Structural magnetic resonance images (MRIs ) showing the structures of interest in this chapter: (a) a sagittal MRI slice showing the dorsal anterior cingulate gyrus (dACC) and structures comprising the ventral medial prefrontal cortex (vmPFC) including the rostral anterior cingulate cortex (rACC), medial frontal gyrus, and subgenual ACC; (b) a sagittal slice showing the amygdala and hippocampus; and (c) a horizontal slice showing bilateral insula (also called the insular cortex)
The studies presented in this chapter utilize a variety of neuroimaging techniques, including functional magnetic resonance imaging (fMRI), positron emission tomography (PET), and single-photon emission computed tomography (SPECT). All three can be equipped to present participants with experimental stimuli such as pictures, sounds, or finger shocks while recording three-dimensional images of brain function. PET can provide measures of cerebral metabolic rate for glucose, regional cerebral blood flow, or receptor occupancy, whereas fMRI measures blood-oxygen-level-dependent (BOLD) signal. Due to its good spatial and temporal resolution, and its ability to gather both functional and structural images in the same session, fMRI has been the most commonly used imaging technique in recent neuroimaging studies of PTSD.
A typical functional image such as Fig. 10.2 shows areas where the amount of brain activity in one experimental condition compared to another condition differs between PTSD and control groups (see Fig. 10.2 caption for more details). For example, this figure shows that amygdala activation in response to fearful versus happy faces is greater in individuals with PTSD compared to trauma-exposed individuals without PTSD. Given the fear conditioning model of PTSD, exaggerated amygdala responsivity is to be expected and indeed has been reported . Interestingly, the general neurocircuitry findings predicted by the fear conditioning model of PTSD have been replicated using a wide variety of experimental paradigms , not limited to fear conditioning per se.
A typical functional image such as Fig. 10.2 illustrates a “difference in differences.” The arrow indicates activation in the right amygdala, where the amount of brain activity in one experimental condition compared to another condition differs between PTSD and control groups. The data analysis steps that yield such an image are as follows: First, within each subject, we statistically compare brain responses in one condition to those of another condition (e.g., amygdala responses to fearful versus happy faces). This comparison yields a statistical image that shows the brain areas that are “activated” significantly more (or less) in one condition versus another. Second, these “difference” images per subject in the PTSD group are statistically compared to difference images of a control group. The result is an image like Fig. 10.2 that can show brain regions that are significantly more (or less) activated in one group than another. The functional magnetic resonance image (fMRI) in Fig. 10.2 displays activation to fearful versus happy facial expressions in the right amygdala (z = 3.14; Montreal Neurological Institute [MNI] coordinates, +22, +2, −14 [arrow]; and z = 3.03; MNI coordinates, +22, 0, −26) that were greater in the PTSD group versus trauma-exposed control group. The bar graph shows signal change in the amygdala in each condition (relative to fixation baseline) for each group. Error bars represent standard error of the mean (Reprinted with permission from )
Functional neuroimaging studies of PTSD commonly include fear conditioning paradigms, symptom provocation tasks, or other emotional and nonemotional tasks. Fear conditioning studies often use mild finger shocks as the aversive US and can record brain activity during acquisition, extinction, and extinction recall phases. Symptom provocation paradigms detect brain activation differences between symptomatic and neutral states in PTSD. One such paradigm is script-driven imagery, in which participants are presented with audio-recorded narratives (“scripts”) of their traumatic and neutral experiences. Other symptom provocation tasks involve exposing participants to trauma-related sounds, odors, or pictures. Trauma-unrelated emotional stimuli include positively and negatively valenced pictures, as well as photographs of faces with varying emotional expressions. Finally, some paradigms implement emotionally neutral cognitive tasks or stimuli or take functional images while participants are at rest. In the following sections, we will describe findings in several brain regions of interest from PTSD studies that have used these paradigms.
The amygdala, a structure involved in fear expression and fear conditioning, is hyperresponsive in individuals with PTSD relative to comparison groups. This general finding has been reported in studies utilizing disparate neuroimaging paradigms including fear conditioning and extinction, symptom provocation, general emotional and neutral stimuli, and resting state.
