Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 Jan 1.
Published in final edited form as: Dev Neurobiol. 2015 Jun 1;76(1):3–18. doi: 10.1002/dneu.22295

Lesions targeted to the anterior forebrain disrupt vocal variability associated with testosterone-induced sensorimotor song development in adult female canaries, Serinus canaria

Melvin L Rouse Jr 1,2,*, Gregory F Ball 1
PMCID: PMC4600635  NIHMSID: NIHMS681578  PMID: 25864444

Abstract

Learned communication is a trait observed in a limited number of vertebrates such as humans but also songbirds (i.e. species in the suborder passeri sometimes called oscines). Robust male-biased sex-differences in song development and production have been observed in several songbird species. However, in some of these species treating adult females with testosterone (T) induces neuro-behavioral changes such that females become more male-like in brain and behavior. T-treatment in these adult females seems to stimulate sensorimotor song development to facilitate song masculinization. In male songbirds it is known that the lateral magnocellular nucleus of the anterior nidopallium (LMAN) plays a modulatory role during song development. LMAN is androgen sensitive and may be a key target of a T-induced recapitulation of a developmental process in adult females. We tested this hypothesis. Adult female canaries were given either a chemical lesion of LMAN or a control sham-surgery. Prior to surgery birds were individually housed for two-weeks in sound-attenuated chambers to record baseline vocal behavior. Post-surgery birds were given one-week to recover before subcutaneous implantation with silastic capsules filled with crystalline-T. Birds remained on treatment for three-weeks (behavioral recordings continued throughout). Birds with a lesion to LMAN had less variability in their song compared to controls. The diversity of syllable and phrase type(s) was greater in sham controls as compared to birds with LMAN lesions. Birds did not differ in song rate. These data suggest that the sustention and conclusion of T-induced sensorimotor song development in adult female canaries requires an intact LMAN.

Introduction

Among temperate zone songbird species, it is often the case that males sing more and have larger associated volumes of the controlling neural substrate for song than females (Nottebohm and Arnold, 1976; see Ball and MacDougall-Shackleton, 2001 for a review). These differences are typically male-biased and quite robust (Arnold, 1996). Sex-based differences in the brain and behavior of temperate zone songbird species have often been accredited to the organizational effects of sex steroids; i.e. steroids acting perinatally to organize the brain in a sex-specific manner (Arnold et al, 1996; Arnold, 2004). However, in some species, such as canaries, testosterone (T) is able to induce male-like patterns of brain and behavior in adulthood (Nottebohm, 1980). In female canaries T treatment in adulthood induces growth in brain regions that control song development and production, and appears to induce a recapitulation of the sensorimotor phase of song development that normally occurs in juveniles (Goldman and Nottebohm, 1983; Vallet et al, 1996) so that the adult T-treated female can eventually produce a more male-like song. However, several song nuclei in adult females increase in size in response to T treatment (Nottebohm, 1980) and the contribution that particular nuclei in the song system make in the induction of these changes in song behavior by T, including the sensorimotor song development process, is not clear.

The lateral magnocellular nucleus of the anterior nidopallium (LMAN) is a cortical-like song control nucleus, a component of the anterior forebrain basal-ganglia pathway of the song system, and is directly implicated in song development in juvenile male zebra finches (Scharff and Nottebohm, 1991; Aronov et al, 2008; Fee and Goldberg, 2011). Converging evidence from multiple research teams indicate that LMAN acts as an endogenous variability generator that serves to guide the trajectory of vocal motor learning in juvenile males (Kao et al, 2005; Olveczky et al, 2005; Thompson et al, 2011). In adulthood, based on studies of male zebra finches, lesions to LMAN have more subtle effects on song production. As reviewed by Brainard and Doupe (2013) lesions to LMAN in adulthood do not have gross effects on song but variability is reduced if one does a careful rendition-by-rendition analysis of song or if one assesses adult song plasticity by a variety of learning tasks. What has not been investigated in detail is whether LMAN is essential for T induced plasticity in the case of adult canaries where T in females seems to result in a re-engagement of sensorimotor song development and the masculinization of song.

We individually housed adult female canaries, recorded their behavior prior to treatment, and then performed stereotaxic surgeries (ibotenic acid lesions or sham controls). After recovery from the lesion-related surgery the birds were treated with T. We hypothesize that an intact LMAN is essential for T to be able to induce a re-engagement of sensorimotor song and induce the masculinization of female song. We also hypothesize that the effects of T on the rate of song production would not be influenced by lesions to LMAN since both the POM and the HVC-RA-nXIIts pathway would be intact and these areas have been directly implicated in T effects on song rate (Ball et al., 2008; Alward et al, 2013). We therefore predicted based on these hypotheses that prior to T-treatment these female canaries would not sing and would also not sing in response to lesion surgery. Further, we predicted that after surgery and T treatment, females with LMAN targeted lesions would sing songs with less syllable variability, i.e. more stereotypy, as compared to sham control birds.

Materials and Methods

Animal Subjects and Experimental Design

13 Female border canaries (Serinus canaria) were housed on short-day length (8L:16D) to place them in a photosensitive state (Hurley et al., 2008). After being held for a period of 6 weeks on this photoperiod birds were transferred to individual sound attenuated chambers (see figure 1A). Birds were recorded daily for 90 minutes just after lights on from the start to the end of the experiment. Birds were placed into individual sound attenuated chambers (41cm × 48cm × 51 cm) and were held there for a period of 14 days to collect baseline data before undergoing stereotaxic surgery for neurotoxic lesions (either saline sham or ibotenic acid lesion) targeted at the lateral magnocellular nucleus of the anterior nidopallium (LMAN). Birds were given a period of 7 days to recover from surgery and to record immediate effects of LMAN lesion. Following this period, birds were then implanted with Silastic™ capsules filled with T. Birds remained on T-treatment for 22 days; final behavior samples and tissue were collected on day 22 of T-treatment. All experimental procedures were approved by the Johns Hopkins University Institutional Animal Care and Use Committee.

Figure 1.

Figure 1

Open in a new tab

Experimental timeline and representative lesion. A) At the beginning of the study birds were transferred from group housing to individual sound attenuated chambers. They remained in these chambers for two-weeks. Birds then underwent stereotaxic surgery (lesion or sham control) and were given one-week to recover. Birds were then T-treated and remained on treatment for three-weeks. Behavioral data was sampled daily throughout the entirety of the study. B) At the end of study lesions quality and volume was assessed as described in the methods. Typical lesions to LMAN were confined to the medio-lateral extent of LMAN and did not damage MMAN. Abbreviations: lateral magnocellular nucleus of the anterior nidopallium (LMAN) and medial magnocellular nucleus of the anterior nidopallium (MMAN)

Stereotaxic Surgery

Bilateral lesions of LMAN were based on stereotaxic coordinates modified from the canary brain atlas (Stokes et al., 1974). Birds were deeply anesthetized with isoflurane gas. The procedure was as follows: anesthesia was induced and maintained using an isoflurane gas (IsoSol isoflurane from Vedco. Inc, St. Joseph, MO) anesthesia system (Isotec 4 from SurgiVet, Inc., Waukesha, WI USA), fitted with a gas scavenging system, vented to a fume hood. LMAN was either neurochemically lesioned using ibotenic acid or sham lesioned with physiological saline. Lesions were done under a surgical microscope with a Hamilton syringe connected to a microinjection unit. Ibotenic acid was dissolved in PBS at a concentration of 1 mg/mL. Approximately 0.25 μl of ibotenic acid and was delivered bilaterally to LMAN. Sham lesioned birds were be subjected to the same procedure but no drug was injected only saline. Coordinates were derived from pilot experiments of photosensitive female canaries. Initial pilot lesions were unilateral using the untargeted hemisphere as a control to locate the focal point and diffusion of chemical lesions. Later pilots were bilateral lesions to test acid volume effectiveness and survivability of the surgery. The coordinates used in this study were as follows: anterior-posterior – 4.25mm, medial-lateral – 1.43mm, dorsal-ventral – 2.40mm.

