Quantification of developmental birdsong learning from the subsyllabic scale to cultural evolution

Abstract

Quantitative analysis of behavior plays an important role in birdsong neuroethology, serving as a common denominator in studies spanning molecular to system-level investigation of sensory-motor conversion, developmental learning, and pattern generation in the brain. In this review, we describe the role of behavioral analysis in facilitating cross-level integration. Modern sound analysis approaches allow investigation of developmental song learning across multiple time scales. Combined with novel methods that allow experimental control of vocal changes, it is now possible to test hypotheses about mechanisms of vocal learning. Further, song analysis can be done at the population level across generations to track cultural evolution and multigenerational behavioral processes. Complementing the investigation of song development with noninvasive brain imaging technology makes it now possible to study behavioral dynamics at multiple levels side by side with developmental changes in brain connectivity and in auditory responses.

Fig. 1.

Adult and juvenile song. (A) Spectral derivatives (4) provide a high-definition alternative to the classic sound spectrogram of a song of an adult (>100-d-old) zebra finch male. The song has a complex yet highly stereotyped structure of short sounds (syllables, designated by letters) arranged in repeating sequences (motifs). The song motifs are aligned below each other, showing the high stereotypy of their performance. Red dashed line indicates point of alignment. (B) Sound spectrogram of the vocalization (subsong) of a juvenile (40-d-old) zebra finch male. The sounds are complex, but not stereotyped. Vocal state keeps changing, but no obvious units, such as syllable types or motifs, can be detected.

Fig. 2.

Vocal and auditory pathways of the songbird brain. (A) The vocal pathway, which consists of a posterior pathway necessary for producing learned song (brain nuclei composing it are shown in dark gray, and connections between them denoted with black arrows); and an anterior pathway necessary for song learning (brain nuclei composing it are shown in white, and connections between them denoted with white arrows). Dashed lines indicate connections between the two pathways. (B) The auditory pathway. Av, avalanche; CLM, caudal lateral mesopallium; CMM, caudal medial mesopallium; CN, cochlear nucleus; CSt, caudal striatum; DM, dorsal medial nucleus; DLM, dorsal lateral nucleus of the medial thalamus; E, entopallium; B, basorostralis; LLD, lateral lemniscus, dorsal nucleus; LLI, lateral lemniscus, intermediate nucleus; LLV, lateral lemniscus, ventral nucleus; MLd, dorsal lateral nucleus of the mesencephalon; LMAN, lateral magnocellular nucleus of the anterior nidopallium; area X, area X of the medial striatum; MO, oval nucleus of the mesopallium; NCM, caudal medial nidopallium; NIf, nucleus interface of the nidopallium; nXIIts, nucleus XII, tracheosyringeal part; Ov, ovoidalis; PAm, paraambiguus; RAm, retroambiguus; RA, robust nucleus of the arcopallium; SO, superior olive; Uva, nucleus uvaeformis. [Reproduced with permission from ref. (Copyright 2004, The New York Academy of Sciences).]

Fig. 3.

Quantification of song development. (A) Features used for continuous description of song (7). Images on each side of the arrows demonstrate extreme values for each feature. (B) An example of an analyzed song rendition (spectrogram). The features shown in A are calculated for each millisecond of song (blue line, amplitude). Syllables (green) are defined by segmenting the amplitude time series with a certain threshold. Each syllable is characterized by its duration and its mean features values. The values for two syllables (in the red frames) are highlighted below. (C) The distribution of two of these mean values (entropy and FM) in the vocalizations of a juvenile (45 d old; Left) and of the same bird at a later developmental stage (90 d old; Right). The variable and unstructured vocalizations of the young bird are reflected in the broad features distribution, where no distinct clusters can be seen. The distribution at an older age reveals the emergence of clusters (syllable types).

Fig. 4.

