Multimodal Signals: Enhancement and Constraint of Song Motor Patterns by Visual Display

See allHide authors and affiliations

Science  23 Jan 2004:
Vol. 303, Issue 5657, pp. 544-546
DOI: 10.1126/science.1091099


Many birds perform visual signals during their learned songs, but little is known about the interrelationship between visual and vocal displays. We show here that male brown-headed cowbirds (Molothrus ater) synchronize the most elaborate wing movements of their display with atypically long silent periods in their song, potentially avoiding adverse biomechanical effects on sound production. Furthermore, expiratory effort for song is significantly reduced when cowbirds perform their wing display. These results show a close integration between vocal and visual displays and suggest that constraints and synergistic interactions between the motor patterns of multimodal signals influence the evolution of birdsong.

Many songbirds combine their songs with elaborate visual displays, including dance-like motions, wing movements, bows, and tail spreading (16). Because the sensorimotor integration of visual and vocal displays has received little attention, it is unclear how visual and vocal displays are coordinated (7, 8). Song in songbirds is a complex learned behavior that requires precise temporal integration of respiratory and syringeal motor systems and muscle systems adjusting resonance properties of the upper vocal tract (911). It is unknown how birds integrate visual displays with the many motor tasks of singing, but elaborate postural changes not only pose a challenge to motor integration, but are also likely to have biomechanical effects on respiratory movements of song.

Song in brown-headed cowbirds is incorporated into a visual display consisting of a fixed sequence of postural changes. Before song, the male erects head and neck feathers (puffing) and, as song begins, spreads the wings, raises them, and moves them from an elevated backward position to a lowered forward position two to three times. Shortly before the end of the song, the wings are folded and a bow is performed. The puffing stage always precedes the song, but the wing display and bow vary greatly in intensity or may be omitted.

Cowbird males sing two to eight different song types (12, 13), which are produced with characteristic respiratory patterns (14, 15). Phonation occurs during three to four expiratory pulses, which are separated by vocally silent inspirations (minibreaths). The initial expiratory pulses give rise to lower frequency sounds, and a loud, higher frequency whistle concludes the song during the terminal expiratory pulse. The wing movements associated with the visual display are likely to affect respiratory effort by either enhancing or impeding song production. Although the detailed pattern of wing motion differs from that during flight, the biomechanical effects of wing up- and downstroke on respiration may be similar (1618).

Are the cowbird wing display and song respiratory patterns coordinated? Wing movements were monitored with three types of transducer systems (19). Respiratory patterns were monitored by measuring subsyringeal air sac pressure. High-speed video recordings of song and display were used to relate the voltage signatures of the transducer systems to the wing movements of the display and the respiratory pattern of the bird's song. All three techniques showed that the wing movements were temporally synchronized with the respiratory sequence of song (Fig. 1) despite variable intensity of the wing display. The lowest wing position occurred near the switch from inspiration to expiration, but temporal alignment between respiration and wing movements was variable (Table 1), with coefficients of variation ranging between 23 and 63%. The upward movement of the wing occurred during the first part of the following expiratory pulse. Phonation during these expiratory pulses (two to four) began 42 to 61 ms after air sac pressure had reached levels typically required for phonation. Airflow measurements indicated that these silent periods during expiration resulted from complete closure of the syringeal valves (Fig. 2) (14). Consequently, the most pronounced wing movements coincided with silent periods during the inspiratory and expiratory phase of respiration, and phonation occurred primarily after the wings were elevated. Prolonged silent periods during expiration have not been observed in any other species (9, 20, 21). Because these other species do not display with their wings during song, the presence of such atypical silent periods during expiration in cowbird song suggests that they are related to the display and do not constitute a respiratory requirement for song production.

Fig. 1.

Cowbird song and visual display are synchronized. (A) Song is illustrated spectrographically (top), with an example (bottom) of physiological recordings: subsyringeal air sac pressure (P; horizontal line indicates ambient pressure), impedance changes associated with the wing display (W), and sound amplitude (A, rectified and integrated). Gray bars indicate the two inspirations (minibreaths) during song. The lowest wing position occurs shortly after inspiration. The numbers indicate the approximate occurrence of highest (1 and 3) and lowest (2) wing positions as depicted in three frames from high-speed video (B).

Fig. 2.

