The Genetic and Neural Circuitry Predictors of Benefit From Manualized or Open-Ended Psychotherapy

The evidence base for limited duration, goal-directed modalities of psychotherapy has grown considerably. Manualized interventions focused on maladaptive cognitions and their resulting affective and behavioral corollaries have been shown to be effective for patients with a wide range of psychiatric disorders, including major depressive disorder, anxiety disorders, and trauma- and stressor-related disorders (1–3). These interventions are attractive in that they are replicable (i.e., many follow protocoled and step-wise approaches to addressing symptoms, thus facilitating clinician training) and are structured to last a limited amount of time, allowing for clinical benefits to be gained and for skills to be consolidated within a shorter time frame. Outcome measures of such modalities can be readily explored, expediting tailoring of therapeutic indications and a standardized technical approach. When successful, treatment can yield long-lasting benefits, particularly when environmental support is in place (4).

Neuroimaging and genetic predictors of response to particular psychotherapeutic interventions have been the subject of recent research (5–7). In a seminal paper (8), the term “therapygenetics” was introduced as a parallel to the term “pharmacogenetics,” within the context of genetic predictors to therapeutic response. This paradigm, which has gathered significant interest in its putative clinical application to stratify individuals for therapy, mostly focused on manualized psychotherapies. As we discuss below, particular genetic polymorphisms may be associated with excessive and maladaptive limbic activity in response to environmental cues, which may limit an individual’s ability to engage in structured, directive treatment (9); rather, additional environmental interventions, concurrent psychopharmacology, or a more open-ended, insight-oriented psychotherapy may be needed. Thus, screening of patients for relevant polymorphisms may be factored into a holistic treatment plan. If psychotherapy is compared with pharmacotherapy, wherein doses are titrated up until effective and length of treatment is often kept for sustainability of effect, it can be argued that manualized, prescriptive psychotherapy may be appropriate for certain individuals and open-ended (potentially longer-term), insight-oriented (e.g., psychodynamic or psychoanalytic) approaches may be better for others.

A number of psychotherapeutic modalities have an established evidence base for a range of psychiatric disorders. For example, Steinert et al., in a large meta-analysis (10), reported equivalent efficacy for psychodynamic therapy compared with other forms of psychotherapy that have an established treatment effect. To assess efficacy, Steinert et al. pooled data from studies evaluating treatment of individuals diagnosed as having anxiety disorders, depressive disorders, eating disorders, personality disorders, substance use disorders, and posttraumatic stress disorder (PTSD). Because different forms of therapy can be helpful for the same diagnoses, the opportunity arises to explore which individual patient factors may inform greater response to particular interventions as well as how to incorporate such data in early treatment planning. We are slowly acquiring an understanding of how particular genetic and neurocircuitry endowments may interact with psychotherapy, resulting in outcomes based on an individual’s ability to engage with and derive benefit from the treatment strategy used. These endowments may not only inform the patient’s ability to work with the techniques being introduced and to internalize the skill set presented, but also to adhere to the longitudinal process of independently building upon and maintaining the acquired benefits once the treatment has been completed (11). In this review, we aimed to highlight some of the advances in this new field of inquiry and discuss how genetic markers may predict responses to different therapy modalities, as well as how clinicians may use this information when considering treatment options. We will also discuss controversies and limitations. For the purposes of this article, we have divided psychotherapy modalities into manualized, prescriptive, time-limited therapies (e.g., cognitive-behavioral therapy [CBT]) and open-ended, nonmanualized, insight-oriented psychotherapies (e.g., psychodynamic and psychoanalytic). We have avoided the terms “short-term” and “long-term,” because treatment length varies even within the same modality and because the technical approach and standardization of the forms of therapy being discussed may be more relevant for this review than a rigid definition of duration.

Early Experience and Development of “Future Priors” Informing Behavior

An individual’s genetic makeup informs temperamental traits, which may lead to greater or lesser engagement with environmental stimuli (12–18). This level of engagement has epigenetic and neuroendocrine bases, with transgenerational considerations factoring into personality constitution, attachment schema, and predisposition to psychopathology (e.g., anxiety disorders, major depressive disorder, and PTSD) (19–22). Such factors may lead an individual to attach immediate and inflexible salience to stimuli, engaging the hypothalamic-pituitary-adrenal (HPA) axis, based on feedback from the amygdala, with ultimate release of cortisol and norepinephrine commensurate with the sensed danger present in the environment (23, 24). Roberts and colleagues (25) studied epigenetic mechanisms through differential DNA methylation patterns of the HPA axis-related gene FKBP5, gauging response to exposure-based CBT for patients with agoraphobia. Decreases in DNA methylation in intron 7 were associated with greater response to therapy, independent of protein expression. The Roberts et al. findings have suggested that decreased DNA methylation at this specific intron site is strongly associated with a decreased level of anxiety following exposure-based CBT, adding to the extant literature that epigenetics and behavior may not be so far removed after all.

