Abstract
Depression is one of the leading causes of mortality, disability, and loss of productivity. The World Health Organization (WHO) ranks depressive disorders as the eleventh cause of disability and mortality (1, 2). The worldwide lifetime prevalence of depression is around 12% (3). In spite of the considerable burden of depression both in terms of prevalence and public health impact, the search for more effective treatments for depression is still ongoing. Emerging evidence suggests that personalizing treatments based on individuals’ biosignature could be the “way forward” (4).
20.1 Introduction
Depression is one of the leading causes of mortality, disability, and loss of productivity. The World Health Organization (WHO) ranks depressive disorders as the eleventh cause of disability and mortality (1, 2). The worldwide lifetime prevalence of depression is around 12% (3). In spite of the considerable burden of depression both in terms of prevalence and public health impact, the search for more effective treatments for depression is still ongoing. Emerging evidence suggests that personalizing treatments based on individuals’ biosignature could be the “way forward” (4).
An increasing number of studies have therefore sought to identify potential biological predictors of clinical outcomes that could guide treatment selection. While the majority of these studies have primarily focused on pharmacological treatments (5, 6), few of them have examined potential correlates of psychotherapeutic outcome. Psychotherapy, specifically cognitive behavioral and interpersonal therapies, has been shown to be particularly efficacious in the treatment of depression. A large meta-analysis concluded that following psychotherapy, 62% of depressed patients no longer met criteria for depression. By comparison, only 48% of depressed patients achieved full remission following care-as-usual, which was defined as interventions other than the psychotherapy received as part of the clinical trial (7). While there is not much evidence for differences in effectiveness across psychotherapies when looking at short-term effects (8), the combined treatment with psychotherapy and medication appears to have superior therapeutic benefits when compared with monotherapy (9, 10). Further, a meta-analysis of eleven studies following up patients for approximately fifteen months showed a small-to-moderate effect size favoring psychotherapy relative to pharmacotherapy (11). These findings indicate that the long-term benefits of psychotherapy may outweigh those of pharmacotherapy. Understanding the neural mechanisms involved in successful psychotherapy has, therefore, substantial clinical relevance for guiding personalized treatments and potentially refining current psychotherapy techniques.
Neuroimaging in psychotherapy is a very new, yet active and growing research area. It holds substantial clinical potential as, in the future, it may provide information on the neural correlates related to the effects of specific therapeutic interventions. To date, studies have integrated psychotherapy and multiple imaging measures by assessing psychotherapy-related brain changes in function of treatment response. Baseline imaging measures have also been examined with the aim of predicting treatment response (12). These studies have provided preliminary insight into the potential neural mechanisms of action of psychotherapy, and may lead to the development of guidelines on how to select treatment for individual patients on the basis of indicators of brain functioning at baseline.
In summary, the application of imaging techniques to study the process and outcome of psychotherapy has the potential to significantly improve our understanding of neural processes underlying changes during psychotherapy and treatment. This review chapter aims to provide up-to-date information on the effects of psychotherapy on the brain and evidence of potential imaging predictors of clinical outcomes.
20.2 Literature Search
In the past decade, an increasing number of studies have integrated imaging techniques into psychotherapy research across psychiatric disorders (13–18). Given that the definition of “psychotherapy” is broad and encompasses a number of therapeutic approaches and techniques, this chapter will focus on three empirically supported and well-established therapeutic psychological interventions for depression: cognitive behavioral therapy (CBT), behavioral activation therapy (BAT), and interpersonal therapy (IPT). These interventions were selected because (1) CBTs are evidence-based treatments and are guided by well-established principles of learning theories, behavioral science and cognitive psychology (19); (2) CBT is one of the most empirically evaluated forms of therapy in relation to emotion regulation and cognitive control; and (3) CBT and IPT are considered to be the gold-standard treatments of depression (20).
