Childhood Maltreatment,Limbic Dysfunction,Resilience,and Psychiatric Symptoms

Neurological Outcomes of Childhood Maltreatment

Studies of the neurodevelopmental outcomes of childhood maltreatment have tended to focus on its impact on the limbic system (Teicher et al., 2003). The major structures of the limbic system consist of the hippocampus (involved in memory formation and retrieval), the hypothalamus (which regulates the autonomic nervous system and thereby the stress response), and the amygdala (involved in emotion processing and regulation). There is evidence that childhood maltreatment is associated with later limbic system dysfunction, particularly limbic system irritability (heightened electrical activity in limbic structures, overreactivity in limbic structures in response to stress, increased prevalence of symptoms suggestive of temporal lobe epilepsy; Teicher, 2000). ¹

In studies using a self-report measure of symptoms of limbic system irritability (e.g., somatic and perceptual distortions, brief hallucinatory events, motor automatisms, and dissociative symptoms), Teicher, Glod, Surrey, and Swett (1993) found positive associations between childhood maltreatment and symptoms of limbic abnormality, and Teicher et al. (2006) found that adults with a history of childhood emotional or physical abuse reported greater limbic system irritability than those with no abuse; moreover, adults with combined childhood emotional and physical abuses reported higher rates of limbic system symptoms than individuals reporting a single type of abuse. In an exploratory diffusion tensor imaging study of young adults, Choi, Jeong, Rohan, Polcari, and Teicher (2009) found that reports of parental verbal abuse during childhood were associated with reduced neural integrity in pathways also characterized by increased limbic system irritability.

There has also been research regarding the possibility that neurobiological outcomes mediate the relationship between childhood maltreatment and later vulnerability to traumatic stress and other psychiatric disorders, particularly as related to aggression (e.g., Lee & Hoaken, 2007). There is some evidence that repeated exposure to childhood trauma is associated with kindling, a process of oversensitizing the limbic system (van der Kolk & Saporta, 1991), and is implicated in lowering the threshold for psychological vulnerability to subsequent traumas (van der Kolk, McFarlane, & Weisaeth, 2006). There is also evidence that limbic system dysfunction is linked to various consequences of childhood maltreatment pertaining to emotional expression and cognition (Grassi-Oliveira, Ashy, & Stein, 2008). For example, limbic system irritability has been found to be positively correlated with psychiatric symptoms of anxiety, depression, and anger-hostility (Anderson, Teicher, Polcari, & Renshaw, 2002).

Affective Responses to Childhood Maltreatment

The immediate outcomes of childhood maltreatment include heightened feelings of shame, feelings that are associated with psychiatric symptoms. In one longitudinal study of children from abusive and nonabusive homes, Stuewig and McCloskey (2005) found that harsh parenting during childhood was associated with shame-proneness during adolescence and that, over time, shame-proneness was associated with depression. In a study looking at childhood maltreatment and internalized shame in adults diagnosed with bipolar disorder, Fowke, Ross, and Ashcroft (2012) found that significantly more participants in their bipolar disorder sample than in their nonpsychiatric sample reported childhood physical abuse, emotional abuse, and neglect; levels of internalized shame were also significantly higher in the bipolar group. Moreover, in the bipolar group, there were significant positive correlations between all of the childhood maltreatment scores and internalized shame. There appears to have been little or no research on the relationship between childhood maltreatment and guilt, and little or no attention to the possibility of gender differences in any such relationship.

Kim, Thibodeau, and Jorgensen (2001) have emphasized the importance of differentiating between shame and guilt, both of which are considered to involve self-evaluative processes related to important behavioral standards and having important implications for “social survival” within particular cultural contexts. They describe shame as involving an inner-directed negative view of the entire self and guilt as involving an outward-directed negative view of specific behaviors in regard to others. In their comprehensive review of relevant studies, they found that shame was much more strongly related to depression than guilt. In their discussion of their findings, they suggest that more research is needed to determine the extent to which sex moderates associations of shame and guilt with depressive symptoms. In the current study, a somewhat different approach was taken, examining the extent to which shame and guilt mediated the relationship between childhood maltreatment and symptomatology.

Evidence has started to emerge in support of a developmental, neurological basis for those two “moral” emotions, guilt and shame. Neuroimaging studies have identified unique activation patterns for shame and guilt that differ from neutral emotion conditions (Michl et al., 2014). Both emotions have been found to be associated with activation in the temporal lobe; however, shame is associated with greater activation in the medial and anterior frontal lobe, whereas guilt is associated with activation in the amygdala and insula.