Cowbird song (spectrogram, top) with subsyringeal air sac pressure (P) and tracheal airflow (FT). During the beginning of the second and third expiratory pulses, the syrinx is kept closed even after air sac pressure reached levels adequate for phonation (gray bars marked by arrows; note zero airflow during these periods).

Table 1.

Time delay from peak inspiration to lowest wing position. Output voltage of transducers captures wing position, with the exception of the accelerometer, which records changes in velocity, explaining the apparent differences between time measurements.

Cowbird 1 Cowbird 2 Cowbird 3 Cowbird 4 Cowbird 5
Method of recording Hall effect Hall effect Impedance change Accelerometer Accelerometer
Average (ms) 36.1 35.9 39.6 26.3 26.1
SEM 2.1 1.3 3.7 5.2 2.5
N 37 42 34 10 42

How do the movements of the visual display affect respiratory effort during song? We monitored the activity of the abdominal expiratory muscles with electromyography (EMG) and as muscle length changes (sonomicrometry) and recorded the resulting subsyringeal air sac pressure. Songs were recorded together with all physiological measurements under two conditions: with the tips of the primaries of the two wings taped such that the wings could not be spread and moved and when the birds were free to display. One bird, which spontaneously did not show a wing display during some of its songs, provided a control for the experimental manipulation.

Although air sac pressure and muscle length changes did not differ between the two conditions, the EMG activity was significantly reduced when birds were allowed to display with their wings (Fig. 3). This was found for all different song types (n = 22) and in all individuals (n = 4). A similar reduction of EMG activity occurred in the bird that spontaneously did not display (Fig. 3B). This confirms that the wing display is the likely reason for the observed difference and excludes the possibility that an atypical posture resulting from the taping of the primaries caused the increased EMG activity. The postural changes associated with the wing movements of the visual display therefore enable cowbirds to generate the same muscle shortening and air sac pressure with reduced electrical activation of the abdominal expiratory muscles. This may be achieved by biomechanical enhancement of muscle force production and thus indicates a synergistic effect of the postural changes on the respiratory movements of the song motor pattern.

Fig. 3.

Respiratory effort was estimated from physiological recordings and compared between no wing display and display conditions. (A) Example of physiological measurements showing abdominal expiratory muscle length changes (ML, relative voltage), subsyringeal air sac pressure (P; horizontal line indicates ambient pressure), and electromyographic activity of expiratory muscles (EMG; downward, original EMG trace rectified and integrated upward). The song is shown spectrographically at the top. The puffing display before song is accompanied by a lengthening of the abdominal muscles (rest length is indicated by stippled line). The gray area indicates the time period over which all measured parameters were integrated. The decrease in EMG activity for displaying males was present in all expiratory pulses. (B) Frequency histogram of ratios of integrated values for same song types between no wing display and wing display. Black bars are songs from the taped condition; array bars are songs from the control bird (a ratio greater than 1.0 indicates that value is higher in the no display condition). Mean ratios were 1.0 ± 0.01 for pressure, 0.98 ± 0.01 for muscle length, and 1.19 ± 0.021 for EMG. The mean ratio for EMG was the only parameter that differed significantly from 1.0 [t (22) = 8.87; P < 0.001].

In sum, the wing display and vocal respiratory movements are coordinated such that silent periods of the song coincide with wing movements that may have adverse mechanical effects on respiration. Silent periods may also serve to compensate for imperfect temporal coordination between vocal and visual displays. Furthermore, despite the variable intensity of the wing display, air sac pressure and airflow of song remained stereotyped. This stereotypy suggests that cowbirds are capable of adjusting the expiratory effort of the song pattern to compensate for the variable synergistic contribution of the wing display. Rapid adjustment of expiratory effort through proprioceptive feedback has been documented for experimentally perturbed respiratory patterns during song in the cardinal (Cardinalis cardinalis) (22). An alternative possibility is that the variable coordination between the display and song motor systems takes place at the motor planning stage.

The motor dynamics of song production in cowbirds suggest that temporal song organization did not evolve independently of the wing display. The evolution of song motor patterns and temporal organization in cowbirds may have been constrained by preexisting flight respiratory-locomotor coordination (1618). This close link between the evolution of vocal and visual communication signals in cowbirds illuminates a previously unknown constraint on the evolution of acoustic signals during multimodal communication in birds (6).

Supporting Online Material

Materials and Methods

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article