Content uploaded by Thomas Frodl
Author content
All content in this area was uploaded by Thomas Frodl on Nov 27, 2014
Content may be subject to copyright.
REVIEW
Brain structure and function in borderline personality disorder
Aisling O’Neill •Thomas Frodl
Received: 10 November 2011 / Accepted: 4 January 2012 / Published online: 18 January 2012
ÓSpringer-Verlag 2012
Abstract The spotlight on borderline personality disorder
(BPD) has been growing in recent years, with the number
of papers discussing potential causes and triggers of the
disorder rapidly on the increase. Also on the increase,
though still lacking sufficient numbers to produce well-
supported hypotheses, are studies employing neuroimaging
techniques as investigative tools in BPD. In this review, we
investigate the current state and findings of neuroimaging
studies in BPD, focusing in particular, on the studies
examining structural, functional, and neurometabolic
abnormalities in the disorder. Some suspected trends in the
data are highlighted, including reductions in the hippo-
campi and amygdalae of BPD patients compared to healthy
controls, exaggerated amygdala activity in BPD patients
when confronted with emotion-related stimulus, and neg-
ative correlations between increases in left amygdalar
creatine and reductions in amygdalar volume, reductions in
absolute N-acetylaspartate concentration in the dorsolateral
prefrontal cortex of BPD patients, and increases in gluta-
mate concentration in the anterior cingulate cortices of
BPD patients. We also discuss the limitations of some of
the current studies including hindrances due to sample
effects and techniques used and the potential of future
neuroimaging research in BPD.
Keywords Borderline personality disorder
Neuroimaging Structural PET fMRI
Neurometabolite Hippocampus Amygdala
A number of research articles have discussed potential
causes and factors which may lead to borderline personality
disorder (BPD), though none has of yet produced any con-
clusive evidence in support of a single-cause theory. In fact,
considering the heterogeneity of the disorder, it is more
likely that a combination of factors is involved in its mani-
festation; each to different degrees within individuals
(Asnaani et al. 2007; Wingenfeld et al. 2010).The main
postulated theories cite the experience of early life trauma
(e.g., childhood abuse or maternal separation), genetics,
neurobiological alterations, or a combination of the above as
being responsible, at least in part, along with external factors
such as environmental and psychosocial stressors, for the
onset of BPD (Goodman et al. 2004; Steele and Siever 2010).
A high correlation between childhood trauma and later
development of BPD has been shown in numerous studies
and it is generally accepted that such stressors are signifi-
cant to the onset of the disorder, though the mechanisms
involved are still under debate (Cohen et al. 2006). Abuse
(physical and/or sexual) and neglect are common child-
hood experiences of adults with BPD, so, more than with
other personality disorders. In one study, of the 358 BPD
patients participating, 91% reported experiences of child-
hood abuse and 92% reported experiences of childhood
neglect; much higher rates than those found amongst the
patients with other personality disorders (Zanarini et al.
1997). In a later study, Zanarini et al. (2002) found a sig-
nificant correlation between the severity of the reported
child abuse and/or neglect of 290 BPD patients and the
overall severity of the disorder.
A. O’Neill T. Frodl
Department of Psychiatry, Institute of Neuroscience,
Trinity College, Dublin 2, Ireland
T. Frodl (&)
Department of Psychiatry, Adelaide and Meath Hospital
Incorporating the National Children Hospital (AMNCH)
and St. James’s Hospital, Dublin, Ireland
e-mail: thomas.frodl@tcd.ie
123
Brain Struct Funct (2012) 217:767–782
DOI 10.1007/s00429-012-0379-4
Over the years, an increasing number of attempts have
been made to produce a more genetic explanation of BPD,
though the research is still far overshadowed by studies
examining the relationship of early life adversity and BPD.
A commonly discussed topic is the heritability of BPD,
which has produced a number of twin studies, sibling
environment/adoption studies, and self-report measures,
though the findings from this approach have been less than
consistent. Self-report measures, involving BPD patients,
usually support an argument for at least a partial herita-
bility, however, it is important to be mindful of the sub-
jective nature of these measures and note that their findings
do not always correspond with the relevant behavioural
data (Williamson 2007; Jacob et al. 2010). The study of
candidate genes is another popular genetic approach taken
in BPD research and has produced a number of genes of
particular interest. Genes currently under investigation
include the 7-repeat polymorphism of the dopamine D4
receptor (DRD4), which has been linked to disorganized
attachment, whilst the combined effect of the 7-repeat
polymorphism and the 10/10 dopamine transporter (DAT)
genotype has been linked to abnormalities in inhibitory
control, both noted features of BPD (Congdon and Canli
2008; Friedel 2004).
Previous BPD research has consisted mainly of studies
exploring the psychological aspects of the disorder from
characteristic behaviours to psychosocial triggers and risk
factors, with early life-stress proving to be strongly asso-
ciated with the occurrences of the disorder. In recent years,
however, appreciation has increased dramatically for the
neurobiological abnormalities which have been associated
with dimensions of personality dysfunction, with findings
from a number of studies confirming the biological
underpinnings of the disorder (Foti et al. 2010; Goodman
et al. 2004). As incidents of early life trauma are so highly
correlated with the occurrence of BPD, the disorder is quite
often classed as being on a spectrum of trauma-related
psychiatric disorders, of which post traumatic stress dis-
order (PTSD) is the core (Schmahl and Bremner 2006).
Neuroimaging research exploring neurobiological abnor-
malities in PTSD has provided the building blocks for
similar research in BPD, with the methodologies used in
PTSD neuroimaging studies such as volumetry of different
brain regions frequently transferred to studies in BPD;
these methodologies themselves often originating in other
better researched areas including Schizophrenia and Alz-
heimer’s disease (Schmahl and Bremner 2006).
The regions of most interest in BPD research are gen-
erally those areas associated with the regulation of normal
behaviours in healthy individuals, which are dysfunctional
in those with BPD (Brambilla et al. 2004). A number of
neuroimaging studies in BPD have reported findings of
reductions in regions of the brain involved in the regulation
of stress responses, emotion, and affect the hippocampus,
the orbitofrontal cortex, and the amygdala, amongst other
areas. Other studies have attempted to examine functional
abnormalities in these same regions either at baseline or
under emotion-inducing conditions, whilst a smaller num-
ber of studies have used magnetic resonance spectroscopy
to explore changes in the concentrations of neurometabo-
lites in certain brain regions of BPD patients, looking
specifically at neurometabolites such as N-acetylaspartate,
creatine, glutamate related compounds, and choline con-
taining compounds.
In this review, we will discuss neuroimaging research
relating to BPD thus far, considering the neurobiological
aspects of the disorder, discussing the previous findings
and their implications for future research.
Initially, a comprehensive literature search was con-
ducted to locate all studies which used neuroimaging
(structural MRI, PET, fMRI, MR-Spectroscopy) to study
individuals with a primary diagnosis of BPD compared to
healthy controls. PubMed, Science Direct, Web of Knowl-
edge and Scopus were searched for publications dating from
the databases’ inceptions through to January 2011 using the
keywords such as: ‘‘borderline personality disorder and
MRI’’, ‘‘borderline personality disorder and (PET or
fMRI)’’, ‘‘borderline personality disorder and spectros-
copy’’. The search was confined to English language arti-
cles. The following study features were used as grounds for
exclusion: Studies containing duplicated datasets (i.e.,
reported the same data in different manuscripts) were
excluded, though meta-analyses which included a number
of datasets were considered. Studies which employed only
screening instruments and not diagnostic tools to confirm
the diagnosis of BPD were also excluded. Two reviewers
(AO’N, TF) independently applied these exclusion criteria
to the databases. Inclusion criteria applied regulated the
reporting of participant demographics, insuring that the
included studies reported the participants with respect to
age, gender, duration of symptoms and the study setting.