A more balanced response to the demands of a complex and changing environment is contingent on integration among cognitive, affective, reward, and interoceptive circuits. How much access individuals have to higher cortical resources as they negotiate their environments and their ability to integrate multisensory input received through subcortical networks are highly individualized, generating responses based on past experience and genetic constitution. What has been termed “structured spontaneity” is the ability individuals demonstrate in shifting focus and allocating energy according to changing circumstances (26, 27). Ideally, these shifts will occur without excessive affective discharge or destructive behavioral overtones; however, when such a balance is not possible, the individual may use pathological grounding tools, which can become a default strategy when faced with dysphoric states (26). After being established, one’s structured templates continue to be updated by life events, constructing a Bayesian model of learning from experience and allowing for “future priors” (28) to be continuously revisited, optimally with increasing flexibility toward stimuli. The term “future priors” can be understood as a bias toward expecting particular outcomes, based on past experience, potentially distorting one’s experience of the environment to confirm these views. Such distortions can be rigid, even taking on delusional proportions.

Subcortical activation of neuroendocrine responses can lead to approach or avoidance behaviors. The ventromedial prefrontal cortex (PFC) maintains inhibitory coupling with the amygdala (29–31) and is also involved in empathic attunement (32, 33). As such, one’s ability to maintain tolerable levels of bottom-up (i.e., subcortical-cortical) emotional arousal, as well as benign interactions with others, is contingent on the information loop drawing from one’s gestalt assessment of bodily state and emotional response and preferential activation of flexible and attuned cortical regions (e.g., ventromedial PFC) instead of areas of more rigid cognitive control (e.g., dorsolateral PFC, when excessively activated) (34, 35). Indeed, individuals showing impaired affective and behavioral regulation may also manifest impaired integration of external stimuli with interoceptive awareness (36). Cognitive control and inflexibility may become survival strategies among individuals who have suffered early life adversity. Closely connected with the amygdala, the insula is involved with visceral reactions to situations (37, 38), creating a temporal structuring of sensed and perceived patterns (39). Because the insula participates in the anticipation and experience of aversive stimuli, “body prediction errors” (40) can occur whenever an uncomfortable bodily state is incongruent with an objective environmental danger, as can occur among individuals with major depressive disorder and anxiety disorders. Thus, interoceptive errors, which may occur with several psychiatric conditions, can skew the lens toward heightened distress even when the individual encounters neutral stimuli, resulting in maladaptive reactions to the environment and the potential for retraumatizing enactments. With melancholic depression, for instance, somatic preoccupation has been linked with decreased connectivity between the insula and cortical areas involved in affective control, which may also contribute to rumination and difficulty shifting attention between spontaneous thoughts (41). As such, misattunement to one’s bodily state, with rigid accompanying cognitions, can perpetuate the dysphoric state.

Dysfunctional integration between the external and internal environments has an impact on the extent to which an individual is willing to engage with stimuli perceived as dangerous. This level of integration is a crucial consideration in exposure-based treatments, which may incorporate active engagement with the environment as part of the prescribed therapy. For instance, it has been shown that the 5G allele of the single nucleotide polymorphism (SNP) rs6295 in the 5-HT1A receptor gene can manifest as decreased self-initiated exposure behaviors and consequent resistance to CBT (9). The decrease in postsynaptic 5-HT1A receptor density results in an increase in glutamatergic drive and is associated with anxiety and depressive symptoms (42). Individuals homozygous for the G allele exhibit fear generalization, heightened awareness to threatening stimuli, and greater avoidance and escape behaviors (9). It is in such instances of predicted heightened reactivity (based on a patient’s individual genetic makeup) that a more flexible and holistic psychotherapeutic approach may be warranted, creating the opportunity to prescribe an open-ended therapy instead of something more time limited, which may predictably be too much for the patient to tolerate.

