Temporal hierarchical control of singing in birds

Science

Volumn 273, Issue 5283, 1996, Pages 1871-1875

  Yu, Albert C ^a   Margoliash, Daniel ^b  

^a UNIVERSITY OF CHICAGO   (United States)

^b UNIVERSITY OF CHICAGO   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

ANIMAL EXPERIMENT; ARTICLE; BIRD; CONTROLLED STUDY; FOREBRAIN; MALE; MESENCEPHALON; NONHUMAN; PREMOTOR CORTEX; PRIORITY JOURNAL; SINGING; SINGLE UNIT ACTIVITY; SPIKE WAVE; VOCALIZATION; VOICE;

EID: 0029746196     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.273.5283.1871     Document Type: Article

References (40)

1. K. S. Lashley, in Cerebral Mechanisms in Behavior, L A. Jeffress, Ed. (Wiley, New York. 1951), pp. 112-135; M. Konishi, in Perception, vol. VIII of Handbook of Sensory Physiology, R. Held, H. W. Leibowitz, H.-L. Teuber, Eds. (Springer-Verlag, Berlin, 1978), pp. 289-309; P. Marler and S. Peters, Science 198, 519 (1977); P. Marler and R. Pickert, Anim. Behav. 32, 673 (1984); H. C. Gerhardt, J. Exp. Biol. 74, 59 (1978); R. R. Hoy and R. C. Paul, Science 180, 82 (1973).
2. K. S. Lashley, in Cerebral Mechanisms in Behavior, L A. Jeffress, Ed. (Wiley, New York. 1951), pp. 112-135; M. Konishi, in Perception, vol. VIII of Handbook of Sensory Physiology, R. Held, H. W. Leibowitz, H.-L. Teuber, Eds. (Springer-Verlag, Berlin, 1978), pp. 289-309; P. Marler and S. Peters, Science 198, 519 (1977); P. Marler and R. Pickert, Anim. Behav. 32, 673 (1984); H. C. Gerhardt, J. Exp. Biol. 74, 59 (1978); R. R. Hoy and R. C. Paul, Science 180, 82 (1973).
3. K. S. Lashley, in Cerebral Mechanisms in Behavior, L A. Jeffress, Ed. (Wiley, New York. 1951), pp. 112-135; M. Konishi, in Perception, vol. VIII of Handbook of Sensory Physiology, R. Held, H. W. Leibowitz, H.-L. Teuber, Eds. (Springer-Verlag, Berlin, 1978), pp. 289-309; P. Marler and S. Peters, Science 198, 519 (1977); P. Marler and R. Pickert, Anim. Behav. 32, 673 (1984); H. C. Gerhardt, J. Exp. Biol. 74, 59 (1978); R. R. Hoy and R. C. Paul, Science 180, 82 (1973).
4. K. S. Lashley, in Cerebral Mechanisms in Behavior, L A. Jeffress, Ed. (Wiley, New York. 1951), pp. 112-135; M. Konishi, in Perception, vol. VIII of Handbook of Sensory Physiology, R. Held, H. W. Leibowitz, H.-L. Teuber, Eds. (Springer-Verlag, Berlin, 1978), pp. 289-309; P. Marler and S. Peters, Science 198, 519 (1977); P. Marler and R. Pickert, Anim. Behav. 32, 673 (1984); H. C. Gerhardt, J. Exp. Biol. 74, 59 (1978); R. R. Hoy and R. C. Paul, Science 180, 82 (1973).
5. K. S. Lashley, in Cerebral Mechanisms in Behavior, L A. Jeffress, Ed. (Wiley, New York. 1951), pp. 112-135; M. Konishi, in Perception, vol. VIII of Handbook of Sensory Physiology, R. Held, H. W. Leibowitz, H.-L. Teuber, Eds. (Springer-Verlag, Berlin, 1978), pp. 289-309; P. Marler and S. Peters, Science 198, 519 (1977); P. Marler and R. Pickert, Anim. Behav. 32, 673 (1984); H. C. Gerhardt, J. Exp. Biol. 74, 59 (1978); R. R. Hoy and R. C. Paul, Science 180, 82 (1973).
6. K. S. Lashley, in Cerebral Mechanisms in Behavior, L A. Jeffress, Ed. (Wiley, New York. 1951), pp. 112-135; M. Konishi, in Perception, vol. VIII of Handbook of Sensory Physiology, R. Held, H. W. Leibowitz, H.-L. Teuber, Eds. (Springer-Verlag, Berlin, 1978), pp. 289-309; P. Marler and S. Peters, Science 198, 519 (1977); P. Marler and R. Pickert, Anim. Behav. 32, 673 (1984); H. C. Gerhardt, J. Exp. Biol. 74, 59 (1978); R. R. Hoy and R. C. Paul, Science 180, 82 (1973).