Of the articles obtained, using the search term ‘‘bor-
derline personality disorder and structural MRI’’, we
carefully analysed 40, excluding ones those did not meet
the needed criteria. Twenty-two articles were found to be
appropriate for this review; articles that focused solely on
specific traits associated with the disorder, solely on
patients with comorbidities, or older articles whose results
had been updated in more recent studies, were excluded.
The search term ‘‘borderline personality disorder MRI
meta-analysis’’ produced two includable meta-analyses.
Using the search terms ‘‘(PET or fMRI) and borderline
personality disorder’’, 100 studies were found. Ten of them
met the inclusion criteria and are reported here; a large
majority of the returned articles did not relate to BPD and
were thus irrelevant, and the exclusion criteria for the
768 Brain Struct Funct (2012) 217:767–782
123
structural search were also employed here. For the search
terms MR-Spectroscopy and BPD, eight articles were
found, four of which are reported in this review, the rest
were excluded based on the previous criteria for the
structural and functional searches.
Structural brain changes
In volumetric studies, the two commonly used analytical
techniques are voxel-based morphometry (VBM) and
manual tracing. Manual tracing involves drawing the
regions of interest (be it the whole brain or its subparts) on
images obtained from brain scans and measuring the vol-
ume enclosed. It allows for the precise identification and
delineation of regions of interest (Irle et al. 2005), though it
can be time consuming and is of more use for larger areas.
In voxel-based morphometry, each brain is mapped to a
template then statistical tests are run across all voxels in the
image to identify differences between testing groups.
Though its results can be hard to validate, studies com-
paring the results of VBM to those of manual tracing or
visual measurements have found relatively good corre-
spondence (Whitwell 2009), though its validity when
dealing with atypical brains (such as those containing
severe pathologies) has been questioned, due to its core
features of normalisation and segmentation to a template
(Mechelli et al. 2005).
The most popular areas for volumetric research in BPD
usually include the neural circuits in the prefrontal cortex,
the limbic system, the anterior cingulate cortex (ACC),
caudate nucleus, the brain areas of the HPA axis, and
other related structures as abnormalities of these areas have
been associated to varying degrees with the psychopa-
thology of BPD. A summary of the reported alterations in
brain structure in BPD can be found in Tables 1and 2.
In general, the brain region which most consistently
showed alterations in studies of BPD patients is the hippo-
campus, with a number of studies finding significant hip-
pocampal volume reductions bilaterally in individuals with
BPD compared to healthy controls (Brambilla et al. 2004;
Schmahl et al. 2003; Tebartz van Elst et al. 2003). Interest-
ingly, Brambilla et al. (2004) found with further examination
that those BPD patients who had a history of childhood
abuse, when compared to healthy controls, still displayed
significant reductions in hippocampal volumes, whilst BPD
patients without such a history did not display significant
reductions. The researchers also found a significant negative
correlation between hippocampal volume of the BPD
patients with a history of childhood abuse and the length of
duration of the abuse (Brambilla et al. 2004). A study by
Driessen et al. (2000) reported similar findings with 16%
reductions in the hippocampal volumes of the BPD patients
being studied, compared to healthy controls, and a negative
correlation being observed between hippocampal volume
and duration of reported early trauma. In the Driessen et al.
(2000) study, however, the negative correlation was only
observed when the BPD and healthy control groups were
considered together. Finally, Zetzsche et al. (2007) also
found smaller grey-matter volume of the hippocampus in
BPD patients compared to controls, with this volume
decrease displaying a positive correlation with aggresive
behavior and number of the previous hospitalizations.
Studies examining volumetric changes in other areas
have been less consistent. With studies concerning the
amygdala, for example, some researchers have argued that
there are notable reductions in volume in patients with
BPD compared to the healthy controls (Schmahl et al.
2003; Tebartz van Elst et al. 2003), whilst others have
failed to show any differences (Brambilla et al. 2004; New
et al. 2007), with yet others reporting an increase in the
grey-matter of the amygdala of BPD patients (Minzenberg
et al. 2008). Driessen et al.’s (2000) study also examined
amygdala volumes in BPD patients, and reported a
reduction of 8% in the BPD patient volumes compared to
those of healthy controls. Another study compared cogni-
tion, hippocampal volumes, and amygdala volumes across
a group of trauma-exposed women with BPD (some with
and some without comorbid PTSD) and a group of healthy
controls. The results showed decreases in both the hippo-
campus and amygdala volumes of the BPD women with
PTSD compared to healthy controls (12 and 33% respec-
tively), and also found cognition to be significantly
impaired in this same group compared to the healthy
controls (Weniger et al. 2009). The hippocampus and
amygdala volumes of the BPD-women without PTSD were
also found to be significantly reduced compared to healthy
controls (11 and 22% respectively), though no significant
differences in cognition were found between these two
groups (Weniger et al. 2009). No significant volumetric
differences were observed between the patients with and
without PTSD, though the authors suggested that a larger
patient-sample group may have shown significant differ-
ences in amygdala size between the two patient groups
(Weniger et al. 2009). An interesting set of results which
contradicted these findings of BPD amygdalar abnormali-
ties found no significant differences in amygdala sizes
between a BPD sample group and a healthy control group
(Zetzsche et al. 2006). However, this same study did find a
significant difference in the bilateral amygdala sizes of
BPD patients with and without comorbid major depression
(MDD). The amygdala volumes of the patients with MDD
were found to be significantly larger than those of the
patients without MDD, and a positive correlation was
found between the volume of the left amygdala of the total
BPD group and depressive symptoms (as measured by the
Brain Struct Funct (2012) 217:767–782 769
123
Table 1 Neurobiological structural changes seen in borderline personality disorder, and sample characteristics of each study
Study Method Sample characteristics; BPD/HC Main results
Size Age Gender Handedness Volume reduction
in BPD
Other results
Driessen et al. (2000)
a
Manual
tracing
21/21 29.9 ±6/29.3 ±6.7 21 (100)/21 (100)
female
18(86)/18(86)
RH
Amygdala, hippocampus
Rusch et al. (2003)
b
VBM 20/21 29.3 ±3.9/28.4 ±6.4 20(100)/21(100)
female
Data
unavailable
Grey matter of left amygdala No reductions in grey
matter of any other
area
Schmahl et al. (2003)
c
Manual
tracing
10/23 27.4 ±7.1/31.5 ±8 10 (100)/23 (100)
Female
10 (100)/20 (87)
RH
Amygdala, hippocampus
Tebartz van Elst et al.
(2003)
d
Manual
tracing
8/8 33.5 ±6.3/30.5 ±5.1 8 (100)/8 (100)
Female
Data
unavailable
Amygdala, hippocampus, left
OFC, right ACC
Brambilla et al.
