Self-esteem – our evaluation of our own worth – is shaped by what other people think of us. It increases when others appreciate and value us, and decreases when we are rejected and start to question our own worth.

Maintaining a positive sense of self is crucial for mental health and well-being. People with low self-esteem are more likely to develop psychiatric conditions, such as anxiety disorders, eating disorders and depression. Despite its importance for mental health, it was not known how the brain accumulates social feedback to determine our self-esteem.

To address this question, Geert-Jan Will of the Wellcome Trust Centre for Neuroimaging at University College London, along with colleagues, developed a computational model that precisely predicts how self-esteem changes from moment to moment as people learn what others think of them.

Activity in the brain was measured while young adults received approving or disapproving feedback from peers who had seemingly viewed their online character profile. After every second or third peer judgment, participants reported their current level of self-esteem.

Unexpected Negative Feedback Hurts Most

The researchers found that self-esteem depended both on whether other people liked the participants and on whether they were liked or disliked more than expected. Self-esteem decreased the most when participants received negative feedback from someone they expected to receive positive feedback from.

The model then identified signals in specific parts of the brain that explain why self-esteem goes up and down according to the feedback received. Moment-to-moment changes in self-esteem correlated with activity in the ventromedial prefrontal cortex, which is a brain region important for valuation.

[caption id=“attachment_92744” align=“aligncenter” width=“680”]Task structure and feedback probabilities Task structure and feedback probabilities.
(A) Participants were provided with a visual cue that indicated which group a rater belonged to (assigned according to their overall disposition to provide approving or disapproving feedback; see panel on the right).
They then made a prediction as to whether the rater would like or dislike them before receiving feedback. After every 2–3 trials, participants were asked to indicate their current level of self-esteem.
(B) Probability of receiving approval or disapproval feedback was dependent on the rater’s group (signaled by a color cue).
Participants received approving feedback in 85%, 70%, 30%, and 15% of the trials. Group colors were randomized across participants.
© 2017 eLife Sciences Publications Ltd. Republished via Creative Commons Attribution license.[/caption]

Will et al. combined the model with responses to questionnaires that assessed psychiatric symptoms, and showed that vulnerable individuals had elevated responses in a part of the brain called the anterior insula. In vulnerable individuals, activity in this region of the brain was strongly coupled to activity in the part of the prefrontal cortex that explained changes in self-esteem.

[caption id=“attachment_92747” align=“aligncenter” width=“680”]neural correlates self-esteem Neuroimaging results plotted separately for high and low ‘interpersonal vulnerability’ participants (based on median split) to facilitate interpretation.
(A) Social prediction errors correlated with activity in a cluster in bilateral ventral striatum extending into sgACC.
(B) Trial-by-trial updates in self-esteem upon receipt of feedback correlated with activity in vmPFC (BA 14m and BA 32pl).
(C and D) Vulnerability modulated the expression of prediction error responses in left anterior insula (extending into inferior frontal gyrus) and insula-vmPFC coupling during self-esteem updates.
Images are thresholded at t > 2.7 (panels A and B), t > 3.6 (panel C) and t > 4.0 (panel D) with no cluster-extent threshold for display purposes. Data are represented as mean ± SEM.
© 2017 eLife Sciences Publications Ltd. Republished via Creative Commons Attribution license.[/caption]

A better understanding of the brain mechanisms that mediate a decline or improvement in self-esteem may help to find more effective treatments for a range of mental health problems. The work was supported by the Wellcome Trust, Medical Research Council, Biomedical Research Council, and Max-Planck-Gesellschaft (UCL Centre for Computational Psychiatry and Ageing Research).

Geert-Jan Will Is a corresponding author Robb B Rutledge Michael Moutoussis Raymond J Dolan Neural and computational processes underlying dynamic changes in self-esteem eLife 2017;6:e28098

© 2017 eLife Sciences Publications Ltd. Republished via Creative Commons Attribution license.

For future updates, subscribe via Newsletter here or Twitter