7. D. Ploog, Brain Res. Rev. 3, 35 (1981); T. J. Seller, Trends Neurosci. 4, 301 (1981).
8. D. Ploog, Brain Res. Rev. 3, 35 (1981); T. J. Seller, Trends Neurosci. 4, 301 (1981).
9. K. Immelmann, in Bird Vocalizations, R. A. Hinde, Ed. (Cambridge Univ. Press, London, 1969), pp. 61-74; R. A. Zann, The Zebra Finch (Oxford Univ. Press, Oxford, 1996).
10. K. Immelmann, in Bird Vocalizations, R. A. Hinde, Ed. (Cambridge Univ. Press, London, 1969), pp. 61-74; R. A. Zann, The Zebra Finch (Oxford Univ. Press, Oxford, 1996).
11. All procedures were approved by an institutional animal care committee. Under pentobarbital and chloral hydrate anesthesia, male zebra finches were implanted with headgear, including electrodes and electronics. After recovery, a bird typically participated in 6- to 8-hour recording sessions every 2 to 3 days We achieved stable chronic recordings in HVc by using bundles of Isonel-insulated microwires (eight recording sites, seven birds) or a custom-built 1-g mechanical microdrive to simultaneously move four Pt-1r electrodes (15 HVc recording sites in two birds, and 22 RA recording sites in four birds). There were no systematic differences noted in the activity patterns of HVc neurons recorded under the two conditions, and the data were combined in all analyses presented here. During a recording session, the bird was attached to a custom-built commutator by a flexible cable, permitting free exploration inside the cage. In some cases we collected data while zebra finches sang spontaneously; otherwise, females or mirrors were introduced into the adjacent half-cages to stimulate directed song [R. Sossinka and J. Böhner, Z. Tierpsychol. 53, 123 (1980)]. During data analysis, female calls could be distinguished from male calls and from male song syllables. After a bird was killed with an overdose of pentobarbital, the position of the recording sites within the target nuclei was confirmed with standard frozen-section histology.
12. S. E. Anderson, A. S. Dave, D. Margoliash, J. Acoust. Soc. Am. 100, 1209 (1996).
13. Single units were isolated (categorized) off-line according to established procedures [M. L. Sutter and D. Margoliash, J. Neurophysiol. 72, 2105 (1994); M. S, Lewicki, Neural Comp. 6. 1005 (1994)]. We constructed MAHs by aligning vocalizations and the related neuronal activity using the syllable onset or offset times, then binning the times of occurrence of action potentials. There were no systematic differences between onset and offset MAHs for either the HVc or RA data; we use onset MAHs in this report. The data shown here are from 40 well-isolated single units in the HVc and 23 well-isolated single units in the RA.
14. Single units were isolated (categorized) off-line according to established procedures [M. L. Sutter and D. Margoliash, J. Neurophysiol. 72, 2105 (1994); M. S, Lewicki, Neural Comp. 6. 1005 (1994)]. We constructed MAHs by aligning vocalizations and the related neuronal activity using the syllable onset or offset times, then binning the times of occurrence of action potentials. There were no systematic differences between onset and offset MAHs for either the HVc or RA data; we use onset MAHs in this report. The data shown here are from 40 well-isolated single units in the HVc and 23 well-isolated single units in the RA.
