days, normally-reared-females were housed either in same-sex group cages in our mixed-sex colony (n=23) or in same-sex group cages in an all-female colony (n=21), and song-naïve females (n=38) were housed in same-sex group cages in an all-female colony. All males used for social cohabitation or song recordings were normally-reared and housed in same-sex group cages in the mixed-sex colony prior to use. All procedures adhered to Canadian Council on Animal Care guidelines and the protocol approved by the Animal Care Committee of McGill University.

Cohabitation conditions. To assess the degree to which female preference is shaped by specific social experiences in adulthood, females were either housed in a cage with a male, thus providing them with an opportunity to mate (‘opposite-sex paired females’), or housed in a cage with another female (‘same-sex paired females’) (Figure 1A). The cages containing opposite-sex paired (OSP) and same-sex paired (SSP) females were both located within the same soundbox but separated from each other by an opaque barrier. This meant that all birds could hear each other, but only see and physically interact with their cage mate. Both cages were provided with nesting material weekly and remained together for two weeks.

Video recording. We filmed each group when the male was first introduced (‘Week 0’), and after the first (‘Week 1’) and second (‘Week 2’) week of cohabitation using a GoPro Hero or Hero 5 camera.  All recordings lasted for at least 2 hours and were recorded at the same time of day within each group. In a few cases (n=18) we recorded birds continuously over 2 weeks. For these birds, we analyzed 2h of video from the three time points listed above.

Preference testing.  To assess preferences for male songs, females were given a two-choice active choice task as described previously [3]. Briefly, females were placed in a cage containing two strings that, when pulled, would each trigger the playback of song from a single male zebra finch through an adjacent speaker (e.g., song of male A for one string and song of male B for the other string). Females were trained to associate pulling strings with sound playback using either conspecific and heterospecific songs or calls from unfamiliar females. Each test consisted of two two-hour sessions. To control for side bias, contingencies switched for the second session of every test. For both sessions, the session started once a female pulled each string three times. Each female was tested on at least three stimulus sets: familiar male vs. unfamiliar male, courtship vs. non-courtship songs from a familiar male, and courtship vs. non-courtship songs from an unfamiliar male. The order of stimulus presentation was randomized within and across testing days for each female. Females were tested on 1-2 stimulus sets per day for a maximum of five consecutive days, including acclimation time. If females did not complete the necessary tests after five days, they were housed in small same-sex groups for at least 2 days before resuming testing. We found that these brief pauses in testing improved motivation to hear song and complete the tests. Females that did not complete any of the three tests after two attempts each were excluded from further analysis (4 normally-reared and 10 song-naive). 

Song Stimuli. Songs were recorded as previously described [4–6]. Briefly, males unrelated to the experimental subjects were recorded in sound attenuating chambers with omnidirectional microphones (Countryman Associates, California) using a sound-activated recordings system (Sound Analysis Pro, SAP; 44.1kHz)[7]. Unfamiliar females not involved in the experiment were presented to the male in a separate cage for brief intervals to collect female-directed courtship song. Songs selected for stimuli were a representative sample of the male’s typical variation in song duration, number of motifs, and introductory notes. All songs were free of noise and female calls. Each stimulus set consisted of 8-15 songs from each male. When females were tested on the songs of two different males, males with similar durations and numbers of motifs were selected as stimulus sets. For familiar male vs. unfamiliar male tests, we used a yoked design: each male’s song served as the familiar song for one set of females (one OSP and two SSP, see Cohabitation conditions), and the unfamiliar song for a different set of females. Therefore, the same stimulus set is used for two sets of females whose experience with the two males differ. All stimulus songs were bandpass filtered (300–10 kHz), normalized by their maximum amplitude, and saved as wav files (44.1 kHz) using custom written code in Matlab (Mathworks, Natick, MA).

Analyses

Pair Behaviors. The three 2h samples of video taken at Weeks 0, 1, and 2, were quantified by individuals blind to the rearing condition of the females. We recorded the duration (in seconds) and number of occurrences of a number of individual and pair bonding behaviors: courtship singing, non-courtship singing, clumping, allopreening, pecking, bill fencing, and nesting behaviors [5,8–11]. Courtship singing by the male was defined by a display of at least two of the following behaviors during singing: orienting towards the female, fluffing of body feathers while flattening feathers on the head, and courtship dancing [5,6,11,12]. We calculated the percent of time spent on each behavior out of two hours for each pair. 

Preference tests. From each test, the total number of playbacks for Stimulus A and Stimulus B during each session was determined. In addition to quantifying the number of pulls for each stimulus, we also performed a bootstrap with replacement (10,000 iterations) to attain a normalized metric of preference strength (‘preference index’) with 95% confidence intervals (CIs). Females were considered to have a significant preference if the mean and 95% CIs were above or below the chance threshold (0.5). We used mixed-effects models with bootstrapped preferences as the dependent variable and cohabitation experience, rearing, and the interaction of experience and rearing as independent variables and individual ID as a random variable. For each rearing and cohabitation experience condition, we also tested whether the distribution was significantly different from chance (0.5). Finally, we also coded responses as either a significant preference or no preference based on the bootstrapped preference and CIs. Preferences were coded as 1 (above 0.5) or -1 (below 0.5) if the CIs did not cross 0.5, and as 0 if the CIs did cross 0.5. We analyzed variation in the pattern of significant preferences across groups using nominal logistic models with rearing, cohabitation experience, and their interaction as independent variables.

We investigated the correlation between preference strength and individual pair bonding behaviors in two ways. First, we performed a correlation analysis for each rearing condition and time point between the preference score and each of the pair bonding behaviors. To gain a greater sense of how combinations of behaviors relate to preference, we also used a principal components analysis (PCA) to investigate the correlations between pair bonding behaviors and partner preference. We ran separate PCAs for each time point (Week 0, 1, 2) and rearing condition and included all pair bonding behaviors and the bootstrapped preference score. We selected the component (in all cases, the first or second PC), with the highest loading of preference. We then compared the loading matrices for each of these components across time and rearing condition. All statistical analyses were completed using JMP Statistical Processing Software (SAS, Cary, NC, USA) or custom-written Matlab code (Mathworks, Natick, MA).

1. Riebel K. 2003 Developmental influences on auditory perception in female zebra finches - Is there a sensitive phase for song preference learning? Anim. Biol. 53, 73–87. (doi:10.1163/157075603769700304)

2. Tchernichovski O, Mitra PP, Lints T, Nottebohm F. 2001 Dynamics of the Vocal Imitation Process: How a Zebra Finch Learns Its Song. science 291, 2564–2569.

3. Barr HJ, Wall EM, Woolley SC. 2021 Dopamine in the songbird auditory cortex shapes auditory preference. Curr. Biol. 31, 4547-4559.e5. (doi:10.1016/j.cub.2021.08.005)

4. Chen Y, Clark O, Woolley SC. 2017 Courtship song preferences in female zebra finches are shaped by developmental auditory experience. Proc. Biol. Sci. 284, 20170054. (doi:10.1098/rspb.2017.0054)

5. Schubloom HE, Woolley SC. 2015 Variation in Social Relationships relates to song preferences and ERG1 expression in a female songbird. Dev. Neurobiol. , 1–36. (doi:10.1002/dneu.)

6. Woolley SC, Doupe AJ. 2008 Social Context–Induced Song Variation Affects Female Behavior and Gene Expression. PLoS Biol. 6, e62. (doi:10.1371/journal.pbio.0060062)

7. Tchernichovski O, Nottebohm F, Ho CE, Pesaran B, Mitra PP. 2000 A procedure for an automated measurement of song similarity. Anim. Behav. 59, 1167–1176. (doi:10.1006/anbe.1999.1416)

8. Campbell DLM, Hauber ME. 2010 Behavioural correlates of female zebra finch (Taeniopygia guttata) responses to multimodal species recognition cues. Ethol. Ecol. Evol. 22, 167–181. (doi:10.1080/03949371003707885)

9. Klatt JD, Goodson JL. 2013 Sex-specific activity and function of hypothalamic nonapeptide neurons during nest-building in zebra finches. Horm. Behav. 64, 818–824. (doi:10.1016/j.yhbeh.2013.10.001)

10. Silcox AP, Evans SM. 1982 Factors affecting the formation and maintenance of pair bonds in the zebra finch, Taeniopygia guttata. Anim. Behav. 30, 1237–1243. (doi:10.1016/S0003-3472(82)80216-9)

11. Zann RA. 1996 The Zebra Finch: A Synthesis of Field and Laboratory Studies. Oxford University Press.

12. Kao MH, Brainard MS. 2006 Lesions of an Avian Basal Ganglia Circuit Prevent Context-Dependent Changes to Song Variability. J. Neurophysiol. 96, 1441–1455. (doi:10.1152/jn.01138.2005)