R

Sw

CU

a

ARRA

KSNSNDPDfMPS

C

MDT

0d

Neuroscience and Biobehavioral Reviews 35 (2011) 688–698

Contents lists available at ScienceDirect

Neuroscience and Biobehavioral Reviews

journa l homepage: www.e lsev ier .com/ locate /neubiorev

eview

uicidal brains: A review of functional and structural brain studies in associationith suicidal behaviour

. van Heeringen, S. Bijttebier ∗, K. Godfrinniversity of Ghent, Department of Psychiatry and Medical Psychology, Unit for Suicide Research, Ghent, Belgium

r t i c l e i n f o

rticle history:eceived 17 May 2010eceived in revised form 2 August 2010ccepted 26 August 2010

eywords:uicideeuroimaginguicidal brain

a b s t r a c t

Evidence of an association between a vulnerability to suicidal behaviour and neurobiological abnormali-ties is accumulating. Post-mortem studies have demonstrated structural and biochemical changes in thebrains of suicide victims. More recently, imaging techniques have become available to study changes inthe brain in vivo. This systematic review of comparative imaging studies of suicidal brains shows thatchanges in the structure and functions of the brain in association with suicidal behaviour are mainlyfound in the orbitofrontal and dorsolateral parts of the prefrontal cortex. Correlational studies suggestthat these changes relate to neuropsychological disturbances in decision-making, problem solving andfluency, respectively. As a consequence, the findings from these studies suggest that suicidal behaviour is

eurobiologyecision-makingrefrontal cortexTI

MRIRI

associated with (1) a particular sensitivity to social disapproval (2) choosing options with high immediatereward and (3) a reduced ability to generate positive future events. Further study is needed to elaboratethese findings and to investigate to what extent changes in the structure and function of suicidal brainsare amenable to psychological and/or biological interventions.

© 2010 Elsevier Ltd. All rights reserved.

ETPECT

ontents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6882. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 689

2.1. Selection criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6903. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 690

3.1. Structural imaging studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6903.1.1. CT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6903.1.2. MRI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6903.1.3. DTI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 694

3.2. Functional imaging studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6943.2.1. SPECT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6943.2.2. PET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6953.2.3. fMRI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 695

4. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 696

Conflicts of interest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

∗ Corresponding author at: University of Ghent, Department of Psychiatry andedical Psychology, Unit for Suicide Research, University Hospital Ghent, Universityepartment of Psychiatry – 1K12F, De Pintelaan 185, 9000 Ghent, Belgium.el.: +32 9 332 58 94; fax: +32 9 332 49 89.

E-mail address: Stijn.bijttebier@ugent.be (S. Bijttebier).URL: http://www.unitforsuicideresearch.be.

149-7634/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.oi:10.1016/j.neubiorev.2010.08.007

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 697. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 697

1. Introduction

As the global annual rate of suicide approximates 15 per 100,000

individuals, it is estimated that one million people worldwidecommit suicide each year (WHO, 2002). Annual rates of non-fatalsuicidal behaviour are 10–20 times higher than those of completedsuicide (Kerkhof, 2000). Suicidal behaviour thus constitutes a major

nd Biobehavioral Reviews 35 (2011) 688–698 689

patiimm

mimorvdh(inbsbr

aiotieas

Electronic search references N = 531

Manual search N = 19

N = 19

N = 512 N = 3

N = 16

Studies excluded ( N = 528): 1. No comparison suicide –

non suicide attempters: N = 499

2. Unclear definition suicidal behaviour:N = 1

Studies meeting selection criteria

N = 22

Fl

C. van Heeringen et al. / Neuroscience a

ublic health problem, and costs at societal and individual levelsre huge. In addition, suicidal behaviour poses a major challengeo clinicians, health care workers and policy makers due to its lim-ted predictability (Hawton and van Heeringen, 2009). Problemsn predicting and thus preventing suicidal behaviour are due to the

ultiplicity of possible causes and a limited insight in predisposingechanisms.Based upon the current state of knowledge a stress-diathesis

odel of suicidal behaviour has been proposed to describe thenteraction between proximal risk factors and such predisposing

echanisms (Hawton and van Heeringen, 2009). The vast majorityf suicidal behaviours are associated with depression, but indeedequire the presence of an additional diathesis or predisposition. Inivo brain imaging is a promising tool for the investigation of theiathesis predisposing to suicidal behaviour, and the last decadeas witnessed a steady increase in the number of such studiesMann, 2005). A first advantage of using brain-imaging techniquess to confirm post-mortem findings of involved brain areas andeurobiological systems such as the serotonin system. Second,rain imaging provides a possibility to demonstrate a biologicalubstrate for neuropsychological changes in relation to suicidalehaviour, thus providing support of a causal interpretation of thiselation.

The findings from brain-imaging studies have not been system-tically reviewed yet. Imaging studies of suicidal behaviour fallnto two broad categories, i.e. the study of the neural correlatesf suicidal behaviour directly, and studies addressing clinical traits

ypically associated with suicidal behaviour, such as aggressive andmpulsive behaviours. The current review focuses on the first cat-gory of studies and therefore aims at the identification, synthesisnd appraisal of structural and functional neuroimaging studies ofuicidal behaviour.

ig. 2. Graphical summary of the reviewed articles: main findings, localized in the lateral l-tryptophan; rCMRglu = regional cerebral glucose metabolism; ↓ = decrease; ↑ = increase

3. Imaging technique: N = 28

Fig. 1. Overview of number of studies included and excluded.

2. Methods

A comprehensive literature search yielded the set of relevantarticles for this review. First, we did an electronic search using theWeb of Science, MEDLINE, PsycINFO (and PubMed) databases forarticles published in English, from 1990 onward. Search terms weresuicidal, suicide, suicide attempt, and imaging, CT, MRI, SPECT, PET,fMRI, DTI. Unpublished studies, case reports or conference abstracts

were not included in this review. Second, the reference lists ofrelevant papers were checked manually for additional relevantpublications not previously identified. Studies, for which electronicfull text was available, were selected if based on commonly used

eft hemisphere. Abbreviations: 5-HT2A = Serotonin-2A; �[11C]MTrp = �-[11C]methyl-; BA = Brodmann area; HL = high lethality; SA = suicide attempter.

6 nd Bio

io

2

rbstdaiopPs(t

3

Ti(

irat

F�

90 C. van Heeringen et al. / Neuroscience a

maging techniques and if individuals were included with a historyf suicidal behaviour.

.1. Selection criteria

Inclusion of studies in this review was based on three crite-ia. First, studies needed to report neuroanatomical differencesetween study groups composed of individuals with a history ofuicide attempts (suicide attempters; SA) and those without a his-ory of such behaviour. Studies comparing two groups of SA withifferent characteristics were also included. Second, taking intoccount the heterogeneity of definitions of suicidal behaviour, stud-es in order to be included needed to provide clear definitionsf suicidal behaviour. Third, studies were included only if Com-uted Tomography (CT), Magnetic Resonance Imaging (MRI), Singlehoton Emission Computed Tomography (SPECT), Positron Emis-ion Tomography (PET), functional Magnetic Resonance ImagingfMRI) or Diffusion Tensor Imaging (DTI) were used as imagingechniques.

