Amygdala–Prefrontal Disconnection in Borderline Personality Disorder

Antonia S New*,1,2, Erin A Hazlett1,3, Monte S Buchsbaum1,3, Marianne Goodman1,2, Serge A Mitelman1,3,

3 1 1,3 1,2 1 Randall Newmark , Roanna Trisdorfer , M Mehmet Haznedar , Harold W Koenigsberg , Janine Flory 1,2 and Larry J Siever

1Department of Psychiatry, Mount Sinai School of Medicine, New York, NY, USA; 2Department of Psychiatry, Mount Sinai School of Medicine,

3 Bronx VA Medical Center, Bronx, NY, USA; Department of Psychiatry and Neuroscience PET Laboratory, Mount Sinai School of Medicine,

New York, NY, USA

Abnormal fronto-amygdala circuitry has been implicated in impulsive aggression, a core symptom of borderline personality disorder

(BPD). We examined relative glucose metabolic rate (rGMR) at rest and after m-CPP (meta-chloropiperazine) with

18fluorodeoxyglucose (FDG) with positron emission tomography (PET) in 26 impulsive aggressive (IED)-BPD patients and 24 controls.

Brain edges/amygdala were visually traced on MRI scans co-registered to PET scans; rGMR was obtained for ventral and dorsal regions of

the amygdala and Brodmann areas within the prefrontal cortex (PFC). Correlation coefficients were calculated between rGMR for

dorsal/ventral amygdala regions and PFC. Additionally, amygdala volumes and rGMR were examined in BPD and controls. Correlations

PFC/amygdala Placebo: Controls showed significant positive correlations between right orbitofrontal (OFC) and ventral, but not dorsal,

amygdala. Patients showed only weak correlations between amygdala and the anterior PFC, with no distinction between dorsal and

ventral amygdala. Correlations PFC/amygdala: m-CPP response: Controls showed positive correlations between OFC and amygdala regions,

whereas patients showed positive correlations between dorsolateral PFC and amygdala. Group differences between interregional

correlational matrices were highly significant. Amygdala volume/metabolism: No group differences were found for amygdala volume, or

metabolism in the placebo condition or in response to meta-chloropiperazine (m-CPP). We demonstrated a tight coupling of metabolic

activity between right OFC and ventral amygdala in healthy subjects with dorsoventral differences in amygdala circuitry, not present in

IED-BPD. We demonstrated no significant differences in amygdala volumes or metabolism between BPD patients and controls.

Neuropsychopharmacology (2007) 32, 1629–1640; doi:10.1038/sj.npp.1301283; published online 3 January 2007

Keywords: amygdala; positron emission tomography; impulsive aggression

INTRODUCTION Halasz et al, 2002; Jinks and McGregor, 1997; McDonald and Mascagni, 1996; Morgan and LeDoux, 1995; Zbrozyna The Prefrontal–Amygdala Circuit and Westwood, 1991). In primates, damage to the lateral The concept that the prefrontal cortex (PFC) controls and PFC causes a loss of inhibitory control in attention tasks inhibits the amygdala and other limbic structures, termed (Dias et al, 1996; Stefanacci and Amaral, 2002), whereas ‘the reptilian brain’, was proposed many years ago (McLean, damage to orbital frontal cortex (OFC) causes a loss of 1955). Abundant preclinical data indicate that areas of the inhibitory control in ‘affective’ processing and increased PFC exert inhibitory control over the amygdala. A series of aggression (Izquierdo et al, 2005). In the macaque monkey, experiments in rats have shown that the medial PFC inhibits whereas both lateral and medial areas within OFC have activity in the basolateral amygdala by stimulating inhibi- strong association with limbic regions, lateral OFC has tory interneurons in the amygdala (Rosenkranz and Grace, specific connections to the amygdala (Carmichael and Price, 1999, 2002). Many other such studies that have shown this 1995). phenomenon in rodents (al Maskati and Zbrozyna, 1989; In human beings, functional brain imaging provides an approach in assessing the relationship between prefrontal and amygdala function by examining the correlation *Correspondence: Dr AS New, Department of Psychiatry, Mount Sinai coefficients between activity in the two structures. Human School of Medicine, Box 1218, One Gustave Levy Place, New York, 18 NY 10029, USA, Tel: + 1 212 241 0193, Fax: + 1 212 824 2302, studies with fluorodeoxyglucose (FDG)-positron emission E-mail: [email protected] tomography (PET) and fMRI have suggested a significant Received 7 March 2006; revised 26 September 2006; accepted 2 correlation between these two structures, but have primarily October 2006 been carried out in healthy subjects. The directionality of Amygdala–prefrontal disconnection AS New et al 1630 these correlations has been inconsistent, with some studies odor was also supported in rat studies of amygdala activity showing positive correlations between areas of PFC and in OFC (Saddoris et al, 2005; Schoenbaum et al, 2003). amygdala (Pezawas et al, 2005), and other studies showing The amygdala has been implicated not only in the negative correlations (Hariri et al, 2000, 2003). Specifically, processing of negative emotion in general, but also more one fMRI study during a perceptual task involving viewing specifically in the production of aggressive behavior in pictures of frightening faces showed significant negative animal studies. Electric stimulation of the lateral nucleus of correlation between rostral anterior cingulate, but negative the amygdala in cats results in predatory attack behavior correlations between amygdala and caudal anterior cingu- (Gregg and Siegel, 2001). In prairie voles, the medial late (Pezawas et al, 2005), whereas another study with a nucleus of the amygdala has been shown to be involved in similar task showed a negative correlation between right the regulation of aggression towards intruders (Wang et al, PFC and amygdala activity (Hariri et al, 2000). Significant 1997). In primates, ablation of the amygdala bilaterally leads negative correlations were also identified between the to increased social affiliation and decreased aggression response of the left amygdala and those of the right OFC (Emery et al, 2001; Meunier et al, 1999). These data have as well as the anterior cingulate (Hariri et al, 2003). An fMRI been taken to demonstrate a central role for the amygdala study of surprised faces show increased activation of the in the production of aggression, and led to the successful right ventral amygdala when the subject interpreted the face use of unilateral amygdalectomy in the treatment of a small negatively, whereas a positive interpretation of the face number of cases of pathological aggression in human beings yielded activation of OFC (Kim et al, 2003). A subsequent (Sachdev et al, 1992). study by the same group showed activation of lateral OFC in response to negative vs positive sentences, whereas medial Borderline Personality Disorder as a Prototype of OFC was activated in response to positive vs negative Emotion Dysregulation sentences (Kim et al, 2004), suggesting that subregions of OFC may play different roles in relation to amygdala Borderline personality disorder (BPD) is an illness, function. characterized by the symptom of emotional dysregulation Studies of patients with affective disorder (Drevets et al, and disinhibited anger, which often leads to aggressive 1992) and post-traumatic stress disorder (Shin et al, 2005) behavior. The model of altered prefrontal–amygdala con- have tended to find negative correlations between specifi- nectivity provides a model for the primary symptom in cally medial PFC and amygdala in patients but not controls, BPD, disinhibition of emotion. To date, however, ours is the suggesting coupling only in psychopathology. Recent data first study in BPD examining the relationship between suggest a role for serotonin in modulating connectivity prefrontal regions and amygdala in BPD. between PFC and the amygdala, with subjects carrying the A number of studies have reported abnormal PFC in ‘short’ allele of the serotonin transporter gene (a gene BPD, as well as in impulsive aggressive subjects with a possibly conferring a risk for mood disorders), showing variety of personality disorders. An 18FDG-PET study repor- greater fMRI amygdala/medial OFC ‘coupling’ during an ted reduced glucose metabolism in BPD patients compared emotional picture viewing task than those with only the to healthy controls in PFC, and anterior cingulate bilaterally ‘long’ allele (Hariri et al, 2006; Heinz et al, 2005; Pezawas (De La Fuente et al, 1997). In response to serotonergic et al, 2005). challenge, specifically impulsive-aggressive BPD patients demonstrate decreased metabolism in anterior cingulate and PFC, compared to controls (New et al, 2002; Siever et al, Subregions of the Amygdala in Emotion 1999b; Soloff et al, 2003). Numerous studies have demonstrated decreased serotonergic responsiveness in impulsive Subregions of the human amygdala have been shown to aggressive patients with personality disorders (Coccaro, have specific functions, roughly organized on a dorso- 1989; Dougherty et al, 1999; New et al, 2004; O’Keane et al, ventral dimension. The ventral amygdala includes primarily 1992; Virkkunen et al, 1994), and impulsive aggression has the basolateral complex (BLC) and the dorsal nucleus of been shown to respond the treatment with SSRIs (Coccaro the Central Nucleus (CN). Whalen et al (2001) suggest that and Kavoussi, 1997). We have recently reported gray matter the human dorsal vs ventral designation within the amyg- reduction in anterior cingulate (BA 24) in a large sample dala provides a means for incorporating numerous results of BPD patients (n ¼ 50) compared with healthy controls from the animal literature offering compelling evidence that (n ¼ 50) (Hazlett et al, 2005). In an analysis of a large the BLC (located ventrally in the human) can be dissociated sample of BPD-(IED) impulsive aggressive subjects, we behaviorally from the central nucleus (located dorsally in found that male subjects with IED-BPD have hypometabo- the human) and is the component of the amygdala lism widely across the frontal lobe compared to healthy predominantly involved in emotion modulation. Further, men, healthy women and women with BPD in response to they note that whereas expressions of fear appear to activate placebo (New et al, under review). This finding extends our the BLC and CN (ie ventral and dorsal amygdala), anger prior finding of decreased rGMR in response to m-CPP in may involve the CN to a lesser degree (than the BLC) IED-BPD in anterior and increase in posterior cingulate because less additional information concerning the stimulus compared to controls in a larger sample (New et al, 2002). is required (Whalen et al, 2001). Furthermore, fMRI studies We find similar evidence of anterior cingulate decreases by this group have shown that specifically the ventral and posterior cingulate increases in activation after seroto- amygdala is in response to emotional stimuli (Kim et al, nergic stimulus. We report these findings in a separate 2003, 2004). The ventral amygdala (basolateral amygdala in manuscript from the present one as they are conceptually rats)–OFC circuit role in associative encoding and aversive quite different. The present study is a correlational analysis,

