History of childhood maltreatment augments dorsolateral prefrontal processing of emotional valence in PTSD

Highlights

• CM history in PTSD influenced left dlPFC activation to facial emotions overlaid with emotion words.

• In PTSD with CM, left dlPFC activation was greater to the fear vs. happy word regardless of face.

• In PTSD with CM, left dlPFC activation to face valence was inversely related to re-experiencing.

Abstract

Posttraumatic stress disorder (PTSD) is characterized by conflicting findings of both increased and decreased amygdala and prefrontal reactivity to threat or trauma stimuli. Childhood maltreatment (CM), a potent risk factor for PTSD, exerts long-lasting influences on threat processing and prefrontal-amygdala function. This suggests that CM history may influence PTSD neural phenotypes related to threat processing. Here, we adapt a well-characterized emotional conflict paradigm to investigate CM effects on both emotional conflict and emotional valence processing within PTSD stratified by task relevance. Forty-two individuals with PTSD (22 reporting extensive CM history (PTSD-CM)) and 20 trauma-exposed healthy controls (TEHCs) underwent functional magnetic resonance imaging while identifying affect of emotional faces (fear and happy) overlaid with a goal-irrelevant emotional distractor word (“FEAR” or “HAPPY”). We examined effects of CM on conflict, conflict adaptation, valence-related activation (fear vs. happy) for goal-relevant (face) and goal-irrelevant stimuli (word), and valence effects in interaction with goal-relevancy (face vs. word). Though no activation differences between groups were observed for conflict contrasts nor for valence effects in the amygdala, CM status interacted with valence processing differences as a function of goal relevance in the left dorsolateral prefrontal cortex (dlPFC). Here, PTSD-CM displayed greater activation relative to PTSD to negative valence when stimuli were goal-irrelevant. CM history also moderated relationships between activation abnormalities and PTSD re-experiencing symptoms. These findings provide initial evidence that CM history augments dorsolateral prefrontal bias to implicitly processed stimulus valence in PTSD.

Introduction

Increased attention to negative valence emotional stimuli signaling a potential threat, i.e. a threat-related attention bias, is a consistently demonstrated characteristic of anxiety disorders (Bar-Haim et al., 2007). However, research in posttraumatic stress disorder (PTSD) has demonstrated inconsistent results characterized by attention bias both towards (Bar-Haim et al., 2007) and away from threat (Bar-Haim et al., 2010), consistent with an exaggerated attention bias variability (Naim et al., 2015). PTSD is also characterized by conflicting findings with respect to amygdala activation, wherein both increased and decreased activation to threat and trauma-related stimuli has been observed in patients (Etkin and Wager, 2007). In combination with the established role of the amygdala in threat detection and responsivity (Costafreda et al., 2008), this variability in amygdala reactivity in PTSD suggests that neural abnormalities in threat processing may be affected by attentional mechanisms governing task relevance and goal-oriented behavior.

A recent meta-analysis of amygdala activation to emotional stimuli demonstrated that amygdala reactivity is heightened during implicit/passive stimulus processing relative to explicit/active conditions, i.e. when processing of the emotional stimulus is incidental and not the focus of task-relevant goals and behavior (Costafreda et al., 2008). This difference in amygdala reactivity as a function of attentional processing may arise as a consequence of increased medial and lateral prefrontal engagement with greater depth of conscious processing, which serves to downregulate amygdala responses (Taylor et al., 2006). Thus, variability in prefrontal-amygdala interactions relating to processing depth of emotional stimuli could explain variability in observed PTSD neural abnormalities. Moreover, recent research suggests that different prefrontal regions modulate different aspects of bottom-up reactivity instantiated in the amygdala. Whereas the medial prefrontal cortex is involved with bottom-up appraisal of stimulus valence, the dorsolateral prefrontal cortex (dlPFC) serves to augment threat or valence-related attentional processes (Comte et al., 2016), which is consistent with the proposed role of the dlPFC in regulating amygdalar emotional reactivity (Delgado et al., 2008, Ray and Zald, 2012) and observations for indirect amygdalar-dlPFC connections (Eden et al., 2015).

The clinical syndrome of PTSD also often comes in the context of major stressors experienced prior to the onset of the disorder-inducing traumatic event. Childhood maltreatment, a particularly damaging form of early life stress, has been described as a robust risk factor for PTSD (Zlotnick et al., 2008) as well as for the adult onset of mood and anxiety disorders (Green et al., 2010). Moreover, early life stress has been shown to impact amygdala-prefrontal structure and function (De Bellis and Keshavan, 2003, Fonzo et al., 2013, Gatt et al., 2010, Miller et al., 2015, Taylor et al., 2006, Tottenham et al., 2011, Tottenham et al., 2010), suggesting that childhood maltreatment may serve as an important determinant of amygdala-prefrontal responses to threat-conveying stimuli within PTSD. Other recent studies have found that altered amygdala and prefrontal responses as a function of maltreatment history can be independent of diagnostic status, i.e. uniform across psychiatric and healthy populations (Grant et al., 2011, van Harmelen et al., 2014, van Harmelen et al., 2013), suggesting that maltreatment effects may be instantiated early in life, persist into adulthood, and augment the expression of neural abnormalities within the PTSD diagnosis.

To address these questions, we re-conceptualized a well-characterized emotional conflict paradigm (Etkin et al., 2006) as a probe of emotional valence processing in interaction with goal relevance (Fig. 1) to parse heterogeneity in PTSD as a function of maltreatment history. Hereafter in this manuscript, the term “valence” refers to the difference between fear and happy stimuli (e.g., word valence refers to the difference between the fear and happy words). More specifically, beyond examining CM effects on conflict detection and adaptation, we were interested in how neural valence processing varied as a function of whether valence was perceived as a distractor (emotion word) or as the target of attention (emotional facial expression). By investigating the paradigm as a function of face and word valence in addition to congruence, we are able to address questions regarding how maltreatment history impacts emotional valence neural processing within PTSD as a function of attentional focus towards emotional stimuli in the service of a behavioral goal, i.e. task relevance. Such an approach may prove useful in addressing existing questions regarding the interaction of goal-relevance with valence processing in PTSD and the neural basis of variability in attention biases to threat.

We tested conflict and valence-processing effects in a group of trauma-exposed healthy controls (TEHCs) as well as a large sample of PTSD patients, of which we separated those self-reporting a moderate-to-severe history of childhood maltreatment (PTSD-CM). This approach enables inference both on the impact of maltreatment history on neural function within the PTSD diagnosis (PTSD-CM vs. PTSD) as well as the effects of diagnosis only on valence processing (PTSD vs. TEHCs). Given compelling cross-species evidence for early life stress promoting increased amygdala reactivity to threat that interferes with goal-oriented behavior (Malter Cohen et al., 2013), we predicted that PTSD-CM relative to PTSD patients would display increased amygdala activation to negative valence in the context of implicit valence processing (emotional word) irrespective of target valence, i.e. a main effect of word valence. Additionally, consistent with the role of the dlPFC in augmenting attention processes in the context of valence-related amygdala reactivity (Comte et al., 2016) and prior findings for maltreatment effects on dlPFC function (Gatt et al., 2010, Marusak et al., 2015), we expected the PTSD-CM group relative to the PTSD group to display altered dlPFC valence-related activation in interaction with goal-relevance (interaction of valence and face/word presentation). Our overall aim is to shed light on substantial diagnostic heterogeneity in amygdala-prefrontal responses as a function of this potent disorder risk factor.