Depression, or major depressive disorder, is a leading burden and the third most common cause of disability worldwide (
1,
2). Psychiatric disorders are also the most costly to treat of any childhood health conditions, with $13.8 billion spent on treatment for psychiatric disorders in the United States in 2011 (
3). By adolescence, one in five youths in the United States will experience at least one severely impairing psychiatric disorder (
4). With a 3.2% prevalence rate of depression across childhood and adolescence in the United States, this is <1% of 3- to 5-year-old children, 1.7% of 6- to 11-year-old children, and 6.1% of 12- to 17-year-old adolescents (
5). Childhood depression is also a precursor to adolescent and adult depression; 72% of children subsequently experience a recurrent depressive episode after initial episode recovery (
6).
Despite the established prevalence of childhood depression, essentially no evidence-based treatments exist for this condition. In one recent meta-analysis that compared psychotherapy effects across age groups, psychotherapy had only a small effect on depressive symptoms in children (g=0.35), yet a medium effect was observed in adolescents (g=0.55), older adults (g=0.66), and middle-aged adults (g=0.77); a large effect was observed in young adults (g=0.98) (
7). A 2017 meta-analysis of cognitive-behavioral therapy (CBT) for childhood depression found a small effect of CBT on depressive symptoms in children (standardized mean difference=−0.41), although it was not superior to wait list or control interventions (
8). In a 2016 meta-analysis, there was inconclusive evidence that psychotherapy was better than no treatment for childhood depression across CBT, family therapy, and psychodynamic therapy modalities (
9). Additionally, while CBT and interpersonal psychotherapy are considered well-established interventions for adolescent depression, no psychosocial interventions are well established for childhood depression. CBT was identified as possibly efficacious (
10).
The evidence for antidepressant medication in childhood and adolescent depression remains mixed. In a network meta-analysis, fluoxetine was the only medication (out of 14) to be superior to placebo in improving depressive symptoms in childhood and adolescent depression (standardized mean difference=0.51; medium effect) (
11). The other 13 medications were not superior to placebo; imipramine, venlafaxine, and duloxetine had a higher rate of adverse event-related discontinuation than placebo. A similar pattern of medication nonsuperiority to placebo and increased adverse event-related discontinuation was found in a separate meta-analysis that examined serotonin-norepinephrine reuptake inhibitors (
12). However, these findings are not without controversy and inconsistencies. Compared with industry-funded trials, trials funded by the National Institute of Mental Health (NIMH) are characterized by lower placebo response rates and larger between-group differences in favor of antidepressant efficacy (
11,
13).
Diagnostic criteria for major depressive disorder can be met in 227 possible symptom combinations, but only 10% of patients possess all nine of the core criteria (
14). Given this significant heterogeneity, research has shifted toward possible traits or endophenotypes. Neuroticism, blunted reward learning, and cognitive control deficits have emerged as three primary endophenotypes of depression (
15). Cognitive control (also referred to as executive function) is a collection of self-regulatory processes required when automatic or proponent thoughts and actions are insufficient to meet environmental demands (
16). Cognitive control subdomains are somewhat heterogeneous, including processes such as working memory, inhibition, flexibility, self-monitoring, emotion regulation, and problem solving (
16–
18). To capture the overarching cognitive control construct, we investigated studies that used any cognitive control measures. Cognitive control is a critical function that is one of the strongest predictors of clinical and functional outcomes across the spectrum of psychopathology (e.g.,
19,
20). Improved identification and treatment of this previously underaddressed component of the disorder offers an opportunity to improve the clinical and functional outcomes of childhood depression. As there is a limited literature on cognitive control in childhood depressive disorders, this perspective will summarize relevant cognitive control findings in adult depression and describe key clinical and research approaches that could harness cognitive control to improve outcomes.
The Role of Cognitive Control in Depression
Cognitive Control Deficits in Depression
Prior meta-analyses have identified small- to large-sized cognitive control deficits as an established component of adult major depressive disorder and other depressive disorders (
21–
23). Although depression is associated with lower cognitive control as a diagnostic group, only approximately 21%−62% of adults with major depressive disorder display cognitive control deficits (
24–
27). In a recent analysis of adult depression, 45% of the sample was labeled “cognitively preserved,” 39% as “cognitively reduced,” and 16% as “cognitively impaired” (
27). Regarding the disease course, small- to medium-sized cognitive control deficits are present during the first depressive episode (
28) and persist after depressive episode remission (
21,
29). Meta-analytic evidence indicates that the time since depression onset and the number of prior episodes are also associated with lower cognitive control (
29). Although healthy control subjects and those with bipolar disorder showed age-related improvements in inhibitory control, depression has been associated with age-related declines in inhibitory control (
30). This potential progression of cognitive control deficit in depression has led to the theoretical conceptualization that each subsequent depressive episode causes further “repetitive scarring” and subsequent cognitive deterioration (
30).
Cognitive Control Predictors of Treatment Outcome
Regarding general treatment predictors, one recent meta-analysis found a medium effect of cognitive control on treatment for older adults with depression across multiple treatment modalities; lower cognitive flexibility was associated with worse treatment response (
31). A qualitative review found a similar effect of lower general cognitive control leading to poor depression treatment, although this review noted response inhibition and semantic organization as the primary implicated cognitive control subdomains (
32). The literature in adult depression is less robust, although lower cognitive control has predicted a worse response to tertiary outpatient care (
33), 3-week treatment with CBT (
34), and a combination of psychodynamic therapy and fluoxetine (
35). In addition, a change in self-reported problem solving was found to predict subsequent change in depressive symptoms across CBT, supportive psychotherapy, and medication interventions in adults (
36).
Regarding antidepressant medication predictors, one meta-analysis in adults and one meta-analysis in older adults found lower cognitive control to be associated with poorer response to antidepressant medication (
37,
38). Interestingly, the cognitive control subdomain differed: Planning and organization was identified in older adults, and initiation and perseveration in all adults. More recently, lower working memory at baseline predicted poorer response to fluoxetine for depression during inpatient hospitalization (
39). In a machine learning analysis of sertraline response in adult depression, cognitive control was one of five pretreatment variables that moderated treatment response (
40). Furthermore, the course of cognitive flexibility and semantic fluency and organization was found to predict the course of depressive symptoms in an antidepressant medication trial (
24).
Childhood Depression Findings
The presence of cognitive control deficits in depression extends downward into childhood and adolescence; a prior meta-analysis identified medium-sized deficits in major depressive disorder within the inhibition, phonemic fluency and organization, and planning domains (
41). In the only studies in adolescent depression, lower sustained attention, inhibitory control, and planning were associated with a poorer response to a fluoxetine trial (
42); baseline self-reported problem solving predicted depressive symptoms across CBT and medication interventions (
43).
It is worth noting that well-done meta-analyses are limited in the ability to examine childhood depression due to the lack of available studies in this age range (e.g.,
44). While the small number of available studies may not show a strong effect, we have consistently found an association between cognitive control and depression at our children’s psychiatric hospital, specifically within outpatient child and adolescent psychiatry (
45), inpatient child psychiatry (
46,
47), and inpatient adolescent psychiatry (
48,
49) programs. However, this may reflect subgroup or phenotype differences between studies, as our patients receive intensive hospital-based treatment and are referred for evaluation due to neurocognitive concerns. For example, in children referred for evaluation during inpatient admission, only 52% of those with clinically significant depressive symptoms also displayed cognitive control deficits (
47). These children were ultimately hospitalized 60% longer than children with elevated depressive symptoms without cognitive control deficits, although the relatively modest rate of deficits in such an intensive setting suggests the rate may be lower in nonhospital treatment settings.
Call to Action
Cognitive control deficits are a core, previously unaddressed component of depression. Preliminary findings show the results also apply to childhood depression. Here, we emphasize clinical and research approaches that can be taken to target these deficits and improve childhood depression outcomes.
Clinical Care
Evaluation.
Neuropsychological evaluation is the application of standardized testing techniques to understand and measure underlying brain-behavior relationships. It is an established, evidence-based clinical service across the spectrum of brain-based disorders (
50). In childhood, neuropsychological evaluation has been shown to improve access to relevant therapies and improve the overall severity of psychiatric symptoms (
51). However, childhood depression is not a common reason for referral. We previously found children with depression were referred for evaluation less often than children with behavioral disorders, despite identifying cognitive control deficits that were specific to children with depression (
46).
Waitlists can limit access to neuropsychology, but options are currently available that primary clinicians can implement. At our hospital, treatment programs now implement a brief cognitive control screener, including the Behavior Rating Inventory of Executive Function–Second Edition, Parent Form (BRIEF-2) and the cognitive control measures of the National Institutes of Health (NIH) Toolbox. Both are established and feasible clinical measures for children (
17,
52,
53), and we have previously found that both measures accurately detect cognitive control deficits specific to childhood depression (
45). A psychometrician or trainee can administer and score these measures in approximately 30 minutes, without requiring a neuropsychologist. This provides critical, immediate data to the primary psychologist or psychiatrist on possible deficits. The team can then integrate the child’s deficits into the treatment plan, or the child can be referred for a formal neuropsychological evaluation as needed.
Recommendation:
•
We encourage primary clinicians and physicians to implement brief and cost-effective cognitive control screening methods into their standard practice and to refer for a more comprehensive neuropsychological evaluation as needed. Neuropsychologists should improve access to care for children with depression, potentially providing brief or tailored evaluations specific to cognitive control.
Treatment.
To our knowledge, cognitive remediation is the only clinically available, evidence-based treatment specifically targeting cognitive control. Cognitive remediation is a systematic intervention that uses learning principles to teach, train, and improve the neurocognitive functions often impaired in various brain-based disorders. A meta-analysis of cognitive remediation in adult depression found significant small to moderate effects for symptom severity and daily functioning as well as moderate to large effects for attention, working memory, and global functioning (
54). Another meta-analysis that examined the effect of various cognitively based therapies (e.g., CBT) found cognitive remediation had a medium-sized effect on repetitive negative thinking in depression (
55). Interestingly, this effect size was second only to rumination-focused CBT, and it was stronger than standard CBT, concreteness training, and mindfulness-based cognitive therapy. Not specific to depression, the cognitive remediation research in childhood is generally consistent with adult findings (i.e., a small-medium effect on cognitive control in preschool, school-aged, and adolescents) (
56–
58).
A cognitive remediation expert working group published an article on the core intervention techniques, which consisted of the presence of a trained therapist, engagement in cognitive exercises, implementation of problem-solving strategies, and procedures to facilitate transfer to real-world functioning (
59). Cognitive remediation programs that incorporate these techniques are considered to be an acceptable, sustainable, and effective intervention in psychiatry (
60). This is uniformly distinct from computer-only cognitive training programs, which despite their potential appeal have limited clinical utility. Various manuals and programs are available using different approaches to cognitive remediation for children and individuals with psychiatric disorders.
We have piloted a cognitive remediation clinic within our neuropsychology program for children with a range of psychiatric disorders, including depression. We have found the service is highly feasible to implement by providers and well received by families and primary clinicians. While our clinic is within neuropsychology, other related providers such as child psychologists, speech/language pathologists, and occupational therapists can implement cognitive remediation.
Recommendation:
•
We are hopeful that cognitive remediation services can be more consistently integrated into the standard of care for childhood depression to target cognitive control deficits.
Research
Evaluation.
Replicating adult studies, cognitive control measures should become a regular component of all treatment/intervention and outcome studies in childhood depression. Historically, one barrier to this additional procedure was the cost and time commitment associated with lengthy in-person testing and involvement by a neuropsychologist. However, feasibility is now drastically improved with parent-report questionnaires such as the BRIEF-2 and computerized measures such as the NIH Toolbox, as clinical trials and studies can reliably assess cognitive control with a research assistant or psychometrician in a short amount of time. Similar to the adult approach and consistent with the NIMH Research Domain Criteria, studies should quickly move away from examining group differences between children with and without depression (
61). Rather, studies should investigate depressive symptoms and cognitive control as continuous constructs and examine their interactions across the spectrum of disease severity (e.g.,
44). Data on the course and responsiveness of cognitive control to depression treatments are critical for improved care. Possible topics for investigation include whether cognitive control deficits precede depressive symptom onset in high-risk samples (e.g., offspring of depressed adults), persist during symptom remission, predict adolescent or adult depression severity, and either predict symptom response to depression treatments or improve after depression treatments.