. Results

As shown in Fig. 1, the search produced 550 publications.wenty-two articles met the inclusion criteria and were thusncluded in this review, categorized as ‘structural imaging studies’CT, MRI, DTI) or ‘functional imaging studies’ (SPECT, PET, fMRI).

In the following part of the paper, studies will be reviewedn chronological order; studies from the same research group areeported together. Table 1 summarizes the sample characteristicsnd detailed results of the reported studies. Figs. 2–5 summarizehe main results in a visual manner.

ig. 3. Graphical summary of the reviewed articles: main findings, localized in the media[11C]MTrp = �-[11C]methyl-l-tryptophan; ↓ = decrease; ↑ = increase; BA = Brodmann area

behavioral Reviews 35 (2011) 688–698

Good quality of included studies has been ensured by the useof strict criteria. Among the selected articles, the good quality ofthe studies by Audenaert et al. (2002), Meyer et al. (2003), Ehrlichet al. (2005), Monkul et al. (2007), Jovev et al. (2008), Pompili etal. (2008) and Jollant et al. (2008, 2010) is noteworthy due to thecontrol for comorbidity, specifically for depression.

3.1. Structural imaging studies

3.1.1. CTNo studies of suicidal behaviour using CT and meeting the selec-

tion criteria were found.

3.1.2. MRIStructural imaging, using MRI, within the suicide research can

be divided into two main fields.

3.1.2.1. White matter hyperintensities. A substantial number ofstudies have demonstrated an association between suicidalbehaviour and white matter hyperintensities (WMH), i.e. deepWMH (DWMH), periventricular hyperintensities (PVH) or subcor-tical gray matter hyperintensities (SCH).

Ahearn et al. (2001) were the first to demonstrate hyperinten-sities associated with a history of SA. As noted in Table 1, SA hadsignificantly more SCH and tended to have more PVH than controls.Ehrlich et al. (2004) reported an increased prevalence of WMH (par-

ticularly PVH) in (psychiatrically hospitalized) children and youthwith unipolar depression and a history of SA. Unipolar depressedsubjects with WMH were 18.6 times more likely to have a historyof SA than subjects with unipolar depression or another psychi-atric disorder with no evidence of WMH. There was no significant

l left hemisphere. Abbreviations: rCMRglu = regional cerebral glucose metabolism;; SA = suicide attempter; HL = high lethality.

C.vanH

eeringenet

al./Neuroscience

andBiobehavioralR

eviews

35(2011)

688–698691

Table 1Neuroimaging of suicidal behaviour—key features of brain-imaging studies of suicide attempters.

Studya Designb Targeted brain regionc Subjectsd Resultse Limitations

Ahearn et al. (2001) MRI: hyperintensities Whole brain 40 unipolar depressed patients (M + F):20 patients have attempted suicide(mean 66.0 yr) and 20 patients havenot attempted suicide (mean 66.4 yr)

Significantly more SCH and PVH in the patientswho attempted suicide

Small sample size; age of first suicideattempt and severity of eachdepressive episode not available

Ehrlich et al. (2004) MRI: hyperintensities Whole brain 153 patients with a primary DSM-III-Ror -IV Axis I diagnosis (M + F), 6–21 yr(mean 14.6 yr)

Patients with unipolar depression and whitematter hyperintensities have increasedprevalence of suicide attempts

No comparison group; no assessmentof comorbidity; small sample size;unclear causality

Ehrlich et al. (2005)* MRI: T2 hyperintensities Whole brain 102 young adult patients with MDD(M + F), mean 26.7 yr

Significant increase in prevalence of PVH (notDWMH) in patients who attempted suicide

Possible selection bias; age andseverity of suicide attempt not takeninto account; unclear causality

Monkul et al. (2007)* MRI: fronto-limbic brainabnormalities

OFC amygdalacingulate hippocampus

34 unipolar depressed patients (F): 7patients have attempted suicide (mean31.4 yr), 10 patients have notattempted suicide (mean 36.5) and 17subjects did not attempt suicide orhave a unipolar depression (mean31.3 yr)

Smaller bilateral OFC gray matter volume andlarger right amygdale volume in patients whoattempted suicide

Possible wrong evaluation of severityof depressive episode; small samplesize; study limited to women

Pompili et al. (2007) MRI: White matterhyperintensities

Whole brain 65 patients with a diagnosis of MDD orBD, without an additional Axis Idiagnosis of schizophrenia or psychosis(M + F): 29 patients have attemptedsuicide (mean 42.17 yr) and 36 patientshave not attempted suicide (mean44.61 yr)

Significantly increased prevalence (4.7 times)of WMH in patients who attempted suicide

Selection bias; retrospectiveassessment of suicide attempts; notable to disclose effects of medicationon suicidality; significant agedifference between experimentalgroups; no differentiation betweensubtypes of WMH

Aguilar et al. (2008) MRI: Whole brain structuralabnormalities

Whole brain OFCAmygdala

37 patients meeting the DSM-IVcriteria for schizophrenia (M): 13patients have attempted suicide (mean37.12 yr) and 24 patients have notattempted suicide (mean 42.65 yr)

Gray matter density reduction in the leftsuperior temporal lobe and left OFC in patientswho attempted suicide. No amygdale volumedifferences between patients who attemptedsuicide and who did not attempt suicide

Only male schizophrenic subjects; agedifference between experimentalgroups; No healthy subjects included

Jovev et al. (2008)* MRI: pituitary gland volume Pituitary gland 20 patients meeting the DSM-IVcriteria for BPD (M + F), mean 17.3 yr

Age and number of parasuicidal behaviours aresignificant predictors of pituitary gland volume

Small sample size; retrospective datacollection of parasuicide history

Pompili et al. (2008)* MRI: deep white matterhyperintensities andperiventricular white matterhyperintensities

Whole brain 99 patients with a diagnosis of MDD,BD-I or BD-II (M + F): 44 patients haveattempted suicide (mean 45.57 yr) and55 patients have not attempted suicide(mean 42.27 yr)

Patients who attempted to commit suicide aremore likely to have higher PVH. The presenceof PVH makes the risk for suicidal behaviour 8times higher

Small sample sizes; DWMH notdiscriminated by location; influence ofmedication not taken into account;unclear causality

Rüsch et al. (2008) MRI: gray and white mattervolumetric differences

Whole brain 105 subjects (M + F): 10 patients had aDSM-IV diagnosis of schizophrenia andhave attempted suicide (mean 30.3 yr),45 patients had a DSM-IV diagnosis ofschizophrenia and have not attemptedsuicide (mean 37.3 yr), 50 healthycontrols were matched with thepatients (mean 36.0 yr)

