Activation of a well-defined brain network in songbirds plans and initiates learned vocal behavior (Nottebohm et al. 1976; Long and Fee 2008). This network for learned vocal control in zebra finch songbirds (Taeniopygia guttata) comprises two neural circuits. A posterior circuit (HVC [proper name], robust nucleus of the arcopallium [RA], and tracheosyringeal subdivision of the hypoglossal nucleus [nXII]) projects to the vocal organ for song stereotypy while an anterior circuit (Area X [proper name], lateral magnocellular nucleus of the nidopallium [lMAN], and dorsolateral anterior thalamic nucleus [DLM]) allows for song plasticity and has a crucial role in vocal learning (Brainard and Doupe 2000; Kao et al. 2005). The posterior and the anterior neural circuits of this vocal control network are interconnected (Fig. 1a). Singing-dependent activation of this vocal control network can be assessed through the expression of activity-dependent genes, notably immediate early genes. (Jarvis and Nottebohm 1997; Jin and Clayton 1997; Kimpo and Doupe 1997; Jarvis et al. 1998, 2000; Whitney et al. 2014).
As adults, male zebra finches produce learned social female-directed songs to court potential mates, social juvenile-directed songs to tutor progeny, and social and non-social undirected song (Derégnaucourt et al. 2013; Chen et al. 2016). RA, Area X, and lMAN of the vocal control network show different activation patterns during social female-directed singing compared to non-social undirected singing (Jarvis et al. 1998). This social context-dependent difference in vocal control network activation is also observed through changes in vocal motor song output (Jarvis et al. 1998), as well as at the level of single neuron firing rates within the vocal control network (Kao et al. 2008). How external social contexts are conveyed to affect the activity of the vocal control network and subsequent singing behavior is not well understood, although current studies implicate dopaminergic influence (Sasaki et al. 2006; Kubikova et al. 2010; Ihle et al. 2015; Gadagkar et al. 2016).
Social context-dependent neural activation patterns of the vocal control network have been primarily attributed to dopaminergic inputs originating in the ventral tegmental area (VTA) (Heimovics and Riters 2005; Hara et al. 2007; Riters 2012). The VTA sends dopaminergic projections to Area X (Lewis et al. 1981; Reiner et al. 2004b; Castelino et al. 2007; Gale et al. 2008), HVC (Appeltants et al. 2000), and RA (Appeltants et al. 2002) in the vocal control network – of which Area X and RA show social-context-dependent cellular activation. The lack of context-dependent modulation of cellular activation in HVC, despite VTA-innervation, suggests a secondary method for conveying social stimuli presence. High levels of circulating oxytocin, a well-conserved nonapeptide that binds promiscuously to a family of G-protein-coupled receptors, correlate with increased social flocking behaviors in zebra finches (Goodson et al. 2009). Some suggest that oxytocin, primarily produced in the paraventricular nucleus of the hypothalamus (PVN), could serve as a neuromodulator of context-dependent behaviors.
Recent work in mammals has highlighted a mechanism for oxytocin-mediated social-context-dependent control of dopamine release from VTA. Oxytocin+ projection neurons, originating in the PVN, synapse directly and selectively onto dopaminergic neurons in the VTA (Hung et al. 2017; Xiao et al. 2017). VTA-projecting oxytocin neurons are more active following social encounters (Hung et al. 2017); subsequently, as oxytocin delivery to the VTA increases, so do the firing rates of dopamine neurons (Xiao et al. 2017). These dopaminergic neurons in the VTA that show oxytocin-mediated activity patterns express the oxytocin and vasotocin 1A receptors (Theofanopoulou et al. 2021). Both receptors are sensitive to oxytocin (Xiao et al. 2017). Altogether, these data suggest that during social encounters, oxytocinergic PVN to VTA projection neurons release oxytocin, which binds to either the oxytocin or vasotocin 1A receptors on dopaminergic VTA neurons, increasing neural firing rates in VTA. This cascade results in further signaling at the ultimate projection sites of these VTA dopamine neurons. Axonal tract-tracing studies performed in avian and mammalian species show that the PVN either uni- or bi-directionally connects with all regions of a brain network implicated in social behavior (Fig. 1c & Table 1). These data suggest that the strict control of synaptic oxytocin release may be dependent on signals originating in one or multiple regions of this social behavior network.