Polymorphisms and Predictors of Response to Psychotherapy

As alluded to above, although there is some degree of standardization for a range of psychotherapeutic modalities, a common definition of short-term psychotherapy remains elusive. In the scope of manualized, validated forms of CBT formally studied in research settings, treatment may last for as little as a single 60-minute session (43–45) or as long as 12 weekly sessions (8). Limited standardization exists for the method of CBT delivery as well, which may range from Internet- or video-based to in-person group or individual sessions (8, 25, 46). Despite these caveats, in this article, we aim to explore the two overall variants of psychotherapy (time-limited manualized versus open-ended insight-oriented) in predicting patient response based on individual genetic makeup. Some of these differences in therapy delivery may be important in considering individuals who will respond or not respond to ostensibly the same manualized form of treatment.

Particular gene polymorphisms may affect neural activation patterns in response to environmental stimuli. Although an extensive review of the biochemical underpinnings of the manifold polymorphisms in neuroscience literature is outside the scope of this article, we briefly define one polymorphism that will be discussed below (i.e., 5-HTTLPR). Several articles competently address pertinent findings of other genetic factors (47–50). Table 1 shows a selected list of polymorphisms that may influence responses to psychotherapy and have been discussed in the literature. Table 1 also illustrates the heterogeneity of psychotherapeutic interventions in the research context, as related to genetic predictors of response. The subtle variances from study to study in Table 1 are important because the generalizability becomes increasingly smaller; even the largest scale and most robustly performed experiments harbor significant and meaningful distinctions. The table also includes studies that investigate other genetic polymorphisms, such as the BDNF gene or HPA-axis related genes, and observed their relationship to manualized therapies in the form of mindfulness-based (51) and exposure-based CBT (52, 53).

The serotonin 5-hydroxytryptamine (5-HT) transporter (5-HTT, also known as SERT) is tasked with sequestering synaptic serotonin into the presynaptic neuron for neurotransmission attenuation and recycling (54, 55) and is transcribed from the 5-HTT gene on the long arm of chromosome 17 (56–58). The 5-HT transporter-linked polymorphic region (5-HTTLPR) locus has been associated with differential sensitivity to the environment (59, 60). More specifically, a repeat-length 5-HTT SNP, rs4795541, exists as either a short (S) allele or a long (L) allele, which constitutes a sequence of 14 (S) or 16 (L) repeats in a 43-base pair insertion and/or deletion, respectively (61). Downstream, these variants can dictate protein effectiveness on the basis of the repeated sequence length. The S allele is less equipped at scavenging synaptic serotonin and attenuating response, and the L allele is better suited to do so (62, 63). These differences have implications for negative reactivity and susceptibility to environmental stressors, which may influence subsequent development of psychopathology (64, 65). For instance, S allele carriers of the 5-HTTLPR polymorphism may show amygdala hyperactivity when presented with emotional stimuli (66), as well as a cingulate cortex-amygdala uncoupling, making it challenging for patients to process their affective responses (67, 68).

The S allele of the 5-HTTLPR polymorphism has been associated with negative affectivity, with a predisposition to anxiety and depression (67, 69). Emotional regulation among healthy individuals is tied intricately to effective amygdala reactivity modulation via connectivity with the pregenual anterior cingulate cortex (70), dorsal anterior cingulate cortex (dACC) (71), and dorsolateral PFC (72). Thus, increased cingulate activity causing decreased amygdala activity enables a better and more likely response to treatment, defining a putative biomarker that may extend from health to disease to treatment prediction and response (70, 71, 73). The 5-HTTLPR S allele has been associated with decreased connectivity between the amygdala and regulatory areas, such as the dACC and the dorsolateral PFC (67). The degree of connectivity between the amygdala and areas of the ACC, mediated by 5-HTTLPR, can be associated with degrees of social dysfunction, with L allele carriers showing a negative correlation between ACC–amygdala connectivity and degree of social dysfunction (74). Aberrant patterns of cingulate cortex hyperactivity can predict response to psychotherapy. Pretreatment subgenual ACC (sgACC) activity in response to negative words has been correlated with residual severity of depressive symptoms and used to classify response and remission (75–77). Indeed, participants with the lowest pretreatment-sustained activation of the sgACC in response to being cued with negative words showed the most improvement after cognitive therapy. The sgACC reactivity may thus interfere with psychotherapeutic treatment efficacy, given its role in emotional response regulation (78–80). Absence of sgACC hyperactivity may contribute to treatment response, as shown by individuals with the lowest pretreatment activity gaining the most benefit from therapy. Both cognitive and psychodynamic modes of psychotherapy have been shown to decrease sgACC hyperactivity, reflecting the emergence of greater cognitive flexibility as a result of successful treatment (81–83). However, hyperactivity of sgACC may predict different outcomes depending on the modality of psychotherapy used. When assayed before treatment, sgACC activity has been shown to predict resistance to CBT but improved response to psychodynamic psychotherapy (81, 84, 85). This finding may show the mediating nature of the 5-HTTLPR S allele, particularly as it informs neurocircuitry activation and an individual’s ability to engage with presented stimuli. Longer-term, open-ended, nonmanualized treatment may allow for a more paced, nondirective, and individually based approach to psychotherapy, potentially circumventing some predictors of nonresponse to more time-limited therapies. In line with this idea, the authors of a recent systematic review of the genetic and neuroimaging predictors of psychiatric treatment response proposed that the ACC may be the most predictive biomarker for psychopharmacologic and psychotherapeutic outcomes (7).