Significantly larger bilaterally inferior frontalwhite matter volumes in patients whoattempted suicide, with significant positivecorrelation between current self-aggression (inpast 6 months) and white matter volume

Small suicidality sample size;significant differences between groupson substance abuse; no generalpsychiatric comparison group

Matsuo et al. (2010) MRI: white matter of theanterior medial regions of thecorpus callosum

Corpus callosum 47 subjects (F): 10 patients with BDwho attempted suicide (mean 36.2 yr),10 patients with BD who have notattempted suicide (mean 44.2 yr) and27 healthy subjects (mean 36.9 yr)

(1) Suicidal BD patients have an inverse partialcorrelation between the anterior genu area andthe total score on the Barratt Impulsivity Scale(2) Smaller anterior corpus callosum predictshigher impulsivity in suicidal BD patients(3) No difference in size of the anterior corpuscallosum areas between suicidal BD patientsand the control groups

Small sample size; purely femalesample; psychiatric medication; lack ofcomparison between patients with orwithout comorbid cluster B disorder;lack of comparison between patientswith or without family history of riskfactors for suicidality

692C.van

Heeringen

etal./N

euroscienceand

BiobehavioralReview

s35

(2011)688–698

Table 1 (Continued )

Studya Designb Targeted brain regionc Subjectsd Resultse Limitations

Lo et al. (2007) DTI: hyperintensities Centrum semiovale 12 subjects (M + F): 6 patients withcarbon monoxide intoxication (suicideattempt) and 6 matched healthysubjects

(1) Suicidal patients have hyperintensities inthe centrum semiovale, next to PVH(2) 4 patients have bilateral globi pallidnecrosis

Small sample size

Audenaert et al. (2001) SPET: Serotonin-2A-receptorfunctioning

Frontal cortex 21 subjects, mean 30.4 yr (M + F): 9patients who have attempted suicideand 12 healthy subjects

Significant decrease in bilateral frontal 5- HT2A

binding index in patients who attemptedsuicide

Impact of alcohol and medication onclearance of the ligand; possible effectof physical trauma on binding index

Audenaert et al. (2002)* SPECT: binding potential Whole brain PFC 40 subjects (M + F): 20 depressedpatients who recently (<7 days)attempted suicide (19–49 yr) and 20healthy subjects (18–50)

Blunted perfusion in the left inferior PFC(during a category fluency task), bilateral gyrustemporalis medius (during a letter fluencytask) and the AC (while performing both aletter fluency and a category fluency task)

Influence of medication; selection bias;at random division of subgroups

van Heeringen et al. (2003) SPECT: Serotonin-2A-receptorfunctioning

PFC 21 subjects (M + F): 9 patients whohave attempted suicide (mean 32.4 yr)and 12 healthy subjects (mean 28.9 yr)

Binding potential in the prefrontal 5-HT2A

receptors is decreased in patients whoattempted suicide, next to higher levels ofhopelessness, harm avoidance andself-transcendence and lower scores onself-directedness and cooperativeness. Inpatients who attempted suicide, prefrontal5-HT2A binding potential correlates negativelywith hopelessness and harm avoidance andpositively with self-directedness andcooperativeness. Also, hopelessness correlatespositively with harm avoidance and negativelywith cooperativeness and self-transcendence

Small sample size; composition ofpatient sample; possible effect ofphysical trauma on binding index;effects of alcohol and medication

Lindström et al. (2004) SPECT: brain serotonin anddopamine transporters

Whole brain 24 subjects, mean 38.8 yr (M + F): 12patients who attempted suicide (5violent and 7 non-violent) and 12healthy matched subjects

No significant 5-HTT or DAT binding potentialdifferences were found. A significantcorrelation between 5-HTT and DAT, next to asignificant positive correlation between 5-HTTand impulsiveness, is found in patientsattempting suicide

Possible type 2 error

Ryding et al. (2006) SPECT: serotonin transporterand dopamine transporter

Whole brain 24 subjects (M + F): 12 patients whoattempted suicide (mean 38.8 yr) and12 matched healthy subjects

Patients, who have attempted suicide, show asignificant correlation between whole brain5-HTT BP and impulsiveness in the rightinferior frontal, bilateral temporal, midbrain,thalamic bilateral basal ganglia and leftcerebellar regions. These patients also showsignificant negative correlation between wholebrain DAT BP and mental energy in thebilateral basal ganglia regions

Amen et al. (2009) SPECT: In vivo brain differences Whole brain PFCsubgenual cingulate

36 subjects (M + F): 12 patientsmeeting DSM-IV criteria for depressionwho committed suicide since the brainimaging (mean 33.8 yr), 12 patientsmeeting DSM-IV criteria for depressionwho did not commit suicide and 12healthy subjects

Patients who attempted suicide, compared tonon-attempter patients, had higher rCBF atrest in the right insular cortex, dorsal AC gyrusand inferior parietal lobule. Patients whoattempted suicide, compared to non-attempterpatients, had perfusion deficits in the leftfrontal lobe, the right thalamus and part of theright medial temporal lobe duringconcentration. Areas of high rCBF were noticedin the right ACC and the left cerebellarpyramid. Patients who attempted suicide havegeneral lower rCBF at rest than healthysubjects

Heterogeneous sample; lack of data forall subjects; medication

C.vanH

eeringenet

al./Neuroscience

andBiobehavioralR

eviews

35(2011)

688–698693

Meyer et al. (2003)* PET: serotonin agonism PFC 69 subjects: 22 patients with MDD(mean 31.0 yr), 18 patients withself-injurious behaviour (mean 31.0 yr)and 29 healthy subjects (mean 31.0 yr)

No significant differences in 5-HT2 BP arefound between patients with self-harmbehaviour and healthy subjects

Indirect measurement of 5-HTconcentrations in the brain

Oquendo et al. (2003) PET: regional brainserotonergic function

Whole brain 25 patients (M + F) meeting DSM-III-Rcriteria of a major depressive episodeand who have attempted suicide: 16patients had a history of high-lethalitysuicide attempts (mean 42.9 yr), 9patients had a history of lo-lethalitysuicide attempts (mean 30.4 yr)

rCMR glucose uptake in PFC, ACC and superiorfrontal gyri is associated with high lethality ofsuicide attempt; lower ventromedialprefrontal activity associated with lowerlifetime impulsivity, higher suicidal intent andhigher-lethality suicide attempts; rCMRglucose uptake in right midcingulate andsuperior frontal gyri correlated negatively withexecutive functions, memory and attention,and positively with language fluency

Small sample size; direct brain injuriesnot ruled out

Leyton et al. (2006) PET: �[11C]MTrp trapping PFC 26 subjects (M + F): 10 patients haveattempted suicide (mean 37.7 yr) and16 healthy subject (mean 35.5 yr)

High-lethality suicide attempters havesignificantly reduced serotonin synthesis inthe OFC and ventromedial PFC (BA 11), next toan elevation in the right paracentral lobule, theleft thalamus, the left middle occipital cortexand the hippocampal gyrus

Effects of drugs on 5-HT transmission;small sample size; imaging techniquepossibly not the best to assess 5-HTneurotransmission

Jollant et al. (2008)* fMRI: neural activity Whole brain 43 subjects (M): 13 patients with apast history of MDD and suicidalbehaviour, 14 patients with a history ofMDD and no suicidal behaviour and 16healthy subjects

(1) Prototypical faces: patients with suicidalbehaviour, compared to patients without suchbehaviour, have greater activity in the rightlateral OFC and lower activity in the rightsuperior frontal gyrus in response to angry(versus neutral) faces(2) Mild faces: patients with suicidal behaviour,compared to patients without such behaviour,have greater activity in the right ACC and theright cerebellum to mild happy versus neutralfaces(3) Neutral faces: healthy subjects, compared tothe patients, have greater activity in the rightcerebellum to neutral faces versus fixationbaseline

Generalizability; small sample size;difficult comparison between groupsdue to lack of matching; medication

Jollant et al. (2010)* fMRI: whole brain andorbitofrontal cortex activity

Whole brain 40 subjects: 13 patients with a pasthistory of MDD and suicidal behaviour(mean 38 yr), 12 patients with a historyof MDD and no suicidal behaviour(mean 43 yr) and 15 healthy subjects(mean 30 yr)

(1) Significantly poorer decision-making insuicide attempters, compared with healthysubjects and patients without a history of asuicide attempt(2) Suicide attempters, compared with patientswithout a history of suicidal behaviour, havedecreased activation during risky versus safechoices in the left lateral OFC (BA47) and leftoccipital cortex (BA19)

Small sample size; differences betweenpatient groups and healthy controls

a Study: *Study of good methodological quality.b Design: fMRI = functional Magnetic Resonance Imaging; MRI = Magnetic Resonance Imaging; PET = Positron Emission Tomography; SPECT = Single Photon Emission Computed Tomography; SPET = Single Photon Emission

Tomography; �[11C]MTrp = �-[11C]methyl-l-tryptophan.c Targeted brain region: AC(C) = anterior cingulate (cortex); DLPFC = dorsolateral prefrontal cortex; OFC = orbitofrontal cortex; PFC = prefrontal cortex.d Subjects: BD = bipolar disorder, BD-I = bipolar depression type I, BD-II = bipolar depression type II; BPD = borderline personality disorder; F = female; M = male; MDD = major depressive disorder.e Results: 5-HT = Serotonin; 5-HT2 = Serotonin-2; 5-HT2A = Serotonin-2A; 5-HTT = the serotonin transporter; BA = Brodmann area; BP: binding potential; DAT = dopamine transporter; DWMH = deep white matter hyperintensities;

PVH: periventricular white matter hyperintensities; rCBF = regional cerebral blood flow; rCMR = regional cerebral glucose metabolism; SCH = subcortical gray matter hyperintensities; WMH = white matter hyperintensities.

694 C. van Heeringen et al. / Neuroscience and Biobehavioral Reviews 35 (2011) 688–698

F in thm ; BA =5

icpwsEPw

oopiMtpfod

3pwucwmma

o

ig. 4. Graphical summary of the reviewed articles: main findings, localizedetabolism; �[11C]MTrp = �-[11C]methyl-l-tryptophan; ↓ = decrease; ↑ = increase

-HT2A = Serotonin-2A.

ncrease in a history of SA in patients with WMH and other psy-hiatric disorders. With regard to localization, DWMH in the rightarietal lobe, but not in the frontal lobe, appeared to be associatedith a significantly higher prevalence of a history of SA. In their

ample of young adults with major depressive disorder (MDD),hrlich et al. (2005) showed a significantly increased prevalence ofVH, most commonly located in the right hemisphere, in patientsith a history of SA as compared to those without such a history.

Pompili et al. (2007) found a significantly increased prevalencef WMH in adults with MDD/bipolar disorder (BD) with a historyf attempted suicide and elevated suicide risk, compared to similaratients without such a history. There was no significant difference

n the frequency of previous suicide or WMH between subjects withDD and BD. In a subsequent study, Pompili et al. (2008) inves-

igated whether WMH are associated with suicidal behaviour inatients with a mood disorder. Attempters and non-attempters dif-ered only in the presence of PVH, with a more common presencef PVH in attempters than in non-attempters. Study groups did notiffer with regard to the presence of DWMH.

.1.2.2. Gray matter volume and density. Monkul et al. (2007) com-ared fronto-limbic brain structures between females diagnosedith a unipolar mood disorder and a history of one or more SA andnipolar females without such a history, to those of female healthyontrols. As indicated in Table 1, the presence of a history of SAas associated with volumetric changes in the bilateral OFC gray

atter and the right amygdala. There were no differences in grayatter volumes between unipolar patients without a history of SA

nd healthy controls.Two studies investigated gray and white matter in the brains

f patients suffering from schizophrenia according to suicidal

e lateral right hemisphere. Abbreviations: rCMRglu = regional cerebral glucoseBrodmann area; SA = suicide attempter; HL = high lethality; WM = white matter;

behaviour. Male patients with a history of SA, compared to thosewithout such a history, showed a significant reduction in gray mat-ter density in specific left hemispheric areas (see Table 1) (Aguilaret al., 2008). Rüsch et al. (2008) found significantly larger whitematter volumes in patients with a history of SA as compared topatients without such a history and to healthy controls (Table 1).No other significant white or gray matter volume differences wereobserved.

The study of Jovev et al. (2008) showed a correlation betweennumber of suicidal behaviours and pituitary gland volume in youngpatients diagnosed with borderline personality disorder and withminimal exposure to treatment.

Matsuo et al. (2010) specifically studied the association betweena history of SA and anterior corpus callosum volumes, but found nodifferences between bipolar patients with a history of SA and thosewithout such a history.

3.1.3. DTIOne study by Lo et al. (2007) showed bilateral PVH and cen-

trum semiovale hyperintensities after deliberate carbon monoxideintoxication.

3.2. Functional imaging studies

3.2.1. SPECTCompared to normal controls, patients with a recent history

of suicide attempts showed a significantly lower 5-HT2A bindingindex (reflecting a decrease in the number and/or in the bindingaffinity of 5-HT2A receptors) (Audenaert et al., 2001). The decreasewas more marked in dorsolateral than in orbitofrontal regions. Thedecrease was also significantly more marked among patients who

C. van Heeringen et al. / Neuroscience and Biobehavioral Reviews 35 (2011) 688–698 695

F edia� n area

utaHie

asacra(

t1

bwn

ibvggssttw

ig. 5. Graphical summary of the reviewed articles: main findings, localized in the m[11C]MTrp = �-[11C]methyl-l-tryptophan; ↓ = decrease; ↑ = increase; BA = Brodman

sed violent methods to attempt suicide (self-injury) than amonghose who attempted suicide by means of self-poisoning. Furthernalysis revealed a significant negative correlation between 5-T2A-receptor binding and levels of hopelessness, one of the most

mportant clinical predictors of suicidal behaviour (van Heeringent al., 2003).

Audenaert et al. (2002) used a neuropsychological split-dosectivation paradigm to assess brain perfusion in SA. They pre-ented a verbal fluency task to depressed patients who had recentlyttempted suicide and to healthy controls. When compared to theontrols, SA showed a blunted increase in perfusion in the left infe-ior PFC, in the gyrus temporalis medius bilaterally, and in thenterior cingulate during specific verbal fluency tasks (see Table 1)Audenaert et al., 2002).

Two studies assessed the binding potential of the serotoninransporter (5-HTT) and dopamine transporter (DAT) using the23I-�-CIT ligand (Lindström et al., 2004; Ryding et al., 2006). Inoth studies binding potentials were compared between patientsith a recent SA and healthy controls, and between violent andon-violent SA. The findings from both studies were negative.

Amen et al. (2009) compared regional cerebral blood flow (rCBF)n the brains of psychiatric patients who committed suicide, withrain rCBF in groups of healthy and non-suicidal depressed indi-iduals. In the baseline condition, i.e. at rest, rCBF in the suicideroup differed from that in the non-suicidal depressed controlroup through a significantly higher perfusion in the right hemi-

phere. During the Continuous Performance Test, the suicide grouphowed perfusion deficits in the left frontal lobe, in additiono a deficit in the right thalamus and part of the right medialemporal lobe, compared with depressed non-suicidal patients,hile clusters of high rCBF were noted in the right anterior cin-

l right hemisphere. Abbreviations: rCMRglu = regional cerebral glucose metabolism;; HL = high lethality; PVH: periventricular white matter hyperintensities.

gulate cortex (ACC; BA 32) and the left cerebellar pyramid. Incomparison with healthy subjects, the suicide group showed agenerally lower rCBF throughout the cortex during the baselinecondition.

3.2.2. PETMeyer et al. (2003) found that depressed subjects with

extremely dysfunctional attitudes had higher 5-HT2 binding poten-tial, compared to healthy subjects, particularly in BA 9.

By measuring relative regional cerebral uptake of fludeoxyglu-cose F 18 (rCMRglu), Oquendo et al. (2003) showed that depressedhigh-lethality suicide attempters had lower rCMRglu in ventral,medial, and lateral PFC compared with low-lethality attempters.This difference was more pronounced after fenfluramine adminis-tration, particularly in the anterior cingulate and the medial frontalgyri (BA 32 and 8) bilaterally, and the right midcingulate and supe-rior frontal gyri (BA 24 and 6). Lower ventromedial PFC activity wasassociated with different characteristics (Table 1). Verbal fluencycorrelated positively with rCMRglu in the same regions.

Leyton et al. (2006) compared high-lethality SA with healthycontrols with regard to the level of their �-[11C] methyl-l-tryptophan (�[11C]MTrp) trapping, an index of serotonin (5-HT)synthesis. High-lethality SA had significantly reduced normalized�[11C]MTrp trapping values in parts of the PFC, while an elevationin �[11C]MTrp trapping was seen in more posterior parts of thebrain (see Table 1). A significant association between ventromedial

�[11C]MTrp trapping and suicidal intent was found.

3.2.3. fMRIJollant et al. (2008) investigated neural reactivity following

exposure to angry and happy versus neutral faces in currently

6 nd Bio

eojteaett

4

ldscksb

Wuamt

moTsowttrith

pbca

rairtsvubtit

aaapmgc

96 C. van Heeringen et al. / Neuroscience a

uthymic men with a history of major depressive disorder, whetherr not with suicidal behaviour (SA and affective comparison sub-ects) and male healthy comparison subjects. The main results ofhis study are summarized in Table 1. In a subsequent study, Jollantt al. (2010) not only confirmed their earlier findings of a compar-tively poor performance on a decision-making task in SA (Jollantt al., 2005), but they also demonstrated a change in activation inhe orbitofrontal cortex and the occipital cortex during disadvan-ageous versus safe choices in SA (Table 1).

. Discussion

Suicidal behaviour constitutes an important public health prob-em and poses a major challenge to health care because ofifficulties in its prediction and thus prevention. This paper reviewstudies of the association between structural or functional brainharacteristics and suicidal behaviour. Such a review is relevant asnowledge of neural characteristics may contribute to our under-tanding, and thus to the prediction and prevention of suicidalehaviour.

The findings from this review can be summarized as follows.hite matter hyperintensities, whether or not in the periventric-

lar area, appear to be associated with the occurrence of suicidettempts in depressed children, adolescents and adults. Limbic grayatter hyperintensities are more commonly found in case of a his-

ory of suicide attempts than if such a history is not present.With regard to inferior frontal and orbitofrontal areas white

atter volumes are increased, while gray matter volumes of therbitofrontal cortex of suicide attempters appear to be reduced.ryptophan trapping, an index of serotonin synthesis, is reduced inignificant correlation with suicidal intent. In persons with a historyf depressive episodes, a history of attempted suicide is associatedith greater activity in the right orbitofrontal cortex in response

o angry (versus neutral) faces than in such persons without a his-ory of suicidal behaviour. In normothymic suicide attempters aeduction in activation in the left lateral orbitofrontal cortex dur-ng disadvantageous (versus safe) choices is found, when comparedo normothymic subjects without a history of suicide attempt andealthy controls.

Regarding the ventral and medial prefrontal cortex the findingsoint at a reduction in tryptophan trapping and glucose uptake,oth correlating negatively with suicidal intent. Ventromedial glu-ose uptake also correlates negatively with lethality of suicidettempts but positively with impulsivity.

Several findings point at a role of the dorsolateral and supe-ior frontal areas in the development of suicidal behaviour. Suicidettempters show a significant reduction in 5-HT2A-receptor bind-ng in the prefrontal cortex, particularly in the dorsolateral area. Theeduction is significantly more marked among violent attemptershan among non-violent attempters. Depressed suicide attemptershow a blunted increase in perfusion during activation using aerbal fluency task. High-lethality attempters show less glucoseptake than low-lethality-attempters and verbal fluency appears toe related to glucose metabolism in these areas. Following exposureo angry faces the right superior frontal gyrus shows less activationn individuals with a history of depression and attempted suicidehan in those with a history of depression but not attempted suicide.