In vertebrates, many social behaviors, including those related to reproductive success, are regulated by the activity of six interconnected brain regions collectively referred to as the social behavior network (Newman 1999; Goodson 2005; O’Connell and Hofmann 2011; Kabelik et al. 2018; Eswine et al. 2019; Horton et al. 2020; Prior et al. 2021). This social behavior network is completely interconnected (Newman 1999). Defined regions of the avian social behavior network are in the cortex (lateral septum [LS]), the thalamus (medial bed nucleus of the stria terminalis [BSTm], medial preoptic area [POM], anterior hypothalamus [AH], and ventromedial hypothalamus [VMH]), and the midbrain (central gray [CG]) subregions of the brain (Fig. 1b). The involvement of the social behavior network in species-specific adaptive, mainly courtship, behaviors seem to be conserved across vertebrates. Specific analyses of this network have been performed in rodents (Wood and Newman 1995; Newman 1999), amphibians (Laberge et al. 2008), reptiles (Sakata et al. 2000; Sakata and Crews 2004; Kabelik et al. 2018), fish (Goodson and Bass 2002), and birds (Edwards et al. 2020). Additionally, the chemo-architecture of the social behavior network shows conservation across vertebrate species (Goodson et al. 2004; Kingsbury et al. 2011). Regions of the social behavior network also include continuous subdivisions of the extended amygdala (Johnston 1923a; Olmos and Ingram 1972). The mammalian extended amygdala (lateral bed nucleus of the stria terminalis [BSTl], BSTm, & medial amygdala) is implicated in managing emotional responses to external stimuli. (Johnston 1923a; Olmos and Ingram 1972). Strong connections to the medial amygdala may prime the social behavior network to deliver the emotional relevance of external stimuli to other networks within the brain in a context-dependent manner. The potential integration of the social behavior network to other defined neural systems is not well established (Kelly 2022). Nonetheless, the involvement of the social behavior network in motivating the context-appropriate performance of learned behaviors is highly plausible.
We hypothesize that the social behavior network may send context-specific information to PVN, facilitating oxytocin-dependent activation of dopaminergic neurons within VTA, which in turn enables context-appropriate song production and neural activation in the vocal control network of adult male zebra finches. In the present study, we test this hypothesis by assessing whether activation patterns of the social behavior network are different during the performance of a learned behavior in non-social and social contexts. Using brain tissue from adult male zebra finches performing either non-social undirected songs or social female-directed songs, we analyzed the transcription of activity marker EGR1 using fluorescence in situ hybridization. We found upregulation of EGR1 mRNA in the AH, BSTm, CG, LS, and VMH of the social behavior network, the BSTl, and the PVN during social singing contexts compared to non-social singing contexts. These data provide correlational evidence that the social behavior network does play a role in the differential production of female-directed versus undirected songs in zebra finches.
Table 1. Afferent and efferent projections between BSTl, PVN, and VTA and the social behavior network. In addition to intra-network axonal projections, regions of the social behavior network receive and make afferent and efferent axonal projections from and to regions outside this network. References to previous tract-tracing experiments are shown for avian and non-avian species. 1 = (Johnston 1923b); 2 = (Olmos and Ingram 1972); 3 = (Montagnese et al. 2004); 4 = (Sawchenko and Swanson 1983); 5 = (Otake and Nakamura 1995); 6 = (Dong et al. 2001); 7 = (Riters and Alger 2004); 8 = (Pfaff and Conrad 1978); 9 = (Prewitt and Herman 1998); 10 = (Balthazart et al. 1994); 11 = (Delville et al. 2000); 12 = (Li et al. 2014); 13 = (Korf 1984); 14 = (Absil et al. 2002); 15 = (Simerly and Swanson 1988); 16 = (Fahrbach et al. 1989); 17 = (Shimogawa et al. 2015); 18 = (Omelchenko and Sesack 2010); 19 = (Ntamati et al. 2018); 20 = (Veening et al. 1982); 21 = (Meibach and Siegel 1977); 22 = (Swanson and Cowan 1979); 23 = (Iyilikci et al. 2017); 24 = (Haglund et al. 1984); 25 = (Risold and Swanson 1997); 26 = (Cádiz-Moretti et al. 2016); 27 = (Hung et al. 2017); 28 = (Xiao et al. 2017); 29 = (Beier et al. 2015); 30 = (Aransay et al. 2015)
Nuclei
|
Afferent Projections
|
Efferent Projections
|
|
|
Aves
|
Non-Aves
|
|
Aves
|
Non-Aves
|
BSTl
|
BSTm
|
|
1,2
|
BSTm
|
|
1,2
|
LS
|
3
|
|
PVN
|
|
4-6
|
POM
|
7
|
|
|
|
|
PVN
|
|
8
|
|
|
|
PVN
|
AH
|
|
4, 8, 9
|
AH
|
10
|
8, 11
|
BSTm
|
|
6
|
BST
|
|
8
|
CG
|
|
12
|
LS
|
|
8
|
LS
|
3, 13
|
|
POM
|
14
|
8, 15
|
POM
|
7, 10
|
4, 7, 9, 10, 15
|
VMH
|
|
8, 16, 17
|
VTA
|
13
|
|
VTA
|
13
|
27-29
|
VTA
|
BST
|
|
29
|
BSTm
|
|
20
|
CG
|
|
18, 19
|
LS
|
13
|
22, 30
|
LS
|
|
21, 29
|
POM
|
7, 10
|
11, 24-26
|
POM
|
7, 10, 23
|
15
|
PVN
|
13
|
|
PVN
|
13
|
29
|
VMH
|
10
|
|
VMH
|
10
|
|
|
|
|