Thursday, July 23, 2020

Cognitive behavioural therapy for depression

1.     Introduction

Major depressive disorder (referred to as “depression” for the remainder of this essay) is defined by the American Psychiatric Association (APA) as depressed mood or lack of interest or pleasure (for most of the day, nearly every day, for at least two weeks), along with a minimum of other symptoms, for example, altered sleep, difficulty in concentrating or suicidality (APA, 2013), and has been identified as one of the leading causes of disability by the World Health Organisation (WHO, 2020). Depression is associated with poorer performance and presenteeism in the workplace (Cocker et al., 2011). Though there are multiple causal factors in suicide, a substantial minority of people with depression die by suicide; the lifetime risk in people with untreated depression has been estimated at 20% (Gotlib & Hammen, 2002). Given this great personal and economic cost, there has been much enthusiasm for finding means for alleviating depression.

Cognitive Behavioural Therapy (CBT) is a form of psychotherapy that focuses on targeting maladaptive cognitions and behaviours in order to improve psychological well-being. It has been used for a number of psychological disorders, including depression. Although the length of a course of therapy of CBT can vary, given that CBT is generally structured and time-limited (e.g. Hazlett-Stevens & Craske, 2002), it is well-suited for controlled empirical investigation. It is also widely available - the National Health Service in the United Kingdom has led substantial efforts to increase the uptake of CBT via the Improving Access to Psychology Therapies program (see Clark et al., 2009). This is turn has led to a substantial number of potential participants for research studies.

In this essay I will provide a critical overview of existing evidence concerning CBT and depression. I will begin by appraising the evidence base for whether it is a successful therapy for treating depression. I will then briefly discuss evidence for the risks associated with CBT, followed by a discussion of relevant neuropsychological research, although in this latter section, I will emphasise that this area of research is in its early stages, and faces inherent barriers to offering a “fundamental”, reductionist account of depression and CBT.


2.     Evidence concerning CBT and depression

CBT: does the evidence suggest that it works for depression?

In the last decade, a number of primary research studies have demonstrated successful use of CBT for treating depression that occurs in the context of other conditions, such as Parkinson’s Disease (Dobkin et al., 2011), chronic obstructive pulmonary disease (Hynninen et al., 2010), and HIV-infected drug users (Safren et al., 2012). This trend in primary research might be interpreted as indicating that the efficacy of CBT for depression per se is a foregone conclusion, although this trend can also be explained by the hypothesis that depression that is co-morbid with other health conditions may respond in a particular way to CBT, and CBT itself may be adapted to accommodate this. However, there is ongoing research interest on the impact of CBT on depression itself.

Indeed, given the quantity of research that has been conducted on CBT, a number of articles in this area are meta-analyses, i.e. quantitative syntheses of the results of multiple previous primary research studies. In appraising existing literature on efficacy, these meta-analyses have tended to focus on randomised control trials, i.e. studies in which patients are randomly assigned to either an experimental group (in this context individual CBT), or one or more comparison/control group(s) – for research of this nature, the term “comparison” is perhaps preferable (Lilienfeld et al., 2015). Following the experimental/comparison intervention, an outcome of interest, in this context depression, is compared between the two groups. By randomly assigning research participants to experimental or comparison conditions, the researcher should avoid systemic bias with regard to the baseline characteristics of the experimental and comparison groups, although inadvertent baseline differences between the groups can still be controlled for statistically. For example, if the experimental group and the comparison group differ in age, this can be entered as a covariate into an analysis of covariance (e.g. Field, 2009).

A review of existing meta-analyses evaluated a range of meta-analyses, appraising them on whether they included only randomised control trials, whether they weighted effect sizes according to sample size, whether moderating variables were included and whether the heterogeneity of effect sizes and outliers was analysed (Butler et al., 2006). Butler et al. found that CBT was somewhat superior to antidepressant medications in treating depression in adults. They also criticised a previous review by Parker et al. (2003) which had suggested that CBT was not as effective as previously suggested in treating depression, on the basis that it had not explicated what criteria they had used to select the primary studies and meta-analyses they reviewed.

More recently, López-López et al. (2019) conducted a systematic review focusing specifically on randomised control trials of CBT in depression. They found that CBT interventions was associated with a greater reduction in depression scores compared to either treatment as usual or waitlist comparison group, and the greatest decrease was observed for face-to-face CBT, compared to multimedia or hybrid CBT. They also examined different components of CBT (e.g. behavioural activation, goal setting, homework), although this analysis was somewhat frustrated by a lack of reporting of the components of CBT interventions in many publications.