15. Pairs of MAHs were adjusted to the same duration with the use of time-warping decimation-interpolation techniques implemented using routines in the Matlab program (MathWorks, Natick, MA). The linear correlation coefficient r for two MAHs representing syllables x and y was calculated as equation presented where x, and y, represent individual histogram bin values, and x̄ and ȳ represent average bin values.
16. The MAHs were separated into two segments: one segment associated with shared note types, and one segment associated with dissimilar note types. The segmentation assumed that neuronal activity preceded the vocalizations by 40 ms for the RA recordings and by 50 ms for the HVc recordings. These values correspond to average multiunit latency preceding singing [J. S. McCasland and M. Konishi, Proc. Natl. Acad. Sci. U.S.A. 78, 7815 (1981)]
17. Spike trains were brought into temporal registry by cross-correlating the corresponding syllables, as follows. Introductory syllables were excluded because for the RA they are not associated with a precise temporal pattern of activity (Fig. 2A). Motif syllables comprising simple harmonic stacks were also excluded because they lack meaningful time-varying frequency modulation, which resulted in unreliable cross-correlations. This left acoustically complex syllables of motifs for analysis. The acoustic records were scored without reference to the associated spike trains to eliminate recordings with acoustic clutter (background cage noises, calls of females). Six additional syllables were eliminated on this basis because less than 10 acoustically uncontaminated exemplars were identified, preventing meaningful statistical analysis. For each of the 15 resultant syllable types (15 neurons, three birds), a "referent" syllable was chosen by manual inspection of spectrographs of the set of exemplar syllables. The spectrograph of each exemplar syllable was then crosscorrelated with the spectrograph of the referent [C. W. Clark, P. Marier, P. Beeman, Ethology 76, 101 (1987)]. We then adjusted the temporal registry of each spike train associated with each exemplar syllable, relative to the spike train associated with the referent syllable, by applying a shift in time (translation) based on the position of the peak in the correlation function. The time shifts were typically quite small (τ = 3.00 ± 2.31 ms, n = 1436 cross-correlations of referent and exemplar syllables), implying that the original manual segmentation was quite accurate; nevertheless, this shift significantly affected the temporal registry of spike trains.
18. An optimal translation of the spike trains was applied on a burst-by-burst basis to minimize the global difference in spike timing. That is, this procedure aligned each spike burst independent of the acoustics of associated notes. The acoustic procedure of (9) failed to further improve the temporal registry of spike bursts when applied on a note-by-note basis. This failure may be the result of the fine temporal resolution of RA neuronal discharge patterns over-whelming inherent limitations in time-frequency resolution in the calculation of spectrographs based on short-time Fourier transformations [G. D. Bergland, IEEE Spectrum 7,41 (1969)]. Nevertheless, it implies that we were not able to quantitatively demonstrate that the timing of each burst pattern was associated with the timing of each note. The RA exhibits a myotopic organization [D. S. Vicario, J. Neurobiol. 22, 63 (1991)], hence the activity of RA neurons may be associated with activation of individual muscles or groups of muscles.
