Neurobiology: Our minds distinguish between various social influences
“This question is all the more relevant in today’s world of social media and the manipulation of opinions, as many people rely on the opinions of others to form their own view,” says Dr. Bahador Bahrami from the LMU Department of Psychology.
Together with Dr. Ali Mahmoodi from the University of Oxford and other researchers, he characterized brain activities that occur if people are socially influenced to change their opinion. The study has now been published in PLOS Biology. “We were able to show that our brain solves social conflicts — that is, differences of opinion — via the same neural machinery that it uses to solve its own internal, subjective conflicts,” summarizes Bahrami. “A specific region of the brain takes two factors into account: how confident we are in our opinion and how polite we are obliged to be toward others.”
Study uses computer-based game and functional magnetic resonance imaging
In their study, the researchers used a computer-based game. Participants in the experiment had to try and remember the position of a dot displayed on a screen. They gave confidence values for their answers. However, they were allowed to revise their guesses after they had seen the answer of a computer or of a virtual ‘partner’ to whom they had been introduced before the experiment. In reality, all answers were provided by computers.
Bahrami’s team tracked the brain activity of all test subjects during the game using functional magnetic resonance imaging (fMRI). This non-invasive method allows areas of the brain with high activity — that is to say, with high oxygen consumption — to be displayed with high spatial resolution.
The study showed: People tended to adjust their answers when their confidence was low, irrespective of whether they thought their partner was human or not. This informational influence was controlled by activity in the dorsal anterior cingulate cortex (dACC) of the brain, a region of the cerebral cortex.
Test subjects also exhibited more conformity toward other opinions if they received confirmation from their communication partner. This normative influence arose only when they believed that their partners were human, as did the correlation with dACC activity. Moreover, the normative influence was associated with stronger functional connections between the dACC and other social processing regions of the brain. This was not the case for the informational influence.
As part of the study, Bahrami and his colleagues also wanted to know what their results meant for AI applications, which are increasingly being used in all kinds of areas. “We established that the human brain only feels the need for politeness when it’s interacting with other people and not with a purportedly artificial (albeit intelligent) agent,” says the LMU researcher. In view of the burgeoning use of artificial intelligence in a wide variety of fields, Bahrami concludes, this is an important topic for further projects.