Acoustic Dimensions Underlying a Type of Nonhuman Primate Conversational Communication
Soltis, Bernhards, Newman
In one study of vocalizations in squirrel monkeys (Saimiri sciureus), standard playback methodology was employed to analyze vocal responses to the “chuck” vocalization, an individually distinct call that is exchanged in conversational format among familiar adult females. We found that subjects responded with a chuck more often to familiar chucks from their own group than they did to unfamiliar chucks from the same or a different species. Subjects sometimes responded to unfamiliar chucks, however. The results of a discriminant function analysis showed that females were most likely to answer unfamiliar chucks that were close in acoustic structure to familiar chucks compared with unfamiliar chucks that were acoustically distinct. Finally, we found that, within social groups, females were more likely to emit a chuck in response to a playback chuck from their affiliated partners than to respond to chucks from unaffiliated group members. In an ongoing study extending this work, we are attempting to determine which aspects of the acoustic structure of the chuck vocalization are salient in individual recognition. Preliminary results show that subjects respond in kind when presented with a playback chuck from familiar group members, but that chuck responses are suppressed when only partial components of the chucks are presented, suggesting that the whole call is the unit of perception in such antiphonal calling behavior. Furthermore, changing the pitch of the chuck calls also suppressed chuck response, suggesting that the frequency of the call is also important to individual recognition.

Hormonal Correlates of Affiliative Behavior in Nonhuman Primates
Soltis, Roberts, Newman
In a second project, we are examining the hormonal correlates of affiliative behaviors in squirrel monkeys. Squirrel monkeys engage in three types of affiliation: mothering, allomothering by other adult females that includes nursing, and same-sex affiliation among adult females. We are investigating cortisol, which is implicated in the stress response, and prolactin, which has been implicated in affiliative behavior. We expected prolactin to be associated with allomothering because of its clear role in lactation, and speculated that it may be associated with same-sex adult friendships because it has also been implicated in affilliative behaviors that do not involve lactation, such as male parenting behavior.

Preliminary results suggest that serum prolactin is positively associated with allomothering and with high levels of same-sex adult affiliation and that serum cortisol is higher in animals with low overall rates of social interaction. The interpretation of results is problematic, however, because capture of animals to draw blood appears to have affected relevant hormone levels. We therefore searched for a noninvasive means of collecting samples for assay and ruled out saliva as a possibility because prolactin was not detectable in saliva. We are presently collecting urine from several mixed-sex groups of squirrel monkeys. We will investigate relationships between urinary cortisol and prolactin and (1) levels of overall affiliation among females, and (2) mothering and allomothering behavior. This preliminary research will lay the groundwork for experiments designed to test causal relationships between hormones and behavior.

Other research completed during the past year used common marmosets (Callithrix jacchus). In this species, adult male-female pairs form extended families, and several cohorts of young may reside together with their biological parents. The offspring and their father engage in extensive play. We hypothesized that opiate activity modulates social play and predicted that administration of morphine would facilitate social play, whereas pretreatment with the morphine antagonist naloxone would block morphine's effects. We studied 11 juvenile marmosets from five families. All subjects had their twin as a play partner, one family had triplets, and two families contained older siblings. Fathers played more extensively in three families containing only the focal twins and parents. Morphine administration was associated with significantly increased social play, including wrestling, biting, chasing, runaway and approach-withdrawal behavior. Morphine's effect on play was attributable to the focal subject, since treatment did not affect play initiated by others. Nonsocial categories of play, such as toy manipulation and locomotor play, and other social behaviors, such as time spent huddled, were unaffected by treatment. Twittering, a vocalization used by young to solicit food, attention, and play, increased after morphine administration. The combination treatment blocked morphine's effects and was associated with significantly increased time spent sitting alone and sleeping. Morphine administration significantly increased overall activity. We conclude that social play is facilitated by opiate activation. Fathers may compensate for fewer sibling play partners in small groups, suggesting that play may have important adaptive developmental consequences for juvenile common marmosets.

Little is known about the regional brain basis of human maternal behavior. In rodents, thalamocingulate circuit lesions impair maternal behavior. In a previous functional MRI (fMRI) study of four human mothers, we found that this same thalamocingulate circuit activates when mothers listen to infant cries (strong elicitors of maternal behavior). During the past year, we completed a larger and more rigorously designed follow-up fMRI study, in which we recruited psychiatrically healthy, right-handed, breastfeeding first-time human mothers with infants four to eight weeks old. We used fMRI to measure their brain activity while they listened to recorded infant cries alternating with white noise control sounds. Consistent with our thalamocingulate circuit hypothesis, the mothers displayed more thalamo-cingulate circuit blood flow (anterior cingulate, mid-cingulate, and anterior thalamic activity) with the infant cries than with the control sounds. Moreover, mothers who reported higher social attachment to people in general displayed higher cry-specific thalamocingulate circuit blood flow.