(2004)
e
Manual
tracing
10/20 29.2 ±9.25/data unavailable 6 (60) female/data
unavailable
Data
unavailable
ACC, hippocampus No volume reduction of
amygdala, caudate,
temporal lobes,
DLPFC
Hazlett et al. (2005)
f
Manual
tracing
50/50 33.2 ±8.5/31.5 ±9.9 23 (46)/20 (40) female Data
unavailable
ACC, posterior cingulate cortex No volume reduction of
whole cingulate,
frontal lobe
Zetzsche et al. (2006)
g
Manual
tracing
25/25 26.1 ±7.1/27.2 ±6.3 25 (100)/25 (100)
female
25 (100)/25
(100) RH
No difference in
amygdala volumes
between whole BPD
group and HC group.
Significantly larger
amygdala volumes in
BPD patients w.
comorbid MDD
compared to BPD
patients w/o MDD
New et al. (2007)
h
Manual
tracing
26/24 BPD males: 35.7 ±7.9
BPD females: 30.7 ±8.6/HC
males: 31.7 ±7.9
HC females: 34 ±11.2
9 (34.6)/9 (37.5)
female
19(73)/19(79)
RH
No volume reduction of
amygdala
Tebartz van Elst et al.
(2007)
i
Manual
tracing
12/10 27.7 ±5.5/26.9 ±4.4 12 (100)/10 (100)
female
Data
unavailable
Amygdala Non-signif. reductions
in hippocampus and
OFC volumes
Zetzsche et al. (2007)
j
Manual
tracing
25/25 26.1 ±7.1/27.2 ±6.3 25 (100)/25 (100)
female
25 (100)/25
(100) RH
Grey matter of hippocampus
Chanen et al. (2008)
k
Manual
tracing
20/20 17.3 ±1.1/19 ±2.2 15 (75)/15 (75)
Female
18 (90)/18 (90)
RH
Right OFC No volume reduction of
amygdala,
hippocampus
770 Brain Struct Funct (2012) 217:767–782
123
Table 1 continued
Study Method Sample characteristics; BPD/HC Main results
Size Age Gender Handedness Volume reduction
in BPD
Other results
Minzenberg et al.
(2008)
l
VBM 12/12 30.3 ±8/
30.7 ±10
5 (41.6)/6 (50)
female
Data unavailable Grey matter of ACC Increased grey matter of
amygdala
Soloff et al. (2008)
m
VBM 34/30 27.5 ±8/
25.6 ±7.7
22 (64.7)/19 (63.3)
female
Data unavailable (BPD women) medial temporal lobe (incl.
Amygdala)
(BPD men) grey matter of ACC
(BPD men) no
reductions in medial
temporal lobe
(BPD women) no
reductions in grey
matter of ACC
Schmahl et al.
(2009)
n
Manual
tracing
25/25 29 ±6.7/
28.5 ±6.06
25 (100)/25 (100)
female
25 (100)/25 (100) RH (w. comorbid PTSD) hippocampus (w/o comorbid PTSD)
no reduction in
hippocampus volume
no volume reduction for
amygdala in either
patient group
Takahashi et al.
(2009b)
o
Manual
tracing
Sample characteristics previously described by Chanen et al. (2008). No signif. difference in
insular cortex vols.
between BPDs and
HCs
Takahashi et al.
(2009a)
o
Manual
tracing
Sample characterstics previously described by Chanen et al. (2008). No differences in cavum
septum pellucidum
between BPD and HC.
Significantly shorter
adhesio interthalamica
in BPD group,
significantly larger 3
rd
ventricle in BPD
Vollm et al. (2009)
o
VBM 7/6 17.3 ±1.1/
19 ±2.2
7 (100)/6 (100) Male 7 (100)/6 (100) RH Grey matter volumes in frontal, temporal,
parietal cortices
Weniger et al.
(2009)
p
Manual
tracing
24/25 BPD w. PTSD:
32 ±7
BPD w/o PTSD:
32 ±5
HC: 33 ±7
24 (100)/25 (100)
Female
BPD w. PTSD: 10
(100)
BPD w/o PTSD: 12
(85.7)
HC: 25(200)RH
Amygdala, hippocampus No signif. differences in
vol. between patients
with/without PTSD
Whittle et al.
(2009)
q
Manual
tracing
Sample characterstics previously described by Chanen et al. (2008). Left ACC
Brain Struct Funct (2012) 217:767–782 771
123
Table 1 continued
Study Method Sample characteristics; BPD/HC Main results
Size Age Gender Handedness Volume reduction
in BPD
Other results
Brunner et al. (2010)
q
VBM 20/20 CC: 20 16.7 ±1.6/16.8 ±1.2
CC: 16 ±1.3
20 (100)/20 (100)
CC: 20 (100)
20 (100)/20 (100)
CC: 20( 100) RH
DLPFC, left OFC No differences found
between BPD group
and clinical control
group
BPD and HC data presented as mean ±SD BPD/HC or number (percentage) BPD/HC
BPD Borderline personality disorder, HC Healthy controls, CC Clinical controls, VBM Voxel-based morphometry, RH Right handed, MDD Major depressive disorder, PTSD Post traumatic
stress disorder, OFC Orbitofrontal cortex, ACC Anterior cingulate cortex, DLPFC Dorsolateral prefrontal cortex, signif. significant, w.with, w/o without
a
All patients had been drug free for at least 7 days at the time of assessment. Nine patients had used psychotropic medications during the 7–1 weeks prior to the assessment. Data regarding
psychotherapeutic treatments were unavailable
b
All patients were free of psychotropic medication for at least the 2 weeks prior to the assessment. Data regarding psychotherapeutic treatments were unavailable
c
Nine of the patients were being treated with psychotropic medication at the time of assessment. Data regarding psychotherapeutic treatments were unavailable
d
All patients were free of psychotropic medication for at least the 2 weeks prior to the assessment. Data regarding psychotherapeutic treatments were unavailable
e
All patients were free of psychotropic medications for at least the 2 months prior to the assessment. Six of the patients had been treated with psychotropic medication in the past, the other four
were drug naı
¨ve. Data regarding psychotherapeutic treatments were unavailable
f
All patients were free of psychoactive medications for at least 6 weeks prior to the assessment. Fourty-three of the 50 patients were drug naı
¨ve. Data regarding psychotherapeutic treatments
were unavailable
g
Twenty of the patients were taking psychotropic medications at the time of the assessment. Nineteen of the patients had been treated with psychotropic medications in the past. Five of the
patients were not being treated with any medications at the time of the assessment, three of whom were drug naı
¨ve. Data regarding psychotherapeutic treatments were unavailable
h
All patients were free of psychotropic medications for at least the 6 weeks prior to the assessment. Twenty-two of the 26 patients were drug naı
¨ve. Data regarding psychotherapeutic
treatments were unavailable
i
All patients were free of psychoactive medications for at least the 2 weeks prior to the assessment
j
Twenty of the patients were taking psychotropic medications at the time of the assessment. Nineteen of the patients had been treated with psychotropic medications in the past. Five of the
patients were not being treated with any medications at the time of the assessment, three of whom were drug naı
¨ve. Data regarding psychotherapeutic treatments were unavailable
k
The patients included were teenagers with first presentation BPD. At the time of the assessment 17 of the patients were unmedicated, whilst 3 were being treated with psychotropic
medications. One patient had been treated with psychotropic medication in the past. Sixteen of the patients had received non-specialised counsellingor psychotherapy in the past
l
All patients were unmedicated at the time of the assessment. Seven of the patients were drug naı
¨ve, whilst five had been free of psychotropic medications for the duration of 4 months to
10 years prior to the assessment. Data regarding psychotherapeutic treatments were unavailable
m
All patients were free of psychoactive medications for the 2–6 weeks prior to the assessment. Data regarding psychotherapeutic treatments were unavailable
n
All patients were free of psychotropic medications for at least the 2 weeks prior to the assessment. Data regarding psychotherapeutic treatments were unavailable
o
All patients were unmedicated at the time of the assessment. Data regarding psychotherapeutic treatments were unavailable
p
Seven of the patients were taking psychotropic medications at the time of the assessment. Data regarding psychotherapeutic treatments were unavailable
q
The patients included were teenagers with first presentation BPD. None of the patients had taken psychotropic medications prior to their current admission to hospital. Nine of the BPD
patients were taking psychotropic medication at the time of the assessment. Four of the CC patients were taking psychotropic medications at the time of the assessment. Data regarding
psychotherapeutic treatments were unavailable
772 Brain Struct Funct (2012) 217:767–782
123
Hamilton rating scale for depression) (Zetzsche et al.
