A Psychophysiological Test of Depression-Related Probability Bias

Dataset associated with unpublished preprint, archived in Aperta with the same title. 

Probability bias is the tendency to think negative events have relatively high probabilities, and positive events have relatively low probabilities. It is robustly associated with depressive symptoms. However, probability bias has only been demonstrated with self-report measures, meaning that probability bias could be an artefact of negative response tendencies in people with depression. We attempted to demonstrate depression-related probability bias using physiological measures. Ninety-seven students read sentences from a computer screen, indicated that they would, would not, or might receive a painful shock. Depressive symptom severity was indeed correlated with skin conductance response (SCR) amplitude upon reading uncertain (‘might’) sentences, and this could not be accounted for by overall SCR responsiveness, learning processes, or intolerance of uncertainty. Unexpectedly however, this correlation was negative, indicating that participants with higher symptom scores tended to show smaller SCRs to uncertain cues. It is possible that more-depressed participants tended to hold pessimistic expectancies before they saw each sentence, and so were less surprised to learn that a shock may be imminent. These results suggest that depression-related pessimism and probability bias are more complex than they appear, and show the importance of combining self-report measures with other methods when studying cognition in depression.