After birds were deeply anesthetized, an incision was made on the skin atop the head and retracted to expose the skull. Anterior-posterior, and medial-lateral coordinates were zeroed from the sinus, which was visualized through the birds’ translucent skull; the zero point was marked on the skull with black ink. In cases where sinus was not visible, a portion of the skull was removed near the posterior portion to reveal the sinus and zero coordinates were taken from there. Small holes were then drilled through both layers of the skull at the points of entry. The holes were just large enough for the syringe needle tip to enter. Dorsal-ventral zero coordinates were then taken from the very top of the exposed meninges. A small tear was then made in the meninges by a sterile small gauge needle and the syringe was slowly lowered into position. Ibotenic acid was then slowly pressure injected slowly into the region of interest (total time approximately 5-minutes) and following that the syringe was raised. The hole in the skull was left unplugged as the skin was then drawn together covering the points of entry. The wound was sutured shut with a sterile thin medical grade filament. Antibiotic ointment was applied around but not directly on the closed wound. The bird was then placed under a heating lamp and O2 was administered before return to the home cage.

Testosterone Implant

Birds were deeply anesthetized with isoflurane gas and a small incision is made (∼1-2 mm) just above the right or left flank. Following incision, a metal probe was used to separate the skin from fat and muscle creating a pocket for the implant to rest. One 12 mm-long Silastic implant filled with T was then inserted. The skin was closed with tissue adhesive (Nexaband liquid bandage Veterinary products laboratory, Phoenix, AZ). This length of implant has been shown previously to be effective in inducing male-like concentrations of T in serum and a masculinization of song behavior (Appeltants et al., 2003; Nottebohm, 1980; Sartor, et al. 2005).

Behavioral Analysis

Behavioral analysis was conducted as follows: sound files (converted from audio/video file format .mp4 to audio only .wav files) were sampled at 22050 Hz with a frequency range of 0 to 11 kHz. Audio files were highpass filtered with audio editing software Goldwave ™ (version 5.55) set to a threshold of 900 Hz to remove low frequency noise (e.g. the sound of the fan/air vent and hum of the light). Sound spectrograms were created for each daily recoding using Avisoft SASlab (Avisoft Bioacoustics, Berlin, Germany). Spectrogram FFT (fast Fourier transform) lengths were set to 512 with an overlap of 75% for the temporal resolution. For each spectrogram the number of songs, song bout duration (in seconds), number of calls, the total time spent singing and/or producing vocalizations (i.e. calls in addition to songs), and other acoustic features (i.e. mean entropy variance, fundamental frequency variance, and energy) for each recording were calculated and exported into an Excel spreadsheet.

Canary song has a characteristic acoustic structure and temporal pattern that distinguishes it from calls and we defined song as being bouts of vocalizations where the total duration was greater than 1.5 seconds of continuous syllables (featuring 5 or more syllables that have a peak amplitude value greater than -22dB) with inter-syllable intervals no longer than 500 milliseconds and a mean entropy value less than W = 0.550. One bird did not sing in response to T-treatment and was removed from the study.

Wiener entropy measures the width and uniformity of an acoustic signal. It is a unitless pure number with 0 indicating a tone (e.g., a uniform sinusoidal wave) and 1 being complete noise (i.e. a non-uniform random signal; Tchernichovski et al., 2000). In juvenile male zebra finch, song entropy variance increases rapidly with the onset of sensorimotor song development, a period marked by significant vocal variability (both in note composition and spectral stability; Tchernichovski et al., 2001; Shank and Margoliash, 2009). High entropy variance indicates more acoustically complex signals and when collapsed across a whole song, particularly for canaries, can indicate higher syllable/phrase diversity (Alward, et al, 2013; Madison, et al, in press). Energy is the sum (i.e. integral) of the squared amplitudes of a sound multiplied by its sampling time (Avisoft SASlab User Manual) and measures the relative ‘loudness’ of a given vocalization.

To determine the stereotypy of these song measures the coefficient of variation (CV = [SD/mean] × 100) was calculated for each bird for each day for each feature of song measured. The coefficient of variation is an established method to approximate the stereotypy of song features and has been used previously in studies of birdsong (Tramontin et al, 1999; Meitzen et al, 2007). Mean values collapsed across the entire song were used to calculate the CV. Higher CV values indicate that the song feature measure is less stable, thus, it is a more variable measure. Lower CV values indicate that the song feature measure is more stable, thus, it is a less variable measure.

When song crystallizes, individual syllable iterations of a given type within a bout of singing and between bouts of singing become highly similar to one another (Waser and Marler, 1977; Marler and Peters, 1982). Crystallized song is characterized by high stereotypy (Marler, 1997). To measure the overall stereotypy of the syllables we followed a novel method used previously by Madison and colleagues (in press). We isolated high amplitude vocalizations (peak amplitude > -14dB; song elements only, calls omitted) on the final day of recording, day 22 of T-treatment, into individual .wav files. Syllables were then randomly re-sampled (maximum 100 syllables), categorized, and template sonograms were made for selected syllable iteration. High amplitude vocalizations (peak amplitude > -14dB; song elements only, calls omitted) on the day L7 (the final day lesion recovery just prior to T-implantation) and days 5, 10, 15, & 20 of T-treatment were then isolated into individual .wav files. All files from this isolation technique were used in this analysis. Binary sonogram templates from the final day were then cross-correlated with sonograms for all song elements above amplitude threshold for day L7 and days 5, 10, 15, & 20 of T-treatment in Avisoft SASlab. Correlations that were above a correlation threshold of r=0.95 were then counted and the ratio of highly correlated (i.e. stereotyped, r > 0.95) syllables to non-stereotyped syllables (i.e. r < 0.95) was calculated for each day of observation (i.e. day L7 and days 5, 10, 15, & 20 of T-treatment).

On the final day of treatment, day 22 of T-treatment, a syllable analysis was done to estimate the repertoire size of individual birds. Song samples for a given individual was aurally and visually inspected using spectrograms produced in Avisoft SASlab. Each unique syllable was marked and tabulated in a spreadsheet.

Finally, on the last day of treatment (day 22 of T-treatment) we randomly resampled 10 individual songs and plotted template sonograms of entire songs. Each song was then cross-correlated with one another in a correlation matrix using the Avisoft correlator module (version 3) in Avisoft SASlab. The average correlation yielded from this matrix was then calculated. These procedures were done for each individual bird. This variable gave a general measure of how similar the songs sang by an individual were which served to give a first-pass analysis of how similar whole songs were to one another.

Histology and Lesion Quantification

At the end of the experiment, birds were deeply anesthetized with isoflurane gas and then euthanized by rapid decapitation and brains were extracted and placed in fixative (5% acrolein). Brains were slightly agitated in acrolein fixative for two hours, rinsed (4 × 15 minutes) in 0.01M Phosphate Buffer-Saline (PBS; pH 7.4) and cryoprotected in 30% sucrose until saturated. Brains were then flash frozen on dry ice and placed in the -70°C until processed for later analysis.

