Early Amygdala Activation and Later Ventromedial Prefrontal Cortex Activation During Anger Induction and Imagery

INTRODUCTION

Anger is a part of everyone’s life. Excessive feelings of anger, irritability and/or hostility, however, are associated with diminished quality of life, relationship difficulties, and adverse health effects such as increased risk for coronary artery disease [e.g., 1–4]. As many as 50% of outpatient psychiatric populations have reported moderate to severe levels of anger and aggression [5].

Anterior paralimbic and limbic brain regions are implicated in neurobiological studies of anger and impulsive aggression. Positron emission tomography (PET) studies have consistently shown increased regional cerebral blood flow in the ventromedial prefrontal cortex (VMPFC) during anger, imagined expression of anger, and exposure to anger in healthy controls [e.g., 6–9]. Individuals with a predisposition to anger and aggression (murderers, violent offenders, and people with antisocial personality disorder [APD]) exhibit decreased activation in the prefrontal cortex compared with healthy controls in PET and single-photon emission computed tomography (SPECT) studies [10–13]. A PET study by our group also found that patients with Major Depressive Disorder (MDD) and anger attacks demonstrate blunted VMPFC activation during anger induction compared to healthy controls and MDD patients without anger attacks [8].

Another brain region that has been implicated in anger and aggression is the amygdala. Among other functions, the amygdala appears to be involved in the initial assessment of threat and facilitation of fight or flight responses [14–17]. For example, following exposure to threatening visual stimuli, patients with APD show reduced potentiation of the startle reflex and impaired aversive conditioning [e.g., 18, 19] comparable to those observed in patients with amygdala lesions [20, 21]. Patients with APD and those with amygdala lesions also exhibit impairments in the processing of fearful facial expressions [e.g., 22–25]. A structural MRI study [26] of patients with temporal lobe epilepsy found that those with comorbid intermittent explosive disorder (IED) had substantially higher rates of amygdala atrophy or amygdala lesions than those without comorbid IED. Finally, a functional MRI study [27] using a memory task demonstrated that participants who scored higher on a scale of antisocial behavior demonstrated reduced amygdala activation while processing negatively valenced words compared with individuals who scored lower on the scale.

Although the amygdala has been consistently found activated in studies involving negative affective, particularly fearful, stimuli [28–31], there is a striking absence of findings involving amygdala activation during anger states in functional imaging studies using anger induction paradigms. One reason for this absence of amygdala activations in anger induction studies may lie in the nature of neural connectivity between the VMPFC and the amygdala, and the time course of amygdala activations. Specifically, there are known bidirectional connections between the prefrontal cortex (especially the medial prefrontal cortex) and the amygdala [32–34], and there is evidence in healthy controls that these two structures are mutually inhibitory, with increased activity in one structure inhibiting activity in the other structure [8, 35, 36]. fMRI studies investigating the temporal characteristics of the amygdala response to affective stimuli have shown that healthy samples exhibit rapid amygdala habituation to emotional stimuli [37–41].

Anger induction paradigms have primarily used PET to measure regional changes in cerebral blood flow associated with anger states. This imaging method is limited in temporal resolution. Changes in regional cerebral blood flow are averaged over longer periods of time (usually >60 seconds) to achieve acceptable signal to noise ratios. If anger induction is associated with an early initial, but not sustained, amygdala activation, it is possible that existing PET anger induction studies may have failed to capture this initial amygdala activation in response to induction of anger. The present study was designed to investigate this hypothesis. We conducted an anger induction experiment with healthy volunteers and assessed changes in VMPFC and amygdala activation using fMRI, which has a better temporal resolution than PET. Based on the above neuroimaging findings we hypothesized a negative correlation between amygdala and VMPFC activation during anger induction and imagery. Specifically, we hypothesized that the amygdala is involved in the early phases of anger experience which is later followed by increased VMPFC activation.

MATERIALS AND METHODS

fMRI data analyses

Data was pre-processed using SPM 5 software (Wellcome Department of Cognitive Neurology, London, UK). All scans showed acceptable levels of motion with less than 3 mm of movement in any direction. fMRI images were motion corrected and spatially normalized to the standardized space established by the Montreal Neurologic Institute (MNI; http://www.bic.mni.mcgill.ca). Images were then smoothed with a three-dimensional, 6 mm FWHM Gaussian filter. Structural images were co-registered with motion corrected functional images and normalized to MNI standardized space.