Regarding limbic areas, a history of suicidal behaviour is associ-ted with larger right amygdala volumes. Suicidal behaviour is alsossociated with increased activity in the right insular cortex, dorsal

nterior cingulate cortex. . ., and with increased tryptophan trap-ing in the hippocampal gyrus and left thalamus. Lethality of theost serious lifetime suicide attempt correlates negatively with

lucose uptake after fenfluramine administration in the anterioringulate gyrus. Increased activity in the right anterior cingulate

behavioral Reviews 35 (2011) 688–698

gyrus is found following exposure to mild happy (versus neutral)faces in case of a history of suicidal behaviour.

Finally, suicidal behaviour appears to be associated withchanges in structural and functional characteristics of the parieto-occipito-temporal areas. Studies have indeed shown decreased graymatter density in the left superior temporal gyrus, increased activ-ity in the inferior parietal lobule, and increased tryptophan trappingin the left middle occipital cortex.

Preceding a discussion of these findings, methodological issuesneed to be addressed. The comparability of findings from differentstudies is limited because of variations in imaging and analytic tech-niques (Cannon et al., 2007). Radioligands differ in their bindingspecificity, and, in spite of the use of Statistical Parametric Map-ping in many studies, anatomical localization of findings is oftenimprecise. Small sample sizes, whether or not due to high dropoutrates, limit the power of studies to detect small group differences(Soloff et al., 2007) or can tend to amplify individual differencesdue to biological heterogeneity (Soloff et al., 2003). Other poten-tial biases may be due to the lack of (e.g. healthy or psychiatric)comparison groups.

Studies, which control for comorbidity, offer the possibility todistinguish between predispositions for comorbid disorders suchas depression and suicide (Audenaert et al., 2002; Meyer et al.,2003; Ehrlich et al., 2005; Monkul et al., 2007; Jovev et al., 2008;Pompili et al., 2008; Jollant et al., 2008, 2010). In some studies,however, patients and controls are not matched for potentiallybiasing characteristics, such as demographic variables, psychiatric(co-) morbidity, nature and severity or chronicity of associateddisorders, treatment, and exposure to risk and protective factors(Pompili et al., 2007; Aguilar et al., 2008; Rüsch et al., 2008). Alcoholdependence, which may lead to changes in brain areas such as thecerebellum, was an exclusion criterion in the majority of reviewedstudies.

Generalizability of findings may be limited due to inclusion ofonly male or female individuals or patients with particular disor-ders such as schizophrenia. In the majority of studies, assessmentof imaging is not blind to behavioural history.

The possibility of publication bias cannot be ruled out and maytherefore limit the interpretability of the current results.

Taking these methodological limitations in mind, the ques-tion as to what extent the findings from this review may helpto understand the development of suicidal behaviour needs tobe answered. Many of the areas involved in suicidal behaviour,as described in this review, are part of emotion-regulating cir-cuits in the brain. Such circuits comprise the prefrontal cortex, theamygdala–hippocampus complex, the thalamus, the basal gangliaand the extensive connections between these areas (Soares andMann, 1997). Lesions in one specific part or disruption of intercon-nections may thereby result in malfunctions in other areas. Thoseabnormalities may trigger the onset of mood disorder and confer abiological vulnerability, which, in combination with environmen-tal stressors, may result in suicidal behaviour (Ehrlich et al., 2004,2005).

The possibility to study clinical and/or neuropsychological cor-relates of brain characteristics constitutes an important advantageof imaging studies over post-mortem studies, and may help tofind an answer to this question by the identification of clusters ofdisturbed brain functions, neuropsychological and/or clinical char-acteristics, and suicidal behaviour. The identification of a biologicalsubstrate for neuropsychological and/or clinical characteristicsrelated to suicidal behaviour thereby supports a causal interpre-

tation of this relationship.

First, the results from this review point at the involvement of theorbitofrontal cortex in the development of suicidal behaviour. Earlystudies in relation to suicidal behaviour have focused on the associ-ation between orbitofrontal dysfunction and impulsive-aggressive

nd Bio

tbBawivcai

rtbdmcesmab(cTtelicepaodmprcirltTf

osHhHe2cHb(pm5nflp

sfi

C. van Heeringen et al. / Neuroscience a

raits. Findings have, however, not been equivocal, which maye due to the multifaceted nature of the impulsivity construct.ehavioural indices of impulsivity, such as response inhibition,re more basic constructs and thus quantifiable in a more reliableay (Mann, 2005). Response inhibition has not yet been studied

n relation to suicidal behaviour, but functional imaging in healthyolunteers indeed shows involvement of the (right) orbitofrontalortex, the cingulate cortex and the inferior parietal lobule (Horn etl., 2003). As described in this review these areas are also involvedn suicidal behaviour.

With regard to the orbitofrontal cortex, attention has moreecently been drawn to the association between orbitofrontal cor-ex dysfunctioning, disturbances in decision-making and suicidalehaviour. A first study showed that violent suicide attemptersiffered from affective controls in their performance on a decision-aking task in that suicide attempters make more disadvantageous

hoices, i.e. choose options with high immediate reward (Jollantt al., 2005). A subsequent functional neuroimaging study indeedhowed that suicide attempters (1) performed worse on a decision-aking task than affective controls and (2) showed reduced

ctivation in the orbitofrontal (and occipital) cortex for the contrastetween risky (disadvantageous) and safe (advantageous) choicesJollant et al., 2010). The insufficient contrast between risky and safehoices prevents advantageous guiding of long-term behaviour.aken together with the findings of increased orbitofrontal activa-ion following exposure to angry faces in suicide attempters (Jollantt al., 2008), these findings suggest that suicidal behaviour is corre-ated with disturbances in the attribution of importance to stimuli,.e. undue importance to signals of others’ disapproval and insuffi-ient importance to risky choices. The development of unbearablemotional pain following perception of signals of others’ disap-roval may be associated with a choice for immediate reward (i.e.lleviation of pain). The association between mental pain and riskf suicidal behaviour has indeed recently been demonstrated inepressed individuals (van Heeringen et al., 2010). While decision-aking in general is associated with the activation of ventromedial

refrontal regions, it appears that the orbitofrontal cortex could beelated with deciding between advantageous and disadvantageoushoices following exposure to particular stimuli, such as angry facesn suicide attempters. A number of reviewed studies point at a cor-elation between orbitofrontal functioning and suicidal intent orethality of suicide attempts. In addition, decision-making appearso be disturbed more in violent than in non-violent attempters.he interpretation of these correlational findings requiresurther study.

A second potential cluster of findings concerns the rolef the dorsolateral prefrontal cortex in the development ofuicidal behaviour. Prefrontal, and in particular dorsolateral 5-T2A-receptor binding correlates significantly with levels ofopelessness, a strong clinical predictor of suicidal behaviour (vaneeringen et al., 2003). Prefrontal (i.e. superior frontal) seroton-rgic activity also correlates with verbal fluency (Oquendo et al.,003). It is of particular interest that dysfunctional attitudes, whichorrelate with hopelessness, are associated with prefrontal 5-T2A-receptor binding (Meyer et al., 2003). A negative correlationetween fluency and hopelessness has been described previouslyMacLeod et al., 1993). Taken together these findings point at aossible role of the dorsolateral prefrontal cortex in the develop-ent of suicidal behaviour: findings of reduced activity, reduced

-HT2A-receptor binding, reduced fluency and increased hopeless-ess appear to cluster in attempted suicide patients. The reduced

uency thereby particularly concerns a reduced ability to generateositive future events.