Brains were sectioned at 30 μm thickness using a cyrostat (Carl Zeiss) and were processed for NeuN immunoreactivity using the following protocol. NeuN is a neuronal marker that clearly and reliably delineates the boundaries of song control nuclei and yields comparable results to more traditional stains like Nissl (Phillmore, et al, 2006). Tissue samples were then processed in random order and the timing of procedure was similar across groups. Brain sections were washed in 0.1 M PBS three times, once in 1% sodium borohydride, then washed three times in 0.1 M PBS, once in 0.5% H2O2 for 1 hour, then the sections were washed three times with 0.1 M PBS. Sections were then incubated for 15 minutes in Avidin blocking solution (Vector laboratories; 1.5ml in 20ml 2% Normal goat serum in 0.3% PBS/T) then washed three times with 0.1 M PBS. Tissue sections were then incubated for 15 minutes in Biotin blocking solution (Vector laboratories; 1.5ml in 20ml 2% PBS/T-NGS) then washed three times with 0.1 M PBS. Sections were then incubated in the primary antibody for neuronal marker N (NeuN) in 20ml 2% Normal goat serum in 0.1% PBS/T at 4°C for 48 hours. The sections were then washed three times in 0.1% PBS/T, then incubated in biotinylated secondary antibody (Vector laboratories; goat anti rabbit IgG, 1:250 in 20ml 2% Normal goat serum in 0.3% PBS/T) for 1 hour, washed three times in 0.1% PBS/T, incubated in Avidin-Biotin horseradish-peroxidase complex (Vectastain ABC, Elite; 1:200 in 20ml 2% Normal goat serum in 0.3% PBS/T) for 1 hour, and then washed three times in 0.1% PBS/T. Antibodies were visualized by incubating the sections with the chromagen nickel-enhanced diaminobenzidine (Sigma Fast DAB) for 6-7 minutes. Brains sections were then placed in.01 M PBS solution and then mounted onto gelatin-coated microscope slides. The slides were open-air dried, rehydrated in 0.01M PBS and then serially dehydrated in ethanol at 30%, 50%, 75%, 95%, 95%, 100% for one minute each and a final step in 100% ethanol for five minutes. The slides were then cleared in xylene (Fisher Scientific) and coverslipped with Permount (Fisher Scientific).

Brain regions of interest were digitized using a bright field light microscope (Zeiss Axioscope, Carl Zeiss, Thornwood NY) with a CCD camera connected to a desktop computer. For each image, the area of the brain region was measured using NIH Image J. One sham control bird was selected at random as a reference to demarcate the boundaries of the lesioned portion of LMAN in the nidopallium. For the reference bird the volume of LMAN was reconstructed combining the areas of subsequent sections with the sampling interval (90 μm) using the formula for a truncated cone (developed by Smith et al., 1995). This method has been used previously in our laboratory (Bernard and Ball, 1995; Bentley et al., 1999; Bernard and Ball, 1997). For the sham control bird the nucleus was reconstructed for both the left and right hemispheres. Digital Images of each section containing LMAN were captured using a bright field light microscope (Zeiss Axioscope, Carl Zeiss, Thornwood NY) with a CCD camera connected to a desktop computer. These images were then made translucent using NIH Image J and were overlaid on top of digital images of birds with LMAN targeted lesions taken from the same microscope and camera.

For these images landmarks such as Area X, MMAN, the relative position of the hyperpallium, and lateral most remaining labeled cells in LMAN were used to determine the sites of LMAN lesion reconstruction (figure 1B). In these composite images, the borders of the remaining portion(s) of LMAN were drawn and the area was measured. Next, the borders of the lesioned portion of LMAN was then drawn using the boundary provided by the reference image and the area was measured. Finally, the boundary of the lesioned area outside of LMAN was drawn and the area measured. Though the majority of lesion outside LMAN was confined to the nidopallium there were some cases where the rostral extent of lesion was tracked dorsally into small portions of the hyperpallium.

LMAN lesion volume was then calculated using the same formula as mentioned above. LMAN targeted lesions were considered “hits” if and only if at least 50 percent of LMAN was lesioned in the combined average of the left and right hemispheres. Birds with lesions greater than 50% were included in the lesion group (n = 5). Birds that had sham surgeries or lesions that were less than 50% of LMAN were included in the sham controls group (n = 7).

Statistical Analysis

To evaluate the song rate data we used a split-plot factorial analysis of variance (ANOVA; 2-way mixed-design with experimental day as the within and lesion target category [i.e. sham control, hit or miss] as the between groups variables). For these repeated measures variables, significance for main effects was corrected using a Greenhouse-Geisser correction for non-sphericity. Though conservative, this correction assures that results are not influenced by variances that are potentially different between all possible pairs of groups in repeated measures ANOVA.

The mean values and coefficients of variation for the song features (entropy variance, fundamental frequency variance, and energy) and syllable stereotypy were evaluated separately for day L7 and days 5, 10, 15, & 20 of T-treatment. These days corresponded with the days in which the birds were exposed to T. To analyze each measure we used a split-plot factorial analysis of variance (ANOVA; 2-way mixed-design with experimental day as the within and lesion target category [i.e. sham control, hit or miss] as the between groups variables). For these repeated measures variables, significance for main effects was corrected using a Greenhouse-Geisser correction for non-sphericity. All post-hoc tests were corrected for multiple comparisons using Bonferonni's correction. The difference in pairwise comparisons and Bonferonni corrected p-values are reported for the effects of lesion target category and the polynomial trend analysis to generally describe the pattern of T-treatment response.

For the final day of singing, the mean number of unique syllables and mean song similarity correlation was tested in separate univariate ANOVAs. For the mean song similarity correlation the log transform was calculated to account for a skewed (i.e. non-normal) distribution of the data. All post-hoc pairwise comparisons were corrected for multiple comparisons using Bonferonni's correction.

Relationships (i.e. correlations) between the behavior on the final day of treatment (day 22 of T-treatment) and lesion size (LMAN lesions and lesions outside of LMAN) were evaluated with two-tailed Pearson correlations. All results were considered statistically significant for α < 0.05. Effect size was calculated using partial eta squared values only for significant results.

Results

Testosterone Induced Increases in Song Rate

We tested for differences in the amount of singing over time. This was measured by the proportion of time spent singing across baseline, lesion surgery recovery, and T-treatment. As expected we found a significant increase in the rate of singing in response to T-treatment; prior to subcutaneous implantation birds did not sing (F(42,294) = 22.019, p < 0.001, ηpartial2 = 0.759; see figure 2). Also as expected, there was no effect of ibotenic acid lesions to the anterior forebrain on the rate of singing (F(1,10) = 1.216, p = 0.296; see figure 2) and there was no interaction between rate of singing and lesion treatment (F(42,294) = 0.926, p = 0.605). However, the complexity of the songs produced varied in observed sonograms as a function of surgical treatment. Birds that received LMAN targeted lesions tended to produce songs that appeared simple, similarly arranged between bouts and lacked syllable diversity (see figure 3A and 3B).

Figure 2.

Figure 2

Open in a new tab

The effect of T-treatment and LMAN lesion on song rate. The above figure is a line graph illustrating the effects of T treatment and LMAN lesion on rate of singing. Prior to T-treatment birds did not sing. During the baseline behavioral recording and during recovery from stereotaxic surgery birds only made calls. After subcutaneous implantation with T-filled implants the birds in both groups produced robust song.

Figure 3.

Figure 3

Open in a new tab

The effect of LMAN lesion on song quality and mean acoustic measures. A) Song spectrograph illustrating typical song in a T-treated sham control female. Frequency is on the y-axis and time is on the x-axis. B) Song spectrograph illustrating typical song in a T-treated female with lesions to LMAN. Frequency is on the y-axis and time is on the x-axis. Birds that received LMAN targeted chemical lesions sang songs that were simpler and less variable in comparison to sham-lesion controls. C) Bar graph showing the effect of lesion on mean song entropy variance. D) Bar graph showing the effect of lesion on mean song energy. Birds with lesions did not differ from sham treated controls for mean song entropy variance or energy.

Acoustic features of song and measures of stereotypy/complexity

Despite the difference in syllable diversity anecdotally noted in the review of sonograms (see figures 3A and 3B), the marginal mean values of the raw acoustic features of song did not differ between the sham lesion controls and birds with LMAN lesions (see figure 3C and 3D). More specifically, there was no main effect of lesion on the mean entropy variance (F(1,10) = 0.410, p = 0.536). However, there was a main effect of treatment day (F(4,40) = 22.576, p < 0.001, ηpartial2 = 0.693). There was no interaction between treatment day and LMAN lesion (F(4,40) = 1.182, p = 0.325). The mean song entropy variance generally increased linearly in response to T-treatment (F(1,10) = 42.440, p < 0.001, ηpartial2 = 0.809). However, there was a significant quadratic trend in the pattern of response over time (F(1,10) = 6.456, p < 0.05, ηpartial2 = 0.392). Bonferroni corrected pairwise comparisons showed that day L7 was different from day 10 (i-j L7-Day10 = -0.033, p <0.01), 15 (i-j L7-Day15 = -0.042, p <0.001), and 20 (i-j L7-Day20 = -0.044, p <0.001) of T-treatment. Day L7 was not different from day 5 of T-treatment (i-j L7-Day5 = -0.010, p = 0.709). Similarly, day 5 was significantly different from day 15 (i-j Day5-Day15 = -0.032, p <0.05) and day 20 (i-j Day5-Day20 = -0.034, p <0.05). Conversely, there was no difference between the final three days of T-treatment: day 10 (i-j Day10-Day15 = -0.009, p = 0.184), day 15 (i-j Day15-Day20 = -0.002, p = 0.999), and day 20 (i-j Day20-Day10 = 0.011, p = 0.106).