Recommendation:
•
All childhood depression research should include measures targeting cognitive control, as well as the other endophenotypes of blunted reward learning and anhedonia.
Treatment.
One novel approach or avenue is to take an experimental therapeutics approach to developing interventions that can specifically target cognitive control in childhood depression. As is well described elsewhere, the NIMH experimental therapeutics approach specifically calls for three steps in treatment development (
62). Target identification involves the identification of a clearly defined mechanism that underlies clinical symptomatology and a novel intervention to modulate the target. Target engagement tests whether an intervention specifically designed for the identified target can in fact modulate the target in a dose-dependent manner. Target validation tests whether engaging this target is associated with actual change in the patient’s observable clinical symptomatology. Cognitive control is neuroanatomically subserved by the frontoparietal or cognitive control network. This includes functional nodes within regions such as the dorsolateral prefrontal cortex, cingulate cortex, and posterior parietal cortex (
63,
64). Highly consistent with the cognitive control deficit endophenotype perspective, major depressive disorder during childhood and adolescence is specifically associated with hypoactivation within these regions during cognitive control demands (
65). Various neuroimaging approaches, as described here with functional MRI, offer established neural mechanisms that underlie cognitive control and can serve as treatment targets in childhood depression.
Emerging brain-based interventions can also be harnessed to target these cognitive control mechanisms. Noninvasive brain stimulation holds tremendous promise in transforming psychiatry, as it takes a precision medicine approach to treatment. One stimulation approach is transcranial magnetic stimulation (TMS). TMS is a noninvasive method of cortical stimulation based on the principle of electromagnetic induction; repetitive TMS (rTMS), the therapeutic application of TMS, involves applying repeated trains of magnetic pulses at a specific frequency in repeated daily sessions over a course of weeks (
66). Depression researchers and clinicians are quite familiar with rTMS, as it has been approved by the U.S. Food and Drug Administration since 2008 for treatment-resistant adult major depressive disorder (
67). It has strong clinical efficacy, including a 29%−62% response rate and a 19%−54% remission rate (e.g.,
68). At least 70 studies to date have also shown cognition-enhancing effects of rTMS (
69,
70), including prior reviews showing effects specific to cognitive control (
71–
74). As shown in one recent systematic review, rTMS is a safe and well-tolerated tool in childhood (
75); preliminary studies have shown cognitive control-enhancing effects of rTMS in children (
76–
78).
However, rTMS represents just one of many options in an experimental therapeutics approach, including highly promising and less intensive procedures such as transcranial direct current stimulation and neurofeedback. As an example, we are currently targeting the working memory mechanism of theta-gamma coupling via theta burst stimulation in teenagers with attention deficit hyperactivity disorder in a target engagement study (NCT03480737).
Recommendation:
•
Adopting the experimental therapeutics approach to treatment development in childhood depression is expected to lead to novel, brain-based treatments that can specifically target cognitive control.
Conclusions
Cognitive control deficits are one of three primary endophenotypes of depression that predict clinical and functional outcomes. These deficits emerge early in depression onset and persist or even worsen throughout the disease course. Consistent with adult findings, cognitive control deficits are a common component of childhood and adolescent depression. These deficits are consistently linked to a poor response to treatment for adolescent, adult, and older adult depression. From a clinical perspective, evaluation and treatment of cognitive control deficits are established clinical practices that can improve such deficits and psychiatric symptoms, yet they are not a primary component of current childhood depression care. More consistent implementation of these clinically available options into the standard of care is likely to improve the historically poor treatment outcomes of childhood depression. From a research perspective, cognitive control measures that are low cost and less time intensive should become a regular component of assessment in ongoing childhood depression research. In addition, the experimental therapeutics approach is likely to lead to novel, brain-based interventions that can specifically modulate mechanisms underlying cognitive control deficits. This enhanced approach to these deficits in childhood depression is expected to improve the chronically poor treatment outcomes of childhood depression.