2006).
The ACC is also known to play a role in the regulation
of emotion and response inhibition (Wingenfeld et al.
2010), and the results of studies investigating volumetric
abnormalities in the region, though few in number, have
been relatively consistent. In a 2003 study by Tebartzan
van Elst et al. (2003) a reduction in ACC volume was
reported in BPD patients compared to the controls. This
finding was supported by a number of later studies which
observed significant reductions in grey-matter in the ACC
of BPD patients (Hazlett et al. 2005; Minzenberg et al.
2008). A 2009 investigation of ACC volumes in adoles-
cents with first presentation BPD and minimal exposure to
treatment found a decrease in left ACC volume of the
patient group compared to healthy controls (Whittle et al.
2009). In addition, the paper also reported a negative cor-
relation between left ACC volume and parasuicidal
behaviours and a positive correlation between left ACC
volume and impulsivity in the patient group (Whittle et al.
2009).
Alterations in glucose metabolism and brain
oxygenation
Other studies examining dysregulation in neural systems
implicated in BPD have used techniques such as positron
emission tomography (PET) or functional magnetic reso-
nance imaging (fMRI) to investigate the activity in certain
areas of interest. In PET studies, a positron emitting
nucleotide is introduced into the body on a biologically
active molecule, its activity in the brain is monitored, and
this information is used to establish data about brain
metabolism, blood flow, or receptor binding. FDG-PET is
the most commonly used form of PET and uses FDG, a
glucose analogue, to measure the rates of regional glucose
uptake. fMRI studies are less-invasive than PET and are
thus becoming a more popular technique, using magnetic
fields and radio frequencies to measure changes in blood
oxygenation which can be used as an indirect measure of
neural activity, blood flow, or volume. Using PET, De la
Fuente et al. (1997) found significant reductions in resting
state glucose metabolism in the premotor areas and the
dorsolateral prefrontal cortex (DLPFC), parts of the ACC
(BA25), as well as the thalamus, caudate and lenticular
nuclei of BPD patients compared to healthy controls.
Interestingly, the BA25, which is connected to the amyg-
dala and hypothalamus, is considered to be an important
centre in the neural circuitry of depression and reductions
in volume of BA25 are suspected to contribute to an
increased risk for depression (Insel 2010). However, a later
resting state PET study by Juengling et al. (2003) produced
results contradictory to those of the De la Fuente study,
reporting glucose hypermetabolism in the frontal and pre-
frontal cortices in patients with BPD relative to controls as
well as significant hypometabolism in the hippocampus
and cuneus, at baseline. Still another study found results
which again varied from those of De la Fuente and Juen-
gling’s studies. Though said study did find glucose hypo-
metabilsm at resting state in the frontal lobe, similar to the
findings of De la Fuente et al., significant hypermetabolism
was also seen in the motor cortex, the medial cingulate
cortex, the ACC, the occipital lobe, and the temporal pole,
amongst other regions, a finding not shown in either of the
other FDG-PET studies mentioned (Salavert et al. 2011).
Though mentioned earlier under the structural findings, the
major findings of the New et al. (2007) related to dys-
functional connectivity between frontal brain areas and the
amygdala in the brains of BPD patients. Using PET, the
researchers found that in the HC group, there was a strong
correlation between activity in the orbitofrontal cortex
(OFC) and activity in the right ventral amygdala (New
et al. 2007). In the BPD group, however, these same sig-
nificant correlations between OFC and amygdala activity
were absent (New et al. 2007). However, little research has
been done to expand on these results.
The results of the fMRI studies, on the other hand, have
been more consistent than those of the FDG-PET studies.
The most common finding amongst these studies is that of
exaggerated activity in the amygdala of patients with BPD
compared to controls during procedures that involve the
processing of emotionally aversive stimuli (Donegan et al.
2003a; Herpertz et al. 2001; Minzenberg et al. 2007a). Of
particular interest is an emotional linguistic study by
Silbersweig et al. (2007), which examined the brain func-
tion of BPD patients compared to controls under conditions
associated with the interaction between behavioural inhi-
bition and negative emotion. In the study, a significant
reduction in activity was seen in the ventromedial
Table 2 Meta analyses of studies examining neurobiological structural changes seen in borderline personality disorder
Study Method Sample BPD/HC Main results
Volume reduction Other results
Hall et al. (2010) MRI – Hippocampus, amygdala No differences in whole brain volume
Nunes et al. (2009) – 104/122 Hippocampus, amygdala
Brain Struct Funct (2012) 217:767–782 773
123
prefrontal cortex (including the medial orbitofrontal cortex
and subgenual ACC) in the BPD patients compared to
controls. A strong correlation was found between
decreasing ventromedial prefrontal activity and decreasing
constraint (impulsivity), whilst a strong correlation was
also found for increased amygdalar ventral striatal activity
and increased negative emotions in the patients (Sil-
bersweig et al. 2007); impulsivity and the experience of
exaggerated negative emotions being diagnostic features
of the disorder. A summary of the functional findings of
studies in BPD can be found in Table 3.
Neurometabolites
Though the number of studies examining structural and
functional abnormalities in BPD is increasing, revealing
more about the dysfunctional frontolimbic brain regions
involved in the neuropathology of the disorder, only a
handful of studies have examined the concentration of
neurometabolites in the brains of patients with BPD.
Magnetic resonance spectroscopy (MRS) is a technique
related to MRI which allows for the study of metabolic
changes in diseases affecting the brain and other organs. It
is being used with increasing frequency in brain research as
it is the only method that allows non invasive in vivo
observation of various neurometabolites, including gluta-
mate (Glu), glutamine (Gln), phosphocreatine (PCr) and
creatine (Cr) (total creatine [tCr]), N-acetylaspartate (total
NAA [tNAA]), and choline-containing compounds (tCho).
Studies examining alterations in NAA concentration have
yet to yield consistent results, though the limited number of
studies is a major obstacle in achieving this goal. The
earliest study of NAA concentrations in BPD, by van Elst
et al. (2001), found a significant 19% reduction of absolute
NAA concentration in the DLPFC of BPD patients com-
pared to controls. Reductions in NAA have been shown to
reflect a state of neuronal damage that often precedes cell
death, thus NAA concentrations are of particular interest as
they could be indicative of neuronal integrity (van Elst
et al. 2001) The researchers also reported non-significant
reductions in striatal Cr concentrations, though no differ-
ences were found in frontal or striatal NAA/Cr and Cho/Cr
ratios. They explained the lack of difference in neuro-
chemical ratios between groups as being due to the non-
significant reductions seen in Cr concentrations; positing
that, contrary to the previous assumptions, in fact, there is
limited usefulness in measuring the relative concentrations
of NAA/Cr and Cho/Cr as the concentration of Cr is not
constant (van Elst et al. 2001). Consequently, a decrease in
both NAA and Cr concentrations may result in false neg-
ative findings. Additional findings of reductions in tNAA
and tCr were produced in a BPD study of the amygdala by
Hoerst et al. (2010b), further opposing the assumption that
Cr is a constant term and providing further evidence sup-
porting a role for tNAA reductions in the mechanisms
underlying the disorder.
In a 2007 pilot study by Tebartz van Elst et al. (2007),
not only was further evidence produced to counter the
traditional view of Cr as a stable neurometabolite, the study
was also the first to discovered a relationship between
neurometabolic abnormalities and structural abnormalities
in a brain-region associated with BPD. The study found a
significant 11–17% reduction in amygdalar volume of
patients with BPD accompanied by a 17% increase in the
Cr concentration of the left amygdala. The Cr concentra-
tion was found to be negatively correlated with amygdala
volume, but positively correlated with measures of anxiety.
The nature of the relationship between these volumetric
and neurometabolic abnormalities, and any others that have
yet to be shown in BPD, are still unclear, though Tebartz
van Elst et al. (2007) suggest that it is likely to be linked to
the pathology and neurochemical abnormalities of the
disorder, specifically in terms of Cr-related processes. PCr
and Cr levels make up the total Cr visible in MRS, and both
play important roles in the phosphate bound energy
metabolism which is essential in the brain. There is evi-
dence to suggest that Cr and PCr/Cr kinase system are
involved in neuronal growth cone activity and axonal
elongation. From this perspective, increases in the Cr sig-
nal in the amygdala could reflect disturbances in local
energy metabolism which in turn may result in amygdalar
volume loss and consequently irregularities in the neuronal
network for affect regulation (Tebartz van Elst et al. 2007).
Alternatively, the observed increases in Cr concentrations
may be a coping strategy of the brain to compensate for a
structurally compromised emotional information process-
ing network, though not enough research exists to confirm
either theory yet (Tebartz van Elst et al. 2007). However,
the findings of the Tebartz van Elst et al. study clash with
those of the previously mentioned Hoerst et al. study,
which not only found an opposing decrease in tCr in the
amygdala but also failed to find a significant correlation
between the neurochemical concentrations and psycho-
metric measures taken. Recently, another separate study by
Hoerst et al. (2010a) found significantly elevated levels of
Glu in the ACCs of patients with BPD compared to that of
controls. Furthermore, the levels of Glu were found to be
positively correlated with the measures of impulsivity
irrespective of diagnosis, whilst other positive correlations
were found between Glu concentrations and levels of dis-
sociation, as well as Glu concentrations and severity of
symptoms within the patient group. It is worth noting,
however, that similar neurochemical abnormalities have
been found in the studies of non-human primates exposed
to early life stress and in studies of adolescents with PTSD,
774 Brain Struct Funct (2012) 217:767–782
123
Table 3 Neurobiological functional abnormalities seen in borderline personality disorder, and sample characteristics of each study
Study Method Sample characteristics; BPD/HC Patient State Results
Size Age Gender Handedness
De la Fuente
et al.
(1997)
a
FDG-PET 10/15 34.2 ±7.2/30.7
b
8 (80)/7
(46.6)
female
10 (100)/15
(100) RH
Resting BPD group showed signif. hypometabolism in the premotor
area, DLPFC, parts of the ACC, thalamus, caudate, and
lenticular nuclei
Herpertz
et al.
(2001)
c
fMRI 6/6 26.2 ±8.1/
27.2 ±4.5
6 (100)/6
(100)
female
6 (100)/6
(100) RH
Viewing emotionally aversive slides BPD group showed activation bilaterally of the amygdala,
medial and inferolateral prefrontal cortex, and fusiform
gyrus
Donegan
et al.
(2003b)
d
fMRI (only
amygdala)
15/15 35 ±11.7/
34.9 ±10
13 (86)/9
(60) female
15 (100)/15
(100) RH
Viewing facial expressions of
emotion
Significantly greater left amygdala activation in the BPD
group
Juengling
et al.
(2003)
e
FDG-PET 12/12 25 ±4/30 ±9 12 (100)/12
(100)
female
Data
unavailable
Resting BPD group showed signif. hypermetabolism in the frontal
and prefrontal cortices, and significant hypometabolism in
the left hippocampus and cuneus
Minzenberg
et al.
(2007b)
f
fMRI 12/12 30.3 ±8/
30.7 ±10
5 (41.6)/6
(50) female
Data
unavailable
Viewing facial expressions of
emotion and making gender
discriminations
In relation to fear stimuli the BPD group showed
exaggerated activation in the amygdala, impaired
activation of ACC
New et al.
(2007)
g
FDG-PET 26/24 BPD males:
35.7 ±7.9
BPD females:
30.7 ±8.6/HC
males:
31.7 ±7.9 HC
females:
34 ±11.2
9 (34.6)/9
(37.5)
female
19 (73)/19
(79) RH
Resting and placebo or resting and
m-CPP
Placebo HC: signif. pos. correl. b/w right OFC and ventral
amygdala
Placebo BPD: weak correl. b/w amygdala and anterior PFC
m-CPP HC: pos. correl. b/w OFC and amygdala regions
m-CPP BPD: pos. correl. b/w DLPFC and amygdala
Silbersweig
et al.
(2007)
h
fMRI 16/14 31.25
i
/23.8
i
15 (93.7)/10
(71.4)
female
15 (93.7)/12
(85.7) RH
Performing emotional linguistic go/
no go task
BPD group showed decrease in ventromedial prefrontal
activity
Koenigsberg
et al.
(2009a)
j
fMRI 18/16 32.6 ±10.4/
31.8 ±7.7
10 (55.5)/9
(56.2)
female
17 (94.4)/14
(87.5) RH
Distancing from/Casual viewing of
images depicting social
interactions
Whilst viewing negative images the BPD group showed
less signal change in the dorsal ACC and intraparietal
sulcus, less deactivation in the amygdala, and increased
activation in the superior temporal sulcus and superior
frontal gyrus
Koenigsberg
et al.
(2009b)
k
fMRI 19/17 34.9 ±11.1/
31.2 ±10.6
7(36.8)/
8(47.05)
female
14(73.6)/
15(88.2)
RH
Viewing emotion inducing images Whilst viewing negative social emotional images, the BPD
group showed exaggerated activation in the amygdala and
visual processing regions
New et al.
(2009)
l
FDG-PET 38/36 30.5 ±8.5/
28.4 ±7.1
16(42.1)/
18(50)
female
32(84.2)/
32(88.8)
RH
Performing aggression provoking
task
BPD group showed hypermetabolism in OFC and
amygdala during provocation; hypometabolism in
anterior, medial, and prefrontal regions during
provocation
Brain Struct Funct (2012) 217:767–782 775
123
Table 3 continued
Study Method Sample characteristics; BPD/HC Patient State Results
Size Age Gender Handedness
Salavert et al. (2011)
m
FDG PET 8/8 35.5 ±9.27/32 ±7.86 6 (75)/5 (62.5)
female
8 (100)/8 (100) RH Resting BPD group showed hypometabolism in the frontal lobe;
hypermetabolism in the motor cortex, medial and anterior
cingulus, occipital lobe, temporal pole, left superior
parietal gyrus and right superior frontal gyrus
BPD and HC data presented as mean ±SD BPD/HC or number (percentage) BPD/HC
BPD Borderline personality disorder, HC Healthy controls, fMRI Functional magnetic resonance imaging, FDG-PET Fludeoxyglucose positron emission tomography, RH Right handed, m-CPP
Meta-chloropiperazine, OFC Orbitofrontal cortex, PFC Prefrontal cortex, ACC Anterior cingulate cortex, DLPFC Dorsolateral prefrontal cortex, signif. significant, correl. correlation, pos.
positive, b/w between
a
SD data unavailable for HC group
b
All patients were free of psychoactive medications for the 2–6 weeks prior to the assessment. Data regarding psychotherapeutic treatments were unavailable
c
All patients were free of psychoactive medications at the time of the assessment. Data regarding psychotherapeutic treatments were unavailable
d
Eleven of the patients were taking psychotropic medications at the time of the assessment. Data regarding psychotherapeutic treatments were unavailable
e
All patients were free of psychotropic medications for at least the 4 weeks prior to the assessment. Data regarding psychotherapeutic treatments were unavailable
f
All patients were unmedicated at the time of the assessment. Seven of the patients were drug naı
¨ve, whilst the remaining five had discontinued treatments with psychotropic medications from
4 months to several years before the assessment. Data regarding psychotherapeutic treatments were unavailable
g
All patients were free of psychotropic medications for at least the 6 weeks prior to the assessment. Twenty-two of the 26 patients were drug naı
¨ve. Data regarding psychotherapeutic
treatments were unavailable
h
Eleven of the patients were taking psychotropic medications at the time of the assessment. Data regarding psychotherapeutic treatments were unavailable
i
SD data unavailable for group
j
All patients were free of psychotropic medications for at least the 2 weeks prior to the assessment. Data regarding psychotherapeutic treatments were unavailable
k
All patients were free of psychotropic medications for at least the 2 weeks prior to the assessment. Data regarding psychotherapeutic treatments were unavailable
l
All patients were free of psychotropic medications for at least the 2 months prior to the assessment. Data regarding psychotherapeutic treatments were unavailable
m
All patients were free of psychotropic medications for at least the 1 month prior to the assessment. Data regarding psychotherapeutic treatments were unavailable
776 Brain Struct Funct (2012) 217:767–782
123
suggesting that these abnormalities are related to the
experiences of trauma and are not BPD specific (Mathew
et al. 2003). A summary of the above findings in neu-
rometabolite studies can be found in Table 4.
Discussion
The articles examined here, though not always in agree-
ment with each other, have produced some interesting
results which shall surely bring us closer to uncovering the
mechanisms underlying BPD. In relation to the studies of
structural abnormalities in patients with BPD, the cited
meta-analyses both concluded that there were significant
reductions in the volumes of the amygdalae and hippo-
campi of the BPD patients analysed. Despite this, indi-
vidual studies appear to have produced more varied results,
e.g., of the 11 studies which examined amygdala volume 6
found significant reductions in volume for the patient
group, 4 found no significant differences between patient
group and control, and 1 found a significant increase in
grey matter of the amygdalae of the patient group. More
difficulties arise with the amygdala in particular as Min-
zenberg et al. (2008) pointed out that a change in either
direction could in theory be linked to different features of
the disorder; for example a smaller amygdala in patients
with BPD could potentially provide an explanation for the
emotional dysregulation associated with the disorder,
whilst an increase in grey-matter concentration could
account for the exaggerated responses to emotional stim-
ulus seen in the amygdala. Despite the reductions in the
amygdalae of BPD patients being a somewhat controver-
sial finding, and many reports of similar amygdalar
reductions existing for other trauma-related disorders, it
Table 4 Neurometabolite changes seen in borderline personality disorder
Study Sample characteristics; BPD/HC MRS details Findings
Size Age Gender Handedness
van Elst
et al.
(2001)
a
12/14 31.6 ±7.1/
30.1 ±3.8
12 (100)/
14 (100)
female
Data
unavailable
2.0 Tesla
2cm
3
voxel size
Std. quadrature head
coil
19% reduction in NAA conc. in DLPFC
Non-significant reductions of 15% in frontal and
16% in striatal Cr conc
NAA/Cr and Cho/Cr ratios showed no differences
Tebartz van
Elst et al.
(2007)
b
12/10 27.7 ±5.5/
26.9 ±4.4
12(100)/
10(100)
Female
Data
unavailable
2.0 Tesla
1.5 cm
3
voxel size
Std. quadrature head
coil
11–17% reductions in amygdalar volume
17% increase in left amygdalar Cr
Left amygdalar Cr conc. pos. correlated with
measures of anxiety
Left amygdalar Cr conc. neg. correlated with
amygdalar volume
Hoerst et al.
(2010a)
c
30/30 29.33 ±7.6/
28.6 ±8
30 (100)/
30 (100)
Female
30 (100)/30
(100) RH
3.0 Tesla
15 920 912 mm
3
voxel size
12 channel receive
only head coil
Increased Glu conc. in ACC
Pos. correlation between Glu conc. and impulsivity
(independent of BPD diagnosis)
Neg. correlation between Glu conc. and BPD
severity (within patient group)
Hoerst et al.
(2010b)
d
21/20 27.24 ±5.5/
28.55 ±8.7
21 (100)/
20 (100)
Female
21 (100)/20
(100) RH
3.0 Tesla
12 910 912 mm
voxel size
12 channel receive
only head coil
Reduction in tNAA and tCr in left amygdala
BPD patients with comorbid PTSD showed lower
levels of tCr compared to Single diagnosis BPD
patients and controls
No significant correlation between neurochemical
conc. and psychometric measurements
BPD Borderline personality disorder, HC Healthy controls, RH Right handed, DLPFC Dorsolateral prefrontal cortex, NAA N-acetylaspartate,
tNAA Total N-acetylaspartate, Cr Creatine, tCr Total creatine, Cho Choline, Glu Glutamate, conc. concentration, neg. negative, pos. positive,
w.with, w/o without
a
All patients were unmedicated at the time of the assessment. All patients were receiving a specialised treatment for BPD which followed the
principles of dialectic behavioural therapy
b
All patients were free of psychoactive medications for at least the 2 weeks prior to the assessment. Data regarding psychotherapeutic
treatments were unavailable
c
All patients were free of psychotropic medications for at least the 3 months prior to the assessment. Data regarding psychotherapeutic
treatments were unavailable
d
All patients were free of psychotropic medications for at least the 14 days prior to the assessment. Data regarding psychotherapeutic treatments
were unavailable
Brain Struct Funct (2012) 217:767–782 777
123
has been suggested that, if it was possible to consistently
find such abnormalities, reductions in the amygdala vol-
umes of BPD patients could be a stronger feature of the
disorder than other structural abnormalities and could
provide important markers for intervention and treatment
(Weniger et al. 2009). In addition, although amygdala
abnormalities are not specific to BPD, observations of
differences in the magnitude of such reductions between
BPD groups with and without severe depressive symptoms,
comorbid MDD, and other disorders may prove to be
useful when distinguishing subtypes of the disorder
(Zetzsche et al. 2006).