Altered-target training. Sound spectrograms of a bird's own song and of the source and target tutor songs (red frames). A juvenile male (43 d old; Upper) is trained with a source tutor song consisting of a single syllable type (AAAA). Once the bird has learned it (at the age of 55 d), training is switched to the target tutor song, in which a new syllable type is alternated with the old one (ABAB). This bird learned the target song on day 90 posthatch.

Fig. 5.

The developmental trajectory of birds trained with altered-target training. (A) Sonograms of songs from different developmental days in two birds trained with AAAA→ABAB paradigms. The new syllable type (syllable B) appears first at one or both bout edges, and later in the center of the bout. Note that the two birds were trained with altered-target paradigms with a different syllable B—a harmonic stack in bird 1 and a broadband sound in bird 2. (B) Raster plots of bout edges for bird 1. Each line represents a song bout. Color indicates goodness of pitch at each millisecond of song. Bouts are arranged by temporal order from top to bottom; x axis, song time in ms; y axis, age in days. Bouts are aligned on the first rendition (Left) of syllable A and on its last rendition (Right). Alignment points are indicated by green arrows. The edges of song bouts are defined as the sounds before the first and after the last rendition of syllable A. Syllable B, which has high goodness of pitch, appears in red. It can be seen that the new syllable (syllable B) appears at the edges of bouts first and then at its center.

Fig. 6.

Evolution of song culture. (A, Upper) Two examples of WT zebra finch songs. (Lower) Two examples of songs of birds reared in isolation (isolates). Colored lines indicate different syllable types within each bird's song. Despite variability among individuals, isolate songs differ from WT songs in syllable structure, duration, and order. (B) Experimental setup. An isolate tutor is kept singly with a pupil, who later serves as a tutor to the next pupil, and so on. (C, Upper) Example of a song of an isolate tutor containing an abnormally long syllable (red frame). (Lower) The same syllable type was shortened in the pupil's song. (D, Upper) Example of a song of another isolate tutor, containing back-to-back renditions of syllable types (marked by letters). (Lower) The pupil copied the syllable types but arranged them in a different, motif-like, order. Reproduced with permission from ref. .

Fig. 7.

Quantification of convergence from isolate to WT song across generations. (A) Distribution of song features in WT songs and the songs’ isolates and their pupils. Axes, the first two principal components (PC1 and PC2) derived from PCA of spectral features. The features distribution of WT songs is shown in purple. Shades of purple indicate the distance from the center of the WT cluster (darker is closer to center). Arrows of the same color indicate a certain learning dynasty (i.e., an isolate tutor and his pupils). (Left) Arrows connect each isolate tutor to its first generation pupils, showing pupils to be more WT-like. (Right) Arrows connect pupils from successive learning generations, showing a continued convergence toward WT-like song features. (B) Examples of a song syllable from an isolate tutor, and of successive generations of its pupils (from top to bottom). An abnormally long part of the syllable (red frames) is gradually shortened over generations. Reproduced with permission from ref. .

Fig. 8.

Differences in BOLD activation to different stimuli. (A) BOLD activation maps for bird's own song (BOS) and repeated syllable (SYLL1) in colony males show stronger activation to BOS. Each map shows average activity in left and right hemisphere 0.5 mm from midline; color scale shows correlation coefficient. (B) Box plot summarizing BOLD activation across stimuli in colony males. Area of activation varies significantly across stimuli with greater activation to songs (ANOVA, P < 0.01). Stimuli: 2 kHz pure tone, conspecific song (CON), BOS, tutor song (TUT), and song syllables (SYLL 1, SYLL 2). All conspecific songs were produced by unfamiliar, colony-raised birds. (C and D) Same as A and B for box-trained males (P < 0.05). (E and F) Same as A and B for isolated males (P = 0.9; the TUT stimulus is absent in F, because isolates were not exposed to a tutor). (G and H) Same as A and B for isolate females (P < 0.05) with an additional conspecific song (CON2) to balance cross-group comparisons. Reproduced from ref. with permission from John Wiley & Sons, Inc. (Copyright 2010).