Neuropsychopharmacology Amygdala–prefrontal disconnection AS New et al 1631 examining the relationship of activity in different brain ipsilateral. In addition, we examined group differences in regions, whereas our replications study closely follows amygdala volume and metabolic activity at baseline and the analysis of our previous publication (New et al, 2002), after m-CPP. Data on regional metabolism in PFC and additionally exploring the role of aggression subtype and cingulate on a subset of patients included in this study sex on the findings. (13 IED-BPD; 13 controls) have been published previously To date, only few neuroimaging studies have evaluated (New et al, 2002). amygdala volume or activity in BPD and none specifically in aggressive subjects. An early study of amygdala volume in BPD showed that total amygdala volume tended to be MATERIALS AND METHODS reduced in female BPD subjects compared to controls Subjects (Driessen et al, 2000). Two subsequent studies also reported decreased amygdala volume in BPD compared to controls in Twenty-six patients (17 men (35.7, SD ¼ 7.9 years), nine relatively small samples (Schmahl et al, 2003; Tebartz van women (30.7, SD ¼ 8.6), range ¼ 20–48; 19 (right-handed), Elst et al, 2003), although a recent study showed no differ- four (left-handed), four (mixed)) meeting DSM-IV criteria ence in amygdala volume compared to controls (Brambilla for BPD and Intermittent Explosive Disorder-modified et al, 2004) and a VBM extension study showed decreases (IED) as defined by the Module for Intermittent Explosive only in the left hippocampus/amygdala complex (Rusch disorder (Coccaro et al, 1998) were included. Patients with a et al, 2003; Tebartz van Elst et al, 2003). A recent larger history of schizophrenia, a psychotic disorder, or bipolar study employing a software package ‘BRAINS’ showed (Type I) affective disorder were excluded. Patients with no difference in amygdala volume in BPD compared to current major depressive disorder were excluded. Patients controls, although those BPD patients with a concurrent with past or current PTSD were accepted into the study, as a major depressive episode had larger amygdala volumes relatively high rate of PTSD in community samples of BPD compared to those without (Zetzsche et al, 2006). has been reported (Swartz et al, 1990). Three (all male) of Functional imaging studies of BPD are also limited in the 26 BPD patients met criteria for current PTSD and one number. One study of six female BPD patients and six female patient met criteria for past PTSD. We studied healthy volunteers (Herpertz et al, 2001) showed that BPD borderline patients with impulsive aggression to find a patients had greater cerebral blood flow (BOLD) signal in more homogeneous group of subjects with severe symp- the amygdala bilaterally during unpleasant pictures com- toms; this resulted, however, in our having a higher portion pared with neutral pictures than healthy controls. Another of male subjects than is usually reported in BPD samples. study reported greater left amygdala activation in BPD All subjects were medication-free 46 weeks (22/27 never- patients to facial expressions of emotion (vs a fixation medicated). Twenty-four age- and sex-matched healthy point) compared with healthy controls (Donegan et al, subjects were also studied (15 men (31.7, SD ¼ 7.9 years); 2003). nine women (34.0, SD ¼ 11.2) range ¼ 21–58;19 (right- Taken together, these studies provide support for a model handed), two (left-handed), three (mixed)). One RH 31- in which the amygdala is linked to emotional processing, year-old male control was inadequately imaged on the but it does not act in isolation; instead functions within a m-CPP day owing to technical difficulties and was used only network of brain regions that together modulate the in baseline analyses. Subjects were screened for severe complex manifestations of emotion. This reciprocal inter- medical or neurological illness, head injury, or past sub- action predicts that if cortical control of the thalamo- stance dependence, as well as substance abuse in the prior amygdala pathway is reduced, emotional responses will be 6 months. All subjects had a negative urine toxicology dysregulated (LeDoux, 1994). Based on this literature, we screen, and females a negative pregnancy test on each scan hypothesized that in BPD, an amygdala uncoupled from the day. Participants provided written informed consent in prefrontal regulation might be associated with loss of accordance with IRB guidelines. Patients were recruited behavioral control. In addition, the numerous data showing through advertisement in local newspapers (90%) and abnormalities in serotonergic function in impulsive aggres- referrals from psychiatric clinics at the Bronx VAMC and sive patients with personality disorders led us to examine Mount Sinai (10%). For the 26 subjects recruited into the the affect of a serotonergic agent on differential amygdala patient group, 164 subjects were screened. Patients were connectivity with the prefrontal cortex. We further hypo- excluded, in order of frequency for not fully meeting BPD, thesized that these changes might be more marked for the current substance abuse, medical problems, pregnancy, ventral than dorsal amygdala. Using 18FDG-PET, we tested and/or current major depression. One subject declined the PFC-amygdala balance theory by comparing inter- participation because of the radioactivity and another regional correlations between all 13 ipsilateral prefrontal declined an intravenous line. In the control group, 121 Brodmann areas and amygdala regions in impulsive candidates responded to advertisement and 97 were exclu- aggressive BPD patients compared with healthy controls ded because of the presence of an Axis I or II diagnosis in as measured at rest and after a serotonergic stimulus. themselves or a first-degree relative. We predicted more robust correlations between medial Diagnoses were made through interviews by a psycho- OFC (BA 11, 12, and 47) and specifically ipsilateral ventral logist using the Structured Clinical Interview for DSM-IV amygdala in healthy controls compared to patients. We Axis I disorders (First et al, 1996) and the Structured examined ipsilateral correlations as evidence suggests that Interview for DSM-IV Personality Disorders (Pfohl et al, the reciprocal PFC–amygdala connections both in non- 1997), respectively followed by a consensus meeting. human primates (Ghashghaei and Barbas, 2002) and in Subjects and (when available with the patient’s consent) a human beings (Di Virgilio et al, 1999) are predominantly family member were interviewed. All patients met DSM-IV

