Overgeneral autobiographical memory predicts higher prospective levels of depressive symptoms and intrusions in borderline patients

ABSTRACT

Overgeneral memory (OGM), the tendency to retrieve categories of events from autobiographical memory instead of single events, is found to be a reliable predictor for future mood disturbances and post-traumatic symptom severity. Patients with borderline personality disorder (BPD) often report co-morbid episodes of major depressive disorder (MDD) and post-traumatic stress disorder (PTSD). Therefore, we investigated whether OGM would predict depression severity and (post-traumatic) stress symptoms in BPD patients. At admission (N = 54) and at six-month follow-up (N ≥ 31), BPD patients completed the Structured Clinical Interview for DSM-IV Disorders, the Assessment of DSM-IV Personality Disorders, the Autobiographical Memory Test, the Beck Depression Inventory—2nd edition (BDI-II), and the Impact of Event Scale. OGM at baseline predicted (a) higher levels of depressive symptoms at follow-up and (b) more intrusions related to a stressful event over and above baseline levels of borderline symptoms, depressive symptoms, and intrusions, respectively. No association was found between memory specificity and event-related avoidance at follow-up. Despite previous findings suggesting that OGM in BPD is less robust than in MDD and PTSD, our results suggest that memory specificity in BPD patients may have some relevance for the course of depressive and stress symptomatology in BPD.

KEYWORDS:

• Overgeneral memory
• borderline personality disorder
• depression severity
• post-traumatic symptom severity
• intrusions

Acknowledgements

We would like to thank all the patients of UPC KU Leuven, Campus Kortenberg, and Psychiatric Hospital Duffel who participated in the study, as well as all staff members of the departments hosting these patients. Special thanks go to Vanessa Aerts, Tom Baeten, Marion Franssen, Dr Philip Haake, MD (passed away), Jos Lemmens, Marianne Nieulandt, and Stef Stuelens, for referring patients for inclusion, and to David Stuers for his contribution in calculating interrater reliability of the AMT output.

Disclosure

No potential conflict of interest was reported by the authors. Although some of the authors are employed by UPC KU Leuven, Campus Kortenberg, or the Psychiatric Hospital, Duffel, we declare that these organisations did not provide any input concerning the study design; the collection, analysis, and interpretation of the data; the writing of the report; and the decision to submit the paper for publication.

Notes

1. To preserve maximum readability, we do not report the results for the proportional indices of specific (%S) and categoric (%GC) memories, unless they are different from the results using the plain number of specific (S) and categoric (GC) memories.

2. For five of them, both the BDI-II and the IES at T2 were missing. One participant completed the BDI-II at T2, but not the IES at T1. Another one completed both the IES and the BDI-II at T2, but had missing data on the SCID-I at T2. BDI-II scores at T2 were missing for two other participants (of which one also had missing data on the AMT at T2).

3. Besides the standard AMT, we also administered a personalised AMT for each respondent, consisting of the 10 most self-discrepant self-guides the respondent formulated. According to Conway, Singer, and Tagini (2004), cues that tap into domains that are highly self-discrepant towards one's actual self-concept evoke more overgeneral memories. However, as reported elsewhere, pAMT administration in BPD patients resulted in less categoric memories and more omissions than standard AMT administration (Van den Broeck, 2015). Also, in this study, no association was found between memory specificity measured with the pAMT at T1 and depression severity and the IES scales at T2.

4. Adding the proportion of specific memories in Block 2 to the equation also resulted in a significant ΔR²,  = .07 with F_change = 4.70, p = .04, with β = −.29, t = −2.17, p = .04, when predicting BDI-II scores at T2.

5. Adding the proportion of specific memories in Block 2 to the equation resulted in a marginally significant ΔR²,  = .096 with F_change = 3.110, p = .052, with β = −.31, t = −2.03, p = .052, when predicting IES total score at T2.

Additional information

Funding

Personnel costs of this research have been funded by UPC KU Leuven Campus Kortenberg, Leuvensesteenweg 517, B-3070 Kortenberg, Belgium, and by the Research Foundation Flanders (Fonds Wetenschappelijk Onderzoek Vlaanderen) [grant number 1900114N].