For each subject, we created contrast images representing each individual condition of interest (i.e., rest before anger, anger induction, anger imagery, rest after anger). For second level random effects analysis, we conducted a flexible factorial analysis using each individual subject’s contrast images. We then modeled the hypothesized time course of amygdala and VMPFC activation during anger induction and imagery (relative to the neutral conditions; Fig. 1). Specifically, we modeled an early increase of amygdala activation during anger induction followed by increased VMPFC activation during anger imagery.

OPEN IN VIEWER

Fig.1

We modeled the hypothesized time course of amygdala and ventromedial prefrontal cortex (VMPFC) activation during anger induction and imagery relative to the neutral conditions. Specifically, we modeled an early increase of amygdala activation during anger induction followed by increased VMPFC activation during anger imagery.

Region of interest (ROI) analyses were then completed for the amygdala (bilateral) as defined by anatomical masks from the Wake Forest University Pick Atlas [44]. To control for multiple statistical comparisons, we maintained a cluster-level false positive detection rate at p < 0.05 with a cluster (k) extent empirically determined by Monte Carlo simulations in the AFNI program AlphaSim (Ward, http://afni.nimh.nih.gov/afni/docpdf/alphasim.pdf). For the VMPFC, spherical ROIs (radius = 5 mm) were centered around coordinates from our previous findings using the same paradigm in PET (+/–8 62 –10) [8].

We also conducted exploratory psychophysiological interaction (PPI) analyses to understand changes in effective connectivity between a functional seed in the VMPFC and the amygdala going from the neutral imagery period to the anger imagery period.

RESULTS

fMRI data

We found significant clusters of activation within the amygdala and VMPFC which demonstrated our hypothesized time course of activation (Fig. 1). Specifically, we found significant bilateral activation in the amygdala that exceeded AlphaSim correction for multiple comparisons to preserve an α<0.05 (MNI coordinates for right = 18, 0, –16, k = 43, Z-score = 3.35; MNI coordinates for left = –24, –6, –12, k = 45, Z-score = 2.95; Fig. 2) that showed early increased activation during the anger induction relative to the neutral induction. However, during anger imagery, there was increased left VMPFC activation (MNI coordinates = –4, 60, –12, k = 3, Z-score = 2.98) that was trending in significance (preserving an α of 0.09 after AlphaSim correction for multiple comparisons). Correspondingly, the amygdala showed decreased activation during the anger imagery period. Our VMPFC finding was a replication of our result from our previous PET study [8]. For a list of all brain regions that demonstrated a similar pattern of increased activation during anger induction and decreased activation during anger imagery (and vice versa), see Table 1.

OPEN IN VIEWER

Fig.2

We found significant bilateral activation in the amygdala during anger induction followed by increased left ventromedial prefrontal cortex (VMPFC) activation during anger imagery. The VMPFC activation was trending in significance.

OPEN IN VIEWER

Table 1

All fMRI activations

Region	MNI Coordinates (x, y, z)	Cluster size (k)	T-score	Z-score
Using the Hypothesized Amygdala Model Contrast
Midbrain	–14, –16, –14	117	3.98	3.80
Parahippocampal gyrus	30, –16, –20	118	3.67	3.52
Superior frontal gyrus	16, 54, 32	62	3.65	3.51
Corpus callosum	–8, 24, 0	41	3.62	3.48
Superior temporal gyrus	54, –22, –2	49	3.37	3.26
Using the Hypothesized VMPFC Model Contrast
Midbrain	–14, –22, –22	130	5.35	4.95
Putamen	32, 0, 2	61	5.34	4.95
Inferior frontal gyrus	–38, 32, –8	83	4.95	4.63
Precentral gyrus	–34, –22, 54	74	4.72	4.44
Medial frontal gyrus	–2, 26, 46	126	4.40	4.16
Sub-gyral	26, 24, 40	72	4.33	4.11
Extra-nuclear	–8, –2, –10	108	4.11	3.92
Precentral gyrus	36, –2, 30	110	4.09	3.90
Inferior frontal gyrus	54, 12, 20	60	4.09	3.90
Culmen	–8, –36, –18	97	3.97	3.79
Inferior frontal gyrus	–26, 30, –10	56	3.93	3.76
Insula	42, 22, –8	52	3.86	3.70
Anterior cingulate	4, 28, 24	43	3.74	3.59
Right cerebrum	16, 6, –18	48	3.63	3.49
Cingulate gyrus	6, –4, 30	48	3.47	3.34
Sub-gyral	–38, 0, –26	58	3.39	3.27

Note: All clusters shown at p < 0.005 and k > 40 voxels.