Studies involving lesions and other forms of damage to the prefrontal cortex have provided further evidence for a neurological basis for these emotions. For example, patients with lesions in the medial prefrontal cortex show diminished emotional experience, including lowered self-conscious emotions that serve a social function, such as guilt, shame, and embarrassment (e.g., Philippi & Koenigs, 2014). In a study conducted by Koenigs and colleagues (2007), patients with focal bilateral damage to the ventromedial prefrontal cortex showed severely diminished embarrassment and guilt, while comparison subjects had intact emotional processing. Moreover, fronto-temporal dementia, which causes degeneration of the medial prefrontal cortex, is tied to a decline in self-conscious emotions that are closely linked to shame and guilt (Lewis, 2010; Michl et al., 2014) such as embarrassment (Philippi & Koenigs, 2014).

Resilience

Not all individuals who experience childhood maltreatment develop psychiatric symptoms; some of these victims appear to be resilient. Resilience refers to individual characteristics (e.g., positive self-esteem and self-control) and external/environmental factors (e.g., family and educational support) that allow individuals to adapt successfully to stressful situations (Luthar, Cicchetti, & Becker, 2000). In a study of 132 college students, Campbell-Sills, Cohan, and Stein (2006) identified trait resilience as a protective factor in the relationship between childhood emotional neglect and present psychiatric symptoms. In a study of 70 PTSD patients, Daniels et al. (2012) found trait resilience to be a significant negative predictor of PTSD symptoms in patients who had reported severe childhood trauma. In a sample of adults who had experienced childhood sexual abuse, Ginzburg et al. (2006) found psychiatric symptoms to be positively correlated with guilt and shame, and negatively correlated with traits of resilience.

Research exploring neurobiological contributors to resilience has indicated the importance of genetic and potential epigenetic mechanisms in response to stress (Feder, Nestler, & Charney, 2009; Southwick & Charney, 2012). Genetic polymorphisms affect limbic reactivity and prefrontal-limbic connectivity, influencing responses to traumatic events (Feder et al., 2009). Brain-imaging studies have confirmed that the prefrontal region is potentially influential in reducing activity in the amygdalae (Hariri, Bookheimer, & Mazziotta, 2000). Animal studies have also shown the importance of this region (Southwick & Charney, 2012). Early experiences coping with stress increase neuroplasticity in the prefrontal cortex, which buffers against the negative effects of future uncontrollable stress (Southwick & Charney, 2012). In addition to prefrontal cortical functioning and capacity to regulate limbic reactivity to stress, moderators of resilience include effective regulation of the hypothalamic-pituitary-adrenal (HPA) axis (which influences hormone secretion in reaction to stress) and the sympathetic nervous system (SNS; which influences the body’s “fight or flight” reaction in response to stress) (Southwick & Charney, 2012).

Russo, Murrough, Han, Charney, and Nestler (2012) have noted that the “study of human resilience is still a mostly phenomenological literature that has only begun to characterize biological factors in resilient individuals that are associated with more successful coping responses” (p. 1475). Based on their review of the scanty relevant literature, they report that developmentally, childhood abuse affects the same regions of the brain shown to contribute to resilience, such as the limbic system. They also found that neurological and genetic resilience has a buffering effect against depression. Thus, resilience may involve psychoneurological processes. The research they summarize on the impact of stress on the limbic system and related processes, taken together with Teicher and colleagues’ (2006) conclusions about the neurological impact of childhood abuse, suggests that childhood abuse may affect the same regions of the brain shown to contribute to resilience, and that any relationship between childhood maltreatment and resilience may have a neurological basis. Previous literature has incorporated resilience, as well as shame and guilt, into studies investigating neurological underpinnings of stress outcomes and psychiatric symptoms; however, the relationship between shame and guilt and resilience has been neglected, particularly from a developmental psychoneurological framework that attends to the issue of potential gender differences in pathways.

Figure 1 provides our proposed model of the relationships among the variables of interest in the current study: childhood maltreatment as a predictor, adult psychiatric symptoms as an outcome, shame and guilt as affective mediators of the maltreatment–psychiatric outcomes relationship, limbic system dysfunction as a neurological mediator of the maltreatment–psychiatric outcomes relationship, and resilience as a moderator of the maltreatment–psychiatric outcomes relationship. This conceptual model of psychiatric outcomes was tested in the current study. Specifically, the present study investigated a conceptual model of psychiatric outcomes in nonclinical samples of men and women reporting childhood psychological or physical maltreatment.