Evidence of increased amygdala recruitment in PTSD has been found during both the acquisition and the extinction learning phases of fear conditioning paradigms, as well as during extinction learning recall. Bremner et al. reported greater left amygdala activation in a PTSD group relative to a control group in a contrast between a fear acquisition condition (in which a shock was paired with a picture of a blue square) and a control condition (in which participants received random shocks in the absence of a CS) . Milad and colleagues used a different fear conditioning paradigm that included an acquisition phase in which two different colored lights (CS+s) were paired with shock and a third color was not (CS−), followed by an extinction phase in which only one of the two CS+s was extinguished [4, 37]. Relative to trauma-exposed healthy control participants, individuals with PTSD showed increased amygdala responsivity to the shock . Furthermore, during late extinction learning, the PTSD group had greater amygdala activation in the contrast between the extinguished CS+ versus the CS−, suggesting impaired late extinction learning . Using the same paradigm, Garfinkel and colleagues found that during next-day recall in a never-conditioned (safety) context, individuals with combat-related PTSD showed increased amygdala activation to the extinguished CS+ versus the CS−, while combat controls without PTSD did not . However, when the extinguished CS+ was then represented in the previously conditioned (danger) context, combat controls showed increased amygdala activation to the CS+ versus the CS−, relative to the PTSD group. In a newer study from Brunetti and colleagues, recently robbed bank clerks underwent single-trial conditioning between a startling noise (US) and pleasant or unpleasant emotional pictures, with increased bilateral amygdala activation to conditioned negative pictures in the PTSD group relative to the non-PTSD group .
Some symptom provocation studies have replicated the basic finding of relatively increased amygdala activation in PTSD. In an early PET study, Rauch and colleagues reported increased amygdala regional cerebral blood flow in a PTSD group during traumatic versus neutral script-driven imagery . However, that study did not include a non-PTSD comparison group, and subsequent studies were needed to draw more firm conclusions. Shin et al. found increased amygdala activation in the traumatic versus neutral script contrast in male combat veterans with PTSD compared to male combat veterans without PTSD . In a task using auditory cue words, St. Jacques and colleagues found increased amygdala activation during retrieval of negative versus positive autobiographical memories in participants with PTSD versus trauma-unexposed control participants .
Amygdala hyperresponsivity in PTSD also has been observed in response to other trauma-related stimuli, such as combat-related sounds, odors, and photographs. Liberzon et al. used SPECT to show increased left amygdala activation to combat noise relative to white noise in combat veterans with PTSD, but not in combat veterans without PTSD or in combat-unexposed control participants . In a similar design but utilizing PET and lacking a non-PTSD control group, Pissiota and colleagues replicated the finding of increased amygdala regional cerebral blood flow in PTSD participants during exposure to combat noises, relative to neutral noise . Vermetten et al. found greater amygdala activation in response to the odor of diesel fuel (a potent reminder of combat for many military veterans) in combat veterans with PTSD, relative to those without PTSD . Morey and colleagues found that participants with PTSD had greater amygdala activation to trauma-related distractor photographs than trauma-exposed healthy control participants . Driessen et al. found greater amygdala responsivity during traumatic versus neutral event recollection in women with borderline personality disorder with comorbid PTSD compared to those without PTSD . Protopopescu et al. found increased amygdala activation in response to viewing trauma-related versus trauma-unrelated words in a PTSD group relative to a trauma-unexposed control group . Brashers-Krug and Jorge showed noncombat and combat-related film clips to combat veterans in an fMRI environment and unexpectedly found that amygdala activation positively predicted PTSD severity during the noncombat film but negatively predicted PTSD severity during the combat film . However, neither subjective ratings of the films nor categorical PTSD diagnoses were reported.
In PTSD, the amygdala is also hyperactivated in response to trauma-unrelated emotional stimuli such as fearful facial expressions. Fearful expressions are essentially predictors of potential threat, and the amygdala is highly responsive to them, even in healthy individuals. Findings of exaggerated amygdala activation to fearful facial expressions in PTSD suggest that functional abnormalities in the amygdala are not specific to trauma-related reminders. For example, Shin et al. found increased amygdala responses to fearful versus happy facial expressions in PTSD relative to trauma-exposed control participants  (see Fig. 10.2). Williams and colleagues used a similar paradigm and found that participants with PTSD had greater amygdala activation than a non-traumatized control group in a fearful versus neutral facial expression contrast . Kemp et al. reported greater right amygdala activation to fearful versus neutral faces in a PTSD group, compared to healthy controls or a group with PTSD and comorbid major depressive disorder . In an fMRI study, Garrett and colleagues found that the left amygdala/hippocampus activated more to emotional faces versus scrambled pictures in traumatized youth with and without current PTSD, relative to non-traumatized control youth .
In an affective priming task using emotional face stimuli, earthquake survivors with PTSD showed increased left amygdala activation compared to controls . Dickie et al. showed photographs of expressive faces to individuals with PTSD in an fMRI environment and found a positive correlation between PTSD symptom severity and left amygdala activation to successfully-remembered fearful faces . El Khoury-Malhame and colleagues found that attentional bias toward threatening words and faces was positively correlated with amygdala activation in a PTSD group but not in a healthy control group . Finally, Brohawn and colleagues found greater amygdala activation in response to photographs of aversive versus neutral scenes in participants with PTSD compared to trauma-exposed control participants .
Interestingly, increased amygdala activation in PTSD can be observed even when emotional visual stimuli are presented below perceptual thresholds, such as in masking paradigms. This has been reported in studies using both trauma-related and trauma-unrelated images. For example, combat veterans with and without PTSD were shown combat-related and combat-unrelated images above, near, and below recognition threshold using backward masking . In that study, the veterans with PTSD had greater amygdala activation than control participants regardless of the content and recognition threshold. In a contrast between backwardly masked fearful and masked happy faces, Rauch et al. found that participants with PTSD had greater amygdala activation compared to trauma-exposed non-PTSD participants . Similarly, Bryant and colleagues found that individuals with PTSD showed greater amygdala activation to masked fearful versus neutral expressions than did healthy trauma-unexposed control participants . Felmingham et al. reported increased amygdala activation in response to masked fearful faces in a PTSD group relative to trauma-exposed and trauma-unexposed controls . Killgore and colleagues used a masked faces paradigm with fMRI to compare healthy controls to an anxiety disorders group including PTSD, panic disorder, and specific phobia. Relative to the control group, participants with anxiety disorders exhibited increased left amygdala activation to both masked fear versus neutral faces and to masked happy versus neutral faces .
Some studies suggest that the amygdala is more active in PTSD versus control groups even during emotionally neutral tasks or at rest . For example, in a SPECT study, Chung et al. reported greater blood flow in the amygdala at rest in participants with PTSD relative to healthy control participants  (but see ). Lanius et al. studied a sample of participants 6–12 weeks after experiencing a psychologically traumatic event and found that the strength of resting-state functional connectivity between the right amygdala and posterior cingulate cortex predicted future PTSD symptom severity . This relationship remained statistically significant after controlling for comorbid depression. Using a neutral auditory oddball task in an fMRI environment, Bryant and colleagues found greater amygdala responses in participants with PTSD relative to a non-traumatized control group . Furthermore, in a study that recruited bank robbery survivors with and without PTSD, all participants showed increased amygdala activation to negative pictures, but only the PTSD group also showed increased amygdala activation to neutral pictures . In a recognition memory task using emotionally neutral images, Whalley et al. found greater left dorsal amygdala activation in the old versus new contrast in PTSD relative to trauma-exposed participants and participants with depression .
Many studies have reported a positive correlation between level of amygdala activation and PTSD symptom severity [39, 41, 44, 48, 54–56, 58, 65, 66, 68, 69]. In addition, PTSD treatment studies  have shown that higher pretreatment amygdala activation predicts less improvement with cognitive behavioral therapy (CBT)  and a positive CBT response is associated with decreased amygdala activation [72, 73].
It should be noted that not all studies report increased amygdala activation in PTSD [74–80]. Amygdala responses habituate rapidly, and averaging amygdala signal over time in PET and SPECT analyses or across fMRI blocks can dilute the initial activation signal. This could lead to inconsistencies across studies. Another caveat to consider is that amygdala hyperreactivity is not specific to PTSD and has also been observed in other anxiety disorders (e.g., specific phobia, social anxiety disorder) . While amygdala hyperreactivity is common across anxiety disorders, PTSD is unique in that this hyperreactivity may reflect a failure of vmPFC inhibition over fear responses and may also be related to increased dACC activation .
Medial Prefrontal Cortex
mPFC structures are involved in fear expression, fear conditioning and extinction, and emotion regulation. mPFC function in PTSD has been examined with fear conditioning, symptom provocation, and emotional and neutral paradigms. In keeping with the well-characterized relationship between amygdala and vmPFC in rodent fear conditioning studies, some PTSD studies have found a negative correlation between amygdala activation and vmPFC activation [36, 41] (but see [49, 50, 82]). Ventral mPFC function in PTSD has been explicitly examined in fear conditioning studies. Several fear conditioning and extinction studies have reported reduced activation or even deactivation within vmPFC structures (such as rostral anterior cingulate cortex [rACC]) in PTSD groups, relative to control groups. Interestingly, in contrast to the hypoactivity of ventral regions of mPFC, more dorsal regions, such as the dorsal anterior cingulate cortex (dACC ), are frequently hyperactive in PTSD. This dissociation has been reported in fear conditioning studies. For example, Bremner and colleagues used PET to demonstrate that, during fear conditioning, women with PTSD had greater dACC activation than trauma-unexposed controls; during extinction, the PTSD group had relatively less subgenual cortex and rACC activation . Brunetti and colleagues found that recently robbed bank clerks with PTSD, relative to those without PTSD, showed increased activation in dorsal mPFC to single-episode conditioning of emotionally negative pictures to an aversive noise .
In addition, Milad and colleagues used fMRI and a 2-day fear conditioning and extinction paradigm to study individuals with PTSD and trauma-exposed healthy control participants [4, 83]. During late conditioning and early extinction, when the CS still signaled danger, the PTSD group showed increased dACC activation, relative to the control group. During late extinction, when the CS should have no longer signaled danger, the PTSD group showed decreased vmPFC activation, relative to the control group. During early extinction recall, the PTSD group showed vmPFC hypoactivation and dACC hyperactivation. This implies decreased inhibitory influence on fear responses from the vmPFC and increased excitatory influence on fear responses from dACC. See Fig. 10.3 for a functional image of early extinction recall. Using the same paradigm, Garfinkel and colleagues found that during both next-day recall in a never-conditioned (safety) context and during next-day renewal in the conditioned (danger) context, combat controls without PTSD showed increased vmPFC activation to the extinguished CS+ versus the CS−, relative to individuals with PTSD . Conducting a resting-state PET scan several days before utilizing that same 2-day fMRI conditioning and extinction paradigm, Marin and colleagues found that resting metabolism for glucose in the dACC positively correlated with PTSD symptoms, and also positively predicted dACC activation and negatively predicted vmPFC and hippocampal activation during extinction recall in PTSD . Another recent study also reported increased dACC activation during extinction recall in PTSD compared to controls .
These functional magnetic resonance images (fMRI) show brain activations during early fear extinction recall, compared to baseline: (a) dorsal anterior cingulate cortex (dACC) activation (yellow) and (b) ventral medial prefrontal cortex (vmPFC) deactivation (light blue). Graphs show brain activity in these two regions during early extinction recall relative to baseline activity of the two groups (error bars indicate standard error of the mean). Numbers following x, y, and z refer to Montreal Neurological Institute (MNI) coordinates. TENC trauma-exposed non-PTSD controls (Reprinted with permission from )
Findings of relatively diminished activation in the vmPFC have been reported in imaging studies involving the presentation of trauma-related material, such as script-driven imagery and trauma-related photos and words. In a symptom provocation study, Whalley and colleagues presented individualized trauma words versus either general trauma words or individualized non-trauma memory words to participants with PTSD as well as depressed and healthy controls. They found that when members of the PTSD group were presented with individualized trauma words, they exhibited increased BOLD response in dorsal mPFC regions including dACC . Bremner and colleagues reported that, compared to female sexual abuse survivors without PTSD, the rACC of female sexual abuse survivors with PTSD failed to activate in response to traumatic versus neutral audio scripts . Using a similar design, Shin and colleagues reported relatively reduced rACC activation in response to traumatic versus neutral scripts in PTSD . Lanius et al. found decreased medial frontal gyrus and rACC activation in participants with PTSD compared to trauma-exposed participants without PTSD in a traumatic versus baseline imagery condition . Using PET, Shin et al. found less activation in medial frontal gyrus during script-driven imagery in male and female Vietnam veterans with PTSD relative to those without PTSD . In a SPECT study, Lindauer et al. found less activation in the medial frontal gyrus in response to traumatic versus neutral scripts in trauma-exposed police officers with PTSD compared to those without PTSD . Using PET during traumatic/stressful scripts, Britton and colleagues found more deactivation in the rACC in combat veterans with PTSD compared with combat veterans without PTSD and combat-unexposed controls without PTSD . Lanius et al. found that participants with PTSD, relative to trauma-exposed participants without PTSD, had decreased ACC activation in response to personalized scripts of trauma-unrelated sad and anxiety-provoking events . In another fMRI study, Lanius and colleagues reported less rACC activation in response to trauma scripts versus baseline in individuals with PTSD compared to those with PTSD and comorbid major depression . In an early PET study, Shin and colleagues found lower regional cerebral blood flow in the rACC of combat veterans with PTSD compared to combat veterans without PTSD during combat versus neutral visual imagery . Bremner and colleagues found lower blood flow in vmPFC in response to combat-related versus neutral audio-visual stimuli in combat veterans with PTSD relative to combat veterans without PTSD . In an fMRI study, Yang and colleagues found relatively reduced rACC activation in a trauma-related versus neutral picture contrast in adolescents with PTSD, relative to trauma-exposed controls . Hou and colleagues showed neutral and trauma-related pictures to survivors of a mining accident with and without PTSD. They found relatively diminished rACC activation in the PTSD versus control group . In the combat-related word versus generally negative word contrast of an fMRI emotional Stroop task, Shin et al. found a lack of rACC activation and increased dACC activation in combat veterans with PTSD, relative to combat veterans without PTSD .
Studies using pictures of facial expressions frequently have shown relatively diminished vmPFC activation in individuals with PTSD. In an fMRI study, Shin et al. showed pictures of fearful and happy facial expressions to participants with PTSD and trauma-exposed control participants . In the fearful versus happy facial expression contrast, the PTSD group showed relatively reduced activations in the rACC, vmPFC, and also dorsal mPFC (see Fig. 10.4). Using an emotional faces interference paradigm, Offringa et al. found that individuals with PTSD showed less rACC activation than trauma-exposed controls . In an fMRI study comparing a PTSD group to healthy trauma-unexposed control group, Williams and colleagues found less activation in the mPFC in the fearful versus neutral expression contrast in the PTSD group . In a fearful versus neutral facial expression contrast, Kemp et al. found that participants with PTSD and comorbid depression had less activation in the mPFC than participants with PTSD only . In an fMRI study, Fonzo et al. presented a face-matching task to women with PTSD as well as trauma-unexposed women. They found increased dACC activation in response to male versus female faces in the PTSD group, and this dACC activation was positively correlated with hyperarousal symptoms . Crozier and colleagues studied maltreated youth with and without PTSD as well as non-maltreated youth controls and found some gender-specific effects. They found that, while looking at fearful versus calm faces, maltreated girls showed deactivation of rostral mPFC relative to control girls, maltreated boys, and control boys . In an fMRI study of subway fire survivors with PTSD, participants underwent a simple same-different judgment task in which task-irrelevant emotional and neutral facial expressions served as distractors . In the fearful versus neutral face contrast, participants with PTSD showed significantly decreased rACC activation compared to healthy, trauma-unexposed control participants. Dickie et al. found a negative correlation between vmPFC activation to forgotten faces and PTSD symptom severity in a subsequent memory paradigm . In a follow-up to that study 6–9 months later, 65% of those participants no longer met the criteria for PTSD diagnosis, and the emotional memory-related increase in subgenual ACC activation was correlated with symptom improvement . In an fMRI study of fearful versus neutral faces, Cisler and colleagues found that PTSD symptom severity in adolescent girls with histories of assault was associated with weakened functional connectivity between perigenual ACC and left amygdala .