Among particular functional and structural characteristics ofuicidal brains this review of studies thus identifies two clusters ofndings, which involve changes in the functions of the orbitofrontal

behavioral Reviews 35 (2011) 688–698 697

cortex and the dorsolateral prefrontal cortex, associated withincreased preference for immediate reward and decreased ability togenerate positive events in the future, respectively. Research out-side the domain of suicidology has demonstrated that the abilityto select an action by considering delays and amount of rewardoutcome is critical for survival and wellbeing (Schweighofer et al.,2007).

In daily life, people indeed make decisions based on predictionof rewards at different time scales. In many instances the choicefor a longer-term positive outcome is more appropriate than achoice for immediate reward. However, the vulnerability to sui-cidal behaviour appears to be associated with choosing immediatereward (possibly the alleviation of emotional pain) in the absenceof the ability to generate future rewards. We have recently indeeddemonstrated an association between emotional pain and suiciderisk in depressed individuals (van Heeringen et al., 2010).

It has recently been shown that the prediction of immediate andfuture rewards differentially recruits cortico-basal ganglia loops(Tanaka et al., 2004a, 2004b). More specifically, it appears that theorbitofrontal cortex is involved in predicting immediate reward,while the dorsolateral prefrontal cortex is involved in future rewardprediction. Taking into account the clearly demonstrated role ofserotonin in the vulnerability to suicidal behaviour, it is worthyto note that serotonin appears to be involved in action selectionby modulating the evaluation of delayed rewards. More particu-larly, experimental data show that delayed rewards have a lowvalue with low serotonin levels leading agents to choose immediateover delayed rewards (Schweighofer et al., 2007). The availability ofserotonin thus appears to correlate with the extent to which futureevents are taken into account when choosing between behaviouraloptions.

Further research is clearly needed to assess the applicabilityof such neuropsychological models to suicidal behaviour. Atten-tion should thereby be given to functional neuroimaging studies ofdecision-making and reward prediction.

Conflicts of interest

The authors have neither financial interest in, nor financial sup-port for writing this review.

References

Aguilar, E.J., Garcia-Marti, G., Marti-Bonmati, L., Lull, J.J., Moratal, D., Escarti, M.J.,Robles, M., Gonzales, M.I., Guillamon, J.C., Sanjuan, J., 2008. Left orbitofrontaland superior temporal gyrus structural changes associated to suicidal behaviorin patients with schizophrenia. Prog. Neuropsychopharmacol. Biol. Psychiatry32, 1673–1676.

Ahearn, E.P., Jamison, K.R., Steffens, D.C., Cassidy, F., Provenzale, J.M., Lehman,A., Weisler, R.H., Carroll, B.J., Krishnan, K.R.R., 2001. MRI correlates of suicideattempt history in unipolar depression. Biol. Psychiatry 50, 266–270.

Amen, D.G., Prunella, J.R., Fallon, J.H., Amen, B., Hanks, C., 2009. A comparativeanalysis of completed suicide using high resolution brain SPECT imaging. J.Neuropsychiatry Clin. Neurosci. 21, 430–439.

Audenaert, K., Goethals, I., Van Laere, K., Lahorte, P., Brans, B., Versijpt, J., Vervaet,M., Beelaert, L., van Heeringen, K., Dierckx, R., 2002. SPECT neuropsychologicalactivation procedure with the Verbal Fluency Test in attempted suicide patients.Nucl. Med. Commun. 23, 907–916.

Audenaert, K., Van Laere, K., Dumont, F., Slegers, G., Mertens, J., van Heeringen, C.,Dierckx, R.A., 2001. Decreased frontal serotonin 5-HT2a receptor binding indexin deliberate self-harm patients. Eur. J. Nucl. Med. 28, 175–182.

Cannon, D.M., Ichise, M., Rollis, D., Klaver, J.M., Gandhi, S.K., Charney, D.S., Manji, H.K.,Drevets, W.C., 2007. Elevated serotonin transporter binding in major depres-sive disorder assessed using positron emission tomography and [C-11]DASB;comparison with bipolar disorder. Biol. Psychiatry 62, 870–877.

Ehrlich, S., Noam, G.G., Lyoo, I.K., Kwon, B.J., Clark, M.A., Renshaw, P.F., 2004. Whitematter hyperintensities and their associations with suicidality in psychiatrically

hospitalized children and adolescents. J. Am. Acad. Child. Adolesc. Psychiatry 43,770–776.

Ehrlich, S., Breeze, J.L., Hesdorffer, D.C., Noam, G.G., Hong, X., Alban, R.L., Davis, S.E.,Renshaw, P.F., 2005. White matter hyperintensities and their association withsuicidality in depressed young adults. J. Affect. Disord. 86, 281–287.

Hawton, K., van Heeringen, K., 2009. Suicide. Lancet 373, 1372–1381.

6 nd Bio

H

J

J

J

J

K

L

L

L

M

M

M

M

M

O

personality characteristics in attempted suicide. J. Affect. Disord. 74, 149–158.van Heeringen, K., Van den Abbeele, D., Vervaet, M., Soenen, L., Audenaert, K., 2010.

98 C. van Heeringen et al. / Neuroscience a

orn, N.R., Dolan, M., Elliott, R., Deakin, J.F.W., Woodruff, P.W.R., 2003. Responseinhibition and impulsivity: an fMRI study. Neuropsychologia 41, 1959–1966.

ollant, F., Bellivier, F., Leboyer, M., Astruc, B., Torres, S., Verdier, R., Castelnau, D.,Malafosse, A., Courtet, P., 2005. Impaired decision making in suicide attempters.Am. J. Psychiatry 162, 304–310.

ollant, F., Lawrence, N.S., Giampietro, V., Brammer, M.J., Fullana, M.A., Drapier, D.,Courtet, P., Phillips, M.L., 2008. Orbitofrontal cortex response to angry faces inmen with histories of suicide attempts. Am. J. Psychiatry 165, 740–748.

ollant, F., Lawrence, N.S., Olie, E., O’Daly, O., Malafosse, A., Courtet, P., Phillips, M.L.,2010. Decreased activation of lateral orbitofrontal cortex during risky choicesunder uncertainty is associated with disadvantageous decision-making and sui-cidal behavior. NeuroImage 51, 1275–1281.