OPEN IN VIEWER

Figure 1.

Childhood maltreatment model of psychiatric outcomes.

The investigation was guided by the following hypotheses:

Method

The current project was a secondary analysis of an available dataset. Protocols of the project were reviewed and approved by the university’s Institutional Review Board. The research was conducted in compliance with all American Psychological Association guidelines and state and federal laws. All surveys were collected anonymously, and all participants included in the dataset had given permission for their anonymous responses to be analyzed in future studies.

Results

Descriptive Statistics

Table 1 provides means and standard deviations for all study variables, separately by gender. Participants reported incidences of parental psychological and physical aggression comparable with samples in previous studies (Teicher & Parigger, 2015). Furthermore, their limbic ratings were comparable with those in previous nonclinical samples (Teicher et al., 1993), as were their reported levels of anxiety, depression, somatization, and hostility (Kellner, 1987).

OPEN IN VIEWER

Table 1.

Descriptive Statistics for Major Study Variables Separately by Gender, with t Test Results.

Major Variables	Females		Males		t
	M	SD	M	SD
Maternal psychological aggression a	7.39	6.49	6.74	6.46	0.92
Paternal psychological aggression a	6.31	5.42	7.66	6.22	−2.09
Maternal physical aggression a	5.40	8.69	5.81	9.02	−0.42
Paternal physical aggression a	3.49	5.97	5.59	8.58	−2.65**
Limbic dysfunction b	53.74	11.83	52.25	14.88	1.04
Trait resilience c	55.83	9.81	54.36	12.74	1.22
Shame index d	2.70	0.63	2.43	0.64	4.03**
Guilt index d	2.68	0.68	2.51	0.66	2.36*
Total psychiatric symptoms e	87.96	16.41	84.24	18.65	1.97*
Anxiety e	22.57	4.49	21.32	4.96	2.45*
Depression e	23.56	5.32	22.32	5.62	2.10*
Hostility e	21.26	4.47	21.20	5.48	0.11
Somatization e	20.56	4.39	19.51	4.47	2.22*

Note. CTSPC = Conflict Tactics Scales: Parent–Child; LSCL = Limbic System Function Questionnaire; RS = Resilience Scale; PFQ = Personal Feelings Questionnaire; SQ = Symptom Questionnaire.

a

CTSPC subscale items range from 0 (never happened) to 6 (happened more than 20 times). Items were summed for a highest possible score of 66 per subscale.

b

LSCL Items range from 0 (low dysfunction) to 4 (high dysfunction). Items were summed for a highest possible score of 132.

c

RS Items range from 0 (low resilience) to 7 (high resilience). Items were summed for a highest possible score of 70.

d

PFQ Items range from 1 (never experience) to 4 (experience almost continuously). Items were summed for a highest possible score of 88.

e

SQ The Symptom Questionnaire has four 17-item scales (Anxiety, Depression, Somatic Symptoms, and Hostility) with each item scoring 1 for false and 2 for true. Highest possible score on the Total Symptom scale is 184; highest possible score on each of the four symptom subscales is 34.

*

p ≤ .05. **p ≤ .01.

As shown in Table 1, t tests revealed significant gender differences on several study variables. Specifically, male participants reported a higher incidence of paternal psychological and physical aggression than females. By contrast, female participants reported higher levels of guilt and shame, as well as higher levels of anxiety, depression, somatization, and total psychiatric symptoms than males.

Correlations Between Different Forms of Parental Aggression and Predicted Outcomes, Mediators, and Moderator

To test Hypothesis 1 (that exposure to childhood maltreatment would be negatively correlated with resilience and positively correlated with limbic system dysfunction, guilt, shame, and psychiatric symptoms), several correlations were performed, separately by gender. As can be seen in Table 2, patterns of correlation varied somewhat by subject gender, type of aggression, and source of aggression.

OPEN IN VIEWER

Table 2.

Correlations Between Childhood Experiences of Aggression and Major Study Variables for Males and Females.

Variables	LD	RES	Shame	Guilt	TSYM	ANX	DEP	HOS	SOM
Females
Maternal psychological	.31**	−.14*	.28**	.27**	.33**	.27**	.32**	.31**	.25**
Paternal psychological	.25**	−.22**	.19*	.19*	.29**	.24**	.28**	.25**	.23**
Maternal physical	.24**	−.12	.1	.17*	.24**	.17*	.24**	.18*	.22**
Paternal physical	.17*	−.21**	.03	.06	.22**	.19*	.18*	.22**	.18*
Males
Maternal psychological	.24**	−.08	.18*	.06	.15*	.1	.15*	.16*	.12
Paternal psychological	.19*	−.07	.12	.06	.15*	.16*	.16*	.1	.08
Maternal physical	.38**	−.19*	.21*	.11	.15*	.09	.15*	.08	.23*
Paternal physical	.30**	−.20*	.17*	.11	.19*	.18*	.15*	.13	.24*

Note. LD = limbic dysfunction; RES = resilience; TSYM = total psychiatric symptoms; ANX = anxiety; DEP = depression; HOS = hostility; SOM = somatic symptoms.

*

p ≤ .05. **p ≤ .01.

Parental aggression and later symptoms

In females, both maternal and paternal psychological and physical aggression were significantly positively correlated with all psychological symptoms and limbic system dysfunction. In males, paternal and maternal physical aggression were significantly positively correlated with somatization but uncorrelated with other psychiatric symptoms, and there were no significant correlations between psychological aggression from either parent and any psychiatric symptom or limbic system dysfunction.

Parental aggression and resilience

In females, paternal and maternal psychological aggression, as well as paternal physical aggression, were significantly negatively correlated with resilience, whereas in males, only paternal and maternal physical aggression showed significant negative correlations with resilience. In females, paternal and maternal psychological aggression were significantly positively correlated with shame, and paternal and maternal psychological aggression, as well as maternal physical aggression, were significantly positively correlated with guilt. By contrast, in males, paternal and maternal physical aggression, as well as maternal psychological aggression, were significantly positively correlated with shame, but no forms of parental aggression were significantly correlated with guilt.

Correlates of Limbic System Dysfunction, Separately by Gender

Findings regarding Hypothesis 2 (that limbic dysfunction would be negatively correlated with resilience and positively correlated with guilt, shame, and psychiatric symptoms) can be found in Table 3. As can be seen, for women, limbic system dysfunction was significantly negatively correlated with resilience and positively correlated with shame, guilt, and all psychiatric symptoms. In males, limbic dysfunction was significantly positively correlated with shame, guilt, and all psychiatric symptoms but was not associated with resilience.

OPEN IN VIEWER

Table 3.

Correlations Between Limbic Dysfunction and Other Major Study Variables.

	LD	RES	Shame	Guilt	TSYM	ANX	DEP	HOS	SOM
Females
LD	—	−.16*	.41**	.30**	.53**	.48**	.43**	.38**	.54**
RES		—	−.27**	−.15*	−.36**	−.30**	−.44**	−.28**	−.22**
Shame			—	.64**	.46**	.46**	.44**	.35**	.37**
Guilt				—	.36**	.37**	.38**	.30**	.20**
TSYM					—	.92**	.91**	.85**	.83**
ANX						—	.79**	.72**	.71**
DEP							—	.70**	.66**
HOS								—	.58**
SOM									—
Males
LD	—	.02	.46**	.37**	.48**	.43**	.43**	.36**	.45**
RES		—	−.18*	−.10	16*	−.11	−.25**	−.09	−.25**
Shame			—	.64**	.27**	.22*	.32**	.19*	.23*
Guilt				—	.26**	.23**	.32**	.15*	.21*
TSYM					—	.92**	.91**	.89**	.86**
ANX						—	.79**	.76**	.72**
DEP							—	.74**	.71**
HOS								—	.65**
SOM									—

Note. LD = limbic dysfunction; RES = resilience; TSYM = total psychiatric symptoms; ANX = anxiety; DEP = depression; HOS = hostility; SOM = somatic symptoms.

*

p ≤ .05. **p ≤ .01.

Regression Analyses

The results of regression analyses run to test Hypothesis 3 (that the relationship between childhood maltreatment and adult psychiatric symptoms would be mediated by limbic system dysfunction, guilt, and shame, and moderated by trait resilience) can be seen in Tables 4 to 6. The analyses are separated by type of parental aggression (mother psychological, mother physical, father psychological, and father physical) and also by gender. In each analysis, total scores for psychiatric symptoms were regressed first on one specific form of parental aggression; resilience scores were added at Step 2; the resilience by parental aggression interaction term was added at Step 3; guilt and shame were added at Step 4; and limbic system dysfunction was added at Step 5.

OPEN IN VIEWER

Table 4.

Linear Multiple Regression Model for the Prediction of Total Psychiatric Symptoms by Maternal Psychological Aggression Mediators and Moderators.

Variable	β	p	R 2	ΔF	p
Females
Step 1			.11	23.33	.001
MPA	.33	.001
Step 2			.21	23.39	.001
Trait resilience	−.32	.001
Step 3			.21	0.58	.45
MPA × Resilience	−.31	.45
Step 4			.33	16.92	.001
Guilt index	.06	.43
Shame index	.34	.001
Step 5			.43	31.58	.001
Limbic dysfunction	.35	.001
Males
Step 1			.02	2.56	.11
MPA	.13	.11
Step 2			.19	30.98	.001
Trait resilience	−.42	.001
Step 3			.20	1.27	.26
MPA × Resilience	−.43	.26
Step 4			.29	8.78	.001
Guilt index	.24	.008
Shame index	.10	.28
Step 5			.32	6.45	.01

Note. MPA = maternal psychological aggression.

OPEN IN VIEWER

Table 5.

Linear Multiple Regression Model for the Prediction of Total Psychiatric Symptoms by Maternal Physical Aggression Mediators and Moderators.

Variable	β	p	R 2	ΔF	p
Females
Step 1			.06	11.01	.001
MPHA	.24	.001
Step 2			.17	24.82	.001
Trait resilience	−.34	.001
Step 3			.17	0.93	.34
MPHA × Resilience	−.33	.34
Step 4			.33	21.14	.001
Guilt index	.06	.44
Shame index	.37	.001
Step 5			.42	31.42	.001
Limbic dysfunction	.35	.001
Males
Step 1			.02	3.27	.07
MPHA	.15	.07
Step 2			.19	29.67	.001
Trait resilience	−.42	.001
Step 3			.20	2.05	.16
MPHA × Resilience	−.58	.16
Step 4			.29	8.92	.001
Guilt index	.23	.01
Shame index	.11	.24
Step 5			.33	7.15	.008
Limbic dysfunction	.23	.008

Note. MPHA = maternal physical aggression.

OPEN IN VIEWER

Table 6.

Linear Multiple Regression Model for the Prediction of Total Psychiatric Symptoms by Paternal Psychological Aggression Mediators and Moderators.

Variable	β	p	R 2	ΔF	p
Females
Step 1			.09	17.46	.001
PPA	.29	.001
Step 2			.18	20.59	.001
Trait resilience	−.31	.001
Step 3			.18	0.02	.90
PPA × Resilience	−.04	.90
Step 4			.33	20.26	.001
Guilt index	.08	.33
Shame index	.36	.001
Step 5			.43	32.76	.001
Limbic dysfunction	.36	.001
Males
Step 1			.02	3.10	.08
PPA	.15	.08
Step 2			.20	31.80	.001
Trait resilience	−.42	.001
Step 3			.20	0.02	.88
PPA × Resilience	.05	.88
Step 4			.30	9.83	.001
Guilt index	.25	.007
Shame index	.11	.88
Step 5			.33	6.41	.01
Limbic dysfunction	.20	.01

Note. PPA = paternal psychological aggression.

As shown in Step 1 of Tables 4 to 6, each type of parental maltreatment contributed significantly to psychiatric symptoms in females; however, in males, only paternal physical aggression contributed significantly to psychiatric symptoms, with maternal physical aggression and paternal psychological aggression making only marginally significant contributions. In both males and females, trait resilience added significantly to variance in psychiatric symptoms at Step 2 of each equation. The resilience by parental aggression interaction term, added at Step 3, did not add significantly to variance explained in any of the equations. At Step 4, in females, shame and guilt added significantly to variance in symptomatology, but the Betas revealed that it was only shame that added significantly to psychiatric outcomes in all four parental aggression analyses. In males, it was guilt that added significantly to variance in psychiatric symptoms in each equation. At the final step, limbic system symptoms added significantly to prediction of psychiatric symptoms in both males and females in all four equations.

