Natural changes in brain temperature underlie variations in song tempo during a mating behavior

Abstract

The song of a male zebra finch is a stereotyped motor sequence whose tempo varies with social context--whether or not the song is directed at a female bird--as well as with the time of day. The neural mechanisms underlying these changes in tempo are unknown. Here we show that brain temperature recorded in freely behaving male finches exhibits a global increase in response to the presentation of a female bird. This increase strongly correlates with, and largely explains, the faster tempo of songs directed at a female compared to songs produced in social isolation. Furthermore, we find that the observed diurnal variations in song tempo are also explained by natural variations in brain temperature. Our findings suggest that brain temperature is an important variable that can influence the dynamics of activity in neural circuits, as well as the temporal features of behaviors that some of these circuits generate.

Figure 1. Method for recording brain temperature in freely behaving zebra finches.

(A) Photographs of the miniature device for brain temperature recordings. (B) Circuit diagram of the device. (C) Voltage output (V_out) produced by the device at a wide range of simulated temperature values (top) and at values in the physiological range (bottom). Simulated temperatures were produced by a commercial thermocouple calibrator. Symbols indicate averages of 10-s recordings; 95% confidence intervals are smaller than the symbols. Red lines are the linear fits to the data.

Figure 2. Brain temperature in male zebra finches rises in response to presentation of a female bird.

(A, B) Schematic diagrams of the implantation of a thermocouple for temperature measurements. (A) A thermocouple was implanted into HVC along the approach angle shown. Yellow ovals indicate major nuclei of the song system, and arrows indicate the motor pathway. (B) In other experiments, a thermocouple was implanted into the hyperpallium, outside of the song system. Dotted line indicates boundary of the hyperpallium. (C) Average amount of singing triggered by presenting a female bird to the male. Error bars are SEM across all birds. (D) Examples of temperature recordings in HVC of bird #3 during individual trials in which a female bird was presented to the male for 5 min. (E) Examples of temperature recordings in the hyperpallium of bird #5 during presentations of a female. (F) Average temperature change across all 8 male birds during presentations of a female. Shaded area is SEM across all birds. (G) Average amount of singing produced in social isolation aligned to the onsets of the epochs of undirected singing. Epochs were defined as singing periods preceded by at least 20 min without singing. (H) Examples of HVC temperature recordings from the same bird as in (D) during epochs of undirected singing. (I) Examples of hyperpallium temperature recordings from the same bird as in (E) during epochs of undirected singing. (J) Average temperature change in all 7 birds during epochs of undirected singing. For all averages, data were first averaged for each bird; mean values were then averaged across all 8 birds.

Figure 3. Song tempo is strongly correlated to brain temperature.

(A) Left: Temperature recording in HVC of bird #2 during a presentation of a female. Right: Spectrograms from the same trial of an undirected motif (at time point 1 indicated on the temperature recording) and two directed motifs – one produced almost immediately after the presentation and another produced >2 min later. Note that the directed motif produced several minutes after the presentation (time point 3) has a faster tempo than the undirected motif (time point 1), while the motif produced immediately after the presentation (time point 2) has a slow tempo similar to the undirected motif. (B) Time-dependent changes of brain temperature (top) and motif duration (bottom) following presentation of a female. Data points are averages in four 15-s bins from 0-1 min, one bin from 1-2 min, and one bin from 2-5 min. Data were first averaged across presentations for each bird; mean values were then averaged across all 8 birds. Error bars are SEM across all birds. Red traces are exponential fits to the data. Baseline duration for each bird is the average duration of undirected motifs; baseline temperature is the average temperature during these motifs. (C) Temperature and motif duration for all motifs produced by the same bird as in (A).

Figure 4. Acoustic variability is not correlated to brain temperature.

(A) Spectrogram of a syllable from bird #7 that contained a harmonic stack and was used for the calculation of fundamental frequency. Dotted lines indicate the automatically detected time interval on which fundamental frequency was calculated. (B) Distributions of fundamental frequency across all directed (red) and undirected (blue) renditions of the harmonic stack shown in (A). Gaussian fits to both distributions are also shown. Note that harmonic stacks during directed singing exhibit less variability. (C) Variability of fundamental frequency measured across different temperature values. All harmonic stacks were sorted into temperature bins (0.2°C wide); the coefficient of variation of the fundamental frequency (CV, standard deviation/mean) was measured within each bin. Lines are linear fits to the data. Baseline temperature is the average temperature recorded during all undirected motifs. (D) CV of the fundamental frequency for all undirected (blue) and directed (red) songs. For directed song, harmonic stacks are separated into those produced within 5 s after the presentation of a female and those produced after 5 s. Note that the variability is already reduced in songs produced immediately upon presentation. In (C) and (D), data shown are average CV values across harmonic stacks from 8 different syllables in 5 birds for directed songs and 7 syllables in 4 birds for undirected songs; error bars are SEM across harmonic stacks.

Figure 5. Brain temperature explains circadian fluctuations in song tempo.

(A) Continuous brain temperature recording in bird #5 during >2 continuous days of social isolation, showing large changes between sleeping a waking hours. (B) Detail showing temperature during waking hours 0–10 of a single recording day. (C) Temperature across all birds (N = 6), averaged in 1-h bins. (D) Durations of all motifs produced by the bird during hours 0–10 of the same recording day as in (B). Each symbol is an individual undirected motif. (E) Motif duration across all birds, averaged in 1-h bins. In all panels, blue lines are linear fits to data from 2–10 h. Baseline values are averages across the first 2 hours of the day, and error bars are SEM across birds. For averages, values were first averaged across days; mean values were then averaged across birds.