Overall, the brain region most consistently found to
display alterations in BPD patients is the hippocampus. The
hippocampus plays an important role in memory consoli-
dation, declarative memory, and is highly sensitive to the
effects of stress, with stress-related increases in glucocor-
ticoid levels being associated with smaller hippocampal
volumes in animal studies (Brambilla et al. 2004). It has
been suggested that volumetric reductions of the hippo-
campus, the most frequently produced result in human
studies, may lead to the neurocognitive deficits, dissocia-
tive symptoms, perceptual distortions, and identity insta-
bility seen in BPD patients (Brambilla et al. 2004).
Although the reductions in regions of the brain known to
play important roles in emotional regulation, processing,
and other functions usually impaired in individuals with
BPD are largely accepted, it is worth noting again that
reductions in the hippocampus, amygdala, and ACC are not
specific to BPD. Reductions in these areas have also been
shown in trauma-exposed individuals, both with and
without psychiatric disorders (most commonly PTSD and
MDD) (Cohen et al. 2006; Macqueen and Frodl 2010; Karl
et al. 2006), and in the same neural structures of non-
human primates that have experienced early life stress
(Bremner 1999; Bremne and Vermetten 2001; Cohen et al.
2006). However in the human studies, the magnitude of
these reductions is generally greater in those with the
psychiatric disorders than in those without. As these
structural abnormalities are not specific to BPD, Wingen-
feld et al. (2010) suggest that these findings in BPD
patients support the theory that early life stress does indeed
have a damaging effect in certain brain regions. However,
the exact cause of the volume reductions observed and
whether the high incidence of BPD patient comorbidity
with PTSD and MDD is due predominantly to trauma-
related aspects of the disorder remain to be seen. It is
important to also acknowledge at this point that the depth
of the interaction between traumatisation and other factors
such as familial/genetic factors, environment, and phar-
macological intervention on long-term neurobiological
changes as yet is not well understood (Bremne and Ver-
metten 2001; Kaufman and Charney 2001). The different
techniques used in the studies (i.e., VBM or manual trac-
ing) did not appear to affect the outcome, though other
methodoligical issues which may have influenced the
results of not only the volumetric studies but also the
functional and spectroscopic studies include the large
disparity seen between sample sizes; gender differences, of
which very little is understood; and the comorbidities of the
participant samples, with some samples consisting of
patients suffering from other disorders in addition to BPD,
which although being true to the general BPD population,
may produce misleading results; whilst other samples
consist of individuals with a sole diagnosis of BPD, which
may not accurately reflect the general BPD population.
Another potential issue worth noting regarding sample
characteristics is the age of the participants. As the brain
goes through maturation, it is known to undergo both
increases and decreases in the white and grey matter of
various areas continuing into the early 20s (Giedd et al.
1999; Paus 2005). Of the studies cited, only the samples of
the Brunner et al. (2010) and Chanen et al. (2008) studies
consisted solely of teenagers with first presentation BPD
(although both the Takahashi et al. studies and the Whittle
et al. study used the same sample as Chanen et al.). The
findings of these adolescent studies were generally in
keeping with those of the adult-sample studies, thus there
does not appear to be any pressing anomalies in the BPD
brain attributable to age or level of brain maturation.
Limitation of this matter is the lack of longitudinal studies
in BPD. An increase of such studies is vital for the
researchers to gain a fuller understanding not only of the
potential effect of brain immaturity on adolescent BPD but
also the long-term effects of the progression of the disorder
and of treatment on the adult brain.
An as yet unexplored explanation of the volumetric
reductions seen in particular brain areas of BPD patients
involves dysregulation on a cellular level. Support for this
theory includes studies which have shown that both chronic
stress and peripheral chronic inflammation are linked to the
down regulation of hippocampal-brain derived neurotro-
phic factor (Karl et al. 2006); a growth factor found in high
concentrations in the hippocampus, amongst other areas. In
addition, increases in inflammatory mediators of the
immune system, linked to chronic dysregulation of the
hypothalamic–pituitary–adrenal axis (HPA axis), which is
itself linked to chronic stress, have been found to increase
atherosclerotic processes in blood vessels (Karl et al.
2006). In turn, these processes can lead to hypertension
which has also been linked to reductions in hippocampal
volume (Wiseman et al. 2004). Studies in PTSD, which
like BPD is strongly related to stressful and traumatic
events, have investigated a potential cellular basis for the
disorder and have found associations between the disorder
and alterations in immune system and cardiovascular
778 Brain Struct Funct (2012) 217:767–782
123
function (Karl et al. 2006). In addition, experimental stress
was shown to result in depressive-like behavior and neu-
ronal changes including atrophy of neurons and downreg-
ulation of neurogenesis (Duman 2002). It is possible that
similar findings could be made in BPD research, however,
such an approach has not yet been explored. Further details
may potentially also be gleaned from post mortem studies,
however these studies for various reasons do not exist in
BPD.
The fMRI studies examining neurobiological functional
abnormalities in BPD patients produced results which were
largely consistent with each other. The most prominent
finding was that of exaggerated activity in the amygdala
whilst passively viewing emotionally aversive slides or the
slides depicting negative facial expressions. The one PET
study that involved an aggression-invoking task produced
similar results, finding hypermetabolism in the amygdalae
and orbitofrontal cortices of the BPD group when pro-
voked. These findings would suggest that individuals with
BPD have different neural dynamics compared to their
healthy counterparts when passively viewing negative
images or being provoked into a negative emotional state
(Koenigsberg et al. 2009a). Indeed in the individuals with
BPD, these abnormal activations may cause impaired uti-
lization of cognitive control regulations leading to the
difficulties in behaviour modulation and in the ability to
regulate emotional reactions during negative emotional
states which characterize the disorder (Silbersweig et al.
2007; Koenigsberg et al. 2009a).
Another PET study which produced prominent results
was that of New et al. (2007). The study found that a
correlation between activity in the prefrontal cortex and the
amygdala seen in HC participants was absent in BPD
participants. New et al. expanded on this and suggested that
the correlation seen in the HC group indicated the intact
coupling between the prefrontal cortex and the right ventral
amygdala. Though the directionality of the associations
between the amygdala and the prefrontal cortex is still
controversial, in this case the directionality of the corre-
lation would indicate that the coupling may be the under-
lying neural substrate responsible for down regulation of
the amygdala in response to aversive stimuli; the absence
of which explaining the failure of those with BPD to down
regulate the amygdala when faced with aversive stimuli
(New et al. 2007).
