Thursday, December 13, 2007

Neural correlates of trust

This is the title of a new paper in PNAS by Krueger et al, that tries to find the neural correlates of conditional and unconditional trust using the sequential, reciprocal trust game. The authors premise is that conditional trust is more costly strategy compared to unconditional trust and might utilize different brain areas as well.

Conditional trust assumes that one's partner is self-interested and estimates the expected value of one's strategy with respect to the benefits of cooperating, the risk of defection, and the future value of past decisions; it causes less balanced goodwill and results in greater variance in cooperative decisions and, therefore, is cognitively more costly to maintain. In contrast, unconditional trust assumes that one's partner is trustworthy and updates the value of one's partner with respect to their characteristics and past performance; balanced goodwill occurs more quickly, allowing the partners to attain high levels of synchronicity in their decisions and, therefore, is cognitively less costly to maintain. In this work, an examination of functional brain activity supports the hypothesis that the preferential activation of different neuronal systems implements these two trust strategies.


The results of their experiments supported their initial hypothesis and they found that while Para Cingulate cortex (PcC) activation was necessary for menatlizing and initial building of trust; later unconditional and conditional trust strategies deployed different brain areas viz Septal Area (SA) and Ventral Tegmental Area (VTA) respectively.

Unconditional trust assumes that one's partner is trustworthy. During the building stage, first movers in the nondefector group showed higher activation in the PcC compared with first movers in the defector group. Through mentalizing, partners of this group verified their prior trustworthy assumption, updated the value of one's partner's strategy with respect to their past performance, and maintained a balanced goodwill toward each other, allowing them to avoid defections. By developing "better" mental models in this early stage, partners in the nondefector group accumulated sufficient mutual goodwill to become socially attached to each other and adopted an unconditional trust strategy.

During the maintenance stage, the nondefector group showed a higher activation in the SA compared with the defector group. Across groups, pairs who showed the highest trust-reciprocate history in their decisions also showed the highest activation in this region. Furthermore, analyses of pre- and postscan behavioral ratings confirmed that only nondefector pairs felt significantly closer to each other and ranked themselves as being more of a partner to the other person after the experiment. Through early mentalizing, partners in the nondefector group must have balanced goodwill more quickly, allowing them to become synchronized in their decision patterns. Brain-to-brain correlations only increased in the SA region for the nondefector group across stages, and only partners in the nondefector group became synchronized in their SA BOLD amplitudes as first movers in adjacent trials of trust games. Synchronization in the SA led to social attachment associated with a significant decrease in activation in the PcC during the maintenance stage. By adopting this cognitively less costly strategy, decision times became significantly faster for the nondefector group across stages of the experiment.

Conditional trust assumes that one's partner is self-interested. During the building stage, first movers in the defector group showed less activation in the PcC compared with the nondefector group. Through less mentalizing in the building stage, partners in this group produced higher errors in the inferences of second movers' goodwill toward them, resulting in less balanced goodwill and, therefore, in less overall trust compared with the nondefector group. More importantly, they started to trust more in the low-payoff games and less in the high-payoff games. This decision pattern implies that defectors were adapting a conditional trust strategy by evaluating the expected value of one's strategy with respect to the risks and benefits of cooperation.

During the maintenance stage, the defector group showed higher activations in the VTA compared with the nondefector group, a region linked to the dopaminergic mesolimbic reward system providing a general reinforcement mechanism to encode expected and realized reward . Across groups, pairs who shared the lowest trust-reciprocate history in their decisions also showed the highest activation in this region. By adopting a cognitively more costly strategy, partners in the defector group showed a significant increase in activation in the PcC over the experiment. Through more mentalizing in this late stage, first movers in the defector group tried to develop more accurate models about the likelihood of their partner's choices so that they could make a more advantageous decision about when to trust. The conditional trust strategy paid off less over time as total earnings decreased for the defector group (but increased for the nondefector group) across stages.


Thus, it seems that SA, based on oxytocin and vasopressin and social bonding is a more cost-effective strategy than the VTA based on dompainerigic system based on reward monitoring.

Sphere: Related Content

No comments: