In our recent fMRI study, we scanned the brains of 62 younger adults (aged 20-33) and 52 older adults (aged 65+) while they participated in a repeated trust game. This social investment game required them to identify which of the two players was trustworthy and which was not. Our observations revealed that older adults found it more challenging to discern the trustworthiness of the players, resulting in lower winnings in the game. Brain areas that are relevant for trust decisions, such as the amygdala, were less active in older people. When they were told how their investments worked out, these brain areas were even more active in older people than in younger people, indicating that they are still responsive - but something impaired their learning process. We found that there is a relationship between dopamine mechanisms in the brain, impaired learning, and loneliness in older people.

We found something even more fundamental when we investigated how subjective experience of loneliness affected these processes. All these age-related learning and brain effects were even more pronounced in lonely older people, including how strongly their dopamine-producing brain areas responded during the task. Non-lonely older people, however, were much more similar to their younger counterparts. It is already known that loneliness and isolation can significantly impair general health and the brain. Here we found yet another factor that should be addressed: if long-term loneliness impairs social cognition, including the relevant brain structures, leading to worse social interactions, what would be the best way to stop this vicious cycle?

Key Findings:

• Older people were in general less trusting and also slower in trust learning
• Computational modeling showed that older adults, particularly if lonely, had a lower learning rate (lower precision of prediction error) and were less open to changes (lower volatility beliefs)
• Older adults' amygdala involvement was lower in trust decisions but higher in outcome evaluation
• Amygdala activity relates to learning parameters and dopaminergic midbrain activity

You can find the full paper here