Comparisons of different methods to train a young zebra finch (Taeniopygia guttata) to learn a song
Highlights
► Social aspects are predominant in the success outcome of song learning in finches. ► The one-to-one tutoring guarantees close-to-perfect song imitation in the Zebra Finch. ► Passive playback of song model is not effective for vocal learning in the Zebra Finch. ► Vocal changes occur in adulthood when the training method is inappropriate.
Introduction
Local dialects or geographical variation provided an early clue that birdsong might be learned. But it was when birds were raised in controlled conditions in the laboratory and their vocalizations were analyzed with the spectrograph that it was clearly demonstrated (Catchpole and Slater, 2008, Zeigler and Marler, 2008). Thorpe (1954) was the first to demonstrate that young chaffinches (Fringilla coelebs) raised in social isolation failed to produce the normal song characteristic of their species. Since Thorpe’s pioneering works in the fifties, young oscine songbirds of several dozens of species have been brought into the laboratory and reared in strictly controlled auditory environments to examine how their experience affected their songs. In nearly all cases, young songbirds deprived of song models develop abnormal songs, although some species-specific song features may be preserved (Brainard and Doupe, 2002). It was discovered that some species learn song only during a sensitive phase of early life whereas others have the possibility to transform their songs during their whole life (Brainard and Doupe, 2002, Catchpole and Slater, 2008). Songbird species differ also in their readiness to copy songs from tape-recordings and from live tutors (for a review, see Baptista and Gaunt, 1997). Some bird species learn conspecific songs from a loudspeaker easily (canaries Serinus canaria: Mundinger, 1995; chaffinches F. coelebs: Thorpe, 1958; white-crowned sparrows Zonotrichia leucophrys: Marler, 1970; nightingales Luscinia megarhynchos: Todt et al., 1979; marsh wrens Cistothorus palustris: Brenowitz et al., 1995; swamp sparrows Melospiza Georgiana: Marler and Peters, 1988; song sparrows Melospiza melodia: Marler and Peters, 1987; red-winged blackbirds Agelaius phoeniceus: Marler et al., 1972). In other species (Treecreepers Certhia spp.: Thielke, 1970; Bullfinches Pyrrhula pyrrhula: Nicolai, 1959), tutoring by tape-recording appears to be virtually ineffective. Passive playback of tape-recordings has several advantages: one can control the timing of sound exposure, the quality and the amount of songs that are broadcast to the young bird (Derégnaucourt, 2011). This method was useful to show that some bird species do learn songs during particular time windows known as sensory phases (Kroodsma and Pickert, 1980). Nevertheless, it has been shown that the timing of song acquisition in birds trained by passive playback could be extended beyond the limits if the young bird is subsequently exposed to live tutors (Baptista and Petrinovitch, 1984).
Several studies show that mere proximity, that is, spatial association, is necessary for song sharing (Hausberger et al., 1995) and social learning (Mann and Slater, 1995). Song acquisition processes also depend on the biology of each species. For example, it is likely that learning from tapes is possible in species that learn their song from their neighbors under natural conditions, with whom they do not interact physically and/or visually (Derégnaucourt, 2011).
Among songbirds, song acquisition in the zebra finch (Taeniopygia guttata) has received a lot of interest. This tiny bird is easy to breed in captivity. Its song is short (1–1.5 s) and very stereotyped. Young male zebra finches learn their song during a sensitive period of early life and it is believed that once the bird is about 90 days old, its song crystallizes and will not change for the remainder of its life (Zann, 1996). Early results suggested that young zebra finches would not learn from passive playback of tape recordings (Immelmann, 1969, Eales, 1989). However, significant learning was observed in subsequent studies despite high inter-individual variability (Bolhuis et al., 1999, Houx and ten Cate, 1999a, Houx and ten Cate, 1999b, Houx et al., 2000, Funabiki and Funabiki, 2009).
Since Immelman’s pioneering study in 1969, it has been assumed that the most important factor in the selection of a zebra finch song model is the close interaction between a young male and the male that rears him, generally his father (Price, 1979, Böhner, 1983). Indeed, this bond will even override species-specificity since zebra finches will learn the song of a Bengalese finch (Lonchura striata) foster father, even if conspecifics share their breeding accommodation. Nevertheless, in controlled conditions, a young male can also learn the song of an adult male with whom he does not have physical contact or social attachment such as what develops between youngsters and the male that is feeding them (Eales, 1985). Still, both visual and vocal interaction between young zebra finches and their tutors are important for the accuracy of song learning (Eales, 1989).