The choice of a comparison group is important: it is easier to demonstrate a more substantial effect of CBT when one compares CBT to a waiting list, rather than comparing CBT to an active intervention (e.g. psychoeducation or social support). Even a given term, such as “treatment as usual”, may describe comparison conditions that are actually quite heterogenous across studies; “treatment as usual” can range from an intervention as minimal as being given a link to a website with information about depression (Clarke et al., 2009), up to an intervention involving antidepressant medication, monitoring, and perhaps referral for specialist psychological services (Williams, 2013). Notwithstanding these methodological issues, the meta-analyses of randomised control trials do largely offer support for the efficacy of CBT in treating depression.

Besides this evidence base arising from randomised control trials, a meta-analysis of non-randomised studies on outpatient individual (and group) CBT found that CBT was effective in reducing depression, although the effect sizes were lower than those observed in randomised control trials  (Hans & Hiller, 2013). These findings sound a cautionary note about assuming that the effect size one observes in a randomised control trial will translate to clinical practice. Hans & Hiller also found a reasonably high drop-out rate; this contrasts with a low risk of attrition bias for nearly half of the randomised control trials observed by López-López et al. (2019).

In the past decade there have been increasing calls to enhance the replicability of empirical research results, i.e. to ensure that the same findings of a study can be produced if the study is run again using the same methods (e.g. Munafò et al., 2017). Different measures can be used to assess depression pre- and post-CBT, such as the Beck Depression Inventory (Beck et al., 1996), the Hamilton Depression Rating Scale (Hamilton, 1960) or the CES-D (Radloff, 1977). Furthermore, studies can take more than one measure of depression, and so the measure which shows the strongest effect size can be selected for publication. Alternatively, researchers may conduct sub-group analysis to identify effects where an overall effect is not evident (e.g. males versus females, younger versus older people, suicidal versus non-suicidal clients, etc.). Munafò et al. refer to the practice of taking a post-hoc observation and publishing it as if this had been a pre-existing hypothesis, which they call “hypothesising after results known.” To take one example, Stangier et al. (2013) found that maintenance CBT was more effective than manualised psychoeducation in preventing reoccurrence or relapse of depression in people who were in remission for at least two months, but this effect was only the case for people at higher risk of reoccurrence. This reported sub-group effect may be one of multiple such analyses they could have conducted. Although their research was registered at, their sub-group analysis is not reported in this registry, suggesting it was a post-hoc comparison, conducted after their data had been collected.

Pre-registration is a particularly useful way to prevent hypothesising after results known. To pre-register a study, researchers make the method of the study, as well as its hypothesised outcomes, available to the research community prior to collecting and analysing their data. Pre-registration is not just for primary research in which novel data is collected from research participants; websites such as PROSPERO ( provide a forum whereby systematic reviews and meta-analyses can be described in advance (including explicit statements of the outcomes to be assessed).

CBT: Are there risks?

Discussion around talking therapies tends to focus more on the effectiveness or efficacy of a therapy in treating depression and less on potential side effects, compared to pharmacological treatments and certainly compared to electro-convulsive therapy (e.g. Ingram et al., 2008). Nonetheless, we should also be mindful that, like any intervention, CBT may have side effects. One study (mostly focused on depression, although some clients were seeking CBT for other issues) estimated that almost half of clients had experienced at least one side effect from CBT, including deterioration of depression, and strains in family relations (Schermuly-Haupt et al., 2018). Schermuly-Haupt et al. note the lack of a generally accepted methodology for assessing the side effects of psychotherapy in general, although a questionnaire for the assessment of potential adverse aspects of psychotherapy has been developed (Parker et al., 2013); again, this was validated with a sample of participants attending a variety of forms of therapy, although CBT was one of the most popular.

Although there is a clear paucity of evidence on this research question, any cost-benefit analysis will be incomplete unless it considers the net benefit of CBT to clients, allowing that there may be negative effects as well as positive. Outcome measures such as the Beck Depression Inventory will likely capture deterioration in overall depression, but side effects such as strain in family relations may be missed by researchers’ primary analyses.

CBT: The search for markers in the central nervous system

Given the fact that some individuals respond better to therapy than others, there has been some interest in the possibility of identifying potential markers/predictors of therapy via functional neuroimaging. However, such efforts should be made with caution, given the heterogeneity of depression, as well as the fact that it is defined and clinically assessed at a cognitive/affective level rather than a neurobiological level. A systematic review found that research on the impact of CBT using brain imaging technologies does not provide convincing evidence that CBT leads to changes in brain function, or at least that such activation is not detectable with existing brain imaging technologies (Franklin et al., 2016). It should be noted that the studies they identified were more likely to use technologies suited to detecting activation (via blood oxygenation) in different brain regions (functional magnetic brain imaging/fMRI) and less likely to use those better suited to identify changes in particular neurochemicals (e.g. positron emission tomography, PET). It may be that more neurochemical-based markers may lead to more interesting results in future. If optogenetics (a technique for targeting specific neural cell types with millisecond accuracy developed in rodent models; Boyden et al., 2005), becomes less invasive and more usable in human trials, this may provide more precise information on this topic in future.

