Naltrexone Alters Responses to Social and Physical Warmth: Implications for Social Bonding

Tristen K. Inagaki; Laura I. Hazlett; Carmen Andreescu

Disclosures

Soc Cogn Affect Neurosci. 2019;14(5):471-479. 

In This Article

Results

Effect of Naltrexone on Neural Activity to Social Warmth

To examine the effect of naltrexone on neural activity to the social warmth task, analyses were constrained to the a priori defined VS and MI mask. VS and MI activity was greater to messages from close others than strangers (left t(76) = 6.138, P < 0.001; right t(76) = 4.493, P < 0.001). As confirmatory evidence of the results corrected at ROI, messages from close others (vs strangers) also lead to greater VS and MI activity corrected at whole brain (Supplementary Materials).

Naltrexone, however, reduced VS and MI activity to the social warmth task as compared to placebo (left t(75) = 1.859, P = 0.034; right t(75) = 2.198, P = 0.016; Figure 1). In other words, naltrexone reduced VS and MI activity to messages from one's own friends, family and romantic partners, consistent with the current hypotheses.

Figure 1.

Neural activity in response to social warmth task. Naltrexone (vs placebo) reduced VS and MI activity to messages from close others (vs strangers). Mean parameter estimates from left- and right-sided activity pictured above.

Effect of Naltrexone on Feelings of Social Connection to Social Warmth

As a manipulation check, feelings of social connection to reading the messages were first assessed ignoring drug condition. Feelings of social connection were indeed greater to messages from close others (M = 4.671, s.d. = 0.871) than strangers [M = 1.966, s.d. = 0.720; F(1,76) = 612.614, P < 0.001], mirroring the pattern from the neural data and confirming the task elicited the intended feelings. The effect of drug on feelings of social connection to the messages was in the expected direction but did not reach statistical significance [F(1, 76) = 2.017, P = 0.080]. For the descriptive statistics, see Supplementary Table 2.

Associations Between Feelings of Social Connection and Neural Activity to Social Warmth

In the placebo group, feelings of social connection were positively related to neural activity to the social warmth task (left: r = 0.325, P = 0.027, 95% CI = [0.027, 0.577]; right: r = 0.290, P = 0.043, 95% CI = [0.006, 0.541]; Figure 2). Thus, greater feelings of social connection were associated with greater VS and MI activity to reading messages from close others (vs strangers). However, there was no association between feelings of social connection and neural activity to the social warmth task in the naltrexone group (left: r = −0.200, P = 0.112, 95% CI = [−0.426, 0.091]; right: r = −0.114, P = 0.245, 95% CI = [−0.370, 0.183]). Further, the correlations in the placebo group were different from the correlations in the naltrexone group (left: z = 2.24, P = 0.013; right: z = 1.71, P = 0.043), suggesting naltrexone erased the subjective experience–brain relationship that was present in the placebo group.

Figure 2.

Associations between feelings of social connection and neural activity in response to the social warmth task. There was a positive correlation between feelings of social connection and activity in the VS and MI to the social warmth task such that greater feelings of social connection to close others were associated with greater VS and MI activity to messages from close others (vs strangers). The same association was absent in the naltrexone group, suggesting naltrexone erased the subjective experience–brain relationship.

Effect of Naltrexone on Neural Activity to Physical Warmth

The effect of naltrexone on VS and MI activity to physical warmth revealed a pattern similar to the social warmth task. Once again mirroring results corrected at the whole brain (Supplementary Materials), there was greater VS and MI activity to the warm than the neutral and cool objects. Importantly, the main effect of temperature was qualified by an interaction [left: F(1, 75) = 5.027, P = 0.014; right: F(1, 75) = 2.624, P = 0.055; Figure 3] such that the difference between VS and MI activity to the warm and cool objects was eliminated by naltrexone. Thus, VS and MI activity was greater to the warm than cool object in the placebo group [left: t(36) = 2.672, P = .006; right: t(36) = 1.991, P = .027], but in the naltrexone group, there was no difference in neural responding [left: t(39) = .182, P = .429; right: t(39) = .165, P = .435]. In other words, naltrexone reduced VS and MI activity to physical warmth. Interactions between drug and temperature for the comparison of the warm to neutral object were not significant for neural results corrected at the ROI threshold (Ps > 0.06).

Figure 3.

Neural activity in response to the physical warmth task. Examining neural responding to the physical warmth task revealed an interaction between drug and temperature. VS and MI activity was greater to holding a warm object compared to holding a cool object in the placebo group, but there was no difference in the naltrexone group. Parameter estimates from left-sided VS and MI activity plotted on the y-axis.

Effect of Naltrexone on Feelings of Social Connection to Physical Warmth

Naltrexone's effect on self-reported feelings was evaluated with a drug (naltrexone vs placebo) × temperature (warm vs cool) interaction. In a replication of prior work (Inagaki and Eisenberger, 2013; Inagaki et al., 2015), a main effect of temperature [F(1,77) = 146.745, P < .001] revealed that the warm object led to greater feelings of social connection (M = 1.612, s.d. = 1.120) than the cool object [M = .135, s.d. = 0.979, t(78) = 12.004, P < 0.001]. Importantly, the main effect was qualified by a marginal interaction with drug [F(1, 77) = 2.721, P = 0.052]. Given the primary hypothesis that naltrexone might affect physical warmth-induced feelings of social connection, the interaction was further interrogated to assess the direction of the effects.

Naltrexone (vs placebo) did not alter feelings of social connection to the cool object [M placebo = 0.117, s.d. = 0.932; M naltrexone = 0.154, s.d. = 1.037; t(77) = 0.168, P = 0.434]. Naltrexone, however, did move feelings of social connection to the warm object in the expected direction, but the reduction did not reach statistical significance [M placebo = 1.792, s.d. = 1.069; M naltrexone = 1.427, s.d. = 1.155; t(77) = 1.456, P = 0.075]. There were no associations between feelings of social connection and neural activity to the physical warmth task (Ps > 0.200).

The effects of naltrexone on ratings of thermal intensity were also evaluated. As expected, the warm stimuli were rated as significantly warmer than the cool stimuli [t(78) = 26.597, P < 0.001]. However, in a replication of our previous findings (Inagaki et al., 2015), there was no effect of naltrexone on perceptions of thermal intensity for either the warm [t(77) = 1.030, P = 0.153] or cool objects [t(77) = 1.007, P = 0.159], suggesting naltrexone did not alter the more objective perceptions of warmth.

Shared Neural Activity: Social and Physical Warmth

The hypothesized shared neural mechanisms between social and physical warmth were evaluated with conjunction analyses. In a replication of our previous findings, the conjunction analysis revealed shared neural activity in the left [x = −42, y = 12, z = −9, t(36) = 4.446, k = 24] and right MI [x = 42, y = 12, z = −12, t(36) = 4.855, k = 27] and caudate, extending into the right VS [x = 3, y = 18, z = 3, t(36) = 4.599, k = 29] to messages from close others (vs strangers) and the warm (vs neutral) object (Figure 4). In those who took naltrexone, however, there was no such overlap. Indeed, the only region to show overlapping activity between social and physical warmth was the left superior temporal gyrus [x = −33, y = 12, z = −21, t(39) = 4.235, k = 17].

Figure 4.

Results for conjunction between social warmth (messages from close others > strangers) and physical warmth (warm > neutral). In the placebo group (panel A), the only regions to show overlapping activity between social and physical warmth were the caudate, extending into the right VS, and left and right MI. In the naltrexone group (panel B), however, the only overlapping activity was in the left superior temporal gyrus. L = left, R = right, VS = ventral striatum, MI = middle insula, STG = superior temporal gyrus.

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....