The role of self-blame and worthlessness in the psychopathology of major depressive disorder

1The influential revised learned helplessness model predicts that vulnerability to major depressive disorder (MDD) is due to a bias to blame oneself for failure in an overgeneralised way resulting in decreased self-worth, hopelessness and depression. Overgeneralised self-blame is associated with excessive self-blaming moral emotions(, e.g. guilt, shame, disgust/contempt towards oneself). This is in contrast to the most widely employed model of depression that claims an overall increase in negative and reduction in positive emotions.

2Recent evidence using experimental probes of moral emotions in remitted MDD has pointed to a relative proneness to feeling disgust/contempt towards oneself with a reduction in disgust/contempt towards others in support of the revised learned helplessness model. The clinical literature, however, has provided contradictory evidence regarding the role of worthlessness and self-blaming emotions in MDD, which the model predicts to be of core pathophysiological importance.

3In support of the model, the combined guilt and worthlessness item in DSM was found to be most distinctive of current MDD compared with a generalized anxiety disorder group. Further support has been provided by the largest transcultural study on MDD, where the feeling of inadequacy (including self-worthlessness) was reported as a consistent symptom of depression. Subsequent studies, however, have reported a wide variation in the consistency of guilt/worthlessness which was most often reported as a single item following DSM. The frequency of guilt/worthlessness in current MDD was found to be between 20% in Australia and Japan, 50% in Benin, and 70–80% in the USA and France. DSM worthlessness was separately reported in another US study as being present in 61% of current MDD patients.

4The clinical assessment of self-blaming emotions has classically been restricted to guilt, which was found only in a subgroup of patients. Although, early studies claimed transcultural variation in the frequency of guilt, more recent evidence suggests that guilt is experienced in a large subgroup of patients across different cultures. This is contradicted by a large study reporting markedly lower frequencies of guilt in Korean compared to US patients with MDD. On a cautionary note, this study used item comparisons of the Hamilton Depression scale without using semi-structured interviews to elicit the information and without reporting how items were translated and culturally adapted.

5The discrepancy in reported frequencies of guilt and worthlessness is likely due to methodological as well as sampling differences. The semi-structured interviews for DSM were designed to provide reliable diagnoses rather than to assess single symptoms or the coherence of symptoms. As a consequence, the criterion threshold for different items on the DSM varies between symptoms rendering a direct comparison and analyses of symptom coherence invalid. Furthermore, the role of self-blaming emotions such as self-disgust/contempt, found to be elevated in MDD using specific instruments of assessment, remains elusive. This is because clinical assessments have solely reported guilt or non-specific reports of self-blame.

6As Jaspers, the founder of phenomenological psychopathology, noted on the analyses of symptom-complexes (, p. 582ff): There are different aspects of the relation of symptoms within a symptom-complex: (1) frequency of symptom co-occurrence, and (2) coherence of symptoms by being related to a common aspect or function. The latter aspect has been emphasized by Schneider when discussing symptoms: “Their connectedness must be due to a normal complex of psychic function, which complex has been affected by the illness”. At the time of this theory, a lack of knowledge about neurobiologically valid models of many higher cognitive functions hampered the success of this approach. Aided by advances in social cognitive neuroscience, we can now aim at isolating symptom-complexes which are likely to be associated with a restricted set of cognitive-anatomical syndromes. The neural architecture underpinning the tendency to overgeneralize self-blaming emotions in MDD has recently been elucidated. This supports the neurobiological validity of self-blaming emotional biases in MDD and has prompted the current study into the phenomenology of associated clinical symptoms.

7Here, we investigated the following hypotheses derived from the revised learned helplessness model: (1) The feeling of inadequacy/worthlessness is a consistent symptom of MDD and co-occurs with other core symptoms when assessed using an instrument designed to assess individual symptoms (AMDP,) rather than those used in DSM validation studies. (2) The type of self-blaming emotion experienced during depressive episodes differs between patients and is not restricted to guilt. (3) Negative emotions towards others are infrequent and do not co-occur with core depression symptoms.

8This study was approved by the South Manchester NHS Research Ethics Committee. All participants gave written informed consent and were compensated for time and travel costs. 132 (37 male) patients with major depressive disorder (MDD), fully remitted for >6 months, were enrolled (n=121 medication-free at time of study) and had no current, as well as no relevant past co-morbid axis-I disorders (see also Supplementary Methods).

9Residual symptoms were assessed using the Montgomery-Åsberg-Depression-Scale (, MADRS) and psychosocial functioning was assessed using the Global-Assessment-of-Functioning (GAF,) Scale (Axis V, DSM-IV). Remitted MDD patients had GAF scores indicating minimal or absent symptoms and high psychosocial functioning (mean=84.4±6.6) and MADRS scores that were well below the cut-off for depression of 10 (mean=1.2±1.6). Their average age was 32.8±12.3 (range 18–65), years of education mean was 16.6±2.4 (range 11–22) and their age at onset ranged from 8 to 52 (mean=21.5±8.6, for further clinical details and cultural background see Supplementary Tables S2 & S3).

