BRIEF REPORTS Emotional Intensity of Idiographic Sad Memories in Depression Predicts Symptom Levels 1 Year Later Jonathan Rottenberg University of South Florida Jutta Joormann, Faith Brozovich, and Ian H. Gotlib Stanford University When cued with generic happy and sad words, depressed individuals have been found to articulate contex- tually impoverished memories of autobiographical events. Although this pattern predicts a worse symptomatic course of disorder in some depressed samples, longitudinal findings with the cue-word paradigm are inconsistent. To address the etiological significance of autobiographical memories outside the cue-word paradigm, the authors used an idiographic interview in which depressed participants generated memories of their happiest and saddest lifetime events. Each memory was coded for detail and emotional intensity. At a 1-year follow-up, participants’ levels of depressive symptoms were reassessed. Lower emotional intensity of saddest memories predicted higher levels of depressive symptoms at follow-up. Several implications for understanding sadness and emotional disclosure in depression are discussed. Keywords: depression, autobiographical memories, emotion, disclosure Major depression is characterized by specific patterns of memory functioning (Burt, Zembar, & Niederehe, 1995). One pattern that has been the focus of considerable research concerns the relative inability of depressed persons to retrieve specific autobiographical memories when they are cued with generic happy and sad words (e.g., Goddard, Dritschel, & Burton, 1996; Swales, Williams, & Wood, 2001; Wil- liams & Dritschel, 1988). The tendency for depressed persons to offer scriptlike, contextually impoverished, autobiographical memories in response to valenced cue words has been labeled overgeneral memory (OM; Williams, 1996). This pattern of memory has been interpreted in terms of a stable affect regulation strategy that allows individuals to minimize the negative affect that might be associated with potentially painful memories, such as those that arise from the experience of childhood adversity (Williams, 1996; Williams, Stiles, & Shapiro, 1999). To gain a better understanding of OM as a vulnerability factor for depression, several investigators have examined the relation between OM and the symptomatic course of depression. These researchers have used the Autobiographical Memory Test (AMT; Williams & Broadbent, 1986) to assess autobiographical memory. In this procedure, participants are asked to retrieve a specific autobiographical memory (i.e., a time and place where something specific happened to them) in direct response to individual cue words that are presented orally or written on cards; words are typically positive (e.g., happy, safe) and negative (e.g., clumsy, lonely) in emotional valence. To date, studies in which this pro- cedure has been used have yielded an inconsistent pattern of findings. In two samples, OM in response to positive but not negative cue words predicted higher levels of depression symp- toms at follow-up, even after controlling for initial levels of depressive symptoms (Brittlebank, Scott, Williams, & Ferrier, 1993; Dalgleish, Spinks, Yiend, & Kuyken, 2001). In contrast, the results of a similar investigation indicated that OM in response to negative but not positive cue words predicted higher levels of depressive symptoms at follow-up (Peeters, Wessel, Merckelbach, & Boon-Vermeeren, 2002). Finally, Brewin, Reynolds, and Tata (1999) found no relation between OM in response to positive or negative cue words and subsequent levels of depressive symptoms after controlling for initial symptom levels. Limitations of the Existing Literature One limitation of previous studies of autobiographical memory is their exclusive reliance on a single procedure, the AMT. The AMT has several features that may cloud efforts to understand the etiological significance of autobiographical memory functioning. For example, there is some indication that the AMT may be sensitive to the degree of impairment of depressed persons (Phil- lips & Williams, 1997), a factor that is likely to vary across samples. Instability of results using the AMT may also reflect differences in implementation of the AMT across studies. Indeed, the results of a recent meta-analysis (van Vreeswijk & de Wilde, Editor’s Note. Michael Eid served as the guest editor for this article.— RJD Jonathan Rottenberg, Department of Psychology, University of South Florida; Jutta Joormann, Faith Brozovich, and Ian H. Gotlib, Department of Psychology, Stanford University. Additional material for this article is on the Web at http://dx.doi.org/ 10.1037/1528-3542.5.2.238.supp We are grateful to Jennifer Champion, Jennifer Hildner, and Kathryn Dingman for their help in conducting this study. This research was sup- ported by National Institute of Mental Health Grant MH59259 awarded to Ian H. Gotlib. Correspondence concerning this article should be addressed to Jonathan Rottenberg, Department of Psychology, University of South Florida, PCD 4118G, 4202 East Fowler Avenue, Tampa, FL 33620-7200. E-mail: jrottenb@cas.usf.edu Emotion Copyright 2005 by the American Psychological Association 2005, Vol. 5, No. 2, 238 –242 1528-3542/05/$12.00 DOI: 10.1037/1528-3542.5.2.238 238