19. An optimal translation of the spike trains was applied on a burst-by-burst basis to minimize the global difference in spike timing. That is, this procedure aligned each spike burst independent of the acoustics of associated notes. The acoustic procedure of (9) failed to further improve the temporal registry of spike bursts when applied on a note-by-note basis. This failure may be the result of the fine temporal resolution of RA neuronal discharge patterns over-whelming inherent limitations in time-frequency resolution in the calculation of spectrographs based on short-time Fourier transformations [G. D. Bergland, IEEE Spectrum 7,41 (1969)]. Nevertheless, it implies that we were not able to quantitatively demonstrate that the timing of each burst pattern was associated with the timing of each note. The RA exhibits a myotopic organization [D. S. Vicario, J. Neurobiol. 22, 63 (1991)], hence the activity of RA neurons may be associated with activation of individual muscles or groups of muscles.
20. S. Nowicki and R. R. Capranica, Science 231, 1297 (1986); F. Goller and R. A. Suthers, Nature 373, 63 (1995); M. W. Westneat, J. H. J. Long, W. Hoese, S. Nowicki, J. Exp. Biol. 182, 147 (1993).
21. S. Nowicki and R. R. Capranica, Science 231, 1297 (1986); F. Goller and R. A. Suthers, Nature 373, 63 (1995); M. W. Westneat, J. H. J. Long, W. Hoese, S. Nowicki, J. Exp. Biol. 182, 147 (1993).
22. S. Nowicki and R. R. Capranica, Science 231, 1297 (1986); F. Goller and R. A. Suthers, Nature 373, 63 (1995); M. W. Westneat, J. H. J. Long, W. Hoese, S. Nowicki, J. Exp. Biol. 182, 147 (1993).
23. E. E. Fetz, in Movement Control. P. Cordo and S. Harnad, Eds. (Cambridge Univ. Press, Cambridge, 1994), pp. 77-88.
24. H. Williams and D. Vicario, J. Neurobiol. 24, 903 (1993).
25. See also E. T. Vu, M. E. Mazurek, Y.-C. Kuo, J. Neurosci. 14, 6924 (1994).
26. N. Tinbergen, Symp. Soc. Exp. Biol. 4, 305 (1950); The Study of Instinct (Oxford Univ. Press, Oxford, 1951).
27. N. Tinbergen, Symp. Soc. Exp. Biol. 4, 305 (1950); The Study of Instinct (Oxford Univ. Press, Oxford, 1951).
28. R. Mooney, J. Neurosci. 12, 2464 (1992).
29. M. Abeles et al., Proc. Natl. Acad. Sci. U.S.A. 92, 8616 (1995); E. Vaadia et al., Nature 373, 515 (1995).
30. M. Abeles et al., Proc. Natl. Acad. Sci. U.S.A. 92, 8616 (1995); E. Vaadia et al., Nature 373, 515 (1995).
31. R. Llinás, Cold Spring Harbor Symp. Quant. Biol 55, 933 (1990).
32. C. E. Carr, W. Heiligenberg, G. J. Rose, J. Neurosci. 6, 107 (1986); C. E. Carr and M. Konishi, Proc. Natl. Acad. Sci. U.S.A. 85, 8311 (1988); E. Ahissar et al., Science 257, 1412 (1992); G. H. Recanzone, M. M. Merzenich, C. E. Schreiner, J. Neurophysiol. 67, 1071 (1992); J. C. Middlebrooks, A. E. Clock, L. Xu, D. M. Green, Science 264, 842 (1994); R. C. de Charms and M. M. Merzenich, Nature 381, 610 (1996); W. Bair and C. Koch, Neural Comp. 8, 1184 (1996).
33. C. E. Carr, W. Heiligenberg, G. J. Rose, J. Neurosci. 6, 107 (1986); C. E. Carr and M. Konishi, Proc. Natl. Acad. Sci. U.S.A. 85, 8311 (1988); E. Ahissar et al., Science 257, 1412 (1992); G. H. Recanzone, M. M. Merzenich, C. E. Schreiner, J. Neurophysiol. 67, 1071 (1992); J. C. Middlebrooks, A. E. Clock, L. Xu, D. M. Green, Science 264, 842 (1994); R. C. de Charms and M. M. Merzenich, Nature 381, 610 (1996); W. Bair and C. Koch, Neural Comp. 8, 1184 (1996).
