During lifetime, 60.7% of men and 51.2% of women experience at least one potentially traumatic event such as being taken hostage or being kidnaped, experiencing or witnessing sexual or physical assault, torture, a terrorist attack, a severe car accident, a natural disaster, war, or the unexpected death of a beloved person (1). A significant proportion of these events already occurs during young age, as 26% of male and 18% of female subjects reported of having experienced at least one traumatic event in a representative community sample (n = 3.021, age 14–24 years) (2). Of those experiencing potentially traumatic events, 10–40% develop psychiatric symptoms of clinical relevance (3–7) such as affective disorders, substance abuse, or post-traumatic stress disorder (PTSD). The latter goes hand in hand with significant impairments of psychological and physical health (1, 7), elevated levels of suicidality (8) as well as with a decreased level of social functioning (9).

The reason why some individuals, who have experienced a traumatic event develop a PTSD while others do not, is still not clearly understood (10). Some factors like an own or family history of psychiatric disorders before the traumatic event took place, the experience of a traumatic event early in lifetime, and/or female gender were shown to be associated with an elevated risk for developing PTSD (3, 11). Furthermore, sociodemographic and ethnologic factors play an important role, when it comes to varying prevalence rates for PTSD (7, 8, 12, 13).

Neurobiological models conceptualize the symptoms of PTSD as correlates of a dysfunctional stress reaction to traumatic events (14). This hypothesis is based on observations such as changes in the hypothalamic–pituitary–adrenal axis in PTSD patients. An abnormal downregulation of cortisol seems to result in sustained, increased levels of corticotropin-releasing hormone, which leads to a higher release of adrenocorticotropic hormone stimulating the release of glucocorticoids in turn (15). The sustained glucocorticoid exposure has been suggested to promote hippocampal atrophy as indicated by neuroimaging studies showing reduced hippocampal volumes in PTSD patients (16). Some initial hypotheses have been brought up linking neurobiological changes to altered mechanisms of extinction and hyperarousal, flashbacks, and avoidance behavior as the core symptoms of PTSD (14). For example, functional imaging studies including stressful tasks (stressful scripts, cues, and trauma reminders) showed a decreased response of the prefrontal cortex and an increased amygdala response in PTSD patients (17). Following this line of thought, it has been postulated that in patients with PTSD, stimulators of the limbic stress system prevail over inhibitors. Such an impaired top-down cognitive control over limbic areas, which is frequently associated with hypoactivation in the dlPFC might also explain flashbacks (18). In PTSD, neutral cues become associated with the traumatic event and may trigger a conditioned fear response. Failure to extinguish the conditioned fear is thought to lead to the persistence of symptoms (19). It has been suggested that the interaction between amygdala and the hippocampus, as two regions that play a critical role in the consolidation of emotionally laden memories, may contribute to the intrusive nature of trauma recollection in PTSD (20).

Diverse pharmacological as well as psychotherapeutic approaches have been suggested for the treatment of PTSD, while the latter represents the gold standard and pharmacological treatment rather constitutes an adjuvant form of symptom control. Following guidelines (21), patients should be offered a trauma-focused psychotherapy such as cognitive behavioral therapy (CBT) (22), eye movement desensitization and reprocessing (EMDR) (23), or prolonged exposure therapy (ET) (24).

CBT helps patients to question and alter dysfunctional trauma-associated cognitions. In vivo or in sensu confrontation with trauma reminders helps patients to overcome their avoidance of situations and thoughts that are no longer dangerous, which leads to habituation and a reorganization of trauma memories finally (25).

During EMDR, patients mentally focus a trauma-associated disturbing image, memory, emotion, or cognition. In comparison to CBT, confrontation is rather short and interrupted by saccadic eye movements that are initiated by the therapist (25, 26). The influence of rapid eye movements is not clearly understood yet. Mechanisms discussed are (1) focusing inner thoughts and feelings on the one hand and external stimuli (e.g., finger moves) on the other, might form an ideal attention level that is necessary for reconsolidation; (2) focusing external stimuli while reprocessing traumatic memories leads to a certain degree of distraction that prevents patients from being overwhelmed by their traumatic memories during the processing stage; (3) moving fingers might induce a certain degree of relaxation (activation of the parasympathetic system), which is then conditionally associated with traumatic memories resulting in releasing effects (25, 27, 28). Thus, besides habituation and conditioning learning processes as in CBT, modulation of attentional processing and facilitation of gaining cognitive control are suggested mechanisms of action.

Prolonged ET is a manualized (24) and effective (29, 30) psychotherapy for the treatment of PTSD. Virtual reality ET is a frequently used version of ET for treating combats, by providing realistic visual, auditory, and even tactile and olfactory cues that can be presented in a graduated manner. Habituation to anxious memories seems to be one of the key mechanisms here.

All treatment forms have in common that traumatic events or memories should no longer be avoided but confrontation on a behavioral (ET) or cognitive (CBT, EMDR) level should take place until a certain degree of fear extinction and habituation has been reached. CBT and EMDR furthermore question trauma-associated dysfunctional cognitions and replace them by more functional ones (top-down processing). Habitation and altered cognitions normally result in reduced feelings of anxiety and physiological arousal. Furthermore, confrontation with aversive stimuli and memories during therapy leads to fear extinction (31). Regarding PTSD, it was argued that extinction may not be based on an erasure of original information, but may rather reflect new learning (32).

Several neuroimaging studies attempted to examine neurobiological aspects of psychotherapeutic interventions in PTSD. The impact of trauma-focused psychotherapy on the stress system is thought to be primarily mediated by learning processes (33, 34), which is expected to be represented by changes in neurobiology. Conditional learning and extinction processes have been linked to the hippocampus and basal ganglia functioning. Biased memory for negative stimuli as in depression, but also PTSD has been suggested to result from an interaction of these regions with a hyperactive amygdala. Increased reciprocal signaling in this network may lead to sustained hyperactivity (18). Furthermore, biased processing of negative information has been linked to hypoactivity of the prefrontal cortex while biased attentional processes were also related to altered ACC activation (18). Psychotherapeutic interventions addressing these processes should, therefore, influence the activation within these areas. Accordingly, previous neuroimaging studies in adults with PTSD after psychotherapeutic intervention found decreased activity in the amygdala and an increased activity in frontal areas (35).

The assessment of neural correlates of successful psychotherapeutic treatment in adolescents with PTSD might allow for tailoring treatments to the special needs and characteristics of this population and has been suggested to have a potential for the identification of neurobiological markers that can serve as predictors for subsequent therapeutic success. We present a systematic review of the current state of the literature regarding neuroimaging and psychotherapy of PTSD.

Results of the present systematic literature review regarding the neurobiology of PTSD treatment indicate that symptom improvement was shown to be accompanied with a pre-to-post-signal reduction in the amygdala. Accordingly, high amygdala activation prior to treatment was associated with subsequent treatment failure or non-response to treatment. Additionally, pre-to-post-signal decrease in the insula was associated with symptom improvement and increased insula activity prior to treatment was shown to be related to treatment resistance. Besides that symptom improvement goes hand in hand with fMRI signal increase in the hippocampus, dACC, PFC, and vACC, elevated dACC activity prior to treatment was shown to be related to subsequent treatment success and was a positive predictor for treatment response. Regarding cerebral networks involved, there were not obvious differences between CBT as the predominant treatment in nine of the studies listed, and ET or EMDR that were applied in three studies each. This may be interpreted in a way that changes in brain functioning rather represent changes in clinical symptoms and improvement than the different techniques that facilitate these changes. This interpretation is furthered by the finding that these therapy forms seemed to be equally effective. Therefore, changes in cerebral networks can hardly be related to distinct types of therapy but rather to elements like habituation, extinction learning, cognitive control, or attention direction.

The insular cortex is involved in emotions, salience, consciousness, motor control, and homeostasis (54, 55). Task-related fMRI studies showed increased activation in the insular cortex in individuals with PTSD, and this activation was positively correlated with the severity of PTSD symptoms (56–58). Neuroimaging studies investigating anxiety disorders also showed greater insular activation (59), pointing to a role in anxiety processing in general. With respect to the latter, the insula can be functionally divided into anterior and posterior parts: anterior insula activation, as shown to relate to salience and error processing (54), has been interpreted as a warning signal that is associated with symptoms of hyperarousal (51), whereas the posterior insular cortex is more strongly linked to the representation of interoceptive/bodily state changes, such as sympathetic arousal during negative affective states (53). Concerning the results of the present review, psychotherapeutic treatment success was associated with pre-to-post-decrease of insula activity especially in anterior parts. This can be interpreted as a correlate of reduced alertness and hyperarousal with traumatic and other emotional stimuli after treatment (55). Regarding therapeutic techniques, these changes could relate to interventions strengthening cognitive control and modulating attention as used in CBT.

Previous fMRI studies consistently showed increased amygdala activation in patients with PTSD compared to controls (60–63) and, with respect to the present review, a reduction of pre-to-post-treatment activity. As the amygdala has been suggested to play a crucial role in the processing of threatening stimuli, fear learning and fear expression (64, 65), the findings at hand might be a neurobiological correlate of a reduced response to threatening stimuli after treatment. Habituation to aversive stimuli as a result of confrontation treatment which is an element of CBT, EMDR, and ET could be a process underlying this observation. Furthermore, the present review highlights that an increased amygdala activity during fear processing prior to treatment was associated with subsequent treatment failure and non-response (48). The authors concluded that those patients, who reveal excessive fear-related amygdala activity, might have problems in regulating upcoming anxiety during confrontation with traumatic memories in therapy, which in turn might be a necessary requirement to experience subsequent habituation. In line with this, Cisler et al. (49) found that an increased amygdala reactivity to both neutral and threat stimuli before treatment was associated with poor treatment response in adolescents with PTSD. However, increased amygdala activity to threat relative to neutral stimuli was related to symptom improvement. The study of Cisler et al. (49) highlights the influence of paradigms and ways of analyses applied.