Neuropsychopharmacology Amygdala–prefrontal disconnection AS New et al 1632 Table 1 Clinical Assessments in IED-BPD and Controls

BDHI* ALS* BIS-7B* CTQ*

Controls (n ¼ 22) 17.7, SD ¼ 8.6 0.30, SD ¼ 0.30 36.0, SD ¼ 18.3 30.3, SD ¼ 12.3 IED-BPD (n ¼ 24) 40.4, SD ¼ 10.1 1.44, SD ¼ 0.58 50.6, SD ¼ 14.4 52.5, SD ¼ 20.3

These are clinical measures for patients and controls. For aggression (BDHI) scores, affective instability (ALS), impulsivity (BIS) and childhood trauma exposure (CTQ) all significant at po0.001. 2-tailed corrected for multiple comparisons. *Groups are significantly different at a level of po0.001.

criteria for BPD, except one subject who met 4/5 criteria Siever et al, 1999a; Soloff et al, 2000). PET–MRI coregistra- needed for a BPD diagnosis by his report and full criteria tion used the algorithm of Woods et al (1993). Brain edges by family-member report. Trait aggression was assessed were visually traced on all MRI axial slices with inter-tracer using the ‘Module for Intermittent Explosive Disorder- reliability of 0.99 on 10 subjects. Modified’ (Coccaro et al, 1998). All subjects completed the Buss-Durkee Hostility Inventory (BDHI) (Buss and Durkee, 1957), the Barratt Impulsivity Scale (BIS-7b) (Barratt, 1965), Regions of Interest Approach the Affective Lability Scale (ALS) (Harvey et al, 1989) and the Childhood Trauma Questionnaire (CTQ) (Bernstein We assessed rGMR within BAs by tracing coronal slices and Fink, 1998). based on a digitized brain atlas with 33 coronal slice All BPD patients had: significant physical and/or maps of BAs defined by microscopic examination of an verbal aggression, meeting criteria for IED (k ¼ 0.92). All entire postmortem brain, a technique detailed elsewhere patients met the ‘impulsiveness’ criterion for BPD (k ¼ 0.78) (Buchsbaum et al, 2001, 2002; Hazlett et al, 2000; Mitelman and 3/27 subjects met the ‘self-damaging’ BPD criterion et al, 2005). To assess the effect of m-CPP on rGMR, (k ¼ 0.90). Controls met none of the above-defined criteria. the dependent measure for PET analyses on drug effect Handedness was determined with the Edinburgh-handed- was expressed as difference scores (m-CPP-placebo) for ness-scale (Oldfield, 1971). Table 1 shows group means rGMR within each BA, calculated by subtracting placebo for symptom domains. counts for each region of interest in each subject from the corresponding rGMR from the m-CPP scan. The amygdala was outlined on coronal MRI sections Procedure using previously published methods (Haznedar et al, 2000) (ICC ¼ 0.82, area measured on three slices at the 25th, 50th, On two separate occasions (1–4 weeks apart), each partici- and 75th percentiles of anteroposterior distance). Outlining pant received m-CPP or placebo in a double-blind counter- of the amygdala began at its largest extent (approximately balanced manner. After an overnight fast, an intravenous the center in the anteroposterior dimension) where clear line was inserted (for blood sampling and injection of boundaries between gray matter and surrounding white m-CPP/placebo/18FDG). 0.08-mg/kg of m-CPP/placebo was matter are visible. At this mid-section, the amygdaloid given by slow push, immediately followed by 5 mCi of complex is roughly elliptical in shape, and anatomical 18FDG. The subject remained resting in a sound-attenuated, margins are defined by the cornu ammonis and the white dimly-lit room for the 35-min tracer-uptake period. matter of gyrus ambiens in the medial aspect, the cornu Following uptake, subjects were positioned in the PET inferius of the lateral ventricle in the ventral aspect, the scanner for a 45-min data-acquisition period. This method temporal lobe white matter laterally, and the gyrus semi- has been described in detail in previous reports (New et al, lunaris in the dorsal aspect. Using an edge contrast- 2002). enhancing technique (gradient filter) (Haznedar et al, PET scans were carried out as described elsewhere 2000), we were able to visualize better the dentate gyrus (Haznedar et al, 1997; New et al, 2002) (GE2048 head- of the hippocampus and boundaries between the hippo- dedicated scanner, resolution 4.5 mm in plane, 5.0 mm campus and the amygdala. The posterior portions of the axially). Fifteen slices at 6.5-mm intervals were obtained in amygdaloid complex were outlined by using the ventricular two sets to cover the entire brain. Slice counts of 1.5–3 M recess, hippocampus, and gyrus semilunaris as reference counts are typical. Scans were reconstructed with a blank points (Figure 1). Anteriorly, the amygdaloid complex and a transmission scan using the Hanning filter. The same gray matter is more heterogeneous and hard to identify. individually molded thermoplastic facemask was used for We outlined from the midsection forward using gradient each scan to minimize head-movement during image filtering and excluded the entorhinal cortex, which may acquisition and to assist in PET/MRI coregistration. PET include the inferior amygdala. The outlining ended at the images were obtained in nanocuries/pixel and standardized first coronal MRI section on which there was visible white as rGMR by dividing each pixel by the mean value for the matter between the amygdala, ambiens, and white matter entire brain (defined by brain-edge from coregistered MRI). of the entorhinal cortex. This procedure may have omitted Although this limits interpretations of single structure the very anterior end of the amygdaloid complex, but it absolute activity, this method is widely used when evaluat- had the advantage of excluding other extraneous structures ing hypotheses related to patterns of metabolic rate across from our analysis. Following the suggestion of Kim et al brain areas and was used in earlier imaging studies of (2003) we divided the amygdala into a top and bottom half, serotonin activation (Mann et al, 1996; New et al, 2002; based on the vertical distance on the mid-coronal slice and