Discussion

The goal of this developmentally based psychoneurological study was to examine the relationships among childhood maltreatment, limbic dysfunction, trait resilience, shame and guilt, and psychiatric symptoms, and to determine whether those relationships varied by gender. It was predicted that shame and guilt and limbic dysfunction would mediate, and resilience would moderate the relationships between childhood maltreatment and later psychiatric symptoms. The findings of this investigation indicated that maternal psychological maltreatment and paternal physical maltreatment were predictive of total psychiatric symptomatology in adulthood, with shame contributing to symptomatology in women and guilt in men. Limbic dysfunction was found to play a significant role in mediating the relationship between childhood maltreatment and psychiatric symptoms, whereas trait resilience was found to play a moderating, protective role in predicting psychiatric symptomatology.

In our preliminary analyses, female participants scored higher than males in anxiety, depression, somatization, guilt, shame, and total psychiatric symptoms. These findings are consistent with other studies showing higher scores in females than in males in somatization (e.g., Ashy, 2003; Teicher et al., 1993), anxiety and depression (e.g., Altemus, 2006; Piccinelli & Wilkinson, 2000), shame and guilt (e.g., Else-Quest, Higgins, Allison, & Morton, 2012; Hoglund & Nicholas, 1995), and a range of internalizing symptoms (e.g., Kramer, Krueger, & Hicks, 2008). Males reported more paternal psychological and physical aggression than females, which is consistent with previous research findings (Ashy, 2003; Ashy & Malley-Morrison, 2000).

The first hypothesis in our developmentally based psychoneurological model, that exposure to childhood maltreatment would be negatively correlated with resilience and positively correlated with limbic dysfunction, guilt, shame, and psychiatric symptoms, was partially supported in both males and females. These correlations are consistent with the results of previous research on childhood maltreatment and resilience in adulthood (e.g., Luthar et al., 2000), as well as research on childhood maltreatment and limbic dysfunction in adulthood (e.g., Anderson et al., 2002; Ashy, 2003; Teicher et al., 2006). The positive association between childhood maltreatment and guilt and shame is also consistent with research showing high levels of negative self-perception and neuroticism in individuals with traumatic experiences (Campbell-Sills et al., 2006; Kling, Ryff, Love, & Essex, 2003).

From a psychoneurological perspective, a possible basis for the link between childhood maltreatment and limbic dysfunction can be found at the biological level, as enhanced activity of stress response systems may cause the brain to follow alternative developmental pathways to manage an environment of heightened stressors (Teicher et al., 2006). Moreover, the negative associations found between maltreatment and resilience can be considered in light of the altered environment experienced by maltreated children; what may constitute resilience in an abnormally stressful environment may not characterize resilience in the normal environment as measured by the Reslience Scale (Wagnild & Young, 1993). Furthermore, the contribution of childhood maltreatment to feelings of guilt and shame may be explained through limbic dysfunction or through behavioral and cognitive mechanisms, such as a negative self-concept and behavioral inhibition, or through internalization of blame for the experienced abuse.

There were several gender-specific findings related to the first hypothesis. Exposure to maltreatment was associated with all psychiatric symptoms in females; however, in males, exposure to maltreatment was associated only with somatization. One possible explanation for this finding is that the connection between abuse and psychiatric symptoms is direct in females but may be indirect in males through other variables such as alcohol and substance abuse (Kessler et al., 1994).

The second hypothesis, that limbic dysfunction would be negatively correlated with resilience and positively correlated with guilt, shame, and psychiatric symptoms, was partially supported, with some variations observed for males and females. The negative association between limbic dysfunction and resilience was supported in females but not in males; this association was not robust, so further study would be appropriate. Resilience has been shown to be a moderator of psychiatric symptoms and PTSD in both males and females with histories of childhood trauma (Campbell-Sills et al., 2006; Daniels et al., 2012), and the amygdala has been implicated in the ability to regulate emotions in response to fearful stimuli (Amorapanth, LeDoux, & Nader, 2000). As the limbic system is involved in emotional regulation, and resilience is conceptualized as the ability to regulate emotions, the limbic dysfunctions might be affecting the core of resilience.

In support of our developmentally based psychoneurological model, the positive relationships among self-reported limbic dysfunction, guilt, and shame are consistent with functional magnetic resonance imaging (fMRI) studies showing activation of the amygdala in response to hostile stimuli (Aleman & Swart, 2008; Sambataro et al., 2006). The association between limbic dysfunction and psychiatric symptoms is consistent with findings from Teicher et al. (2006) indicating comparatively large effect sizes for anxiety, depression, and hostility in individuals exposed to childhood traumas compared with individuals not reporting such exposure.