However, this association is far from straightforward, particularly when the complexity of patients’ environments is taken into account. Several studies of the impact of the 5-HTTLPR on response to psychotherapy have revealed discrepancies (47, 86–88). In a study by Eley et al. (8), 359 children (ages 6–13) diagnosed as having anxiety disorders engaged in 4–12 sessions of family and/or group and manual-based CBT. Follow-up data were predominantly collected immediately after conclusion of the CBT and again 6 months later (with some data collected at 3 or 12 months). Roughly 20% more of the children with the SS-allele genotype showed complete remission of anxiety symptoms at follow-up, compared with the children who had SL and LL genotypes (78.4% versus 58.4%, respectively). Eley et al. concluded that the children with two copies of the 5-HTTLPR S allele had better outcomes following CBT for anxiety disorder and that this genotype could predict a change in symptom severity from pretreatment to follow-up. This study propelled further studies of therapygenetics to better generalize these findings. A later study also showed better responses among individuals with panic disorder and agoraphobia who were carriers of the S allele, after 1 week of exposure-based behavioral therapy, with complete remission observed among some at the 2-week mark (86).

Many features of the study by Eley and colleagues are noteworthy but, unfortunately, have been difficult to replicate (89), including the nonmedicated status of the participants, which has been a significant confounding variable in attempts to broaden these data to elderly or other patient groups. In an attempt to validate the findings of Eley et al. more largely, Bockting et al. (89) genotyped and randomly assigned 187 adults in remission from recurrent depression to receive either eight weekly sessions of CBT or treatment as usual; primary outcome was time to recurrence of depression, and patients were assessed prospectively over about 5.5 years. Bockting et al. were unable to corroborate the Eley et al. findings; roughly 75% of both treatment groups experienced depression relapse during follow-up (which was considerably longer than in the Eley et al. study), with no association seen with the 5-HTTLPR alleles.

Thus, the S allele of the 5-HTTLPR polymorphism seems to act as a double-edged sword. Some findings have indicated it is associated with improved response to manualized, time-limited psychotherapy, possibly stemming from the individual’s ability to more immediately engage with environmental stimuli, and thus with proposed treatments. For instance, the 5-HTTLPR polymorphism was assayed for its effect on psychological interventions for patients with poststroke depression (90). The presence of a single S allele was associated with a significantly better response to treatment compared with those with no S allele. It was argued that carriers of the S allele are more likely to view environmental stimuli with a negative bias, have heightened emotional reactivity, and develop negative information processing earlier in life, but conversely may also derive benefit from their more intensive attention to environmental stimuli. In the Kohen et al. study, in response to a nine-session brief pleasant events, problem-solving intervention, patients carrying the 5-HTTLPR S allele showed improvements compared with those carrying the L allele variants.