A particularly large and ambitious brain imaging study employed fMRI (including a replication dataset) and appeared to identify different “biotypes” associated with depression (Drysdale et al., 2017). The fact that the authors did not simply conflate all patients into a single biotype suggests an allowance for the heterogeneity of depression. Furthermore, the biotypes were predictive of responsiveness to transcranial magnetic stimulation therapy, suggesting this could perhaps be a promising avenue for predicting responsiveness to CBT. Unfortunately, a subsequent attempt at replication did not find evidence that was as convincing (Dinga et al., 2019).

Of course, “locating” depression at a sub-individual level (e.g. reduced activity in a specific brain region) risks ignoring the social and environmental factors associated with depression; for example, socioeconomic status is a predictor of depression (Lorant et al., 2003). Even for one-on-one therapy sessions with individuals, discussion will often focus on inter-individual factors (e.g. marital relationship, problems with one’s boss) or the impact of broader social/economic/cultural factors (e.g. poverty, systemic racism). Nonetheless, a more nuanced understanding can allow for the aetiological impact of psychosocial adversity to be realised at a neurobiological level (e.g. Gianaros & Manuck, 2017).



Research on the impact of CBT on depression has been ongoing for decades. This has generated a substantial evidence base demonstrating the efficacy and effectiveness of CBT for treating depression. However, even within the last decade, the standards by which empirical psychology is appraised have risen, and I would contend this is particularly true with regard to replicability. As much research on CBT and depression turns from the general question of “whether it works or not” per se, and begins to focus on issues such as treatment for co-morbid depression in the context of other conditions, as well as the question of online versions of CBT compared to face-to-face delivery, it is important that researchers harmonise their methodologies to allow for easier comparison between studies, and increase the use of pre-registration, transparency and data-sharing so that researchers can build on one another’s efforts more rapidly.

Despite this success is demonstrating the efficacy and effectiveness of CBT in treating depression, I wish to sound some cautionary notes with regard to this research area. Firstly, one must be cautious about assuming that the effect sizes observed in randomised control trials will generalise to the clinic. Secondly, although there has been some research explicitly addressing potential negative side effects, this has been much more limited in scope, and has often been conducted on a variety of different psychotherapies, rather than focusing on CBT itself. Lastly, although there has been research enthusiasm in searching for neurological correlates of depression and associated “biomarkers” or biological predictors of responsiveness to treatment (including CBT), given the complexity of both depression and the human central nervous system, understanding how they interact is a mammoth task, and research in this area is still in its early stages.


Related articles: 

I felt a funeral in my brain
If you like this post, follow me on Twitter: @ArchivePsych


American Psychiatric Association (2013). Diagnostic and statistical manual of mental disorders (DSM-5®). Washington, DC: American Psychiatric Publishing.

Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Beck depression inventory (BDI-II). San Antonio, Texas: The Psychological Corporation.

Boyden, E. S., Zhang, F., Bamberg, E., Nagel, G., & Deisseroth, K. (2005). Millisecond-timescale, genetically targeted optical control of neural activity. Nature Neuroscience8(9), 1263-1268.

Butler, A. C., Chapman, J. E., Forman, E. M., & Beck, A. T. (2006). The empirical status of cognitive-behavioral therapy: A review of meta-analyses. Clinical Psychology Review26(1), 17-31.

Clark, D. M., Layard, R., Smithies, R., Richards, D. A., Suckling, R., & Wright, B. (2009). Improving access to psychological therapy: Initial evaluation of two UK demonstration sites. Behaviour Research and Therapy47(11), 910-920.

Clarke, G., Kelleher, C., Hornbrook, M., DeBar, L., Dickerson, J., & Gullion, C. (2009). Randomized effectiveness trial of an internet, pure self-help, cognitive behavioral intervention for depressive symptoms in young adults. Cognitive Behaviour Therapy38(4), 222-234.

Cocker, F., Martin, A., Scott, J., Venn, A., Otahal, P., & Sanderson, K. (2011). Factors associated with presenteeism among employed Australian adults reporting lifetime major depression with 12-month symptoms. Journal of Affective Disorders135(1-3), 231-240.

Dinga, R., Schmaal, L., Penninx, B. W., van Tol, M. J., Veltman, D. J., van Velzen, L., ... & Marquand, A. F. (2019). Evaluating the evidence for biotypes of depression: Methodological replication and extension of Drysdale et al. (2017). NeuroImage: Clinical22, 101796.

Dobkin, R. D., Menza, M., Allen, L. A., Gara, M. A., Mark, M. H., Tiu, J., ... & Friedman, J. (2011). Cognitive-behavioral therapy for depression in Parkinson's disease: A randomized, controlled trial. American Journal of Psychiatry168(10), 1066-1074.