Likewise, there was no main effect of lesion on the mean song energy (F(1,10) = 0.332, p = 0.577). However, there was a main effect of treatment day (F(4,40) = 11.039, p < 0.01, ηpartial2 = 0.525). There was no interaction between treatment day and LMAN lesion (F(4,40) = 0.262, p = 0.719). The mean song energy generally increased linearly in response to T-treatment (F(1,10) = 14.154, p < 0.01, ηpartial2 = 0.586). Bonferroni corrected pairwise comparisons showed that day L7 was different from day 20 (i-j L7-Day20 = -0.086, p <0.05) of T-treatment. Day L7 was not different from day 5 of T-treatment (i-j L7-Day5 = -0.005, p = 0.999), day 10 (i-j L7-Day10 = -0.038, p = 0.140), or day 15 (i-j L7-Day15 = -0.065, p = 0.053). Day 5 was significantly different from day 20 (i-j Day5-Day20 = -0.081, p <0.05). However, day 10 did not differ from any day of treatment; day 5 (i-j Day10-Day5 = 0.032, p = 0.145), day 15 (i-j Day10-Day15 = -0.027, p = 0.460), or day 20 (i-j Day10-Day20 = -0.048, p = 0.165). Likewise, there was no difference between the final two days of T-treatment: day 15 and day 20 (i-j Day20-Day15 = 0.021, p = 0.999).

Though the marginal means did not differ by lesion target category the mean coefficient of variation (CV) for each song feature tended to be lower (i.e. less variable) in birds with LMAN targeted lesions (see figure 4). More specifically, birds with LMAN lesions had significantly lower entropy variance CV than sham controls (F(1,10) = 12.123, p < 0.01, ηpartial2 = 0.548). Likewise, there was a main effect of treatment day (F(4,40) = 6.715, p < 0.05, ηpartial2 = 0.402). There was no interaction between treatment day and LMAN lesion (F(4,40) = 1.977, p = 0.176). The mean song entropy variance CV generally increased linearly in response to T-treatment (F(1,10) = 20.464, p < 0.01, ηpartial2 = 0.672). However, there was also a significant quadratic trend in the pattern of response over time (F(1,10) = 15.355, p < 0.01, ηpartial2 = 0.606). Bonferroni corrected pairwise comparisons showed that day L7 was different from day 10 (i-j L7-Day10 = -27.395, p <0.01), 15 (i-j L7-Day15 = -27.199, p <0.01), and 20 (i-j L7-Day20 = -30.619, p <0.001) of T-treatment. Day L7 was not different from day 5 of T-treatment (i-j L7-Day5 = -17.098, p = 0.705). Similarly, day 5 was not different from any other day of treatment: day 10 (i-j Day5-Day10 = -10.296, p = 0.999), day 15 (i-j Day5-Day15 = -10.101, p = 0.999) and day 20 (i-j Day5-Day20 = -13.521, p = 0.999). Likewise, there was no difference between the final three days of T-treatment: day 10 (i-j Day10-Day15 = 0.196, p = 0.999), day 15 (i-j Day15-Day20 = -3.420, p = 0.999), and day 20 (i-j Day20-Day10 = 3.224, p = 0.999).

Figure 4.

Figure 4

Open in a new tab

Song stereotypy as measured by the coefficient of variation. A) A line graph illustrating the effect of lesion on the level of stereotypy (i.e. coefficient of variation [CV]) for the acoustic measure song entropy variance. B) A line graph illustrating the effect of lesion on the level of stereotypy (i.e. the CV) for the acoustic measure song energy. Birds with lesions to LMAN sang songs that were significantly more stereotyped than sham controls. Birds in the lesion group generally had lower CVs than sham control animals. Lower CV indicates higher stereotypy, i.e. less variability or greater stability in the signal.

As with entropy variance CV, birds with LMAN lesions had significantly lower song energy CV than sham controls (F(1,10) = 11.770, p < 0.01, ηpartial2 = 0.541; figure 4A). Likewise, there was a main effect of treatment day (F(4,40) = 16.427, p < 0.001, ηpartial2 = 0.622). There was no interaction between treatment day and LMAN lesion (F(4,40) = 1.789, p = 0.203). The mean song energy CV generally increased linearly in response to T-treatment (F(1,10) = 30.129, p < 0.001, ηpartial2 = 0.751). However, there was also a significant quadratic trend in the pattern of response over time (F(1,10) = 22.356, p < 0.001, ηpartial2 = 0.691). Bonferroni corrected pairwise comparisons showed that day L7 was different from day 10 (i-j L7-Day10 = -54.821, p <0.001), 15 (i-j L7-Day15 = -66.429, p <0.001), and 20 (i-j L7-Day20 = -68.698, p <0.001) of T-treatment. Day L7 was not different from day 5 of T-treatment (i-j L7-Day5 = -18.798, p = 0.652). Similarly, day 5 was not different from any other day of treatment: day 10 (i-j Day5-Day10 = -36.023, p = 0.180), day 15 (i-j Day5-Day15 = -47.631, p = 0.184) and day 20 (i-j Day5-Day20 = -49.901, p = 0.109). Likewise, there was no difference between the final three days of T-treatment: day 10 (i-j Day10-Day15 = -11.608, p = 0.999), day 15 (i-j Day15-Day20 = -2.269, p = 0.999), and day 20 (i-j Day20-Day10 = 13.877, p = 0.736).

We tracked the stereotypy of individual syllables and as expected we found that birds with LMAN lesions had significantly larger ratio of songs above correlation threshold compared to sham controls (F(1,10) = 8.617, p < 0.05, ηpartial2 = 0.463; figure 5A). Likewise, there was a main effect of treatment day (F(4,40) = 17.556, p < 0.001, ηpartial2 = 0.637). There was no interaction between treatment day and LMAN lesion (F(4,40) = 0.218, p = 0.782). The ratio of matched syllables generally increased linearly in response to T-treatment (F(1,10) = 29.087, p < 0.001, ηpartial2 = 0.744). However, there was also a significant quadratic trend in the pattern of response over time (F(1,10) = 10.682, p < 0.01, ηpartial2 = 0.516). Bonferroni corrected pairwise comparisons showed that day L7 was different from day 10 (i-j L7-Day10 = -0.397, p <0.01), 15 (i-j L7-Day15 = -0.447, p <0.01), and 20 (i-j L7-Day20 = -0.479, p <0.01) of T-treatment. Day L7 was not different from day 5 of T-treatment (i-j L7-Day5 = -0.152, p = 0.543). Similarly, day 5 was not different from any other day of treatment: day 10 (i-j Day5-Day10 = -0.246, p = 0.142), day 15 (i-j Day5-Day15 = -0.296, p = 0.100) and day 20 (i-j Day5-Day20 = -0.327, p = 0.074). Likewise, there was no difference between the final three days of T-treatment: day 10 (i-j Day10-Day15 = -0.050, p = 0.886), day 15 (i-j Day15-Day20 = -0.032, p = 0.999), and day 20 (i-j Day20-Day10 = 0.082, p = 0.142).

Figure 5.

Figure 5

Open in a new tab

Syllable stereotypy, repertoire size, and song similarity. A) A line graph showing the effect of lesion on the ratio of syllables that matched exemplar templates (i.e. above an r = 0.95 correlation threshold). Birds with LMAN lesions had significantly greater syllable stereotypy (i.e. more highly correlated individual syllables) compared to sham controls. B) Bar graph showing the effect of lesion on the number of unique syllables. Similar to syllable stereotypy, it was found that birds with LMAN lesions had fewer of unique syllables than sham controls. C) The effect of lesion on the overall similarity of songs. Birds with LMAN lesions sang songs that were overall more similar to one another compared to sham controls.