The third hypothesis, which was the most direct test of our integrated developmentally based psychoneurological model, predicted that resilience, guilt, shame, and limbic dysfunction are intervening variables between childhood maltreatment and adult psychiatric symptoms; this hypothesis was supported by the results of the linear regression analyses. In females, maternal and paternal psychological aggression, shame, limbic system dysfunction, and resilience contributed significantly to the variance in psychiatric symptoms, with shame as a major predictor of symptoms and a mediator of the childhood aggression–symptoms relationship. In males, paternal physical aggression, guilt, limbic system dysfunction, and resilience contributed significantly to the variance in psychiatric symptoms, with guilt as a major predictor of symptoms and a mediator of the maltreatment–symptoms relationship. Based on the regression analysis, limbic system dysfunction, guilt, and shame can be seen as risk factors for psychiatric symptoms (with some variation by gender), while, consistent with previous research (Campbell-Sills et al., 2006), resilience can be seen as a protective factor.

Of particular interest are the findings concerning shame and guilt, which provide some support for a partially gender-specific pathway from childhood maltreatment to psychiatric symptomatology. These findings make sense in regard to the literature, indicating that abused females tend to express their distress through internalizing symptoms and abused males tend to express their distress more through externalizing symptoms. The findings also have clinical implications, suggesting that child and adolescent clinicians might do well to address symptoms of extreme shame or guilt in an effort to subvert the development of psychiatric symptoms, particularly in traumatized youth. Moreover, Sable, Danis, Mauzy, and Gallagher (2006) found that shame and guilt were barriers to the reporting of sexual assault in college students; shame and guilt may play similar roles in maltreated children.

The nature of the sample recruited for this study has both negative and positive implications. On one hand, finding support for hypotheses derived from our developmentally based psychoneurological model in a relatively advantaged nonclinical sample testifies to the power of these relationships in a sample with access to environmental protective factors that are not typically available to severely maltreated children (Garmezy, 1991). On the other hand, inclusion of a higher risk sample would have provided additional confirmation of the generalizability of the proposed pathways between various levels of maltreatment and developmentally based psychoneurological outcomes.

Like most studies, the present study has limitations that point to the need for future research. For example, our version of the widely used childhood maltreatment measure included only physical and psychological maltreatment from parents to child. Other forms of childhood maltreatment, including neglect, sexual trauma, witnessed violence between parents, sibling and peer aggressions, and abuses by authority figures outside the home, are potential areas for future research. Future tests of the model might consider the extent to which social class and other demographic variables moderate the relationships described in our current model and the extent to which gender differences vary by culture.

The risk of recall bias in the use of CTS might be considered another limitation, and indeed has been identified as a limitation in at least one critical review of literature analyzing studies of childhood maltreatment as a predictor of intimate partner violence (Gil-Gonzalez, Vives-Cases, Ruiz, Carrasco-Portino, & Alvarez-Dardet, 2008). However, the CTS has been administered numerous times in studies all around the world and has gained consistent support for its validity. Given that the findings of this study are largely consistent with the predictions of the conceptual model, this suggests that recall bias is not a unique problem for this study and does not call the findings of this study into question.

Another limitation of the project was its technical inability to specify and assess dysfunctions in particular limbic areas or other areas such as the prefrontal cortex. Childhood emotional maltreatment has been found to be linked to changes in the medial prefrontal cortex (mPFC) morphology. In addition, child emotional maltreatment has been linked to amygdala hyperactivity and increased risk for the development of psychopathology (van Harmelen et al., 2014). Future research should use imaging techniques and physiological instruments to confirm limbic dysfunction through direct assessments of the brain. Moreover, resilience is a relatively new area of research, and multilevel analysis may be needed to understand the complex phenomena that determine individual levels of resilience and their relevance to various psychiatric disorders.

Despite these limitations, these findings provide a basis for some optimism regarding the development of individuals who have been maltreated in childhood. Even if maltreatment has the power to affect neurological functioning, the development of resilience can help counter the long-term negative effects of maltreatment. Indeed, the findings of this study may have implications for health care professionals. The mediating role of resilience suggests that treatment of individuals who have experienced childhood traumas should focus on the psychological resources available to the individual in addition to his or her vulnerabilities.