Overgeneral autobiographical memory in adolescents at risk for depression

Abstract

Autobiographical memory functioning is implicated in the course and onset of depression in adults (Williams et al., 2007), and there is preliminary evidence that adolescents with a diagnosis of depression have a bias towards retrieving overgeneral autobiographical memories (Kuyken & Howell, 2006; Kuyken, Howell, & Dalgleish, 2006; Park, Goodyer, & Teasdale, 2002). In two independent studies we asked whether adolescents at risk for depression exhibit autobiographical memory deficits. In the first community study of 179 adolescents, risk was operationalised as higher scores on neuroticism. We found that neuroticism was associated with greater retrieval of categoric memories to negative cue words and that severity of depressive symptoms mediated this relationship. In the second study, groups of formerly depressed (n=15) and never depressed adolescents (n=15) were matched on age, gender, depressive symptoms and verbal fluency, and again compared on their autobiographical memory functioning. Mirroring the findings from Study 1 we found that risk, operationalised here as a history of depression in remission, was associated with greater retrieval of categoric memories to negative cue words. Taken together, these studies show that risk for depression is associated with a greater tendency to retrieve categoric memories to negative cue words and suggest that autobiographical retrieval may be implicated in onset of depression in adolescence.

Keywords:

• Adolescence
• Autobiographical memory
• Cognition
• Neuroticism
• Depression
• Risk

Acknowledgements

We are grateful to the staff and students at the participating schools. Rachel Day provided research assistance. Willem Kuyken was part funded by the UK National Institutes of Health Research. Tim Dalgleish was funded by the UK Medical Research Council.

Notes

Some of the questionnaires contained missing data. Rather than use data imputation methods, we elected to use a conservative approach of using the subsets of the dataset for each analysis where complete data on the relevant measures were available. We report the n for each analysis.