For the final day of treatment we inspected each song and tabulated the number of unique syllables per bird. We found that birds with LMAN lesion had significantly fewer of unique syllables on the final day treatment compared to sham controls (F(1,10) = 18.481, p < 0.01, ηpartial2 = 0.653; see figure 5B). We also assessed the overall similarity of songs by cross-correlating entire sonograms of individual songs on the final day of treatment. We found that birds with LMAN lesions had more similar songs on the final day of treatment compared to sham controls (F(1,10) = 9.327, p < 0.05, ηpartial2 = 0.483; see figure 5C).

Correlations between lesion specificity and song stereotypy/complexity

We wanted to evaluate the effect of lesion size and specificity on the each of the behavioral measures observed. Pearson correlations were run on all birds with anterior forebrain lesions (regardless of LMAN lesion size; i.e. original sham controls were removed). On the final day of treatment (day 22 of T-treatment) there were significant correlations between the percent of LMAN lesioned and the song stereotypy and complexity measures (entropy variance CV, syllable energy CV, number of unique syllables, and song similarity; refer to table 1 for results; figure 6). Conversely, there were no significant correlations between the lesion volume outside of LMAN and the stereotypy or complexity measures (refer to table 1 for results; figure 6).

Table 1.

Correlations between lesion specificity and size with song feature stereotypy and complexity.

n = 7 Coefficient of Variation Day 22: Entropy Variance Coefficient of Variation Day 22: Energy Number of Unique Syllables on Day 22 Song Similarity on Day 22
Percent of LMAN Lesioned Pearson Correlation -.774* -.947** -.965** .853*
Sig. (2-tailed) 0.041 0.001 0 0.015
Lesion Outside LMAN Volume (mmˆ3) Pearson Correlation 0.482 0.329 0.2 0.093
Sig. (2-tailed) 0.273 0.471 0.667 0.844
Open in a new tab
*

Correlation is significant at the 0.05 level (2-tailed).

**

Correlation is significant at the 0.01 level (2-tailed).

Figure 6.

Figure 6

Open in a new tab

The correlation of lesion size and specificity on the stereotypy of song features. Scatterplots show the overall distribution and regression line comparing LMAN lesion size with the coefficient of variation of song features. It was found that the greater the lesion to LMAN the lower the song feature coefficient of variation, number of unique syllables, and the greater the song similarity (the lower the CV the less variability; A-D).

Discussion

In this study we evaluated the role of LMAN in T-induced masculinization of adult female canary song. LMAN is a clear potential target of androgen action based on the high density of androgen receptors in this nucleus that has been observed in studies of several songbird species including canaries (Arnold et al., 1976; Arnold and Gorski, 1984; Balthazart et al, 1992; Bernard et al, 1999; Fusani et al, 2000; Gahr, 2000; Ball et al, 2004). It is also essential for song development in juvenile males based primarily on studies of zebra finches. Sensorimotor song development is supported by input from LMAN to RA, though not to the exclusion of nucleus HVC, which is also critical for guiding song development in addition to singing overall (Olveczky et al., 2005). To test the role of LMAN we collected baseline song data and confirmed previous findings that adult female canaries do not sing in isolation in the absence of T. We then performed stereotaxic surgeries on all of the birds; some received chemical lesions targeted toward LMAN and others underwent sham surgeries. Birds were then treated with T after recovery from surgery. T induced significant increases in the rate of song and song output was male-like in syllable repertoire in sham-operated controls. This finding replicates previous observations and further confirms that the anterior forebrain (specifically in this case LMAN, the nidopallium, and the hyperpallium) does not control the rate of singing (Riters and Ball, 1999). Recent studies in canaries suggest that T action in the POM plays a key role in promoting increases in song rate and that actions in the forebrain probably regulate aspects of song quality (Alward et al., 2013).

The masculinization of female canary song is characterized by increases in song complexity and stereotypy over time. As previously stated, T-induced increases in the rate of singing and there was no difference in song rate for lesion and sham control birds. Likewise, the mean value of the acoustic measures (entropy variance and energy) increased over time and there was no statistical difference between groups. Further, by the final days of treatment there was no difference in the mean values of song entropy variance and energy. Likewise in addition to a significant linear trend, there was a significant quadratic trend. All of these measures indicate that these acoustic features were stabilizing, or rather, had already stabilized by the end of study.

That said, birds with LMAN lesions tended to sing songs that were simpler, repetitive, and more stereotyped compared to sham operated controls. To evaluate this anecdotal observation of the individual sonograms we measured the level of stability (i.e. stereotypy) of the acoustic features of song by calculating the coefficient of variation for song entropy variance and energy. The coefficient of variation gives a broad measure of the level of stereotypy of a given variable. The lower the coefficient of variation the more stereotyped the song feature; the higher the coefficient of variation, the less stereotyped the song feature. Birds with LMAN lesions sang songs with lower coefficients of variation for syllable entropy variance and energy in comparison with sham-operated controls, indicating that these features were more stable and less variable in the lesion group.

These findings suggest that though all the birds saw an increase in male-like singing in response to T (song rate and the mean feature values increase linearly over time) the way in which these features vary within and between songs is more stereotyped in birds with LMAN lesions compared to sham-operated controls. This result is consistent with what would be expected if LMAN were functioning in a way that is similar to what has been shown to occur during the initial stages of song development in juvenile male zebra finches. However, one thing that must be noted is that for the lesion group, though they sang at a rate that was male-like and the mean acoustic features increased, it is not clear that song was fully masculinized as the songs appeared to be simpler and repetitive, which is closer to typical female canary song than male (Pesch and Guttinger, 1985; Vallet, Kreutzer, and Gahr 1996). In addition, because birds were generally not singing prior to T-treatment and were treated systemically with T, for the lesion group, linear changes in mean acoustic feature value over time may be more related to changes in the vocal musculature (i.e. the syrinx). The syrinx is a known target of androgen activity and the actions of T, in addition to the development of syringeal muscle gestures early in response T-induced singing, may account for the linear increase in mean entropy variance and energy; particularly, for the lesion group (Lieberburg and Nottebohm, 1987). However, our analysis does not exclude the possibility that there are subtle neural plasticity-related song development changes that occurred over time in the lesion group that could also account for this. We will return to this point later.

Furthermore, we found that the individual syllables produced by birds with LMAN lesions during T-treatment were highly similar to one another, more so than sham controls. Additionally, birds with LMAN lesions had a smaller repertoire size compared sham controls. Finally, the syllable arrangement(s) of songs tended to be more similar to one another in LMAN lesion birds compared to shams as noted by higher song similarity correlations compared to sham controls. Male canary song, unlike zebra finch song (the most commonly studied birdsong), is characterized by relatively large syllable repertoires that change across seasons. Likewise, the ways in which these syllables are arranged vary between and sometimes within individual bouts of singing (Markowitz et al, 2013). These findings suggest that LMAN may play a role in the generation and/or development of new syllables. One of the distinguishing features between male and female canary song, recognized previously in the literature, is repertoire size; females who were observed singing spontaneously in the absence of T had very small, acoustically plastic repertoires (Pesch and Guttinger, 1985; Vallet, Kreutzer, and Gahr, 1996). Female canaries with a lesion to LMAN had smaller repertoires yet more stereotypy compared to sham controls. This means that LMAN may play a role the T-associated increase in repertoire due to removal of the vocal-motor variability/plasticity necessary for vocal experimentation.

T-induces male-like singing in female canaries after prolonged exposure and is likely acting in LMAN to help facilitate masculinization of syllable repertoires of T-treated females. This masculinization of syllable repertoires is mostly likely related the variability induced in the individual iterations of syllables. By not being able to alter the acoustic structure of the already limited number of syllables produced (i.e. the increased syllable stereotypy in LMAN lesion animals) new or modified syllables cannot be introduced thereby constraining the ability of T to masculinize song. Likewise, with smaller syllable repertoires the possible number of arrangements is lower meaning that there is inherently a limited capacity to create different arrangements of notes for songs in birds with small repertoires. This suggests that LMAN is not necessarily responsible for this aspect of song in canaries.