Drysdale, A. T., Grosenick, L., Downar, J., Dunlop, K., Mansouri, F., Meng, Y., ... & Schatzberg, A. F. (2017). Resting-state connectivity biomarkers define neurophysiological subtypes of depression. Nature Medicine23(1), 28-38.

Field, A. (2009). Discovering statistics using SPSS: And sex and drugs and rock 'n' roll (3rd Edition). London: Sage.

Franklin, G., Carson, A. J., & Welch, K. A. (2016). Cognitive behavioural therapy for depression: systematic review of imaging studies. Acta neuropsychiatrica28(2), 61-74.

Gianaros, P. J., & Manuck, S. B. (2010). Neurobiological pathways linking socioeconomic position and health. Psychosomatic Medicine72(5), 450-461.

Gotlib, I. & Hammen, C. (2002). Handbook of depression. New York: Guilford Press.

Hamilton, M. (1960). A rating scale for depression. Journal of Neurology,  Neurosurgery and Psychiatry, 23(1), 56-62.

Hans, E., & Hiller, W. (2013). Effectiveness of and dropout from outpatient cognitive behavioral therapy for adult unipolar depression: A meta-analysis of nonrandomized effectiveness studies. Journal of Consulting and Clinical Psychology81(1), 75-88.

Hazlett-Stevens, H., & Craske, M. G. (2002). Brief cognitive-behavioural therapy: Definition and scientific foundations. In F. W. Bond & W. Dryden (Eds.), Handbook of brief cognitive behavioural therapy (pp. 1-20). Chichester: John Wiley & Sons.

Hynninen, M. J., Bjerke, N., Pallesen, S., Bakke, P. S., & Nordhus, I. H. (2010). A randomized controlled trial of cognitive behavioral therapy for anxiety and depression in COPD. Respiratory Medicine104(7), 986-994.

Ingram, A., Saling, M. M., & Schweitzer, I. (2008). Cognitive side effects of brief pulse electroconvulsive therapy: a review. The Journal of ECT24(1), 3-9.

Lilienfeld, S. O., Sauvigné, K. C., Lynn, S. J., Cautin, R. L., Latzman, R. D., & Waldman, I. D. (2015). Fifty psychological and psychiatric terms to avoid: a list of inaccurate, misleading, misused, ambiguous, and logically confused words and phrases. Frontiers in Psychology6, 1100.

López-López, J. A., Davies, S. R., Caldwell, D. M., Churchill, R., Peters, T. J., Tallon, D., ... & Lewis, G. (2019). The process and delivery of CBT for depression in adults: a systematic review and network meta-analysis. Psychological medicine49(12), 1937-1947.

Lorant, V., Deliège, D., Eaton, W., Robert, A., Philippot, P., & Ansseau, M. (2003). Socioeconomic inequalities in depression: A meta-analysis. American Journal of Epidemiology157(2), 98-112.

Munafò, M. R., Nosek, B. A., Bishop, D. V., Button, K. S., Chambers, C. D., Du Sert, N. P., ... & Ioannidis, J. P. (2017). A manifesto for reproducible science. Nature Human Behaviour1(1), 1-9.

Parker, G., Fletcher, K., Berk, M., & Paterson, A. (2013). Development of a measure quantifying adverse psychotherapeutic ingredients: The Experiences of Therapy Questionnaire (ETQ). Psychiatry Research206(2-3), 293-301.

Parker, G., Roy, K., & Eyers, K. (2003). Cognitive behavior therapy for depression? Choose horses for courses. American Journal of Psychiatry160(5), 825-834.

Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement1(3), 385-401.

Safren, S. A., O'Cleirigh, C. M., Bullis, J. R., Otto, M. W., Stein, M. D., & Pollack, M. H. (2012). Cognitive behavioral therapy for adherence and depression (CBT-AD) in HIV-infected injection drug users: a randomized controlled trial. Journal of Consulting and Clinical Psychology80(3), 404-415.

Schermuly-Haupt, M. L., Linden, M., & Rush, A. J. (2018). Unwanted events and side effects in cognitive behavior therapy. Cognitive Therapy and Research42(3), 219-229.

Stangier, U., Hilling, C., Heidenreich, T., Risch, A. K., Barocka, A., Schlösser, R., ... & Weck, F. (2013). Maintenance cognitive-behavioral therapy and manualized psychoeducation in the treatment of recurrent depression: a multicenter prospective randomized controlled trial. American Journal of Psychiatry170(6), 624-632.

World Health Organisation (2016). Depression: Fact sheet. Retrieved 21/07/2020 from

Williams, C., Wilson, P., Morrison, J., McMahon, A., Andrew, W., Allan, L., ... & Tansey, L. (2013). Guided self-help cognitive behavioural therapy for depression in primary care: a randomised controlled trial. PloS One8(1), e52735.