Neuropsychopharmacology Amygdala–prefrontal disconnection AS New et al 1633

Figure 1 Tracing the amygdala. (a) Coronal MRI: anterior–posterior dimension of amygdala. (b) Sobel-gradient filter to enhance gray/white boundaries. (c) Amygdala outlined. applied the MRI-traced template to resliced and coregis- correlations between right or left ventral amygdala and OFC tered coronal PET slices. Volume is expressed as absolute (BA 11, 12, 47) (Table 2a). In addition to OFC, controls volume in mm3 and relative to whole brain volume. The showed positive correlations between subgenual BA 25 data were assessed for movement artifact by measuring and ventral amygdala on the right and between BA 44 and the ratio of the area of the middle PET slice (obtained ventral amygdala bilaterally. Controls showed no significant by a radial edging algorithm which draws the edge at 62% positive correlations between any frontal BAs and dorsal of the maximum value) to the area of the middle MRI amygdala in either hemisphere. Patients showed no signi- slice (obtained by hand-tracing); head-movement during ficant positive correlations between frontal BAs and either the longer scan would tend to blur the image and enlarge dorsal or ventral amygdala at rest. In patients, there were the area of activity. No group difference for movement significant negative correlations between the dorsal and artifact was detected (normals ¼ 0.96770.031, patients ¼ ventral amygdala and BAs 6, 8, 9, 10, 32, and 46 bilaterally. 0.96470.037, t ¼ 0.30). To test for group differences in the correlation matrices for ventral/dorsal amygdala with frontal BAs for each hemi- 2 Statistical Analysis sphere, a Kullback’s w test was conducted (Kullback, 1967). Controls differed significantly from patients in correlations Pearson’s correlation coefficients were employed for frontal between prefrontal BAs and dorsal and ventral amygdala in BAs and regions of the amygdala with significance level of both hemispheres (right ventral amygdala: df ¼ 91, Kull- 2 po0.05. The Kullback’s w test for correlational matrices back’s w2 ¼ 236.853, po0.0001; right dorsal amygdala: was employed to test group differences in correlational df ¼ 91, Kullback’s w2 ¼ 193.939, po0.001; left ventral matrices (Kullback, 1967). The Kullback test provides one amygdala, df ¼ 91, Kullback’s w2 ¼ 194.290, po0.001; left single p-value for comparing two correlation matrices each dorsal amygdala, df ¼ 91, Kullback’s w2 ¼ 184.355, po0.001) from a separate group of subjects. This avoids the Type I (see Figure 2). The most striking group differences occurred error associated with multiple correlation testing because in correlations between right ventral amygdala and orbital only one comparison is made. Univariate tests can then be BAs 11, 12, and 47. Figure 3 shows scatter plots of rGMR for supplied as post hoc tests to locate major sources of matrix correlations between these orbital BAs in the right hemi- differences. Mixed-factorial repeated-measures ANOVAs sphere and ventral amygdala. were employed to examine group differences in amygdala volume and metabolism. In addition, Pearson’s correlation coefficients were used to test correlation between amygdala m-CPP-placebo condition: correlations PFC and amyg- activity and the ALS, BIS, BDHI and CTQ scores. All dala. In response to m-CPP, controls showed significant correlations are reported for po0.05, two-tailed. positive correlations between ventral amygdala and BA 12 ANOVAs examining amygdala volume and function and 25 on the left. In addition, controls showed significant were conducted both including and excluding the subjects positive correlations for m-CPP-placebo rGMR between meeting PTSD, as PTSD has been associated with both dorsal amygdala and left BA 11 and 12 and right BA 11 and increased (Protopopescu et al, 2005; Shin et al, 2004), and 25. In controls, the frontal pole correlated negatively with decreased amygdala activity (Britton et al, 2005); amygdala dorsal amygdala in response to m-CPP in right BA 6 and 8 volume in PTSD has been shown not to be different from (see Table 2b). Patients, in contrast had positive correla- controls (Bremner, 2002; Wignall et al, 2004). tions between dorsal amygdala and dorsolateral PFC BA 44,45 and 46 on the left and between ventral amygdala and dorsolateral PFC in BA 44 and 45. In addition, patients RESULTS showed a positive correlation between BA 12 and left ventral Correlations PFC and Amygdala amygdala and a negative correlation between dorsal amygdala and BA 8 on the left. Placebo condition: correlations PFC and amygdala. The Also for m-CPP-placebo, controls differed significantly correlations between the ventral amygdala and orbitofrontal from patients in correlations between the prefrontal BAs BA 11, BA 12, and BA 47 were positive and significant in and ipsilateral dorsal and ventral amygdala in both hemi- controls on the right; whereas on the left, they were positive spheres (right ventral amygdala: df ¼ 91, Kullback’s but not significant in BA 11, BA 12, and BA 47 (Table 2a). w2 ¼ 190.04, po0.001; right dorsal amygdala: df ¼ 91, Kull- In contrast, the IED-BPD group showed no significant back’s w2 ¼ 248.71, po0.001; left ventral amygdala, df ¼ 91,

Neuropsychopharmacology Amygdala–prefrontal disconnection AS New et al 1634 Table 2 Correlations Frontal Brodmann Areas and Amygdala: Placebo and m-CPP-Placebo Condition

BA6 BA8 BA9 BA10 BA11 BA12 BA24 BA25 BA32 BA44 BA45 BA46 BA47

(a) Placebo Control Right dorsal amygdala À0.43* À0.34 À0.31 0.02E 0.39E 0.39E À0.08 0.34 0.07E 0.27 0.22 0.12E 0.37 Left dorsal amygdala À0.17 À0.20 À0.15 À0.03E 0.07 0.17 À0.24 0.01 À0.27E 0.35E 0.24 À0.11 0.17 Right ventral amygdala À0.46* À0.19 À0.20 0.11 0.46* 0.54*E À0.09 0.44* 0.19E 0.50*E 0.32E 0.09E 0.49* Left ventral amygdala À0.15 À0.24 À0.27 À0.16 0.29 0.37 À0.44* 0.19 À0.34 0.52* 0.37 À0.06 0.36

BPD Right dorsal amygdala À0.50* À0.55* À0.57* À0.56*E À0.26E À0.28E À0.20 À0.24 À0.63*E À0.29 À0.28 À0.52*E À0.15 Left dorsal amygdala À0.62* À0.59* À0.60* À0.63*E À0.29 À0.20 À0.21 À0.33 À0.58*E À0.36E À0.25 À0.58* À0.32 Right ventral amygdala À0.61* À0.62* À0.59* À0.41* À0.07 À0.02E À0.32 À0.03 À0.61*E À0.31E À0.28E À0.53*E 0.02 Left ventral amygdala À0.54* À0.56* À0.45* À0.42* 0.05 0.08 À0.10 À0.08 À0.51* 0.08 0.29 À0.38 0.11

(b) m-CPP-placebo Control Right dorsal amygdala À0.57*E À0.55*E À0.33 À0.28 0.42*E 0.40E À0.19 0.44*E 0.16 0.08 0.27 À0.10 0.35 Left dorsal amygdala À0.17 À0.45* À0.02 À0.06 0.45* 0.49* À0.08 0.32 À0.25 0.25 0.31 0.03 0.39 Right ventral amygdala À0.29 À0.21 À0.15 0.30E 0.21 0.21 À0.36 0.36 À0.02 0.18 0.07 À0.22 0.28 Left ventral amygdala À0.11 À0.17 À0.03 À0.09 0.37 0.44* À0.18 0.43* 0.16 0.36 0.18 À0.05 0.28

BPD Right dorsal amygdala 0.05E 0.03E À0.15 À0.18 À0.35E À0.32E À0.13 À0.37E À0.18 À0.28 À0.21 À0.19 À0.23 Left dorsal amygdala À0.38 À0.45* À0.28 À0.22 0.23 0.43 À0.01 0.10 À0.13 0.44* 0.62* 0.20* 0.15 Right ventral amygdala À0.21 À0.25 À0.36 À0.31E À0.28 À0.16 À0.17 0.36 À0.22 À0.36 À0.09 À0.26 À0.27 Left ventral amygdala À0.12 À0.24 À0.03 0.03 0.30 0.40* À0.16 0.26 À0.02 0.54* 0.54* 0.36 0.28

Relative glucose metabolism (rGMR) correlations between frontal lobe brodmann areas and dorsoventral amygdala divisions for placebo (a) and m-CPP-placebo (b). Asterisks denote significant Pearson correlation coefficients (po0.05). Diamonds denote significantly different correlations for patient group from control group for a particular region of interest, (po0.05, two-sided). Red demonstrates significant positive correlation coefficients between indicated brain regions, whereas bold indicates significantly negative correlation coefficients between brain regions.

Figure 2 Correlations orbital floor and amygdala: placebo condition, This is a visual map of significant positive and negative as well as non-significant correlations between rGMR in frontal Brodmann areas on the orbital surface of the brain and rGMR in ipsilateral amygdala in healthy subjects on the left and in impulsive aggressive borderline personality disorder subjects on the right. Significant positive correlations at po0.05 are shown in red and significant negative correlations in blue.

