The introduction to this thesis demonstrates that depression research has tended to focus solely on biological or psychological factors, which has resulted in a dualistic understanding of the disorder. For example, chapter two presents a meta-analysis of 34 studies using the well-validated psychological treatment of group cognitive behaviour therapy for depression, none of which assessed biological measures of depression, and none that studied psychological and biological functions simultaneously. Evidence from two fields of research clearly shows that both biochemical and cognitive functioning are involved in depression. In order to advance depression research, there is a need to integrate these key variables into a cohesive theory.
In 1989, Free and Oei proposed such a theory named the Circular Processes model. This model posits that key biochemical and cognitive processes underlying depressed mood interact with one another in a reciprocal or circular manner. This thesis was designed to test the Circular Process model in two samples of depressed adults. Three major hypotheses were tested: that there would be relationships among key biochemical and cognitive measures in depressed patients, that the variables would produce interaction effects in their prediction of depression levels, and that changes would be demonstrated in both cognitive and biochemical variables in patients who respond to treatment with cognitive therapy. Four studies were conducted to test the above hypotheses.
In study one, relationships among cognitive variables and urinary catecholamines were investigated in a sample of 84 moderately depressed outpatients. No significant bivariate correlations were found. However, significant cognitive-by-biochemical interactions were found in the prediction of depression in these adults, and added between 7 and 13% further variance in depression scores after the biochemical and cognitive main effects were accounted for. This warranted further investigation of relationships during the treatment period.
In study two, sixty-seven participants completed 24 hours of group cognitive behaviour therapy (GCBT) over a 12-week period, during which time 60% of the sample demonstrated a clinically reliable improvement in mood. Onethird of the sample was also prescribed antidepressant medication throughout the GCBT, however medication use was not significantly related to outcome in the analyses. There were no significant cognitive-by-catecholamine interactions using mid-treatment variables. However, end-of-treatment interaction terms were significantly related to mood change over the 12 months following treatment. Negative automatic thoughts by catecholamine interaction terms accounted for nearly 60% of the variance in mood change scores after the main effects were accounted for. Further, it was revealed that all of the cognitive measures changed significantly more in treatment responders than in non-responders. In contrast, changes in catecholamine measures were not related to treatment response. This could have been due to the fact that only one of the three major monoamine systems (noradrenergic) was measured in studies one and two. Also, the sample comprised of moderately depressed outpatients may not have demonstrated marked biochemical dysfunctions. Studies three and four were designed to address these possibilities.
In study three, relationships among cognitive and biochemical variables were investigated in a sample of more severely depressed hospital patients. Dopaminergic and serotonergic metabolites were assessed in addition to the noradrenergic metabolites. At intake, there was a significant correlation between adrenaline clearance and negative automatic thoughts. However, no interactive models were significantly predictive of depression.
In study four, relationships were investigated during treatment with cognitive therapy. Forty-three patients received 24 hours of GCBT over a four-week period, and 44% of the sample achieved clinically reliable mood change. A majority of the sample (91%) was prescribed antidepressant medication, although medication use and type of medication were not related to treatment response. The interactions of end-of-treatment adrenergic metabolites (adrenaline, metadrenaline and VMA) and dysfunctional attitudes were significantly related to mood change over the six months follow-up period. Adrenergic functioning moderated the relationship between attitudes at the end of treatment and subsequent mood change. Higher adrenergic metabolite clearance was interpreted as a greater stress response or agitation, which influenced the link between attitudes and mood change. Interestingly, there were no significant interactions involving the other monoamine systems (noradrenergic, dopaminergic or serotonergic). In repeated measures analyses, all the cognitive variables were significantly related to treatment response, as were adrenaline and metadrenaline levels. The other biochemical measures were unrelated to treatment outcome. The serotonin metabolite 5-HIAA changed significantly more in responders than in non-responders over the six months follow-up period.
Taking the results of the four studies into account, support for the Circular Process model of depression was mixed. While there was some evidence that interactions between cognitive and biochemical variables were related to depression levels cross-sectionally, and to subsequent mood changes, these relationships were not consistently found across analyses. Adrenergic function was related to response to GCBT for depression, and moderated the relationship between dysfunctional attitudes and subsequent mood change. The main effects revealed some important information. At mid-treatment and end-of-treatment, biochemical main effects were in the same order of importance in terms of their prediction of subsequent mood change. Noradrenergic and dopaminergic effects proved stronger than serotonergic effects. These results supported the monoamine theories of depression, as well as raising the issue of potential attenuation of monoamine dysfunction due to multiple medication use in the hospital patients. At mid-treatment, there were no main effects for cognitive variables, but at the end of treatment, dysfunctional attitudes main effects accounted for an average variance of 25% in subsequent mood change, while automatic negative thoughts and thoughts of hopelessness accounted for mean variances of 19% and 17% respectively.
The model was modified on the basis of these findings, and extended to include pathways for future research. For example, the link between stressful events, adrenergic function, cognition, and mood requires empirical testing. It was recommended that specific symptoms (or symptom clusters) be used as the criterion for analysis rather than global depression scores, to better reflect the links between a particular monoamine transmitter and a particular group of symptoms. For example, noradrenaline function and energy / interest symptoms; dopamdne function and drive / motivation symptoms; and serotonin function and symptoms of impulsivity.
This research has contributed to the understanding of depression by integrating two previously disparate theories of depression into a psychobiological model, and testing the model in two groups of depressed patients. It is the first research using GCBT for depression to assess cognitive and monoamine functions simultaneously during treatment and follow-up periods. There was evidence that GCBT altered adrenergic functioning, and this was consistent with previous studies reporting links between response to CBT for depression and other biological measures such as thyroid function and sleep EEG parameters. Future research will no doubt build stronger bridges between the psyche and biology, and thus improve our understanding of the phenomena of depression.