Nonlinear Biomechanical Resonances in Birdsong

TL;DR

The study tackles how biomechanics constrain and enable rapid song production in birds by combining a nonlinear, three-cavity respiratory model with empirical recordings from canaries. It shows that singing pushes the respiratory system into a nonlinear resonance, broadening the amplified-frequency bandwidth to cover all syllabic rates and achieving about $94.22\%$ of the theoretical maximum magnification, far beyond the linear-case value of roughly $55.3\%$. The work reveals a tight coupling between neural motor commands and body mechanics, suggesting that sexually selected displays can exploit biomechanical optimization across behavioral states. These findings imply a general principle where nonlinear biomechanical resonances support efficient, high-rate performance in both vital and display behaviors, with broad implications for understanding motor control and signal evolution.

Abstract

Evolution has shaped animal bodies, yet to what extent biomechanical systems impose constraints and provide opportunities across different behaviors remains unclear. In birds, quiet breathing operates at a resonance of the respiratory biomechanics, but song, a behavior thought to be shaped by strong sexual selection, requires much higher breathing rates. Combining physiological recordings with a nonlinear biomechanical model, we show in canaries (Serinus canaria) that song production drives the system into a nonlinear regime that broadens the frequency range of amplified responses. This enhancement encompasses the full range of syllabic rates, with an average magnification of ~94% of the theoretical maximum. Thus, birds sing at a resonance, indicating that rapid song rhythms evolved to operate under shifting natural frequencies of the respiratory biomechanics. Our results illustrate a shared optimization strategy across behavioral states, reveal a deep connection between neural and biomechanical dynamical parameters and show that sexually selected displays may still rely on optimization strategies.

Figures (3)

• Figure 1: Avian respiratory system and its timescales across behaviors. (a) Sound spectrogram of a typical canary song composed of trills at different rates, each syllable produced within a respiratory cycle. (b) Distribution of respiratory rates during quiet breathing (in gray, right axis) and song production (in black, left axis). Respiratory frequency during song production expands a broad range, largely departing from quiet breathing rates. Data from five individuals is used (see number of samples in Table S1 in the SM supplement). (c) The respiratory system consists of air sacs and a rigid lung. Muscles force the rib cage in or out, thus compressing or expanding the air sacs that act as bellows and ventilate the lung. (d) Elements of the model. A forced piston changes the volume of the cavities representing the posterior (shaded in green, PS) and anterior (shaded in blue, AS) air sacs. Aerodynamic valves direct the flow.
• Figure 2: Respiratory dynamics during song production. Model prediction (blue line) of measured air sac pressure patterns (black line) using recorded expiratory electromyographic activity (EMG) as input. Pressure units are in terms of quiet respiration amplitude, which is normalized to $[-1, 1]$. Top panel shows the spectrogram of the recorded sound.
• Figure 3: Birds breathe and sing at resonances of the biomechanics. Magnification curves for quiet breathing (dashed gray) and singing (solid black) reveal a nonlinear enhancement of the bandwidth during song production. Black dots indicate the modal production rate of each syllable type (see Materials and Methods and Fig. S3 in the SM supplement). The bottom panel displays the marginal distribution of syllable rates during singing (black, left axis) and the distribution of quiet breathing rates (gray, right axis). Number of samples are displayed in Table S1 in the SM supplement. Black arrows in the lower panel indicate the highest syllabic rate observed, which defines a magnification level (marked by a black arrow in the top panel). This magnification level sets the bandwidth, shown by horizontal arrows.