Kullback’s w2 ¼ 169.10, po0.001; left dorsal amygdala, amygdala would be disrupted, we also tested simple group df ¼ 91, Kullback’s w2 ¼ 194.40, po0.001) (see Figure 4). differences in amygdala volume and metabolism.

Amygdala volume. Because of evidence for specificity of Group Differences in Amygdala Volume and ventral amygdala in human emotion, we divided the Metabolism amygdala into dorsal and ventral components. A group Although our hypothesis in this study was that the coupling (control, patient) Â hemisphere (R, L) Â region (dorsal, of metabolic activity between frontal BAs and ventral ventral) repeated-measures ANOVA with dorsal and ventral

Neuropsychopharmacology Amygdala–prefrontal disconnection AS New et al 1635

Figure 3 Scatterplots right hemisphere OFC/ventral amygdala. Scatterplots of correlations of orbital Brodmann areas and amygdala rGMR for right BA 11, 12, 47 and ventral amygdala illustrating distribution in controls and in IED-BPD.

Figure 4 Correlations orbital floor and amygdala: m-CPP- Placebo Condition. This is a visual map of significant positive and negative as well as nonsignificant correlations between rGMR for m-CPP minus placebo in frontal Brodmann areas on the orbital surface of the brain and rGMR for m-CPP minus placebo in ipsilateral amygdala in healthy subjects on the left and in impulsive aggressive borderline personality disorder subjects on the right. Significant positive correlations at po0.05 are shown in red and significant negative correlations in blue. amygdala volume as the dependent variable showed no 13657198 mm3 for women, 13977208 mm3 for men; left main effect of group or interaction with group. The same hemisphere: 11987254 mm3 for women, 13657225 mm3 result was found with relative amygdala volume (amygdala for men. This is within the range for normal human volume/whole brain volume). It should be noted that the amygdala volumes found by others ranging from 1050 to total amygdala volumes found in our healthy subjects fall 1600 mm3 for the right amygdala and 1140 to 1400 mm3 for well within the range of that found by others. Specifically, the left amygdala (Convit et al, 1999; Driessen et al, 2000; we found total amygdala volumes for healthy control Szabo et al, 2003). subjects (mathematically identical of the sum of the top Similarly, when the same analysis was conducted exclud- and bottom for each hemisphere) to be: right hemisphere: ing subjects meeting criteria for PTSD, no significant effect

Neuropsychopharmacology Amygdala–prefrontal disconnection AS New et al 1636 Table 3 Amygdala Volume Amygdala activity. In patients, but not controls, we observed a negative correlation between aggression and affective 3 Absolute volume mm (SD) Controls (n ¼ 23) IED-BPD (n ¼ 26) lability scores for placebo rGMR in right ventral and dorsal amygdala (BDHI: ventral, r ¼À0.67, dorsal r ¼ 0.63; ALS: Right dorsal amygdala 919 (155) 889 (141) ventral, r ¼À0.67, dorsal r ¼À0.68). For drug-placebo, in Left dorsal amygdala 888 (158) 861 (153) controls but not patients, we observed negative correlations Right ventral amygdala 1004 (140) 926 (149) between ALS scores and the ventral amygdala (controls Left ventral amygdala 956 (162) 927 (194) r ¼À0.64) (Bonferroni corrected significance, po0.002, two-tailed) (between group test of correlation differences Relative to whole brain Â 1000, mean, SD was not significant). Right dorsal amygdala 0.11 (0.02) 0.10 (0.02) Left dorsal amygdala 0.10 (0.01) 0.10 (0.02) DISCUSSION Right ventral amygdala 0.12 (0.02) 0.11 (0.02) Left ventral amygdala 0.11 (0.01) 0.11 (0.02) Implications of Group Differences in Fronto-Amygdala Correlations Absolute volume in mm3 and relative to whole brain volume for dorsoventral amygdala divisions. The most striking finding that we report is the highly significant normal-BPD group differences in correlation patterns between frontal BA and ipsilateral amygdala metabolic activity (m-CPP minus placebo as well as the resting condition). Healthy controls, both at rest and in response of group in relative volume was detected. The groups to a serotonergic probe, show positive correlations bet- did not differ in whole brain volume (controls mean 1190, ween OFC (BA 11, 12) and right ventral amygdala, as was SD ¼ 150 mm3, patients 1180, SD ¼ 125 mm3, t ¼ 0.04, predicted from the model of the ventral amygdala as the p ¼ NS) (see Table 3). component of human amygdala most closely associated with frontal lobe emotion modulation (Kim et al, 2003; Amygdala metabolism: placebo. A group (control, Somerville et al, 2004; Whalen et al, 1998). This correla- patient) Â hemisphere (R, L) Â region (dorsal, ventral) tional approach is based on the assumption that significant repeated-measures ANOVA with placebo rGMR in the correlations may reveal an important functional relation- dorsal/ventral amygdala as the dependent variable showed ship between the structures as discussed by Katz et al no main effect of group or significant interaction with (1996). Correlations in healthy controls support the idea of group. Similarly, there was no interaction involving group intact coupling between PFC, particularly the ventrolateral for rGMR in the whole amygdala. The same result was region (BA 11, 47 and more dorsally in BA 44), and right found when excluding the subjects with PTSD. ventral amygdala, a tight coupling which may be the neural substrate for downregulation of the amygdala in response to aversive stimuli. The absence of such tight coupling in BPD A group (con- Amygdala metabolism (m-CPP-placebo). patients, indicated by the lack of significant correlations trol, patient) Â hemisphere (R, L) Â region (dorsal,ventral) between OFC and amygdala, suggests a disconnect between repeated-measures ANOVA with rGMR m-CPPÀplacebo OFC and amygdala, which may explain the failure of BPD difference scores within regions of the amygdala as the patients to downregulate the amygdala in response to dependent variable showed no interaction involving group. aversive stimuli. Furthermore, the BPD patients appear to The same result was found when excluding the subjects have lost anatomic specificity of ventral vs dorsal amygdala, with PTSD. showing nonsignificant correlations with OFC and modest negative correlations between amygdala and widespread Correlations of volume/metabolism with clinical symp- regions of the frontal lobe. toms. To examine the relationship between metabolism and The directionality of correlations between amygdala and symptom dimensions, we examined correlations between PFC is inconsistent in the literature. Our studies are volume and metabolism and clinical measures of impulsiv- consistent with findings of Pezawas et al, in which rostral ity, aggression, and affective instability. areas of anterior cingulate showed positive correlations with amygdala, at least under certain conditions (Kim et al, 2004; Amygdala volume. We observed a negative correlation Pezawas et al, 2005). However, other studies have demon- between BIS score and right relative ventral amygdala strated negative correlations between amygdala and medial volume in controls (n ¼ 22, r ¼À0.45, po0.001), but not OFC (Hariri et al, 2000, 2003). Previous fMRI studies patients (n ¼ 23, r ¼À0.27, p ¼ NS) (between-group test of showing negative correlations between amygdala and OFC correlation differences was not significant). In patients, we have employed event-related designs showing acute reac- found a negative correlation between impulsivity scores and tions to aversive stimuli (Dougherty et al, 2004; Shin et al, left ventral amygdala volume (n ¼ 23, r ¼À0.42, po0.004), 2005). Our study, in contrast, employs 18FDG-PET, which not present in controls (n ¼ 22 r ¼À0.08, p ¼ NS) (between- reflects an average of activity of a 30-min epoch. As FDG group test of correlation differences was not significant). uptake appears to reflect metabolic activity at axon These correlations survive Bonferroni correction for multi- terminals (Sokoloff, 1982), positive correlations would arise ple comparisons for which a p-value o0.004, two-tailed, between areas that receive input causing firing and its area is significant. of efferent connection (Katz et al, 1996). A limitation of the