The amygdala and the hippocampus have both been suggested to play critical roles in the consolidation of emotionally laden memories in healthy adults. The interaction between these two hyper-responsive regions may contribute to the intrusive nature of trauma recollection in PTSD (20). Prolonged exposure to stress such as in patients with PTSD induces high glucocorticoid levels, which seem to harm the hippocampus (impairment of neurogenesis and loss of dendritic spines) (66). Structural imaging studies showed that PTSD patients have reduced hippocampal volume compared to HC (67–70). Whether the smaller hippocampal volumes in PTSD patients are associated with functional abnormalities still remains unclear. fMRI studies that compared PTSD patients with HC revealed both increased (71) and decreased hippocampal activation (72). The present review found pre-to-post-treatment increase of the hippocampus activity. It has been speculated that this might relate to improved short- and long-term consolidation and memory processes, improved encoding and recognition of episodic memory as well as an improved evaluation of experiences (35, 66). Furthermore, increased hippocampus activity might point toward improved identification of safe contexts, improved learning, and alteration of existing memories (20). Habituation and exposition as integral parts of the majority of treatments applied may be linked to these changes observed.

In the context of the processing of traumatic events, the ACC has been implicated in helping modulate fight/flight reactions to perceived threats (45). Altered activity in the ACC after successful psychotherapy of PTSD was, therefore, interpreted as a correlate of improved conflict monitoring and emotion regulation (37). The ACC can be subdivided functionally in ventral and dorsal parts: the vACC has been proposed to play a particular role in emotion processing, automatic emotion control, and social decision making (37, 73). The vACC was shown to be interconnected with the amygdala, nucleus accumbens, hypothalamus, and anterior insula and is assumed to be involved in assessing the salience of emotion and motivational information (74). The dACC has been proposed to be involved in cognitive processing or appraisal of emotion and the saliency of the source of social information (37, 73). The dACC is particularly connected with the prefrontal cortex and parietal cortex as well as the motor system and the frontal eye fields, making it a central node for processing top-down and bottom-up stimuli and assigning appropriate control over other areas in the brain (74). Following the results of the present review, pre-to-post-enhancement of dorsal and ventral ACC activity might be a correlate of improved cognitive processing of trauma-associated cognitions, better appraisal of trauma-associated emotions as well as in improved emotion control, all of these being processes that are directly addressed particularly in CBT. The present review found that treatment success was rather associated with an elevated pre-treatment activity in the dACC, which in turn might relate to better preconditions regarding cognitive control and emotion regulation.

Frontal brain areas have been strongly associated with PTSD in the literature before. The majority of studies revealed an increased pre-to-post-treatment activity in the PFC, which might be interpreted as a correlate of improved top-down regulation of emotional systems as well as a potential correlate of improved processing of risk and fear, decision making and emotion regulation (20, 75) as established during psychotherapy, particularly CBT. Changed prefrontal functioning might also relate to findings from structural neuroimaging studies that found reduced volumes in prefrontal brain regions (75–77) in PTSD patients. In a similar vein, results of a connectivity study of King et al. (42) showed increased connectivity to dlPFC regions within the central executive network after successful PTSD treatment, which was interpreted as a potential correlate of improved emotion regulation. The dlPFC was shown to play a critical role in working memory. Regarding PTSD, it was argued that extinction may not be based on an erasure of original information, but may rather reflect new learning (32). Increased activation in the dlPFC in the context of successful psychotherapy might represent a correlate of newly gained control of unwanted memories (47).

Post-traumatic stress disorder patients after successful treatment point toward a predominantly increased pre-to-post-treatment activity in “cognitive brain areas” like the PFC, dACC, and hippocampus when being compared to PTSD waitlist patients (35, 40, 42). Decreased pre-to-post-treatment activity in “emotional brain areas” like the insula and amygdala was more pronounced in within-group analyses (37, 53) and compared to patients of the waitlist control group (35, 40). Those findings underscore the importance of taking into consideration the control group applied when interpreting data of treatment studies.

Among others the impact of the review presented here might, therefore, be limited due to cross-study variations in control groups applied, different trauma histories, subject characteristics, and paradigms applied as this makes it very difficult to draw valid conclusions regarding subgroup and overall results.

Neurocircuitry models emphasize reduced activation of frontal brain areas and elevated activity of limbic regions in patients with PTSD and a loss of top-down inhibition is thought to be one of the main factors resulting in impaired extinction of traumatic memories or insufficient modulation of affect (e.g., re-experiencing) (20). The present review was able to show that brain activation results partly fit into this picture with relatively consistent signal increases in the dACC as a frontal brain area, while activity in key areas related to emotional processing like the anterior insula and amygdala rather decreased from pre-to-post-psychotherapy. This can be seen as an improved bottom-up and top-down regulation in response to successful treatment. Future studies should investigate whether abnormal activity patterns in patients with PTSD indeed normalize after treatment and become comparable to those of HC.

JS and KM contributed substantially to the work, analyzed and interpreted the data, and drafted the manuscript. BA and PP contributed substantially to the work, interpreted the data, and revised the manuscript critically for important intellectual content. All the authors approved the final version to be published and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf and declare: no support from any organization for the submitted work and no other relationships or activities that could appear to have influenced the submitted work.