The counterargument, and what may help explain some of the discrepant findings, is that heightened reactivity may lead to avoidance of potentially threatening situations. Also, even in instances of successful psychotherapeutic interventions, the ability an individual with the S allele has of independently building on the skills gained, and hence maintaining the benefits of treatment over time, is more difficult to establish. There are nuances to the different study methodologies, even those with ostensibly similar manualized interventions. It is our argument that greater environmental structuring may be beneficial for some individuals with genetic alleles rendering them more negatively reactive, and that it is the interstices of the manualized treatment that may factor into some of the contrasting results. Open-ended, insight-oriented psychotherapies may be more geared toward engaging with patients in a more paced manner, fostering rapport, and addressing resistances before incorporating more directive, technical approaches. Although there is space in everyday practice for such an approach in other forms of therapy, the time-limited and structured nature of psychotherapy as used in research protocols may not have such leeway. Promoting more gradual immersion into the psychotherapy process, meeting the patient where he or she is, creates an atmosphere of safety and containment, with less of a demand to immediately engage in protocoled and directive therapeutic work. Although direct studies gauging genetic predictors of response to psychodynamic therapy are lacking, it has been shown that this form of treatment can lead to decreased limbic hyperactivity (81, 91, 92), an increase in 5-HT1A receptor binding in the orbitofrontal and ventral prefrontal cortices (93), and an increase in SERT density in the medial PFC (94).

The Case for Open-Ended Psychotherapies

Despite some polymorphisms conferring higher risk of adverse outcomes, the results of psychotherapy may depend fundamentally on the environment in which one is situated. In effect, there is a dynamic relationship between genetic risk and environmental input. This relationship can make particular genetic endowments more or less adaptive, with the potential for or protection against psychopathology deriving from the balance between genetics and environment, something Belsky (95, 96) termed “differential susceptibility.” In recognition of this dynamic quality, there has been a shift away from labeling allelic variants as “vulnerability genes” and toward labeling them “plasticity genes.” This concept is an important grounding point for the rest of this article, however, which posits that taking into account individual genetic constitution may allow for prescription of a psychotherapy treatment plan that will include more psychosocial and/or environmental interventions. The trend toward time-limited and manualized treatments, which rely substantially on the patient’s ability to quickly integrate the tools being modeled in a practical manner, may not be entirely realistic in instances where distress tolerance is more limited. In such cases, instead of dismissing the utility of psychotherapy, there may be a rationale for extending the treatment frame to something more open ended, allowing for the integration of therapeutic skills to occur gradually, tailored to the patient’s tolerance, while recommending or providing help in environmental restructuring to maximize the potential for therapeutic success. When attempting to rewrite emotional memories by revisiting the patient’s symptoms and conflicts, with the hope of superimposing a more flexible neural signature, there is a period of lability during which reconsolidation can tend toward more adaptive or less adaptive directions (97). Maintenance of extinction recall depends on prefrontal and anterior cingulate functioning, with effective inhibitory input onto the amygdalae when re-presented with emotionally salient stimuli (98). Genetic variants leading to decreased coupling may hinder engagement in and retention of the work of psychotherapy (68). On the other hand, individuals with the L allele of 5-HTTLPR have been shown to exhibit more robust coupling, with the potential ability to retain some of the benefits gained from psychotherapy (70, 73, 87, 99, 100). As such, extinction retention has been argued to be an endophenotype that mediates genetic variability of 5-HTT expression (101).

Ideally, the results of genetic testing would facilitate referrals to particular modalities of psychotherapy, either open ended or manualized, depending on the patient’s ability to safely interact with the proposed therapy without becoming overwhelmed. Time-limited and manualized treatment modalities may provide neither sufficient grounding to allow for optimal revisiting of distressing templates nor the level of integration of techniques required for independent elaboration on the benefits of therapy. Technique aside, rapport has been argued to be the crux of success for psychiatric interventions on both the pharmacological and psychotherapeutic levels (102, 103); hence, taking the time to foster the relationship has an intangible quality, which in itself is important for outcome and needs to be heeded. For individuals with certain polymorphisms, concerted efforts at optimizing environmental support may help mitigate anxious traits and the tendency for maladaptive responses to stimuli. Such differential susceptibility has been extensively argued and should be factored into the decision to prescribe and advocate for an open-ended or manualized psychotherapy for a patient (95, 96, 104, 105).