Neuropsychopharmacology Amygdala–prefrontal disconnection AS New et al 1637 correlational approach using FDG is that it does not discri- between the ventral and dorsal amygdala in BPD. There was minate between inhibitory and excitatory connections. no relationship between aggression scores and left ventral Another reason for inconsistencies in the literature amygdala volume in controls in part because the variance in arise from the specific brain regions examined. Findings aggression measures within the controls group was very of a negative correlation between the regions of PFC and limited. amygdala have been found with the rostral anterior This study has a number of limitations in design and cingulate (Pezawas et al, 2005), as well as in ventral regions analysis. The lack of group differences in activation in of PFC (Hariri et al, 2000, 2003). Our finding of positive amygdala may relate to the study design, which did not correlations between particularly lateral OFC and ventral employ a behavioral task. As there was a pharmacological amygdala in healthy subjects (not found in our borderline challenge, the interaction with an impulse control task group) is consistent with regional specificity of the would have required four scans instead of two, and this connection between amygdala and PFC shown in primate would have provided logistical and subject volunteerism studies (Carmichael and Price, 1995), and some human problems. The responsivity of the amygdala may not be studies however (Kim et al, 2004). adequately probed with a resting scan or with a pharmaco- The ventral amygdala has been suggested to have greater logic challenge. Future studies involving behavioral provo- interconnection with OFC whereas the dorsal amygdala has cation during functional scanning will elucidate this. In greater interconnection with the cingulate and other limbic addition, our statistical analysis employed a broad assess- areas (Aggleton and Saunders, 2000). In our data, cingulate ment of correlations between all prefrontal Brodmann areas left 24 and right 25 were both significantly correlated areas in each hemisphere as well as the ventral and dorsal with ventral but not dorsal amygdala, consistent with the amygdala. However, as functional imaging studies, present general pattern of greater ventral amygdala–cortical links, and future, are producing a large number of activation areas but not with a distinction between cingulate and OFC from exploratory mapping, the full presentation of the connections with the amygdala. correlation matrix seemed justified to provide findings We show no significant group differences in rGMR bet- for other exploratory imaging studies. ween BPD patients and controls in the amygdala, which may Another limitation of the present study is the lack of suggest that the primary abnormality in BPD relates to the precision of our top-bottom parcellation of the amygdala. failure of the PFC to ‘come on line’ in response to amygdala The amygdala ROI has a volume of about 1300 mm3 in activation. Group differences in correlations in response our tracings. The full-width half-maximum resolution of to serotonergic stimulation show healthy controls with our scanner measured at the center of the ring is 4.5 mm positive correlations between orbitofrontal areas (BA 11, in plane and 6.5 mm in the z-axis. This yields a volume of bilaterally; left BA 12) and dorsal amygdala (Table 2b). This 131 mm3 and the amygdala volume is thus approximately is consistent with evidence that the dorsal amygdala (central 10 times as great. The mean height of our amygdala tracing nucleus) is rich in neuronal fibers positive for serotonin on the coronal ROI is 14.3 mm, and encompasses two PET and serotonin-transporter in primates (Freedman and Shi, slices. Nevertheless, top and bottom amygdala values are 2001). Borderline patients show positive correlations bet- one-half as high, and partial volume effects limit our statis- ween dorsal and ventral amygdala and dorsolateral BAs. tical power to detect differences. In addition, our division into top and bottom half is geometrical and does not weight Implications of Lack of Group Differences in Amygdala the lateral basal portion. MRI images provide inadequate Volume and Activity resolution to easily trace individual amygdala nuclei, again reducing statistical power. Consistent with larger studies in the literature employing ‘BRAINS’ to delineate the amygdala, our hand-tracing of amygdala in a large sample show no difference in amygdala ACKNOWLEDGEMENTS volume between BPD and controls. Although some studies have shown group differences, these studies have employed This research was supported by NIMH Grants MH566606 exclusively female subjects, where as we include both men to Dr Siever, MH067918 to Dr New, and MH60023 to and women and have selected BPD subjects with impulsive Dr Buchsbaum, and by the VA Medical Research Program aggression. We explored our data to examine whether (Career Development Award) to Dr New and a NARSAD entering sex into the ANOVA would yield significant results Independent Investigator Award to Dr Hazlett. This work and found no main effect of sex or interaction between was also supported in part by a grant (5-M01 RR00071) for sex and diagnosis in amygdala volume. the Mount Sinai General Clinical Research Center from the Clinical correlation with volume show some laterality in National Center for Research Resources, at the NIH. that inverse correlations were seen in patients between right ventral amygdala and measures of impulsivity, but in left ventral amygdala in controls. Prior studies of amygdala REFERENCES volume decreases in BPD, however, have shown a bilateral F effect. Clinical correlations in amygdala activity revealed Aggleton JP, Saunders RC (2000). The amygdala what’s hap- pened in the last decade? In: Aggleton JP (ed). The Amygdala. A inverse correlations with affective instability and aggression Functional Analysis. Oxford University Press: Oxford. pp 1–30. with right ventral and dorsal amygdala in BPD, whereas in al Maskati HA, Zbrozyna AW (1989). Cardiovascular and motor controls, a negative correlation between right ventral components of the defence reaction elicited in rats by electrical amygdala and affective lability was observed. This supports and chemical stimulation in amygdala. J Auton Nerv Syst 28: our evidence that there is a loss of anatomic specificity 127–131.