In juvenile male songbirds that are learning to sing vocal experimentation is fundamental; it is in this experimentation that song is refined and honed. Plastic song is characterized by high phonologic variability and lacks the stereotypy of adult song (see Marler and Peters, 1982a for detailed studies in sparrows). Zebra finch song is learned through a generative process wherein a proto-syllable(s), repeated over time, transitions to a heterogenous set of syllables that are spectrally differentiated (Elliot et al, 2014; Tchernichovski et al., 2001). It is not clear if juvenile canaries engage in a similar generative process to initially learn song, nor is it altogether clear if this process is repeated across seasons as the adult animal adds or drops certain notes/phrases from the repertoire. Prior to song crystallization if LMAN is inactivated song becomes highly stereotyped and lacks acoustic variability (Elliot et al, 2014). The aforementioned coefficient of variation results replicates this finding. LMAN lesion in adult female canaries and inactivation in juvenile male zebra finch yield similar behavioral results. The variability that LMAN introduces to vocal output is necessary for juvenile song development; in open-ended learners like canaries, however, the potential role of LMAN in adult seasonal and T-induced plasticity in syllable and song repertoires is less clear.

Interestingly, it was found that the relative size of the LMAN lesion was significantly negatively correlated with the coefficients of variation for entropy variance, energy, fundamental frequency, and syllable repertoire. Likewise, LMAN lesion size was significantly positively correlated with the mean song similarity (i.e. correlation). Conversely though, the size of non-specific collateral damage in the nidopallium and hyperpallium was not correlated with any of these measures. These findings tell us that LMAN specifically is the most likely driver of vocal variability in T-induced song development and masculinization in adult female canaries. By only finding an effect in LMAN specific lesions T thus appears to be acting directly in LMAN to increase the vocal variability needed to learn and crystallize male-like song in females that do not usually sing.

There are several lines of experimental studies that have analyzed the role of LMAN in song development; however, most of these studies are limited in scope. Most studies were done in male juvenile zebra finches strictly limiting interpretation to the developmental context of a closed ended learner. However, there are a large number of songbird species where the vocal repertoire changes as the animal ages; with each new breeding season stimulating the addition or deletion of new syllables and the re-composition of familiar songs (e.g. adult canaries and European starlings; Nottebohm et al, 1986; Eens, 1997). The data presented in this study is illustrative in this regard; it suggests that LMANs role, particularly for canaries, does not necessarily change in adulthood versus early development. Likewise these findings suggest that T, in open-ended learners, is acting in LMAN (and the anterior forebrain pathway) to increase the vocal plasticity associated with seasonal/age related changes in syllable/song repertoires.

As discussed by Benton and colleagues (1998), LMANs role in adult patterns of behavior may depend on the sensitivity of the song circuit (and song behavior) to photoperiod and seasonal changes in T-activity. In adult male white-crowned sparrows, lesions to LMAN disrupted the return of many previously crystallized elements of song and the destabilization of song elements that do return during photoperiod-induced increases in song activity (Benton et al. 1998). Under normal condition, juvenile white-crowned sparrows learn song from a tutor and the song of an adult bird closely matches that of its tutor and the song that he first crystallized (Benton et al. 1998). White-crowned sparrows are seasonal breeders and across seasons the song of an individual adult male white-crowned sparrow remains largely consistent. This is important to note because it suggests that LMANs function may differ by species and more importantly by seasonal patterns of song behavior in adulthood. These data differ significantly from what was originally found in zebra finches, which are also closed ended learners, but are opportunistic breeders. It was originally thought that adult zebra finches exhibit no significant effects of LMAN lesion on singing (Bottjer et al. 1984; Morrison and Nottebohm 1993; Scharff and Nottebohm 1991). However, more recent data demonstrates that there is a more nuanced, time and context dependent effect of LMAN ablation in adult male zebra finch. For instance, LMAN lesion in adult male zebra finch significantly reduces the moment-by-moment variation observed in the undirected song of adult male zebra finch (Kao and Brainard, 2006; Thompson et al, 2011). These findings, in toto, are important because it suggests that LMAN is critical to the maintenance of adult patterns of song activity in closed-ended learners. These findings also demonstrate that LMANs function must be contextualized to species-specific patterns of brain, behavior, and steroid sensitivity.

T is essential for the actual crystallization of song; in the absence of T, castrate male sparrow song remains highly variable and lacks stereotypy when compared with age yoked T-treated castrates and control birds (Marler et al. 1988). In adult female canaries T increases song rate, masculinizes the output, and crystallizes song. However, the mechanism(s) by which LMAN and the anterior forebrain pathway open or close T-induced song development (i.e. masculinization) and crystallization is not well understood, especially for adult song development. In white crowns early treatment with T leads to early crystallization, however, the song is abnormal and close to that of males who were raised in isolation with no access to a tutor suggesting that T must act in a particular window for song to develop and crystallize correctly (Whaling et al, 1998). Song development can start in males prior to endogenous increases in T; however, T is required for song to crystallize and if exposed early in development can hasten song crystallization (Marler et al, 1988; Whaling et al 1998). These findings suggest that without an intact LMAN prior to the endogenous rise in T when song development starts and after rising T-levels crystallize behavior, song will most likely not develop correctly and the “adult” song will feature elements that are not typical. This hypothesis possibly explains why LMAN lesioned adult male white crowns sing a plastic-like juvenile song and early T-treated juvenile male white crowns sing an isolate-raised adult-crystallized song.

That said, we cannot exclude the possibility that the effects observed in this study are due to the preserved signaling of HVC to RA as well as the parallel feedback loops that send information back to HVC via medial MAN, nucleus uvaformis (UVA), and nucleus interfacialis (NIf). This possibility, nevertheless, would not necessarily contradict our hypothesis for LMANs role as a source of variability in the development of song. Because T is still able to fully activate and induce growth in other song nuclei the basic mechanics of singing are intact, meaning that the ability to sing is preserved; however, song either crystallizes sooner because there is no LMAN to introduce variability, or, song crystallizes at the same rate only it is masked by the effect of a smaller repertoire. A third alternative combines these two ideas. It is possible that song crystallizes sooner in animals with lesions because LMAN induced variability plays a role in the generation of new notes in Canaries and because there is a severely limited repertoire, songs are refined faster. This possibility would indicate that the variability LMAN injects into developing canary song is required for the full development of a male-like syllable repertoire. However, the latter two possibilities are less likely to be true because birds with an LMAN show signs of song stereotypy immediately after T-treatment (i.e. lower song coefficients of variation and increased syllable stereotypy).

More generally, the current data also suggests multiple roles and locations for T action, action that is timed to occur in such a way to optimize song development. Rising levels of circulating T increases the motivation to sing by acting directly in the POM (i.e. increased song rate) but acts directly in the song system to reduce variability overtime and crystallize song (Riters and Ball, 1999; Alward et al, 2013). Initially the song that is produced via stimulation of the POM is variable because the bird is in an early phase of song development ostensibly due to LMAN activity in RA. However, over time LMAN to RA inputs are silenced thus driving more stereotyped vocal-motor activity via HVC to RA inputs. There are many biochemical changes that T induces in LMAN and the anterior forebrain; however, there is a narrowing list of candidate mechanisms for how T closes song development. These candidates may be the same in males and as in T-treated females; however, this has yet to be tested.

Likewise one point that is also not made clear with the current data set is if the effects observed in these studies are solely due to T or perhaps are due to androgenic or estrogenic metabolites (Saldanha, et al, 1999; Saldanha, et al, 2000; Soma, et al, 2003). In the case of T induced recapitulation of song development it most likely that T or its androgenic metabolites that are acting to increase vocal plasticity and modifications to song/syllable repertoires. LMAN, unlike HVC, does not express the estrogen receptor (neither ERα nor ERβ), however, it does robustly express the androgen receptor (Balthazart, et al, 1992; Bernard, et al, 1999). Furthermore, it has been shown previously that males tend to have greater brain aromatase activity and synaptic expression than females, thus, T effects in females may be more likely due to androgenic effects and not estrogenic (Schumaker and Balthazart, 1986; Foidart, et al, 1994; Peterson, et al, 2005).