If fairly complete song imitation in the zebra finch has never been obtained with passive exposure to a song model, such results have been obtained using operant conditioning with song as a reward (Adret, 1993, Tchernichovski et al., 2001, Derégnaucourt et al., 2005).
There might be several reasons why passive playback failed so far to induce significant song learning in the zebra finch. First, the song playback sessions might be not done at the appropriate moment for the birds. Second, the playback quota might not be sufficient to capture the bird’s attention. On several occasions investigators have emphasized that very good imitation is possible even when a model was presented a small number of times (Petrinovitch, 1985, Hultsch and Todt, 1989, Peters et al., 1992). In zebra finches, in an operant conditioning context when the bird has to peck a key to trigger song playback, 40 playbacks of the song motif per day, lasting a total of 30 s, result in significant learning. More exposure leads to less complete imitation (Tchernichovski et al., 1999).
In our study we decided to test these hypotheses. Accordingly, we decided to play back song at moments that are predictable for the birds, namely, just after the lights switch on in the morning and just before the lights switch off in the evening. Also, we restricted the quota of daily song exposure. We will compare these results with those of birds trained with a procedure of operant conditioning, and also those of live-tutored birds.
Hormonal status is closely correlated to the end of the sensitive period for song learning and the crystallization of a stereotyped song in both open-ended (Marler et al., 1987) and age-limited song learners like the zebra finch (Pröve, 1983). At dph 101, we measured testosterone levels and calculated correlations with song imitation, song stereotypy and singing activity. Early castration (Arnold, 1975a, Bottjer and Hewer, 1992) and exogenous testosterone (Korsia and Bottjer, 1991) both disrupt song learning in the zebra finch. In the same species, several studies suggest that high testosterone concentrations reduce the potential for vocal plasticity (Korsia and Bottjer, 1991, Williams et al., 2003). We therefore expected that birds with low levels of testosterone would be less likely to have a stereotyped song. Also, castration of male finches decreases song rates, but normal singing activity is re-established after the castrated birds have been injected with testosterone (Arnold, 1975b, Pröve, 1974). In our study, we checked whether there could be a link between testosterone levels and singing activity at dph 100.
After dph 101, the birds were placed in a common aviary with conspecifics. Several weeks later, their song was recorded again to quantify eventual vocal changes.
Section snippets
Study subjects
Sixty-two male zebra finches from 16 different breeding pairs of our colony were used for this study. Breeding pairs were maintained in separate cages (100 × 50 × 50 cm) and nest control was made daily. The chicks were raised by both parents until dph 10 when the father was removed. The mother and the chicks were moved to another room in our institute, to avoid any imprinting of the young on male song. The breeding cages were maintained on LD 12:12 until the chicks fledged.
Experimental groups
On day 35, each chick was
Song imitation
We observed significant differences among tutoring regimes (Fig. 2A, one-way ANOVA, F(2, 59) = 15.691, P < 0.0001). Birds trained with a live tutor learned better than birds trained with operant conditioning (Tukey HSD, P = 0.005) and passive playback (P < 0.001). Birds trained with operant conditioning did also learn better than birds trained with passive playback (P = 0.004). Close-to-perfect imitation of the song model was observed in finches trained with live tutoring and operant conditioning but not
Discussion
These results indicate that young male zebra finches learn poorly from passive tape playback even when timing and quota of song exposure are controlled. Better results are obtained when the bird can self-elicit tape playback using an operant conditioning procedure, despite a high inter-individual variability. One-to-one live tutoring is clearly the most effective method to get a close-to-perfect imitation of a song model. Moreover, the less the bird learned during the training period, the more
Acknowledgements
We would like to thank Arthur Leblois for having organized the workshop and for giving us the opportunity to participate in this special issue. We thank Wolfgang Goymann, Ingrid Schwabl and Monika Trappschuh for hormone analysis. We thank Nicole Geberzahn, Ofer Tchernichovski and Olga Fehér for their valuable comments on a previous draft of the manuscript. We thank Benjamin Treuhaft for reviewing our English manuscript. This research was supported by the Max-Planck Society.
References (71)
Operant conditioning, song learning and imprinting to taped song in the zebra finch
Anim. Behav.
(1993)- et al.
Social interaction, sensitive phases and the song template hypothesis in the white-crowned Sparrow?
Anim. Behav.
(1984) - et al.
Song development in the white-crowned sparrow: social factors and sex differences
Anim. Behav.
(1986) Song learning in the zebra finch (Taeniopygia guttata): selectivity in the choice of a tutor and accuracy of song copies
Anim. Behav.