Neuropsychopharmacology Amygdala–prefrontal disconnection AS New et al 1638 Barratt ES (1965). Factor analysis of some psychometric measures dyadic social interactions in rhesus monkeys (Macaca mulatta). of impulsiveness and anxiety. Psychol Rep 16: 547–554. Behav Neurosci 115: 515–544. Bernstein D, Fink L (1998). Childhood Trauma Questionnaire: A First M, Spitzer R, Gibbon M, Williams J (1996). Structured Retrospective Self-report. Psychological Corp: San Antonio, TX. Clinical Interview for Axis I Disorders–Patient Edition. New York Brambilla P, Soloff PH, Sala M, Nicoletti MA, Keshavan MS, Soares Biometrics Research, New York State Psychiatric Institute: New JC (2004). Anatomical MRI study of borderline personality York, NY. disorder patients. Psychiatry Res 131: 125–133. Freedman LJ, Shi C (2001). Monoaminergic innervation of the Bremner JD (2002). Neuroimaging studies in post-traumatic stress macaque extended amygdala. Neuroscience 104: 1067–1084. disorder. Curr Psychiatry Rep 4: 254–263. Ghashghaei HT, Barbas H (2002). Pathways for emotion: interac- Britton JC, Phan KL, Taylor SF, Fig LM, Liberzon I (2005). tions of prefrontal and anterior temporal pathways in the Corticolimbic blood flow in posttraumatic stress disorder during amygdala of the rhesus monkey. Neuroscience 115: 1261–1279. script-driven imagery. Biol Psychiatry 57: 832–840. Gregg TR, Siegel A (2001). Brain structures and neurotransmitters Buchsbaum MS, Hollander E, Haznedar MM, Tang C, Spiegel- regulating aggression in cats: implications for human aggression. Cohen J, Wei TC et al (2001). Effect of fluoxetine on regional Prog Neuropsychopharmacol Biol Psychiatry 25: 91–140. cerebral metabolism in autistic spectrum disorders: a pilot Halasz J, Liposits Z, Meelis W, Kruk MR, Haller J (2002). study. Int J Neuropsychopharmacol 4: 119–125. Hypothalamic attack area-mediated activation of the forebrain Buchsbaum MS, Nenadic I, Hazlett E, Spiegel-Cohen J, Fleischman in aggression. Neuroreport 13: 1267–1270. MB, Akhavan A et al (2002). Differential metabolic rates in Hariri AR, Bookheimer SY, Mazziotta JC (2000). Modulating prefrontal and temporal Brodmann areas in schizophrenia and emotional responses: effects of a neocortical network on the schizotypal personality disorder. Schizophr Res 54: 141–150. limbic system. Neuroreport 11: 43–48. Buss AH, Durkee A (1957). An inventory for assessing different Hariri AR, Drabant EM, Weinberger DR (2006). Imaging genetics: kinds of hostility. J Consult Psychol 21: 343–348. perspectives from studies of genetically driven variation in Carmichael ST, Price JL (1995). Limbic connections of the orbital serotonin function and corticolimbic affective processing. Biol and medial prefrontal cortex in macaque monkeys. J Comp Psychiatry 59: 888–897. Neurol 363: 615–641. Hariri AR, Mattay VS, Tessitore A, Fera F, Weinberger DR (2003). Coccaro EF (1989). Central serotonin and impulsive aggression. Neocortical modulation of the amygdala response to fearful Br J Psychiatry 155: 52–62. stimuli. Biol Psychiatry 53: 494–501. Coccaro EF, Kavoussi RJ, Berman M, Lish J (1998). Intermittent Harvey P, Greenberg B, Serper M (1989). The affective lability explosive disorder-revised: development, reliability, and validity scales: development reliability and validity. J Clin Psychol 45: of research criteria. Comp Psychiatry 39: 368–376. 786–793. Coccaro EF, Kavoussi RJ (1997). Fluoxetine and impulsive Hazlett E, Buchsbaum M, Jeu L, Nenadic I, Fleischman MB, aggressive behavior in personality disordered subjects. Arch Shihabuddin L et al (2000). Hyporfrontality in unmedicated Gen Psychiatry 54: 1081–1088. schizophrenia patients studied with PET during performance of Convit A, McHugh P, Wolf OT, de Leon MJ, Bobinski M, De Santi S a serial verbal learning task. Schizophr Res 43: 33–46. et al (1999). MRI volume of the amygdala: a reliable method Hazlett EA, New AS, Newmark R, Haznedar MM, Lo JN, Speiser LJ allowing separation from the hippocampal formation. Psychiatry et al (2005). Reduced anterior and posterior cingulate gray Res 90: 113–123. matter in borderline personality disorder. Biol Psychiatry 58: De La Fuente JM, Goldman S, Stanus E, Vizuete C, Morlan I, Bobes 614–623. J et al (1997). Brain glucose metabolism in borderline Haznedar M, Buchsbaum M, Metzer M, Solimando A, Speigel- personality disorder. J Psychiatr Res 31: 531–541. Cohen J, Hollander E (1997). Anterior cingulate gyrus volume in Di Virgilio G, Clarke S, Pizzolato G, Schaffner T (1999). Cortical glucose metabolism in autistic disorder. Am J Psychiatry 154: regions contributing to the anterior commissure in man. 1043–1045. Exp Brain Res 124: 1–7. Haznedar M, Buchsbaum M, Wei T, Hof PR, Cartwright C, Dias R, Robbins TW, Roberts AC (1996). Dissociation in prefrontal Bienstock C et al (2000). Limbic circuitry in patients with autism cortex of affective and attentional shifts. Nature 380: 69–72. spectrum disorders studied with positron emission tomo- Donegan NH, Sanislow CA, Blumberg HP, Fulbright RK, Lacadie C, graphy and magnetic resonance imaging. Am J Psychiatry 157: Skudlarski P et al (2003). Amygdala hyperreactivity in border- 1994–2001. line personality disorder: implications for emotional dysregula- Heinz A, Braus DF, Smolka MN, Wrase J, Puls I, Hermann D et al tion. Biol Psychiatry 54: 1284–1293. (2005). Amygdala–prefrontal coupling depends on a genetic Dougherty DD, Rauch SL, Deckersbach T, Marci C, Loh R, variation of the serotonin transporter. Nat Neurosci 8: 20–21. Shin LM et al (2004). Ventromedial prefrontal cortex and Herpertz S, Dietrich T, Wenning B, Krings T, Erberich S, Willmes amygdala dysfunction during an anger induction positron K et al (2001). Evidence of abnormal amygdala functioning in emission tomography study in patients with major depres- borderline personality disorder: a functional MRI study. Biol sive disorder with anger attacks. Arch Gen Psychiatry 61: Psychiatry 50: 292–298. 795–804. Izquierdo A, Suda RK, Murray EA (2005). Comparison of the Dougherty DM, Bjork JM, Huckabee H, Moeller F, Swann A (1999). effects of bilateral orbital prefrontal cortex lesions and amygdala Laboratory measures of aggression and impulsivity in women lesions on emotional responses in rhesus monkeys. J Neurosci with borderline personality disorder. Psychiatry Res 85: 315–326. 25: 8534–8542. Drevets WC, Videen TO, Price JL, Preskorn SH, Carmichael ST, Jinks AL, McGregor IS (1997). Modulation of anxiety-related Raichle ME (1992). A functional anatomical study of unipolar behaviours following lesions of the prelimbic or infralimbic depression. J Neurosci 12: 3628–3641. cortex in the rat. Brain Res 772: 181–190. Driessen M, Herrmann J, Stahl K, Zwaan M, Meier S, Hill A et al Katz M, Buchsbaum MS, Siegel Jr BV, Wu J, Haier RJ, Bunney Jr (2000). Magnetic resonance imaging volumes of the hippo- WE (1996). Correlational patterns of cerebral glucose metabo- campus and the amgydala in women with borderline personality lism in never-medicated schizophrenics. Neuropsychobiology 33: disorder and early traumatization. Arch Gen Psychiatry 57: 1–11. 1115–1122. Kim H, Somerville LH, Johnstone T, Alexander AL, Whalen PJ Emery N, Capitanio J, Mason W, Machado C, Mendoza S, Amaral (2003). Inverse amygdala and medial prefrontal cortex responses D (2001). The effects of bilateral lesions of the amygdala on to surprised faces. Neuroreport 14: 2317–2322.

