Human-like brain hemispheric dominance in birdsong learning

Abstract

Unlike nonhuman primates, songbirds learn to vocalize very much like human infants acquire spoken language. In humans, Broca's area in the frontal lobe and Wernicke's area in the temporal lobe are crucially involved in speech production and perception, respectively. Songbirds have analogous brain regions that show a similar neural dissociation between vocal production and auditory perception and memory. In both humans and songbirds, there is evidence for lateralization of neural responsiveness in these brain regions. Human infants already show left-sided dominance in their brain activation when exposed to speech. Moreover, a memory-specific left-sided dominance in Wernicke's area for speech perception has been demonstrated in 2.5-mo-old babies. It is possible that auditory-vocal learning is associated with hemispheric dominance and that this association arose in songbirds and humans through convergent evolution. Therefore, we investigated whether there is similar song memory-related lateralization in the songbird brain. We exposed male zebra finches to tutor or unfamiliar song. We found left-sided dominance of neuronal activation in a Broca-like brain region (HVC, a letter-based name) of juvenile and adult zebra finch males, independent of the song stimulus presented. In addition, juvenile males showed left-sided dominance for tutor song but not for unfamiliar song in a Wernicke-like brain region (the caudomedial nidopallium). Thus, left-sided dominance in the caudomedial nidopallium was specific for the song-learning phase and was memory-related. These findings demonstrate a remarkable neural parallel between birdsong and human spoken language, and they have important consequences for our understanding of the evolution of auditory-vocal learning and its neural mechanisms.

Fig. 1.

Schematic side views of the songbird (A and B) and human (C) brain. (A) Regions depicted in a light shade show increased neuronal activation when the bird hears song. The NCM and CMM regions are assumed to contain the neural substrate for tutor song memory. (B) Nuclei HVC, Av, RA, LMAN, and Area X show increased neuronal activation when the bird is singing. (C) In the human brain, Broca’s area is most importantly involved in speech production, whereas Wernicke’s area is mainly involved in speech perception and understanding. Modified from Moorman et al. (20). Area X, area X of the striatum; Av, avalanche; CLM, caudolateral mesopallium; CMM, caudomedial mesopallium; CN, cochlear nucleus; DLM, medial subdivision of the dorsolateral nucleus of the anterior thalamus; DM, dorsomedial subdivision of the nucleus intercollicularis of the mesencephalon; L1, L2, and L3, subdivisions of field L; LLD, lateral lemniscus, dorsal nucleus; LLI, lateral lemniscus, intermediate nucleus; LLV, lateral lemniscus, ventral nucleus; LMAN, lateral magnocellular nucleus of the anterior nidopallium; LMO, lateral oval nucleus of the mesopallium; MLd, dorsal part of the lateral nucleus of the mesencephalon; NIf, interfacial nucleus of the nidopallium; nXIIts, tracheosyringeal portion of the nucleus hypoglossus (nucleus XII); Ov, nucleus ovoidalis; PAm, nucleus para-ambiguus medullaris; RA, robust nucleus of the arcopallium; RAm, nucleus retroambiguus medullaris; SO, superior olive; Uva, nucleus uvaeformis; VTA, ventral tegmental area.

Fig. 2.

Photomicrographs of juvenile zebra finch brains showing Zenk immunostaining. Representative images at the level of the NCM, HVC, and hippocampus (HP) are shown for the silence, novel, and tutor stimulus groups. (Scale bar: 0.2 mm.)

Fig. 3.

Zenk expression in the brains of juvenile and adult male zebra finches. (Left to Right) Mean number of Zenk-immunopositive neurons per square millimeter is shown for the different brain regions. (Upper) Results for juvenile male zebra finches are shown. There is left-sided dominance in the NCM of juveniles in response to tutor song but not to novel song or silence. In the HVC, there is left-sided dominance irrespective of the stimulus presented. The levels of neuronal activation in the hippocampus did not differ between stimulus groups or hemispheres. (Lower) Results for adult male zebra finches. The mean activation level is higher in the NCM of adult zebra finches that were exposed to tutor song than to silence, but there is no difference between the hemispheres. There is left-sided dominance in the HVC, irrespective of the stimulus presented, similar to the juveniles. The levels of neuronal activation in the hippocampus did not differ between stimulus groups or hemispheres. Black bars represent the left hemisphere, and gray bars represent the right hemisphere. Error bars represent the SEM.

Fig. 4.

Correlation between lateralization ratio and degree of song similarity. (A) Lateralization ratios ([L − R]/[L + R]) were calculated for each subject from the number of Zenk-immunopositive cells per square millimeter in the NCM of juvenile zebra finches and were correlated with song similarity scores. In the birds that were exposed to tutor song (▲), the correlation was significant (solid line, plotted by linear regression, R² = 0.67). There were no significant correlations in the novel (○, dotted line) or silence (, dashed line) group. (B) Spectrograms of juveniles that produced a good or poor imitation of their tutor’s song. (Lower Left) Juvenile had a song similarity of 73.3% with its tutor. (Lower Right) Juvenile had a song similarity of 49.4% with its tutor.

Fig. 5.

Photomicrographs of adult zebra finch brains showing Zenk immunostaining. Representative images at the level of the NCM, HVC, and hippocampus (HP) are shown for the silence and tutor stimulus groups. (Scale bar: 0.2 mm.)