Although, nowadays, CBT is considered to be an umbrella term for a wide range of interventions, the common premise is that maladaptive or dysfunctional beliefs and biased information processing contribute to the development and maintenance of depressive symptoms. Based on this model, addressing maladaptive thoughts is a first step toward reducing the risk for relapse (21). The goal of BAT is to help patients engage more in rewarding behaviors while also reducing withdrawal and avoidance (22). BAT has been found to be as effective as CBT to reduce depressive symptoms and prevent relapse (23, 24). IPT is a short, present-oriented, form of psychotherapy that views interpersonal issues as the primary trigger for the development and maintenance of psychological distress (25, 26).
We reviewed existing publications that included both neuroimaging and clinical outcome measures to evaluate the efficacy of psychotherapy. We restricted our search to those studies with adult samples and who described their participants as suffering from depression. We searched PubMed, Scopus, Ovid, and Cochrane for articles containing the terms “depression,” “psychotherapy,” “cognitive therapy,” or “behavioral therapy” combined with terms referring to widely used structural and functional neuroimaging techniques including “MRI,” “fMRI,” “DTI,” “photon emission,” “positron emission,” or “spectroscopy.” We additionally included electroencephalogram (EEG) measures as they provide neural indices of cognitive processes of relevance for psychotherapy (e.g., cognitive control and self-monitoring). We also reviewed studies discussed in Weingarten et al. and Fournier et al.’s systematic and meta-analysis reviews on neuroimaging and psychotherapy in psychiatry (27, 28). Although this search has no claims of being fully exhaustive, it aimed to provide comprehensive evidence of the state of the science in the field of imaging and psychotherapy.
For ease of reading we present our findings in two different sections. The first section examines the effects of psychotherapy on the brain, and the second section discusses the imaging predictors of psychotherapy response. Table 20.1 summarizes the primary findings of this review.
20.3 Effects of Psychotherapy on the Brain in Depression
To put our findings into context, it is important to provide first a brief overview of the structural and functional brain networks associated with mood disorders. Frontal (orbitofrontal, dorsolateral, and ventromedial), limbic (amygdala, hippocampus, and insula), and anterior cingulate cortex (ACC) are closely connected key emotion-processing regions (29, 30). Fear processing and anhedonia are prevalent in depression, and have been linked to poor functioning and connectivity within and between basal ganglia, striatum, and para/hippocampal regions (28, 31). The well-established evidence of the reduced connectivity between the prefrontal, cingulate, and limbic-striatal structures (32) supports the hypothesis of a fronto-limbic disconnection leading to the mood dysregulation observed in mood disorders (33). This disconnection is hypothesized to disrupt cognitive control or “top-down” processes, and may lead to increased emotional or “bottom-up” activity (34). There is general agreement that psychotherapy may strengthen the cognitive “top-down” network by teaching individuals to implement effective problem-solving and coping skills to manage stressful situations (28).
Overall, the majority of the studies retrieved in our systematic review have focused on CBT, and to a lesser extent on BAT and IPT. To date, most imaging studies of psychotherapy have used fMRI and positron emission tomography (PET), and only few studies adopted sMRI, magnetic resonance spectroscopy (MRS), and EEG techniques. In the following section, we will discuss available imaging findings illustrating the effects of psychotherapy on functional, metabolic, and structural brain measures.