34. C. E. Carr, W. Heiligenberg, G. J. Rose, J. Neurosci. 6, 107 (1986); C. E. Carr and M. Konishi, Proc. Natl. Acad. Sci. U.S.A. 85, 8311 (1988); E. Ahissar et al., Science 257, 1412 (1992); G. H. Recanzone, M. M. Merzenich, C. E. Schreiner, J. Neurophysiol. 67, 1071 (1992); J. C. Middlebrooks, A. E. Clock, L. Xu, D. M. Green, Science 264, 842 (1994); R. C. de Charms and M. M. Merzenich, Nature 381, 610 (1996); W. Bair and C. Koch, Neural Comp. 8, 1184 (1996).
35. C. E. Carr, W. Heiligenberg, G. J. Rose, J. Neurosci. 6, 107 (1986); C. E. Carr and M. Konishi, Proc. Natl. Acad. Sci. U.S.A. 85, 8311 (1988); E. Ahissar et al., Science 257, 1412 (1992); G. H. Recanzone, M. M. Merzenich, C. E. Schreiner, J. Neurophysiol. 67, 1071 (1992); J. C. Middlebrooks, A. E. Clock, L. Xu, D. M. Green, Science 264, 842 (1994); R. C. de Charms and M. M. Merzenich, Nature 381, 610 (1996); W. Bair and C. Koch, Neural Comp. 8, 1184 (1996).
36. C. E. Carr, W. Heiligenberg, G. J. Rose, J. Neurosci. 6, 107 (1986); C. E. Carr and M. Konishi, Proc. Natl. Acad. Sci. U.S.A. 85, 8311 (1988); E. Ahissar et al., Science 257, 1412 (1992); G. H. Recanzone, M. M. Merzenich, C. E. Schreiner, J. Neurophysiol. 67, 1071 (1992); J. C. Middlebrooks, A. E. Clock, L. Xu, D. M. Green, Science 264, 842 (1994); R. C. de Charms and M. M. Merzenich, Nature 381, 610 (1996); W. Bair and C. Koch, Neural Comp. 8, 1184 (1996).
37. C. E. Carr, W. Heiligenberg, G. J. Rose, J. Neurosci. 6, 107 (1986); C. E. Carr and M. Konishi, Proc. Natl. Acad. Sci. U.S.A. 85, 8311 (1988); E. Ahissar et al., Science 257, 1412 (1992); G. H. Recanzone, M. M. Merzenich, C. E. Schreiner, J. Neurophysiol. 67, 1071 (1992); J. C. Middlebrooks, A. E. Clock, L. Xu, D. M. Green, Science 264, 842 (1994); R. C. de Charms and M. M. Merzenich, Nature 381, 610 (1996); W. Bair and C. Koch, Neural Comp. 8, 1184 (1996).
38. C. E. Carr, W. Heiligenberg, G. J. Rose, J. Neurosci. 6, 107 (1986); C. E. Carr and M. Konishi, Proc. Natl. Acad. Sci. U.S.A. 85, 8311 (1988); E. Ahissar et al., Science 257, 1412 (1992); G. H. Recanzone, M. M. Merzenich, C. E. Schreiner, J. Neurophysiol. 67, 1071 (1992); J. C. Middlebrooks, A. E. Clock, L. Xu, D. M. Green, Science 264, 842 (1994); R. C. de Charms and M. M. Merzenich, Nature 381, 610 (1996); W. Bair and C. Koch, Neural Comp. 8, 1184 (1996).
39. To construct an eMAH, we defined a canonical song: the most common number of introductory syllables, the most common sequence of syllables within a motif, and the most common number of motifs. The eMAH was derived from concatenation of individual MAHs corresponding to each syllable type within its specific context of the canonical song-for example, all spikes corresponding to syllable E in the first motif or all spikes corresponding to syllable E in the second motif (Fig. 1B). Individual MAHs were calculated starting 50 ms before their corresponding syllable.
40. J. J. Gilpin manufactured the devices for chronic recording. A. S. Dave collected the data for one of the HVc birds. We thank A. S. Dave, S. E. Anderson, and J. A. Kogan, who provided valuable advice and assistance on aspects of the data analysis. M. Konishi and P. S. Ulinski provided useful critiques of the manuscript. Supported by a grant from the Whitehall Foundation (M91-05). A.C.Y. was supported by an NIH predoctoral fellowship (1 F31 MH10151).