Long memory in song learning by zebra finches

Abstract

Young songbirds use memorized tutor songs as templates to shape their own songs. This process requires control of voice by auditory feedback. We prevented zebra finches from hearing their own vocalizations by exposure to loud noise after 35 d of age, before which they had been reared with song tutors from birth. When the noise stopped at 102-200 d of age, the birds sang unstable and noisy song syllables that did not resemble the tutor syllables. The similarity to the tutor syllables steadily increased until the time of song crystallization approximately 30 d later. These findings show that the memory of tutor syllables survives auditory perturbations during the period when it is normally recalled and that zebra finches can use the memory well after the normal period of song development. The temporal order of syllables resembled the tutor model only in birds released from the noise before 80 d of age but not in older birds. Thus, different schedules and processes may govern the learning of syllable phonology and syntax.

Figure 4.

Time course of changes in similarity in syllable structure. a, All the birds in both control and experimental groups were reared by their biological (zebra finch) or foster (Bengalese finch) parents until 35 d of age. Then, the control birds were individually isolated to develop their song from memory. The Exp-birds were exposed to continuous broadband noise in individual sound isolation boxes between 35 and 129 ± 27 d of age. After their release from the noise, all of these birds developed syllables resembling the tutor model. Open symbols show the similarity score at each sampling day in the controls, and closed symbols show that in the Exp-birds. The same symbols and lines are used for multiple pupils raised by the same tutor, including two Bengalese and five zebra finch tutors. b, Summary of similarities in syllables. This figure plots the means and SDs of the curves in a to show the general trends. Although the Exp-birds had a lower mean similarity at the time of song crystallization than the controls, the similarity increased at almost the same rate as in the controls.

Figure 1.

Changes in hearing thresholds after release from noise. Hearing thresholds of Exp-birds (n = 14) to pure tones (0.5, 1, 2, 4, and 8 kHz) were measured behaviorally and compared with those in control birds (n = 9). Plotted (symbols and bars) are the means and SDs. Symbols are explained in the inset. A2, 2 d after release from noise. The speed of recovery was different for different frequencies: the mean threshold for 1 kHz was different from that of the controls on the day of release from the noise but no longer so 2 d after release, whereas the mean thresholds for 2 and 4 kHz took 4 d to become indistinguishable from those of the controls.

Figure 2.

Songs recorded shortly after release from noise. Birds heard tutor songs until 35 d of age and were exposed to noise for different durations. Songs recorded within 3 d after release from noise consisted of syllables and silent intervals spaced more or less regularly. The syllables were noisy and lacked most of the elaborate sound patterns contained in the tutor songs. These abnormalities appeared regardless of the age of release from noise: 75 d for Green 18, 106 d for Purple 20, 109 d for Gold 1, and 122 d for Black 1.

Figure 3.

Gradual emergence of memorized tutor syllable structures. Unstable and noisy syllables became more stable and acquired distinct forms with time after release from noise. a, Bengalese tutor song. b, Noisy song of a zebra finch pupil sung 2 d after release (A2), the mean similarity score (%) seldom becomes 0 because of resemblances by chance. c, Song of the same bird 7 d after release (A7). d, Thirty-eight days after release. Note how the mean similarity increases with time. e, Song of a zebra finch pupil that heard the same Bengalese tutor, but not exposed to noise. f, Song of a zebra finch for comparison.

Figure 5.

Learning of syllable sequences. a, Song of Bengalese 1, which tutored all the birds in this figure until 35 d of age. b, The control bird closely copied not only the tutor's syllables but also their sequences. c, This zebra finch also reproduced the tutor syntax quite well after exposure to noise from 35 to 75 d of age. d, Noise exposure from 35 to 112 d of age decreased this bird's ability to reproduce the tutor syntax, although the bird did copy syllable phonology quite well. Letters above syllables (such as a, b, c) indicate the pupils' copies of the tutor's syllables A, B, C, respectively. The pupils' syllables that do not resemble any of the tutor's syllables are indicated by other letters, such as u. The numbers, such as 0.276, indicate the degree of similarity to the tutor syntax, 1.0 being perfect copying.

Figure 6.

Inability of experimental birds to reproduce tutor syllable sequences. a, The mean similarity score of the Exp-birds was significantly lower than that of controls and similar to the random level (p = 0.9474). Dashed line indicates random level. b, The rate of syllable repetitions was different between the controls and Exp-birds. c, Removal of repeated syllables reduces the syntactical differences. Circles are means, and error bars are SDs. d, Similarity to tutor syntax declines sharply when birds were released from noise after 80 d of age. All 14 birds shown in d were raised by the same foster parent (Bengalese 1). Open circles, Control birds; filled circles, Exp-birds; crosses, Exp2-birds.