20.3.1 fMRI
Most fMRI studies of psychotherapy employed task-based protocols requiring the explicit or implicit processing of emotional stimuli. The primary outcome measures of these studies are typically the pre- to post-psychotherapy changes in either brain activation or functional connectivity. It is noteworthy mentioning that, unless otherwise specified, the findings here below refer to treatment-related changes in brain functioning in patients with depression. When compared to their pretreatment imaging measures, CBT-treated patients showed reduced activation in the medial prefrontal cortex and ventral ACC during the evaluation of negative stimuli (35) and increased activation in these regions in response to positive stimuli (stimuli included self-referential and non-self-referential cues)(36, 37). CBT was also associated with increased activation in the amygdala, caudate nucleus, and hippocampus, which are regions involved in emotion, reward processing, and emotional memory, respectively (38). Interestingly, prior to CBT, depressed patients were found to have stronger fronto-cingulate connectivity compared to healthy controls (39). The CBT-related improvement in depressive symptoms correlated positively with the reduction in fronto-cingulate connectivity (39). This finding suggests that these brain regions might underlie some of the dysfunctional thoughts targeted as part of CBT. This is also partially in line with findings showing that a decrease in connectivity between the dorsal anterior and subgenual ACC regions is linked to lower levels of self-reported worry in anxious patients (40). CBT also led to an increase in connectivity between the amygdala and frontoparietal regions during a task evaluating feelings and thoughts triggered by emotionally salient stimuli (37). These findings are consistent with previous evidence that low connectivity in the fronto-limbic network predicts poor response to both psychotherapy and antidepressant medication (41, 42).
One of the few studies examining the neural predictors of BAT outcomes found that having been treated with BAT was associated with decreased activation in the prefrontal, cingulate and paracingulate, caudate nucleus, fusiform, and cerebellar regions in response to sad stimuli (43). The frontal regions have noteworthy a high clinical relevance and play an important role in cognitive control and information processing (44). For example, the orbitofrontal cortex has been implicated in affective processing, decision-making, and suicidal thoughts (45, 46). As Dichter et al. (43) pointed out, this finding stands in contrast with the previous pharmacological literature showing an increase in prefrontal activation alongside symptom remission (47, 48). While differences in task protocols could contribute to such inconsistent results, these findings could also suggest that psychotherapy and medication lead to symptom remission by targeting different brain networks. For instance, in Dichter et al.’s study, psychotherapy induced changes in activation in the pars triangularis, which is important for language and motor control but is not directly involved in emotion regulation or cognitive control.
20.3.2 PET and SPECT
CBT-related serotonin changes were measured with positron emission tomography (PET) using a serotonergic 5-HT1B receptor-selective radioligand (49). CBT was associated with a 33% reduction in serotonergic binding potential in the dorsal brain stem. This finding is not surprising given that this brain region includes the raphe nucleus, a key node in the serotonin pathway (49). The authors argued that a reduction in serotonergic binding may reflect the downregulation of the inhibitory 5-HT1B receptor and the increase in serotonin release to emotion-processing areas such as the prefrontal and limbic regions. It is noteworthy mentioning that all the patients involved in this study responded successfully to CBT, thus suggesting a strong link between CBT outcomes and serotonin production. A (18 F)-2-fluoro-2-deoxy-d-glucose–PET study by Goldapple et al. (50) compared CBT to paroxetine, a selective serotonin reuptake inhibitor. CBT-treated patients displayed an increase in glucose metabolism in the prefrontal, cingulate, and hippocampal regions, which are brain areas associated with mood regulation (50).
Few studies have examined the impact of IPT on functional brain measures. IPT was associated with increased glucose metabolism in the left insula compared to paroxetine (51). Further, IPT-treated individuals showed increased brain blood flow (measured with single-photon emission computed tomography or SPECT) in posterior cingulate brain regions when compared to the serotonin-norepinephrine reuptake inhibitor venlafaxine (52). It is noteworthy that this study did not detect the changes in activation in cingulate or frontal regions observed in the fMRI psychotherapy studies discussed in the previous section.
The divergence of results may be due to the selection of regions of interest included in the analyses. It is also important to remember that the fMRI and PET techniques measure different physiological processes. While fMRI findings are task-specific, PET measures are closely related to the brain metabolism at rest.
