Is depression associated with reduced optimistic belief updating?

Abstract

When asked to evaluate their probability of experiencing a negative life event, healthy individuals update their beliefs more following good news than bad. This is referred to as optimistic belief updating. By contrast, individuals with depression update their beliefs by a similar amount, showing reduced optimism. We conducted the first independent replication of this effect and extended this work to examine whether reduced optimistic belief updating in depression also occurs for positive life events. Replicating previous research, healthy and depression groups differed in belief updating for negative events (β = 0.71, 95% CI: 0.24, 1.18). Whereas healthy participants updated their beliefs more following good news than bad, individuals experiencing depression lacked this bias. However, our findings for positive events were inconclusive. While we did not find statistical evidence that patterns of belief updating between groups varied by valence (β = -0.51, 95% CI: -1.16, 0.15), mean update scores suggested that both groups showed largely similar updating for positive life events. Our results add confidence to previous findings that depression is characterized by negative future expectations maintained by reduced updating in response to good news. However, further research is required to understand the specificity of this to negative events, and into refining methods for quantifying belief updating in clinical and non-clinical research.

Keywords: belief updating; cognitive biases; depression; optimism.

Figure 1.

Hypothesized differences between absolute mean updates in the likelihood of life events according to desirability and valence of life event, for healthy controls and participants with moderate to severe levels of depression.

Figure 2.

Main task procedure.

Figure 3.

Example of desirable versus undesirable categorizations in positive and negative life events. For negative life events initial overestimations are desirable, whereas initial underestimations are undesirable. For positive life events, initial underestimations are desirable, whereas initial overestimations are undesirable.

Figure 4.

Consort flow diagram for the healthy and depression groups.

Figure 5.

Mean update scores by group, desirability and valence. Error bars represent standard errors. For negative life events (left panel) we replicated Korn et al.'s [11] findings of reduced optimistic updating in individuals experiencing depression; whereas participants in the healthy group showed greater updates for desirable versus undesirable information, participants in the depression group showed similar levels of updating for desirable and undesirable information. For positive life events, we hypothesized that the effect of optimism bias would be consistent with those observed for negative life events. As illustrated by the red dashed bar, mean update scores did not support our hypothesis; while participants in the healthy group showed greater updating for desirable versus undesirable information for negative life events, they showed a similar change in beliefs for desirable and undesirable information for positive life events, that is, no optimistic belief updating for positive events.

Figure 6.

Mean initial estimates for negative and positive life events by group. Error bars represent standard errors. Participants in the depression group initially estimated their chances of experience a negative life event as greater, and chances of experiencing a positive life event as lower, than participants in the healthy group.

Figure 7.

Mean scaled update scores (updates divided by estimation errors) by group, valence and desirability. Error bars represent standard errors. For negative life events (left panel), whereas participants in the healthy group showed greater updates for desirable versus undesirable information, participants in the depression group showed greater updates for undesirable versus desirable information. For positive life events (right panel), both the healthy and depression groups showed greater updates for desirable versus undesirable information.