As described earlier, presence of the S allele of the 5-HTTLPR has been associated with both resistance to and benefit from psychotherapy (8, 86, 106). These contradictory findings highlight some of the key methodological differences in ostensibly the same form of manualized treatment. For example, studies that have reported greater treatment response among patients with the S allele have focused on more tailored psychosocial interventions (8, 90), in this way capitalizing on patients’ greater sensitivity to environmental stimuli by providing greater support. In contrast, individuals with the L allele of this polymorphism may benefit more from manualized, time-limited therapies, which rely more on the “teaching of skills” approach, which can be subsequently expanded on independently by patients who may not require as robust a social structure. A notable instance of such differences has been described by Wannemüller and colleagues (46). In their report, one session of exposure therapy (120–140 minutes) was provided to unmedicated individuals who had different forms of specific phobia, assessing effects of the 5-HTTLPR genotype on postintervention effects (immediately and after 7 months). Whereas subjective fear decreased immediately postexposure in the different cohorts (LL, LS, and SS), a marked difference in symptom reporting was observed at the 7-month follow-up. Namely, LL carriers continued to show a decrease in their fear response, indicating they were able to independently use the skills learned during the therapy session. In contrast, there was a return of fear among the SS carriers, nearing preintervention levels. As such, an argument is made for the need to tailor treatment according to the patient’s ability to internalize what is being presented in psychotherapy. We may imagine that, technique being standardized, treatments either fit or do not, but an expansion of what is provided to account for the environmental impact postintervention may be crucial to consolidate effects and optimize therapeutic benefit. For instance, in the study by Eley et al. (8), in which the S allele was shown to confer greater response to psychotherapy, parental training was a core element of the intervention, helping to enrich the child’s posttreatment environment. A paramount point is that immediate symptomatic improvement (e.g., fear decrease) is not necessarily predictive of symptom levels at follow-up. Moreover, follow-up in and of itself has been unstandardized and, naturally, may lead to disparate results when considering that postintervention periods have varied from 2 weeks to 5.5 years (86, 89). As such, extended protocols of more manualized treatments, both in terms of duration and environmental restructuring (when possible), may be a potential adaptation to genetic information obtained about individual patients and their ability or inability to work within particular therapeutic frames. Arguably, more open-ended and less directive models of therapy (e.g., psychodynamic and/or psychoanalytic) may allow for a more gradual immersion into a dialogue about symptom burden and intrapsychic conflict. Indeed, the allowance of greater numbers of weekly sessions in insight-oriented therapies helps shape the provider as an increasingly central figure in the patient’s life, heightening the transference element and allowing for a more individualized processing to take place. At the same time, the less restrictive time frame and attentiveness to how much the patient can tolerate may increase compliance and limit dropout. This concept is somewhat in line with a meta-analysis of 33 articles on treatment for specific phobia, which indicated better long-term improvement with multiple psychotherapy sessions, as opposed to one-session interventions (107).

Cost-Effectiveness, Limitations, and Future Directions

The association of individual genetic markers with nonresponse to manualized treatment represents a new possibility to guide clinicians in selecting psychotherapies for individual patients and will need further study. For comparison, the field of pharmacogenomics, although exciting in its prospects and having yielded some replicable findings informing treatment considerations, has had its share of conflicting evidence, leaving practitioners uncertain as to how such data should be incorporated in daily clinical practice (108). Therapygenetics is an even newer and less developed field, making it challenging to know exactly how it fits into the puzzle of mental health treatment options and whether the data it reveals are cost-effective and worth exploring. Insurers are inconsistent in coverage provided for pharmacogenomic testing, in some instances openly citing the lack of an evidence base. As a result, it would likely be financially onerous for a patient to obtain genetic data to help prospectively inform the best type of psychotherapy. In one study, the nerve growth factor SNP rs6330 was assayed to determine the utility of predicting response to CBT (109). The authors determined that for every five patients assessed, one significantly more accurate prediction of outcome could be made.