Neuropsychopharmacology Amygdala–prefrontal disconnection AS New et al 1639 Kim H, Somerville LH, Johnstone T, Polis S, Alexander AL, Shin Saddoris MP, Gallagher M, Schoenbaum G (2005). Rapid LM et al (2004). Contextual modulation of amygdala respon- associative encoding in basolateral amygdala depends on sivity to surprised faces. J Cogn Neurosci 16: 1730–1745. connections with orbitofrontal cortex. Neuron 46: 321–331. Kullback S (1967). On testing correlational matrices. Appl Statist Schmahl CG, Elzinga BM, Vermetten E, Sanislow C, McGlashan 16: 80–85. TH, Bremner JD (2003). Neural correlates of memories of LeDoux JE (1994). Emotion, memory and the brain. Sci Am 270: abandonment in women with and without borderline personality 50–57. disorder. Biol Psychiatry 54: 142–151. Mann JJ, Malone KM, Diehl DJ, Perel J, Nichols TE, Mintun MA Schoenbaum G, Setlow B, Nugent SL, Saddoris MP, Gallagher M (1996). Positron emission tomographic imaging of serotonin (2003). Lesions of orbitofrontal cortex and basolateral amygdala activation effects on prefrontal cortex in healthy volunteers. complex disrupt acquisition of odor-guided discriminations and J Cereb Blood Flow Metab 16: 418–426. reversals. Learn Mem 10: 129–140. McDonald AJ, Mascagni F (1996). Cortico-cortical and cortico- Shin LM, Orr SP, Carson MA, Rauch SL, Macklin ML, Lasko NB amygdaloid projections of the rat occipital cortex: a Phaseolus et al (2004). Regional cerebral blood flow in the amygdala and vulgaris leucoagglutinin study. Neuroscience 71: 37–54. medial prefrontal cortex during traumatic imagery in male and McLean PD (1955). The limbis system (‘visceral brain’) and female Vietnam veterans with PTSD. Arch Gen Psychiatry 61: emotional behaviour. Arch Neurol Psychiatry 73: 130–134. 168–176. Meunier M, Bachevalier J, Murray EA, Malkova L, Mishkin M Shin LM, Wright CI, Cannistraro PA, Wedig MM, McMullin K, (1999). Effects of aspiration versus neurotoxic lesions of the Martis B et al (2005). A functional magnetic resonance imaging amygdala on emotional responses in monkeys. Eur J Neurosci 11: study of amygdala and medial prefrontal cortex responses to 4403–4418. overtly presented fearful faces in posttraumatic stress disorder. Mitelman SA, Buchsbaum MS, Brickman AM, Shihabuddin L Arch Gen Psychiatry 62: 273–281. (2005). Cortical intercorrelations of frontal area volumes in Siever LJ, Buchsbaum M, New A, Spiegel-Cohen J, Wei T, Hazlett E schizophrenia. Neuroimage 27: 753–770. et al (1999a). d,l-fenfluramine response in impulsive personality Morgan MA, LeDoux JE (1995). Differential contribution of disorder assessed with 18F-deoxyglucose positron emission dorsal and ventral medial prefrontal cortex to the acquisition tomography. Neuropsychopharmacology 20: 413–423. and extinction of conditioned fear in rats. Behav Neurosci 109: Siever LJ, Buchsbaum MS, New AS, Spiegel-Cohen J, Wei T, Hazlett 681–688. EA et al (1999b). d,l-fenfluramine response in impulsive New AS, Hazlett EA, Buchsbaum MS, Goodman M, Reynolds D, personality disorder assessed with [18F]fluorodeoxyglucose Mitropoulou V et al (2002). Blunted prefrontal cortical positron emission tomography. Neuropsychopharmacology 20: 18fluorodeoxyglucose positron emission tomography response 413–423. to meta-chloropiperazine in impulsive aggression. Arch Gen Sokoloff L (1982). The radioactive deoxyglucose method theory, Psychiatry 59: 621–629. procedures, and applications for measurement of glucose New AS, Trestman RF, Mitropoulou V, Goodman M, Koenigsberg utilization in the central nervous system. Adv Neurochemistry HH, Silverman J et al (2004). Low prolactin response 4: 1–82. to fenfluramine in impulsive aggression. J Psychiatr Res 38: Soloff PH, Kelly TM, Strotmeyer SJ, Malone KM, Mann JJ (2003). 223–230. Impulsivity, gender, and response to fenfluramine challenge in O’Keane V, Maloney E, O’Neil H, O’Connor A, Smith C, Dinan TG borderline personality disorder. Psychiatry Res 119: 11–24. (1992). Blunted prolactin response to d-fenfluramine in socio- Soloff PH, Meltzer CC, Greer PJ, Constantine D, Kelly TM (2000). A pathy: evidence for subsensitivity of central serotonergic fenfluramine-activated FDG-PET study of borderline personality function. Br J Psychiatry 160: 643–646. disorder. Biol Psychiatry 47: 540–547. Oldfield RC (1971). The assessment and analysis of handedness: Somerville LH, Kim H, Johnstone T, Alexander AL, Whalen PJ the Edinburgh inventory. Neuropsychologia 9: 97–113. (2004). Human amygdala responses during presentation Pezawas L, Meyer-Lindenberg A, Drabant EM, Verchinski BA, of happy and neutral faces: correlations with state anxiety. Munoz KE, Kolachana BS et al (2005). 5-HTTLPR polymor- Biol Psychiatry 55: 897–903. phism impacts human cingulate–amygdala interactions: a Stefanacci L, Amaral DG (2002). Some observations on cortical genetic susceptibility mechanism for depression. Nat Neurosci inputs to the macaque monkey amygdala: an anterograde tracing 8: 828–834. study. J Comp Neurol 451: 301–323. Pfohl B, Blum N, Zimmerman M (1997). Structured Clinical Swartz MS, Blazer DG, George LK, Winfield I (1990). Estimating Interview for DSM-IV Personality (SIDP-IV). American Psychia- the prevalence of borderline personality disorder in the tric Press: Washington, DC. community. J Pers Disor 4: 257–272. Protopopescu X, Pan H, Tuescher O, Cloitre M, Goldstein M, Szabo CA, Lancaster JL, Xiong J, Cook C, Fox P (2003). MR Engelien W et al (2005). Differential time courses and specificity imaging volumetry of subcortical structures and cerebellar of amygdala activity in posttraumatic stress disorder subjects hemispheres in normal persons. AJNR Am J Neuroradiol 24: and normal control subjects. Biol Psychiatry 57: 464–473. 644–647. Rosenkranz JA, Grace AA (1999). Modulation of basolateral Tebartz van Elst L, Hesslinger B, Thiel T, Geiger E, Haegele K, amygdala neuronal firing and afferent drive by dopamine Lemieux L et al (2003). Frontolimbic brain abnormalities in receptor activation in vivo. J Neurosci 19: 11027–11039. patients with borderline personality disorder: a volumetric Rosenkranz JA, Grace AA (2002). Cellular mechanisms of magnetic resonance imaging study. Biol Psychiatry 54: 163–171. infralimbic and prelimbic prefrontal cortical inhibition and Virkkunen M, Rawlings R, Tokola R, Poland RE, Guidotti A, dopaminergic modulation of basolateral amygdala neurons Nemeroff C et al (1994). CSF biochemistries, glucose metabo- in vivo. J Neurosci 22: 324–337. lism, and diurnal activity rhythms in alcoholic violent offenders, Rusch N, van Elst LT, Ludaescher P, Wilke M, Huppertz HJ, Thiel fire setters, and healthy volunteers. Arch Gen Psychiatry 51: T et al (2003). A voxel-based morphometric MRI study in female 20–27. patients with borderline personality disorder. Neuroimage 20: Wang Z, Hulihan TJ, Insel TR (1997). Sexual and social experience 385–392. is associated with different patterns of behavior and neural Sachdev P, Smith JS, Matheson J, Last P, Blumbergs P (1992). activation in male prairie voles. Brain Res 767: 321–332. Amygdalo-hippocampectomy for pathological aggression. Aust Whalen PJ, Rauch SL, Etcoff NL, McInerney SC, Lee MB, Jenike NZ J Psychiatry 26: 671–676. MA (1998). Masked presentations of emotional facial expres-

Neuropsychopharmacology Amygdala–prefrontal disconnection AS New et al 1640 sions modulate amygdala activity without explicit knowledge. Woods R, Mazziota J, Cherry S (1993). MRI-PET registration with J Neurosci 18: 411–418. automated algorithm. J Comput Assist Tomogr 17: 536–546. Whalen PJ, Shin LM, McInerney SC, Fischer H, Wright CI, Rauch Zbrozyna AW, Westwood DM (1991). Stimulation in prefrontal SL (2001). A functional MRI study of human amygdala responses cortex inhibits conditioned increase in blood pressure and to facial expressions of fear versus anger. Emotion 1: 70–83. avoidance bar pressing in rats. Physiol Behav 49: 705–708. Wignall EL, Dickson JM, Vaughan P, Farrow TF, Wilkinson ID, Zetzsche T, Frodl T, Preuss UW, Schmitt G, Seifert D, Leinsinger G Hunter MD et al (2004). Smaller hippocampal volume in patients et al (2006). Amygdala volume and depressive symptoms in with recent-onset posttraumatic stress disorder. Biol Psychiatry patients with borderline personality disorder. Biol Psychiatry 60: 56: 832–836. 302–310.

Neuropsychopharmacology