Statistical learning in songbirds: from self-tutoring to song culture

Abstract

At the onset of vocal development, both songbirds and humans produce variable vocal babbling with broadly distributed acoustic features. Over development, these vocalizations differentiate into the well-defined, categorical signals that characterize adult vocal behaviour. A broadly distributed signal is ideal for vocal exploration, that is, for matching vocal production to the statistics of the sensory input. The developmental transition to categorical signals is a gradual process during which the vocal output becomes differentiated and stable. But does it require categorical input? We trained juvenile zebra finches with playbacks of their own developing song, produced just a few moments earlier, updated continuously over development. Although the vocalizations of these self-tutored (ST) birds were initially broadly distributed, birds quickly developed categorical signals, as fast as birds that were trained with a categorical, adult song template. By contrast, siblings of those birds that received no training (isolates) developed phonological categories much more slowly and never reached the same level of category differentiation as their ST brothers. Therefore, instead of simply mirroring the statistical properties of their sensory input, songbirds actively transform it into distinct categories. We suggest that the early self-generation of phonological categories facilitates the establishment of vocal culture by making the song easier to transmit at the micro level, while promoting stability of shared vocabulary at the group level over generations.This article is part of the themed issue 'New frontiers for statistical learning in the cognitive sciences'.

Keywords: birdsong; categorical signal; development; isolate song; song culture; statistical learning.

Figure 1.

Transition from continuous to categorical signal over development. (a) Sonogram of a juvenile zebra finch (day 40) showing highly variable, undifferentiated syllables. (b) Adult song of the same zebra finch showing stable syllable types falling into distinct categories. Introductory syllables are denoted by i; letters stand for song syllables that are repeated in a fixed order (ABCDE…). (c) Development of syllable types (categorical signal) in a WT bird. Dots represent song syllables: mean frequency modulation (y axis) is plotted against syllable duration (x axis, both normalized). Bottom panel represents distribution of syllables at the subsong stage, top panel at adulthood.

Figure 2.

Self-tutoring in zebra finches. (a) ISO siblings are raised in social isolation without any song input; their vocalizations are recorded continuously. (b) ST birds are also socially isolated and recorded during the entire song development, but in addition, they learn to peck on a key (red button) that induces playback of a song randomly selected from their own recent vocalizations, repeated recursively over development.

Figure 3.

Adult songs of ISO and ST (ST) sibling pairs. Sonograms of three pairs are shown (a–c). Songs of the ISO brothers are above their ST siblings'. Song motifs are underlined in yellow, song syllables in blue. ST motifs are more stable (repeated in the same way every time), composed of more syllable types and are generally longer. ST syllable durations are similar to the WT durations; ISO durations tend to be too short or too long.

Figure 4.

Syllables form tight clusters in WT and ST birds but not in ISOs. (a) Cluster formation in an ISO bird (left column), his ST brother (right column). Dots represent song syllables: mean frequency modulation is plotted against syllable duration (both normalized). The top panels represent adult songs, the bottom ones the beginning of development, and the middle row a stage in early development when ST birds have already begun to produce stable and tightly clustered syllables not present in ISOs. (b) Median Euclidean distance between neighbouring syllables. Over song ontogeny, syllables of the same type become more similar in every group, but this happens faster and earlier in self- and WT birds while the ISOs never from tight clusters.

Figure 5.

Distribution of ISO, ST and WT song features. (a) Song features of ST songs (green lines) are more WT-like (blue lines) than ISO song features (red lines): their cumulative distributions lie in-between the WT and the ISO distributions. (b) KS distance for combined acoustic features for the three possible comparisons between ISO, ST and WT songs.