Perhaps the most important consideration with regard to allelic predictability, however, is effect size. Genetic associations in and of themselves are intrinsically small, and many factors have a role in the downstream functional manifestations apart from just one sequence variant in a gene or its surrounding regions. Genetic associations typically require numbers in the thousands to reach true significance statistically, let alone clinically, which remains widely prohibitive in most research settings (110). Furthermore, the issue becomes how to define treatment success, which ranges from complete remission of active primary symptoms to relapse prevention or prophylaxis in high-risk groups. Further study will be needed if such direct and causal explorations of therapygenetics are to provide clinical benefit. As reflected in Table 1, some challenges exist in the methodological comparisons and generalizability of this work. As mentioned earlier, the table represents a nonexhaustive and selected body of highly visible and highly cited peer-reviewed papers within the field of therapygenetics. One can glean the disparities that exist, which portend difficulties in comparing one study to the next. For instance, the length of psychotherapy has ranged from a single 1-hour session (44) to 10 weekly 2-hour sessions (49). The type of psychotherapy studied has largely included cognitive-behavioral, with one mindfulness-based variant included, but no psychodynamic or more frequent and lengthier modalities. The psychosocial aspects of the interventions delivered have been explored only on the basis of the authors’ presentations in their publications and have been explicitly addressed only by a single study of patients with poststroke depression (90), in which environmental support was provided and emphasized. Primary outcome measures have been highly heterogeneous, as have the lengths of the follow-ups in which the success of the treatment interventions have been evaluated. Length of follow-up is perhaps one of the most important factors when considering the sustainability of treatment efficacy, yet it has ranged from immediately postintervention to 5.5 years of follow-up, with results including positive and negative predictive value for genetic testing identified by different studies. With such a backdrop, the argument emerges that durability of treatment may be optimized by a priori individualization of treatment made on the basis of likelihood of response deriving from genetic data. In instances where manualized, time-limited treatments may be the only option, provision of environmental support may help consolidate clinical benefits for individuals who otherwise may have been selected for longer-term or open-ended treatment.

Candidate gene approaches, such as the presently discussed studies focusing on 5-HTT gene variation, should be complemented by genome-wide association studies, such as the Rayner et al. study (111) regarding CBT response by patients with anxiety and depression. In that work, the authors performed genome-wide association meta-analyses of symptoms subsequent to completion of CBT. Even in this largest study to date, the authors concluded that heritability of therapy outcome was smaller than predicted, alluding to the heterogeneity and complexity of the traits under investigation. Our current understanding is that the upstream and nonmutation-based means of regulating and influencing protein expression is through epigenetics. Rather than assaying and correlating polymorphisms or allelic variations (e.g., genetic mutations), epigenetic mechanisms use other factors to modulate genetic transcription (e.g., DNA methylation patterns). Roberts and colleagues (112) studied methylation patterns in the serotonin transporter promoter region before and after administration of CBT. They found that at one particular site, responders to the therapy had an increase in methylation, whereas nonresponders had the opposite, suggesting that treatment responsiveness was associated with this epigenetic mechanism (the direction of which—causal versus effectual—remained to be determined). In a study of the monoamine oxidase A (MAOA) gene with a cohort of patients diagnosed as having panic disorder, Ziegler et al. (113) found that methylation patterns were distinct according to disorder severity and in responsiveness to CBT. Namely, panic disorder severity was negatively associated with MAOA methylation, and successful CBT was associated with an increase in methylation (113). In another attempt to bridge epigenetic modifications with behavioral corollaries (114), transcranial magnetic stimulation was used to target the medial prefrontal cortex in addition to provision of an exposure-based CBT and psychoeducation. Those authors analyzed the MAOA epigenetic relationship with methylation patterns in various loci of the gene. Similarly, Reif et al. (115) probed MAOA with respect to a long allele polymorphism in the promoter region (uVNTR) of patients diagnosed as having panic disorder with agoraphobia and found that carriers were more resistant to exposure-based CBT, which also suggested the potential efficacy of tailoring psychotherapy interventions with this resistance in mind. Such multipronged approaches to treatment, integrating insights from genetics, psychotherapy, and neuromodulation, are exciting directions for further study.

Conclusions

The fields of therapygenetics and therapy neuroimaging remain in early development. Much work needs to be done to standardize neuroimaging techniques and protocols to be able to harmoniously share and compare the work of many talented researchers. Likewise, validation or large reproduction of studies of genetic predictors of manualized psychotherapeutic treatment success has been deemed almost prohibitively challenging because of the large amount of data needed and the lack of methodologically standardized and compatible outcome objectives or diagnoses from which to work with or toward. The efforts have nonetheless yielded impressive converging ideas, genetic foci, and neural correlates from which a basic neuroscientific and biologic mechanism can be explored. With concerted efforts and collaborations, further developments should allow for more clarity about how clinicians can use this information in treating patients. As the goal of individualizing patient care turns to predictors of treatment response to psychotherapeutic interventions, we are left to tailor a manualized, time-limited therapy model that best meets the needs of the patient. In contrast, open-ended, insight-oriented modalities remain largely untapped from a research perspective, offering possibilities for further exploration of how genetic and imaging markers may guide and predict success. More to the point, the predictable nonresponders to manualized treatment are perhaps better suited from the outset for longer, more intensive forms of psychodynamic and psychoanalytic therapy.