The functional magnetic resonance image shows fearful versus happy activation in the rostral anterior cingulate (arrow) (z = 3.61; Montreal Neurological Institute [MNI] coordinates, −10, +38, −12) that was greater in the control group than in the PTSD group during the same study as Fig. 10.2. The bar graph shows fMRI activation in the rostral anterior cingulate (MNI coordinates, −10, +38, −12) in each condition (relative to fixation baseline) for each group. Error bars represent standard error of the mean (Reprinted with permission from )
Studies using trauma-unrelated emotional stimuli other than facial expressions have revealed similar results. Using PET, Bremner et al. studied the retrieval of deeply encoded negative versus neutral valence words in PTSD participants, relative to trauma-unexposed healthy control participants . The PTSD group showed relatively decreased regional cerebral blood flow in the subgenual anterior cingulate cortex and rACC . In another PET study, the same group administered an emotional Stroop task to women with and without PTSD and found that those with PTSD had reduced regional cerebral blood flow in the rACC relative to those without PTSD . Diener et al. used a trauma-unrelated cognitive stressor (mental arithmetic during white noise) and found that the stress-related pain threshold increases in PTSD participants correlated with increased activation in rACC . In a PET study of blood flow responses to aversive versus neutral pictures , Phan et al. found reduced regional cerebral blood flow in the vmPFC in combat veterans with PTSD, relative to trauma-unexposed controls but not relative to combat veterans without PTSD . Hayes and colleagues reported relatively increased dACC activation in combat veterans with severe PTSD symptoms compared to low-symptom veterans during trauma-unrelated emotional distraction .
Some studies utilizing neutral stimuli or resting-state imaging have found reduced vmPFC activation or increased dACC activation in PTSD. For example, Semple et al. used PET to compare regional cerebral blood flow in individuals with PTSD and a history of substance abuse to trauma-unexposed healthy control participants. They found that during a neutral valence auditory continuous performance task, the PTSD/substance abuse group had lower regional cerebral blood flow in medial frontal gyrus/ACC . Jovanovic and colleagues used a nonemotional go/no-go task to demonstrate decreased BOLD response in the vmPFC of traumatized women with PTSD relative to traumatized women without PTSD; this decrease was also associated with fear-potentiated startle in a separate fear conditioning task . In another fMRI study, Moores et al. found reduced ACC activation in PTSD participants relative to trauma-unexposed healthy control participants during a neutral-word verbal working memory updating task . Bryant et al. found increased dACC activation in a PTSD group compared to a trauma-unexposed control group during a neutral valence auditory oddball paradigm . Shin et al. reported increased dACC activation during a neutral valence interference task in PTSD participants, relative to trauma-exposed participants without PTSD . In a PET study, Shin et al. found increased resting metabolic rate for glucose in the dACC of combat veterans with PTSD as well as in their combat-unexposed monozygotic twins. In addition, dACC activation in the combat-unexposed twins was positively correlated with the combat-exposed co-twin’s PTSD symptom severity . These findings suggest that increased resting dACC activation could be a familial risk factor for PTSD. Osuch et al. used PET to acquire functional brain images of recent car accident survivors, most of whom did not go on to develop PTSD. They found that, during rest, the car accident survivors had greater regional cerebral blood flow in rACC than trauma-unexposed controls, suggesting a possible protective effect of rACC activation against the development of PTSD .
Not all studies have reported reduced vmPFC activation in PTSD, however [43, 52, 59, 65, 108, 109]. One potential reason for inconsistent findings is that the normally reduced vmPFC activation seen in PTSD may be limited to conscious processing of stimuli [71, 110].
Several studies have reported that activation in vmPFC regions negatively correlates with symptom severity in PTSD [36, 41, 50, 54, 76, 94, 97, 111] (but see [52, 112, 113]). Several treatment studies have reported a relationship between increased vmPFC activation and PTSD symptom improvement [69, 72, 73, 114, 115] (but see ).
The hippocampus plays a critical role in normal declarative, contextual, episodic, emotional, and spatial memory [116–119]. Abnormal hippocampal function in PTSD may be associated with deficits in any or all of these processes [56, 74, 99, 120–122]. Deficits in declarative memory may be an important link between hippocampal abnormalities and symptoms of PTSD, such as inability to recall important aspects of the traumatic event .
PTSD researchers have studied hippocampal function using experimental fear conditioning paradigms, trauma-related stimuli, emotional but trauma-unrelated stimuli, and nonemotional tasks. Across these paradigms, hippocampal findings have been mixed. Evidence from experimental fear conditioning paradigms suggests that individuals with PTSD may have an impaired ability to use both internal and external contextual cues to identify safe contexts and respond appropriately . For example, Milad and colleagues found that individuals with PTSD fail to extinguish fear responses in novel contexts and showed decreased hippocampal activation during extinction recall . Using the same conditioning and extinction paradigm, Garfinkel and colleagues found that during next-day fear renewal in the conditioned (danger) context, combat controls without PTSD showed increased left hippocampus activation to the CS− versus the extinguished CS+, relative to individuals with PTSD . In a separate study, Marin and colleagues found that in a resting-state PET scan conducted 4 days before conditioning and extinction, metabolic rate for glucose in dACC negatively predicted hippocampal activation during extinction recall in PTSD .