(1983)- et al.
Song learning with audiovisual compound stimuli in zebra finches
Anim. Behav.
(1999) - et al.
Song acquisition in European starlings: a comparison of the songs of live-tutored, tape-tutored, untutored, and wild-caught males
Anim. Behav.
(1993) Song learning in zebra finches: some effects of song model availability on what is learnt and when
Anim. Behav.
(1985)- et al.
Song learning in captive European starlings, Sturnus vulgaris
Anim. Behav.
(1992) - et al.
Latent song type memories are accessible through auditory stimulation in a hand-reared songbird
Anim. Behav.
(2002) - et al.
Testosterone and corticosterone during the breeding cycle of equatorial and European stonechats (Saxicola torquata axillaris and S. t. rubicola)
Horm. Behav.
(2006)
Auditory-visual stimulus pairing enhances perceptual learning in a songbird
Anim. Behav.
Song tutor choice by zebra finches in aviaries
Anim. Behav.
Effects of acoustical stimulation and deprivation on song development in red-winged blackbirds (Agelaius phoeniceus)
Anim. Behav.
Behaviour-genetic analysis of canary song: inter-strain differences in sensory learning, and epigenetic rules
Anim. Behav.
Song and female choice in zebra finches: a review
Adv. Stud. Behav.
A procedure for an automated measurement of song similarity
Anim. Behav.
Female canaries are sexually responsive to special song phrases
Anim. Behav.
Using robots to understand animal behavior
Adv. Stud. Behav.
Testosterone decreases the potential for song plasticity in adult male zebra finches
Horm. Behav.
The effects of castration on song development in zebra finches (Poephila guttata)
J. Exp. Zool.
The effects of castration and androgen replacement on song, courtship, and aggression in zebra finches (Poephila guttata)
J. Exp. Zool.
Social interaction and vocal development in birds
Castration and antisteroid treatment impair vocal learning in male zebra finches
J. Neurobiol.
What songbirds teach us about learning
Nature
Brain space for learned song in birds develop independently of song learning
J. Neurosci.
Bird Song: Biological Themes and Variations
Birdsong learning in the laboratory, with especial reference to the song of the zebra finch (Taeniopygia guttata)
Interact. Stud.
How sleep affects the developmental learning of bird song
Nature
The influences of visual and vocal interaction on song learning in zebra finches
Anim. Behav.
Factors limiting song acquisition in adult zebra finches
Dev. Neurobiol.
Early condition, song learning, and the volume of song brain nuclei in the zebra finch (Taeniopygia guttata)
J. Neurobiol.
Birdsong memory: a neural dissociation between song recognition and production
Curr. Biol.
Incomplete and inaccurate vocal imitation after knockdown of FoxP2 in songbird basal ganglia nucleus area X
PLoS Biol.
Social influences on song acquisition and sharing in the European starling (Sturnus vulgaris)
Song sharing reflects the social organization in a captive group of European starlings (Sturnus vulgaris)
J. Comp. Psychol.
Cited by (62)
A neuroethological view of the multifaceted sensory influences on birdsong
2024, Current Opinion in NeurobiologyCuriosity constructs communicative competence through social feedback loops
2023, Advances in Child Development and BehaviorMultisensory processes in birds: From single neurons to the influence of social interactions and sensory loss
2022, Neuroscience and Biobehavioral ReviewsCitation Excerpt :In an attempt to control for these factors, a focus on basic perceptual mechanisms indicated that, in nightingales (Luscinia megarhynchos L.), juveniles can be tutored with a tape but their vocal learning performance is improved when strobe flashes are tuned to be synchronous with song notes (Hultsch et al., 1999). Other studies have addressed the issue by examining song acquisition in young birds that are allowed either to visually and acoustically interact with a live tutor, or to other conditions that include interacting only acoustically with a tutor or passively hearing the adult model (Chen et al., 2016; Derégnaucourt et al., 2013; Varkevisser et al., 2022). Tutoring for as little as one day in zebra finches is sufficient to highlight the benefit of multimodal information on song learning: interacting both acoustically and visually with tutors made songs more similar to the tutor’s songs than interacting only acoustically or passively hearing the adult models (Chen et al., 2016).
Multimodality during live tutoring is relevant for vocal learning in zebra finches
2022, Animal BehaviourThe neurobiology of innate and learned vocalizations in rodents and songbirds
2020, Current Opinion in NeurobiologyContingent parental responses are naturally associated with zebra finch song learning
2020, Animal Behaviour