Figure 7.

Figure 7

Open in a new tab

The correlation of the amount of non-specific damage on the stereotypy of song features. Scatterplots show the overall distribution and regression line comparing the size of the non-specific damage with the coefficient of variation of song features. There were no significant correlations between the size (i.e. volume) of the lesion outside of LMAN and song features meaning that LMAN alone and not the surrounding nidopallium (or hyperpallium in some cases) is likely to play a role in the stereotyping (i.e. sensorimotor song development) of song elements (A-D).

Acknowledgments

Funding: NIH/NINDS RO1 35467 to GFB; partial support is provided by SSTC PAI P7/17 from Belgian Science Policy Office to GFB.

Footnotes

Conflict of Interest: The authors have nothing to disclose

References

  1. Alward BA, Balthazart J, Ball GF. Differential effects of global versus local testosterone on singing behavior and its underlying neural substrate. Proceedings of the National Academy of Sciences of the United States of America. 2013;110:19573–19578. doi: 10.1073/pnas.1311371110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Appeltants D, Ball G, Balthazart J. Song activation by testosterone is associated with an increased catecholaminergic innervation of the song control system in female canaries. Neuroscience. 2003;121:801–814. doi: 10.1016/s0306-4522(03)00496-2. [DOI] [PubMed] [Google Scholar]
  3. Arnold AP. Sex chromosomes and brain gender. Nature Reviews Neuroscience. 2004;5(9):701–708. doi: 10.1038/nrn1494. [DOI] [PubMed] [Google Scholar]
  4. Arnold A, Gorski R. Gonadal-steroid induction of structural sex-differences in the central nervous system. Annual Review of Neuroscience. 1984;7:413–442. doi: 10.1146/annurev.ne.07.030184.002213. [DOI] [PubMed] [Google Scholar]
  5. Arnold AP, Nottebohm F, Pfaff DW. Hormone concentrating cells in vocal control and other areas of the brain of the zebra finch (Poephila guttata) Journal of Comparative Neurology. 1976;165:487–512. doi: 10.1002/cne.901650406. [DOI] [PubMed] [Google Scholar]
  6. Arnold AP, Wade J, Grisham W, Jacobs EC, Campagnoni AT. Sexual differentiation of the brain in songbirds. Developmental Neuroscience. 1996;18:124–136. doi: 10.1159/000111400. [DOI] [PubMed] [Google Scholar]
  7. Aronov D, Andalman AS, Fee MS. A specialized forebrain circuit for vocal babbling in the juvenile songbird. Science. 2008;320:630–634. doi: 10.1126/science.1155140. [DOI] [PubMed] [Google Scholar]
  8. Ball GF, Auger CJ, Bernard DJ, Charlier TD, Sartor JJ, Riters LV, Balthazart J. Seasonal plasticity in the song control system - Multiple brain sites of steroid hormone action and the importance of variation in song behavior. Annals of the New York Academy of Sciences. 2004;1016(1):586–610. doi: 10.1196/annals.1298.043. [DOI] [PubMed] [Google Scholar]
  9. Ball GF, Hulse SH. Birdsong. American Psychologist. 1998;53:37–58. doi: 10.1037//0003-066x.53.1.37. [DOI] [PubMed] [Google Scholar]
  10. Ball GF, Macdougall-Shackleton SA. Sex differences in songbirds 25 years later: what have we learned and where do we go? Microscopy research and technique. 2001;54(6):327–334. doi: 10.1002/jemt.1146. [DOI] [PubMed] [Google Scholar]
  11. Balthazart J, Foidart A, Wilson EM, Ball GF. Immunocytochemical localization of androgen receptors in the male songbird and quail brain. Journal of Comparative Neurology. 1992;317:407–420. doi: 10.1002/cne.903170407. [DOI] [PubMed] [Google Scholar]
  12. Bentley GE, Van't Hof TJ, Ball GF. Seasonal neuroplasticity in the songbird telencephalon: A role for melatonin. Proceedings of the National Academy of Sciences of the United States of America. 1999;96:4674–4679. doi: 10.1073/pnas.96.8.4674. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Benton S, Nelson DA, Marler P, DeVoogd TJ. Anterior forebrain pathway is needed for stable song expression in adult male white-crowned sparrows (Zonotrichia leucophrys) Behavioural brain research. 1998;96(1):135–150. doi: 10.1016/s0166-4328(98)00011-4. [DOI] [PubMed] [Google Scholar]
  14. Bernard DJ, Ball GF. Two histological markers reveal a similar photoperiodic difference in the volume of the high vocal center in male European starlings. Journal of Comparative Neurology. 1995;360:726–734. doi: 10.1002/cne.903600415. [DOI] [PubMed] [Google Scholar]
  15. Bernard DJ, Ball GF. Photoperiodic condition modulates the effects of testosterone on song control nuclei volumes in male European starlings. General and Comparative Endocrinology. 1997;105:276–283. doi: 10.1006/gcen.1996.6829. [DOI] [PubMed] [Google Scholar]
  16. Bernard DJ, Bentley GE, Balthazart J, Turek FW, Ball GF. Androgen Receptor, Estrogen Receptor α, and Estrogen Receptor β Show Distinct Patterns of Expression in Forebrain Song Control Nuclei of European Starlings. Endocrinology. 1999;140(10):4633–4643. doi: 10.1210/endo.140.10.7024. [DOI] [PubMed] [Google Scholar]
  17. Bernard DJ, Eens M, Ball GF. Age- and behavior-related variation in volumes of song control nuclei in male European starlings. Journal of Neurobiology. 1996;30:329–339. doi: 10.1002/(SICI)1097-4695(199607)30:3<329::AID-NEU2>3.0.CO;2-6. [DOI] [PubMed] [Google Scholar]
  18. Brainard MS, Doupe AJ. What songbirds teach us about learning. Nature. 2002;417:351–358. doi: 10.1038/417351a. [DOI] [PubMed] [Google Scholar]
  19. Brainard MS, Doupe AJ. Translating Birdsong: Songbirds as a Model for Basic and Applied Medical Research. Annual Review of Neuroscience. 2013;36:489–517. doi: 10.1146/annurev-neuro-060909-152826. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Eens M. Understanding the complex song of the European starling: An integrated ethological approach. Advances in the Study of Behavior. 1997;26:355–434. [Google Scholar]
  21. Elliott KC, Wu W, Bertram R, Johnson F. Disconnection of a basal ganglia circuit in juvenile songbirds attenuates the spectral differentiation of song syllables. Developmental neurobiology. 2014;74(6):574–590. doi: 10.1002/dneu.22151. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Fee MS, Goldberg JH. A hypothesis for basal ganglia-dependent reinforcement learning in the songbird. Neuroscience. 2011;198:152–170. doi: 10.1016/j.neuroscience.2011.09.069. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Foidart A, De Clerck A, Harada N, Balthazart J. Aromatase-immunoreactive cells in the quail brain: effects of testosterone and sex dimorphism. Physiology & behavior. 1994;55(3):453–464. doi: 10.1016/0031-9384(94)90100-7. [DOI] [PubMed] [Google Scholar]
  24. Fusani L, Van't Hof T, Hutchison JB, Gahr M. Seasonal expression of androgen receptors, estrogen receptors, and aromatase in the canary brain in relation to circulating androgens and estrogens. Journal of Neurobiology. 2000;43:254–268. [PubMed] [Google Scholar]
  25. Gahr M. Neural song control system of hummingbirds: Comparison to swifts, vocal learning (songbirds) and nonlearning (suboscines) passerines, and vocal learning (budgerigars) and nonlearning (dove, owl, gull, quail, chicken) nonpasserines. Journal of Comparative Neurology. 2000;426:182–196. [PubMed] [Google Scholar]