Our exploratory PPI analyses revealed decreased coupling between the VMPFC functional seed and the left amygdala (MNI coordinates = –22, –8, –14, k = 4, Z-score = 2.27, p < 0.05; Fig. 3) when participants were experiencing anger (anger imagery period) compared to experiencing a neutral emotion (neutral imagery period).

OPEN IN VIEWER

Fig.3

Exploratory PPI analyses revealed decreased coupling between the VMPFC functional seed and the left amygdala when participants were experiencing anger (anger imagery period) compared to experiencing a neutral emotion (neutral imagery period).

DISCUSSION

Previously, most studies using autobiographical scripts to induce emotion were done in PET. This study successfully brought an emotion-induction paradigm from PET to fMRI and was able to use the greater temporal resolution of fMRI to clarify the temporal relation between the amygdala and VMPFC in the experience of anger. We had hypothesized that the lack of amygdala activation reported in neuroimaging studies could be due to an early but not sustained involvement of the amygdala during anger. We found that there was indeed an early increase of bilateral amygdala activation during anger induction, which was then followed by increased left VMPFC activation during the anger imagery period.

We also found negative correlations between self-rated anger and activation in the VMPFC. In other words, as self-rated anger increased, activation in the VMPFC decreased. These results support the idea that there is mutually inhibitory relationship between the amygdala and VMPFC [35, 36, 45]. Specifically, the amygdala’s early activation suggests that the amygdala might be more involved in the initial experience of anger while the VMPFC is more involved in the evaluation of that anger and in the inhibition of an impulsive response.

A unique feature of this study is that subjects were asked to provide self-relevant, autobiographical information for the anger induction and imagery paradigm. Most prior research has investigated the neural correlates of anger only when participants are viewing standard angry faces. Our use of self-relevant stimuli might have better ecological validity in terms of potentially studying the more realistic experience of self-generated and personally relevant anger processes. A clear future direction is to investigate the time course of amygdala and VMPFC activation during self-generated anger in psychiatric populations that have difficulty controlling their anger.

One limitation of the present study is that it was not designed to assess sex differences in anger. Future studies would be desirable to detect those differences. Also, the current sample scored slightly lower than average on anger reactivity; however, these scores did not significantly correlate with VMPFC or amygdala activation. Future work is needed to understand if this discrepancy was related to the fact that participants were asked to think of their most anger-provoking events and generate anger that may be an atypical experience for them in real life.

Overall, our study combined the advantages of using an actual event that has affected behavior with the superior temporal resolution of fMRI. Our findings suggest that it is important to look at the specific time course of amygdala activation when trying to understand the neural correlates of anger, which may not be possible with PET imaging. We also found that the VMPFC modulates early amygdala activation in healthy controls, and this modulation results in decreased self-rated anger.

CONFLICTS OF INTEREST

Dr. Dougherty has served as a consultant to and received honoraria from Medtronic, received travel support from Roche, and has received grant/research support from Medtronic, Cyberonics, Eli Lilly, and Roche.

Dr. Rauch has participated in research funded by Medtronic and Cyberonics. He has also received royalties for APPI and Oxford University Press.

Dr. Deckersbach’s research has been funded by NIMH, NARSAD, TSA, OCF, Tufts University and the Depression and Bipolar Disorder Alternative Treatment Foundation. He has received honoraria, consultation fees and/or royalties from the MGH Psychiatry Academy, BrainCells Inc., Systems Research and Applications Corporation, Boston University, the Catalan Agency for Health Technology Assessment and Research, the National Association of Social Workers Massachusetts, the Massachusetts Medical Society, Tufts University, NIDA, NIMH, and Oxford University Press. He has also participated in research funded by NIH, NIA, AHRQ, Janssen Pharmaceuticals, The Forest Research Institute, Shire Development Inc., Medtronic, Cyberonics, Northstar, and Takeda.

Ms. Chou and Dr. Buhlman have no conflicts of interest to report.