Sexually dimorphic activation of dopaminergic areas depends on affiliation during courtship and pair formation

Abstract

For many species, dyadic interaction during courtship and pair bonding engage intense emotional states that control approach or avoidance behavior. Previous studies have shown that one component of a common social brain network (SBN), dopaminergic areas, are highly engaged during male songbird courtship of females. We tested whether the level of activity in dopaminergic systems of both females and males during courtship is related to their level of affiliation. In order to objectively quantify affiliative behaviors, we developed a system for tracking the position of both birds during free interaction sessions. During a third successive daily interaction session, there was a range of levels of affiliation among bird pairs, as quantified by several position and movement parameters. Because both positive and negative social interactions were present, we chose to characterize affiliation strength by pair valence. As a potential neural system involved in regulating pair valence, the level of activity of the dopaminergic group A11 (within the central gray) was selectively reduced in females of positive valence pairs. Further, activation of non-dopaminergic neurons in VTA was negatively related to valence, with this relationship strongest in ventral VTA of females. Together, these results suggest that inhibition of fear or avoidance networks may be associated with development of close affiliation, and highlight the importance of negative as well as positive emotional states in the process of courtship, and in development of long-lasting social bonds.

Keywords: courtship; dopamine; social behavior; video tracking.

Figure 1

Schematic of system for automatic tracking of individual birds during interaction sessions. Lower left and right panels are images obtained by two webcams positioned outside the front and side of a cage containing a female and a male zebra finch. Software could identify and discriminate between female and male birds (outlined in gray and orange, respectively) based on color differences. In the side view, detection of female bird is prevented by the obscuring male. Positions detected by both cameras were combined to obtain three-dimensional coordinates of both birds in the cage (upper panel, axis labels indicate dimensions in mm).

Figure 2

Information from tracking system used to obtain pair valence, and relation of courtship behaviors to valence. (A) A Principal Component Analysis (PCA) was performed on five behavioral features of interactions of female and male birds. Results from the first two principal components, which accounted for 88% of the variance, are plotted for each pair. Mappings of each behavioral factor are overlaid projected within this plane, with the length of the line equal to the PC1 and PC2 components. The five behavioral features are: X1—% time spent jumping between perches, X2—% time both birds on same perch, X3—average distance between birds when both on same perch, X4—distance moved when jumping between perches, X5—% time both birds on opposite perches. Symbols for pairs which had positive and negative valences, illustrated in (B), are labeled with “+” and “−,” respectively. The final calculated valence of each pair is coded with a grayscale map, from black = −1 to white = +1. B. Relationship of PC1 and PC2 to the five behavioral features used. (C) Presence of female and male birds on either perch in the cage for each video frame. Video frames when only the female or male was on perch 1 (pch1) or perch 2 (pch2) are indicated with gray and orange, respectively. Frames when both female and male were on the same perch are indicated by magenta. Upper and lower perch position plots represent data from day 3 recordings of positive valence (0.38, indicated by “+” to left of symbol in A) and negative valence (−0.24, indicated by “−” to left of symbol in A) pairs. (D) Total time singing by each male during first and third day interaction sessions. Dotted line indicates equal duration of singing on both days. Most males sang more during first day of interaction with a female. The valence of each male during the day 3 session is indicated by a gray scale map, with -1 to 1 indicated by black to white. The amount of singing in the third session was negatively related to valence. (E) Fighting during male singing bouts in the first session was associated with low valence during the third session. Typical aggression features associated with low, medium, and higher valence indicated by black, medium gray, and light gray were female retreat, female attack, and male attack/female nonretreat, respectively. Valence scores were significantly lower during day 3 in pairs that fought in day 1 (∗).

Figure 3

Schematic diagram of dopaminergic areas in which expression of c-Fos was quantified, and example of colocalization. (A) From top (caudal) to bottom (rostral), sections included A11 in CG (black) and A8 (gray), A10 in VTA, A11 in SCI, and A14 (black) and A15 (gray). Expression in A8 was not quantified, but is included for reference to (B). (B) Image of coronal section containing A11 in CG (black) and A8 (gray) Expression in A8 was not quantified due to difficulty in consistently defining its border. (C) Region outlined in white box in (B) is expanded to illustrate labeling of dopaminergic neurons with tyrosine hydroxylase (blue) and nuclear expression of c-Fos protein (brown). Scale bars in (A–C) indicate 2 mm, 250 μm, and 50 μm, respectively.

Figure 4

Relationship of pair valence to expression of c-Fos in dopaminergic and non-dopaminergic neurons. (A) In dopaminergic neurons in central gray (CG/PAG), the level of expression of c-Fos was negatively correlated with valence in females (filled symbols) but not males (empty symbols). Dotted lines represent linear fit to female data. (B) In non-dopaminergic neurons in VTA, for females there was a stronger negative relationship of valence to c-Fos expression in ventral than dorsal VTA, and for males this negative relationship was similar in both regions. Because the level of expression in dorsal VTA was about double that in ventral VTA, the scale of y-axis indicates separate percentages for both ventral (lower values) and dorsal (higher values) regions.