Studies involving trauma-related stimuli have shown evidence of relatively decreased hippocampal activation in PTSD. An early PET study by Bremner et al. found that, compared to trauma-exposed controls, women who were diagnosed with childhood sexual abuse-related PTSD showed greater deactivation in right hippocampus while listening to trauma-related scripts versus neutral scripts . In a later fMRI study, veterans of the wars in Iraq and Afghanistan showed decreased hippocampal activation while encoding trauma-related images compared to veterans without PTSD . Bremner and colleagues found evidence of relatively decreased hippocampal blood flow during the retrieval of trauma-related and trauma-unrelated emotionally valenced words in a PTSD group compared to a trauma-exposed control group . Improvement of PTSD symptoms has been associated with increases in hippocampal activity during script-driven imagery . However, one study using trauma reminders did not find decreased hippocampus activation in PTSD. Specifically, St. Jacques and colleagues showed increased hippocampal recruitment during cued retrieval of negative versus positive autobiographical memories in a PTSD group compared to a trauma-unexposed control group .
Some studies using trauma-unrelated emotional stimuli have found evidence of increased hippocampal activation in PTSD. Whalley and colleagues examined changes in brain activation during the recognition of neutral target images in emotional versus neutral contexts . They found that the PTSD group showed greater hippocampal recruitment for emotional versus neutral contexts than the control group. An fMRI study comparing maltreated youth with and without PTSD to non-maltreated youth controls found that while looking at fearful versus scrambled faces (matched with face stimuli for spatial frequency and luminance but with no recognizable content), maltreated boys showed activation of left hippocampus relative to control boys, maltreated girls, and control girls . In an emotional memory task, Brohawn and colleagues found that, compared to trauma-exposed controls, participants with PTSD showed an exaggerated hippocampal response to negative pictures that were subsequently remembered versus forgotten. They also found that the encoding of negative pictures was associated with greater hippocampal activation than the encoding of neutral pictures in PTSD participants .
Hippocampal function in PTSD has also been examined during nonemotional tasks and while participants are at rest. Bremner and colleagues found that the left hippocampus failed to activate in participants with PTSD, compared to trauma-exposed controls, during the encoding of neutral verbal passages . These differences were independent of differences in hippocampal volume. Another study found a lack of hippocampus recruitment in participants with PTSD versus trauma-unexposed controls during a working memory updating task . Molina and colleagues found diminished hippocampal glucose metabolism at rest in PTSD participants compared to trauma-exposed controls, indicating a possible differential pattern of baseline activation in this region . Astur and colleagues found a negative correlation between hippocampus activation and PTSD symptom severity during a virtual spatial navigation task . On the other hand, increased activation of the left hippocampus has been reported in PTSD participants completing a declarative memory task [127, 128]. Individuals with PTSD have also exhibited increased hippocampal blood flow during the resting state . A few functional neuroimaging studies have shown evidence for both increased and decreased hippocampal activation in PTSD. In a PET study by Shin and colleagues, individuals with PTSD showed diminished regional cerebral blood flow in the hippocampus during word-stem completions for deeply versus shallowly encoded words. However, after collapsing across deeply and shallowly encoded conditions, the PTSD group showed greater regional cerebral blood flow in the hippocampus compared to controls . In addition, Werner and colleagues found that, during a face-occupation pairing task, participants with PTSD showed increased hippocampal activation during encoding but decreased parahippocampal activation during retrieval, as compared to trauma-unexposed controls .
There is some evidence that activation of the hippocampus relates to symptom severity. Osuch et al. found that in participants with chronic PTSD, hippocampal activation was positively correlated with symptom severity during a script-driven imagery task . Shin et al. reported that hippocampal activation collapsed across conditions was positively correlated with symptom severity in the PTSD group . In contrast, Dickie et al. found that improvement of PTSD symptoms was positively correlated with increases in hippocampal activity while viewing fearful versus neutral face s . In a functional connectivity analysis using fearful versus neutral faces, Cisler and colleagues found connectivity strength between left parahippocampus and vmPFC positively associated with severity of assault exposure in adolescent girls and connectivity strength between right parahippocampal gyrus and left middle frontal gyrus positively associated with PTSD symptom severity .