20.3.3 MRS
Two proton magnetic resonance spectroscopy (1 H-MRS) studies measured changes in gamma-aminobutyric acid (GABA) concentrations in the occipital cortex of depressed patients. The selection of this region was based on studies showing that antidepressant treatments such as serotonin reuptake inhibitors and electroconvulsive therapy were associated with an increase in occipital GABA concentrations (53, 54). CBT was not found to alter occipital GABA levels in either of these studies (55, 56). This suggests that, while previous pharmacological treatments targeted the GABAergic pathways, CBT interventions may not affect this brain circuit.
20.4 Imaging as a Predictor of Psychotherapy Response
Functional and structural neuroimaging have, so far, shown that psychotherapy targets fronto-limbic and fronto-cingulate networks. As part of this review, we also examined whether these same regions could predict the likelihood of response to CBT, BAT, and IPT.
20.4.1 fMRI
Hypoactivity in the ACC (35) and sustained hyperactivity in the dorsolateral prefrontal cortex during an emotional task were associated with favorable response to CBT (57). In the latter study, CBT nonresponders showed increased activity in the right amygdala at baseline, which may indicate that an overactive amygdala interferes with CBT. Alternatively, it is possible that some of the patients included in this study suffered from comorbid anxiety. If this were the case, the poor response to CBT may be due to the fact that CBT targeted depression instead of anxiety. Indeed, the amygdala is involved in fear processing and shows increased reactivity in response to stressful events (58).
Increased resting state functional connectivity between subcallosal cingulate regions, the frontal operculum, and ventromedial prefrontal cortex were found to be predictors of a positive response to CBT (59). The same study also found that, before CBT, responders activated a network of brain regions similar to that of healthy volunteers. In another study, baseline hypoactivity in the frontal inferior triangle and right superior frontal gyrus and hyperactivity in the middle frontal and left superior frontal gyrus predicted better CBT outcome (60). Elevated functional activity in the cingulate, paracingulate, temporal, and striatal regions was also found to be a predictor of a better response to CBT (37, 38, 41, 43, 50, 61). More specifically, individuals with high cingulate activation in response to sad stimuli prior to CBT were more likely to respond to CBT (56). Fu et al. argued that ACC hyperactivity, and reduced amygdala-hippocampal hypoconnectivity may be the key predictors for a positive response to CBT (20, 62).
Decreased connectivity between the anterior insula and the middle temporal regions was associated with severe anhedonia and symptom severity prior to CBT treatment. Intriguingly, hypoconnectivity between these regions was also linked to increased likelihood of CBT response (22). ACC activation in response to a reward-processing task was found to be a positive predictor of BAT outcome (63). This finding is not surprising given that BAT encourages patients to engage in activities that give them a sense of purpose. The success of BAT may therefore require high reward responsiveness on a neural level (64).
20.4.2 PET
In line with previously reviewed structural and functional imaging findings, the ACC glucose metabolism was found to be a predictor of positive CBT response (41, 65, 66). High baseline glucose metabolism in subcallosal cingulate (67) and limbic-subcortical areas (65), and reduced metabolism in the insula (68) predicted a better outcome following CBT (65). By comparison, individuals with high glucose metabolism in the insula were more likely to remit in response to medication (70). The findings related to the insula are intriguing since this brain region plays a role in emotional self-awareness and processing of subjective feeling states (69). However, it has yet to be determined whether having low glucose metabolism in the insula means having poor self-awareness and reduced ability to process feelings. An alternative interpretation of this finding may be that the insula contributes to the successful assimilation and implementation of CBT strategies.
Related posts:
Chapter 1 – Brain Imaging Methods in Mood Disorders
Chapter 2 – Neuroanatomical Findings in Unipolar Depression and the Role of the Hippocampus
Chapter 7 – Functional Connectome in Bipolar Disorder
Chapter 8 – Magnetic Resonance Spectroscopy Investigations of Bioenergy and Mitochondrial Function in Mood Disorders
Chapter 14 – Electrophysiological Biomarkers for Mood Disorders
Chapter 15 – Magnetoencephalography Studies in Mood Disorders
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