ovev, M., Garner, B., Phillips, L., Velakoulis, D., Wood, S.J., Jackson, H.J., Pantelis, C.,McGorry, P.D., Chanen, A.M., 2008. An MRI study of pituitary volume and para-suicidal behavior in teenagers with first-presentation borderline personalitydisorder. Psychiatr. Res. Neuroimaging 162, 273–277.

erkhof, A.J.F.M., 2000. Attempted suicide: patterns and trends. In: Hawton, K.,van Heeringen, K. (Eds.), The International Handbook of Suicide and AttemptedSuicide. John Wiley & Sons Ltd., Chichester, pp. 49–64.

eyton, M., Paquette, V., Gravel, P., Rosa-Neto, P., Weston, F., Diksic, M., Benkelfat,C., 2006. �-[11C]Methyl-l-tryptophan trapping in the orbital and ventralmedial prefrontal cortex of suicide attempters. Eur. Neuropsychopharmacol. 16,220–223.

indström, M.B., Ryding, E., Bosson, P., Ahnlide, J.A., Rosén, I., Träskman-Bendz, L.,2004. Impulsivity related to brain serotonin transporter binding capacity insuicide attempters. Eur. Neuropsychopharmacol. 14, 295–300.

o, C.P., Chen, S.Y., Chou, M.C., Wang, C.Y., Lee, K.W., Hsueh, C.J., Chen, C.Y., Huang,K.L., Huang, G.S., 2007. Diffusion-tensor MR imaging for evaluation of the effi-cacy of hyperbaric oxygen therapy in patients with delayed neuropsychiatricsyndrome caused by carbon monoxide inhalation. Eur. J. Neurol. 14, 777–782.

acLeod, A.K., Rose, G.S., Williams, J.M.G., 1993. Components of hopelessness aboutthe future in parasuicide. Cognit. Ther. Res. 17, 441–455.

ann, J.J., 2005. What does brain imaging tell us about the predisposition to suicidalbehavior. Crisis 26, 101–103.

atsuo, K., Nielsen, N., Nicoletti, M.A., Hatch, J.P., Monkul, E.S., Watanabe, Y., Zunta-Soares, G.B., Nery, F.G., Soares, J.C., 2010. Anterior genu corpus callosum andimpulsivity in suicidal patients with bipolar disorder. Neurosci. Lett. 469, 75–80.

eyer, J.H., McMain, S., Kennedy, S.H., Korman, L., Brown, G.M., DaSilva, J.N., Wil-son, A.A., Blak, T., Eynan-Harvey, R., Goulding, V.S., Houle, S., Links, P., 2003.Dysfunctional attitudes and 5-HT2 receptors during depression and self-harm.Am. J. Psychiatry 160, 90–99.

onkul, E.S., Hatch, J.P., Nicoletti, M.A., Spence, S., Brambilla, P., Lacerda, A.L.T.,Sassi, R.B., Mallinger, A.G., Keshavan, M.S., Soares, J.C., 2007. Fronto-limbic brainstructures in suicidal and non-suicidal female patients with major depressivedisorder. Mol. Psychiatry 12, 360–366.

quendo, M.A., Placidi, G.P.A., Malone, K.M., Campbell, C., Keilp, J., Brodsky, B., Kege-les, L.S., Cooper, T.B., Parsey, R.V., Van Heertum, R.L., Mann, J.J., 2003. Positron

behavioral Reviews 35 (2011) 688–698

emission tomography of regional brain metabolic responses to a serotoner-gic challenge and lethality of suicide attempts in major depression. Arch. Gen.Psychiatry 60, 14–22.

Pompili, M., Ehrlich, S., De Pisa, E., Mann, J.J., Innamorati, M., Cittadini, A., Montagna,B., Iliceto, P., Romano, A., Amore, M., Tatarelli, R., Girardi, P., 2007. White matterhyperintensities and their associations with suicidality in patients with majoraffective disorders. Eur. Arch. Psychiatry Clin. Neurosci. 257, 494–499.

Pompili, M., Innamorati, M., Mann, J.J., Oquendo, M.A., Lester, D., Del Casale, A.,Serafini, G., Rigucci, S., Romano, A., Tamburello, A., Manfredi, G., De Pisa, E.,Ehrlich, S., Giupponi, G., Amore, M., Tatarelli, R., Girardi, P., 2008. Periventricularwhite matter hyperintensities as predictors of suicide attempts in bipolar dis-orders and unipolar depression. Prog. Neuropsychopharmacol. Biol. Psychiatry32, 1501–1507.

Rüsch, N., Spoletini, I., Wilke, M., Martinotti, G., Bria, P., Trequattrini, A., Bonaviri,G., Caltagirone, C., Spalletta, G., 2008. Inferior frontal white matter volume andsuicidality in schizophrenia. Psychiatr. Res. Neuroimaging 164, 206–214.

Ryding, E., Ahnlide, J.A., Lindström, M., Rosén, I., Träskman-Bendz, L., 2006.Regional brain serotonin and dopamine transporter binding capacity in suicideattempters relate to impulsiveness and mental energy. Psychiatr. Res. Neu-roimaging 148, 195–203.

Schweighofer, N., Tanaka, S.C., Doya, K., 2007. Serotonin and the evaluation of futurerewards—theory, experiments, and possible neural mechanisms. Ann. N. Y. Acad.Sci. 1104, 289–300.

Soares, J.C., Mann, J.J., 1997. The anatomy of mood disorders—review of the structuralneuroimaging studies. Biol. Psychiatry 41, 86–106.

Soloff, P.H., Meltzer, C.C., Becker, C., Greer, P.J., Kelly, T.M., Constantine, D., 2003.Impulsivity and prefrontal hypometabolism in borderline personality disorder.Psychiatr. Res. Neuroimaging 123, 153–163.

Soloff, P.H., Price, J.C., Meltzer, C.C., Fabio, A., Frank, G.K., Kaye, W.H., 2007. 5HT2A

receptor binding is increased in borderline personality disorder. Biol. Psychiatry62, 580–587.

Tanaka, S.C., Doya, K., Okada, G., Ueda, K., Okamoto, Y., Yamawaki, S., 2004a. Differ-ent cortico-basal ganglia loops specialize in reward prediction at different timescales. Adv. Neural Inf. Process. Syst. 16, 701–708.

Tanaka, S.C., Doya, K., Okada, G., Ueda, K., Okamoto, Y., Yamawaki, S., 2004b. Pre-diction of immediate and future rewards differentially recruits cortico-basalganglia loops. Nat. Neurosci. 7, 887–893.

van Heeringen, C., Audenaert, K., Van Laere, K., Dumont, F., Slegers, G., Mertens, J.,Dierckx, R.A., 2003. Prefrontal 5-HT2a receptor binding index, hopelessness and

The functional neuroanatomy of mental pain in depression. Psychiatr. Res. Neu-roimaging 181, 141–144.

WHO, 2002. World Report on Violence and Health, 1st ed. World Health Organiza-tion, Geneva.