  26. Goldman SA, Nottebohm F. Neuronal production, migration, and differentiation in a vocal control nucleus of the adult female canary brain. Proceedings of the National Academy of Sciences of the United States of America-Biological Sciences. 1983;80:2390–2394. doi: 10.1073/pnas.80.8.2390. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Kao MH, Brainard MS. Lesions of an avian basal ganglia circuit prevent context-dependent changes to song variability. Journal of neurophysiology. 2006;96(3):1441–1455. doi: 10.1152/jn.01138.2005. [DOI] [PubMed] [Google Scholar]
  28. Kao MH, Doupe AJ, Brainard MS. Contributions of an avian basal ganglia-forebrain circuit to real-time modulation of song. Nature. 2005;433:638–643. doi: 10.1038/nature03127. [DOI] [PubMed] [Google Scholar]
  29. Lieberburg I, Nottebohm F. High-affinity androgen binding proteins in syringeal tissues of songbirds. General and comparative endocrinology. 1979;37(3):286–293. doi: 10.1016/0016-6480(79)90002-9. [DOI] [PubMed] [Google Scholar]
  30. Marler P. Three models of song learning: Evidence from behavior. Journal of Neurobiology. 1997;33:501–516. [PubMed] [Google Scholar]
  31. Marler P, Peters S. Developmental overproduction and selective attrition - new processes in the epigenesis of birdsong. Developmental Psychobiology. 1982;15:369–378. doi: 10.1002/dev.420150409. [DOI] [PubMed] [Google Scholar]
  32. Marler P, Peters S. Structural-changes in song ontogeny in the swamp sparrow Melospiza-georgiana. Auk. 1982;99:446–458. [Google Scholar]
  33. Marler P, Peters S, Ball GF, Dufty AM, Wingfield JC. The role of sex-steroids in the acquisition and production of birdsong. Nature. 1988;336:770–772. doi: 10.1038/336770a0. [DOI] [PubMed] [Google Scholar]
  34. Meitzen J, Moore IT, Lent K, Brenowitz EA, Perkel DJ. Steroid hormones act transsynaptically within the Forebrain to regulate neuronal phenotype and song stereotypy. Journal of Neuroscience. 2007;27:12045–12057. doi: 10.1523/JNEUROSCI.3289-07.2007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Nottebohm F. Testosterone triggers growth of brain vocal control nuclei in adult female canaries. Brain Research. 1980;189:429–436. doi: 10.1016/0006-8993(80)90102-x. [DOI] [PubMed] [Google Scholar]
  36. Nottebohm F, Arnold AP. Sexual dimorphism in vocal control areas of songbird brain. Science. 1976;194:211–213. doi: 10.1126/science.959852. [DOI] [PubMed] [Google Scholar]
  37. Nottebohm F, Nottebohm ME, Crane L. Developmental and seasonal-changes in canary song and their relation to changes in the anatomy of song-control nuclei. Behavioral and Neural Biology. 1986;46:445–471. doi: 10.1016/s0163-1047(86)90485-1. [DOI] [PubMed] [Google Scholar]
  38. Olveczky BP, Andalman AS, Fee MS. Vocal experimentation in the juvenile songbird requires a basal ganglia circuit. Plos Biology. 2005;3:902–909. doi: 10.1371/journal.pbio.0030153. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Peterson RS, Yarram L, Schlinger BA, Saldanha CJ. Aromatase is pre-synaptic and sexually dimorphic in the adult zebra finch brain. Proceedings of the Royal Society B: Biological Sciences. 2005;272(1576):2089–2096. doi: 10.1098/rspb.2005.3181. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Phoenix C, Goy R, Gerall A, Young W. Organizing action of prenatally administered testosterone propionate on the tissues mediating mating behavior in the female guinea pig. Endocrinology. 1959;65:369–382. doi: 10.1210/endo-65-3-369. [DOI] [PubMed] [Google Scholar]
  41. Riters LV, Ball GF. Lesions to the medial preoptic area affect singing in the male European starling (Sturnus vulgaris) Hormones and Behavior. 1999;36:276–286. doi: 10.1006/hbeh.1999.1549. [DOI] [PubMed] [Google Scholar]
  42. Saldanha CJ, Clayton NS, Schlinger BA. Androgen metabolism in the juvenile oscine forebrain: A cross-species analysis at neural sites implicated in memory function. Journal of neurobiology. 1999;40(3):397–406. doi: 10.1002/(sici)1097-4695(19990905)40:3<397::aid-neu11>3.0.co;2-6. [DOI] [PubMed] [Google Scholar]
  43. Saldanha CJ, Tuerk MJ, Kim YH, Fernandes AO, Arnold AP, Schlinger BA. Distribution and regulation of telencephalic aromatase expression in the zebra finch revealed with a specific antibody. Journal of Comparative Neurology. 2000;423(4):619–630. doi: 10.1002/1096-9861(20000807)423:4<619::aid-cne7>3.0.co;2-u. [DOI] [PubMed] [Google Scholar]
  44. Schumacher M, Balthazart J. Testosterone-induced brain aromatase is sexually dimorphic. Brain research. 1986;370(2):285–293. doi: 10.1016/0006-8993(86)90483-x. [DOI] [PubMed] [Google Scholar]
  45. Shank SS, Margoliash D. Sleep and sensorimotor integration during early vocal learning in a songbird. Nature. 2009;458:73–U74. doi: 10.1038/nature07615. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Smith GT, Brenowitz EA, Wingfield JC, Baptista LF. Seasonal-changes in song nuclei and song behavior in Gambels white-crowned sparrows. Journal of Neurobiology. 1995;28:114–125. doi: 10.1002/neu.480280110. [DOI] [PubMed] [Google Scholar]
  47. Soma KK, Schlinger BA, Wingfield JC, Saldanha CJ. Brain aromatase, 5α-reductase, and 5β-reductase change seasonally in wild male song sparrows: Relationship to aggressive and sexual behavior. Journal of neurobiology. 2003;56(3):209–221. doi: 10.1002/neu.10225. [DOI] [PubMed] [Google Scholar]
  48. Suthers RA, Vallet E, Kreutzer M. Bilateral coordination and the motor basis of female preference for sexual signals in canary song. Journal of Experimental Biology. 2012;215:2950–2959. doi: 10.1242/jeb.071944. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Suthers RA, Vallet E, Tanvez A, Kreutzer M. Bilateral song production in domestic canaries. Journal of Neurobiology. 2004;60:381–393. doi: 10.1002/neu.20040. [DOI] [PubMed] [Google Scholar]
  50. Tchernichovski O, Mitra PP, Lints T, Nottebohm F. Dynamics of the vocal imitation process: How a zebra finch learns its song. Science. 2001;291:2564–2569. doi: 10.1126/science.1058522. [DOI] [PubMed] [Google Scholar]
  51. Tchernichovski O, Nottebohm F, Ho CE, Pesaran B, Mitra PP. A procedure for an automated measurement of song similarity. Animal Behaviour. 2000;59:1167–1176. doi: 10.1006/anbe.1999.1416. [DOI] [PubMed] [Google Scholar]
  52. Thompson JA, Basista MJ, Wu W, Bertram R, Johnson F. Dual pre-motor contribution to songbird syllable variation. The Journal of Neuroscience. 2011;31(1):322–330. doi: 10.1523/JNEUROSCI.5967-09.2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Tramontin AD, Wingfield JC, Brenowitz PA. Contributions of social cues and photoperiod to seasonal plasticity in the adult avian song control system. Journal of Neuroscience. 1999;19:476–483. doi: 10.1523/JNEUROSCI.19-01-00476.1999. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Vallet E, Kreutzer M, Gahr M. Testosterone induces sexual release quality in the song of female canaries. Ethology. 1996;102:617–628. [Google Scholar]
  55. Waser MS, Marler P. Song learning in canaries. Journal of Comparative and Physiological Psychology. 1977;91:1–7. doi: 10.1037/h0077303. [DOI] [PubMed] [Google Scholar]

RESOURCES