10We assessed 132 patients using a phenomenological psychopathology-based instrument (AMDP) translated from German, adding new items to assess moral emotions (n=94 patients). In accordance with the SCID-I, we asked patients about the worst two weeks of their last and most severe episode. Importantly all symptoms were measured on the same 4 point scale (0=absent, 1=mild/minimal, 2=moderate, 3=severe) without pre-defining different diagnostically relevant thresholds for different symptoms as is done on the SCID-I. English translations of symptom labels correspond to published symptom label translations. Instructions for ratings were based on definitions in the German version. In addition to the existing standard items of feelings of inadequacy and guilt, we developed additional items to assess moral emotions more systematically (the Moral Emotion Addendum to the AMDP, see Supplementary Methods). This was based on our previous work on experimental probes of moral emotions and their distinct neural correlates. Inter-rater reliability for the AMDP and moral emotion items were very high (Supplementary Table S4).

11All analyses were carried out using SPSS21 (www.spss.com) at p=.05, 2-sided. Symptom ratings were transformed into two categories: absent to mild (0 and 1) vs. moderate to severe (2 and 3). Hierarchical cluster analysis (binary Euclidean distance, Ward method) was employed to identify symptom coherence.

12Fig. 1 shows that feelings of inadequacy (feeling worth less than others or feeling that one fails) very closely co-occurred with hopelessness and depressed mood. Together these symptoms were the most consistent, occurring in more than 90% of patients. This core cluster closely co-occurred with another cluster of symptoms of high consistency around 90%: lack of drive (this item includes lack of energy), affective rigidity (i.e. an inability to respond emotionally including anhedonia), blunted affect, and social withdrawal.

13Fig. 2 shows that self-blaming emotions occurred in most patients (82%). Interestingly, self-disgust/contempt (46%) was slightly more frequent than guilt (39%), followed by shame (20%). Anger or disgust towards others was experienced by 26% of patients. The cluster analysis revealed that guilt as measured on the AMDP and on the moral emotion addendum closely co-occurred with self-disgust/contempt and clearly separated from other moral emotions. Furthermore, the cluster comprising guilt and self-disgust/contempt occurred most closely with core depressive symptoms (hopelessness and feelings of inadequacy), whilst the other moral emotions showed only very distant relationships with core depressive symptoms.

14Consistent with the learned helplessness model, feelings of inadequacy and hopelessness were part of the core syndrome of major depression. Self-blaming emotions were very common but self-disgust/contempt was slightly more frequent than guilt. Guilt and self-disgust most closely co-occurred with core depressive symptoms, whilst negative emotions towards others were infrequent and not associated with core symptoms.

15Our results are in keeping with the finding that feelings of inadequacy are a consistent feature of MDD, whilst feelings of guilt are more variable. This contrasts with another study reporting guilt in over 80% of MDD patients which is similar to our overall frequency for self-blaming emotions. These differences are likely due to the way guilt was assessed.

16Population-based studies using the Composite International Diagnostic Interview yielded only around 20% of patients with current MDD fulfilling the DSM worthlessness/guilt criterion. DSM-IV-TR (, p. 350) states that “Blaming oneself for failing to meet occupational or interpersonal responsibilities as a result of depression is very common and, unless delusional, is not considered sufficient to meet this criterion.” In practice it is very difficult to assess whether self-blame and inadequacy is felt because of being ill or because of failing generally, which explains the variation and probably underreporting of the item in DSM-based studies.

17The frequency of other-blaming emotions in our study is similar to previous reports for MDD with anger attacks and hostile depression and to the frequency of anger/irritability in current MDD. In keeping with our findings, observed feelings of worthlessness to be common in patients with outwardly directed anger. Unlike our study, however, studies of depression with anger attacks did not exclude patients with low psychosocial functioning outside of depressive episodes, and in fact showed an association of anger attacks with personality disorders. Patients with subsyndromal and mixed bipolar disorders, however, were not rigorously excluded from our study and may have inflated the frequency of other-blaming emotions, given the higher likelihood in these patients.

18Symptom assessment was retrospective and could have led to an underestimation of the overall frequency due to recall biases, but this is unlikely to have biased patients towards particular emotions relative to others. Assessing remitted MDD patients, however, allowed us to exclude co-morbidity with more rigor than by studying current MDD during which co-morbid personality disorders are harder to detect.

19Feelings of inadequacy, hopelessness and self-blaming emotions are closely associated with depressed mood and are associated with high distress. This calls for sensitive probes of these symptoms in clinical assessment tools. Future studies are needed to confirm these results in patients with current MDD and to investigate how distinctive self-blaming emotions are of MDD without co-morbid personality and relevant axis-I disorders as studied here.

20The authors have no conflict of interest to declare.

21Jorge Moll, John F. William Deakin, and Roland Zahn designed the study. Karen Lythe, Jennifer Gethin, Sophie Green and Roland Zahn carried out the study. Karen Lythe, Jennifer Gethin and Roland Zahn analyzed the data. Roland Zahn wrote the first draft of the paper. All authors significantly contributed to and have approved the final manuscript.

22RZ was funded by a University of Manchester Stepping Stones Fellowship & an MRC clinician scientist fellowship (G0902304). SG and JG were funded by MRC and EPSRC PhD studentships respectively. The funders had no role in study design; in the collection, analysis and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication.

23The project was funded by an MRC Clinician Scientist Fellowship (to RZ, G0902304), EPSRC (to JG), and MRC (to SG) PhD studentships. We are grateful to Clifford Workman for phone-screening participants and Rolf-Dieter Stieglitz for commenting on the manuscript.