The other PET studies examined patients only during a
resting state, though the results of these studies produced
very little consistency. One reason for the lack of consis-
tency amongst these PET studies could be that at resting
state, the patients’ emotional state is not known thus the
range of neural activity could be vast, whereas in the
studies involving specific tasks or images, the patients’
emotional states can be more accurately predicted and the
resulting images compared with more reliability. This
would suggest that when investigating neurobiological
functioning in BPD patients, studies which employ emo-
tion-related stimuli provide a more reliable measure than
resting state studies; a fitting theory considering that the
characteristic features of the disorder include emotional
dysregulation and irregular emotional response behaviours.
Another potential cause of inconsistencies in functionality
studies is sample characteristics, specifically gender dif-
ferences. Gender effects are particularly relevant in
investigations of emotional states such as aggression, as
aggression in particular is known to be directed differently
in males and females and this could result in different
neural activation between the genders (Schmahl et al.
2003; Schmahl and Bremner 2006). Another potentially
confounding variable in functional studies is the past and
present treatment (including both pharmacological and
psychotherapeutic) received by the patients. Such treat-
ments have been shown to affect regional cerebral func-
tioning to a certain extent in different psychiatric disorders
(Salavert et al. 2011), however studying untreated BPD
patients can be very difficult due to the patient predispo-
sition in BPD to self harm or attempt suicide. Although
some patients may have never received treatment of any
kind, as we can see from Table 3the majority of patient
participants in the studies cited here had at least a history of
treatment with psychotropic medications, with others being
treated with psychotropic medications at the time of the
assessment. Unfortunately, very little data were reported by
the studies regarding details of the psychotherapeutic
treatment of the patients, which is in itself a major limi-
tation of the research. Nevertheless, Salavert et al. (2011)
suggest that although the influence of pharmacological and
psychotherapeutic treatments cannot be ruled out conclu-
sively, it is accepted that both treatments tend to improve
the neurofunctional deficits observed in the disorder. Thus,
finding significant differences in functioning between BPD
patients and healthy controls, despite the potential influ-
ence of various treatments, gives credence to the effects
found (Salavert et al. 2011). Nonetheless, these studies still
confirm the importance of the areas of interest established
by structural studies, e.g., the ACC and hippocampus
(Schmahl and Bremner 2006).
The neurometabolite studies, however, have thus far been
inconclusive purely because of the small number of studies
that tackle the subject. The fewer the number of studies, the
less comparisons can be made and the weaker the conclusion
drawn. The few studies in existence have investigated con-
centrations of neurometabolites such as glutamate, gluta-
mine, phosphocreatine and creatine, N-acetylaspartate, and
choline-containing compounds. These metabolites have
been described as markers for and associated with a number
of neuronal features and occurrences such as neuronal
Brain Struct Funct (2012) 217:767–782 779
123
integrity (tNAA), energy-dependent functions (Cr), and
demyelination (Cho), though these associations are still
under debate (Jung et al. 2002). Though still requiring fur-
ther validation, these studies have produced some interesting
results. For example, the study by Tebartz van Elst et al. in
2007 found two interesting irregularities in the BPD patients;
increases in left amygdalar creatine and reductions in amy-
gdalar volume (Tebartz van Elst et al. 2007). Adding to this,
the researchers also found a negative correlation between
left amygdalar creatine concentration and amygdalar
volume and a positive correlation between these creatine
concentrations and psychometric measures of anxiety. A
reduction in left amygdalar creatine levels was then also
produced in a later study, lending support to the above Te-
bartz van Elst et al. findings, though the later study did not
find any correlations between neurometabolite concentra-
tions and any psychometric measurements (Hoerst et al.
2010b). These findings tie in very well with the structural
studies which have found reductions in amygdala volumes in
BPD patients, though much more spectroscopy research is
needed to support these results and further any theories. The
results of the other Hoerst et al. (2010a) study from the same
year are in keeping with the theory that elevated levels of Glu
in the ACC are significantly associated with both severity of
BPD symptoms and subjective impulsivity ratings, the latter
independent of a BPD diagnosis. These findings, along with
the finding of increased levels of Glx in non human primates
exposed to early life stress, also support the suggested
associated between HPA axis activation and heightened
glutamate neurotransmission in the prefrontal cortex
(Mathew et al. 2003).
As the research stands, it is evident that much more
concrete findings are needed to gain a firm understanding
of the neurobiological underpinnings of the disorder. The
abnormalities found thus far in the volumetric and func-
tional studies overlap dramatically with those found in
studies of individuals who have experienced trauma, those
with trauma-related disorders, and those with MDD. The
neural abnormalities found in individuals with PTSD in
particular are very similar to those seen in the BPD studies
cited here. As both disorders share a range of etiological
factors and symptomology, the shared biological features
are perhaps an inevitability. Studies examining brain
abnormalities in BPD patients with and without comorbid
attention deficit hyperactivity disorder (ADHD), however,
remain largely absent. This is indeed surprising considering
that there is a clear overlap of symptoms between the
disorders (including emotional instability and impulsivity)
and that around 60% of adults with BPD have a lifetime
history of ADHD or ADHD symptoms (Fossati et al.
2002). This lack of data regarding BPD and comorbid
ADHD is a severe limitation of the previous research and it
is strongly recommended that if possible the issue be
addressed in future studies, along with the further study of
other prominent comorbidities. By furthering the study of
BPD and its comorbid disorders, it may be possible to
establish biological markers to distinguish the disorders
from each other. An additional goal would be to establish
markers to differentiate the subtypes of the disorder.
Another direction for studies examining abnormalities in
terms of potential markers would be to investigate abnor-
malities in first presentation BPD patients, as these could
lead to effective methods for intervention and treatment.
As yet, the studies examining the neurobiology of BPD
have been solely cross-sectional, which as mentioned ear-
lier, is a major limitation in the quest to expand our
knowledge of the disorder. Longitudinal studies examining
the course of the disorder from the first presentation into
adulthood and equally as importantly examining the
ongoing effects of pharmacological and psychotherapeutic
intervention are vital to gain a better understanding of the
disorder and for indexing and improving current treat-
ments. To achieve these goals employing a combination of
volumetric, functional, and spectroscopic methods, cross-
sectionally and longitudinally, though expensive and time
consuming, may prove to be essential, as alone their find-
ings may only provide snippets of the broader picture.
References
Asnaani A, Chelminski I, Young D, Zimmerman M (2007) Heter-
ogeneity of borderline personality disorder: do the number of
criteria met make a difference? J Pers Disord 21(6):615–625.
doi:10.1521/pedi.2007.21.6.615
Brambilla P, Soloff PH, Sala M, Nicoletti MA, Keshavan MS, Soares
JC (2004) Anatomical MRI study of borderline personality
disorder patients. Psychiatry Res 131(2):125–133. doi:10.1016/
j.pscychresns.2004.04.003
Bremne JD, Vermetten E (2001) Stress and development: behavioral
and biological consequences. Dev Psychopathol 13(3):473–489
Bremner JD (1999) Does stress damage the brain? Biol Psychiatry
45(7):797–805. doi:10.1016/s0006-3223(99)00009-8
Brunner R, Henze R, Parzer P, Kramer J, Feigl N, Lutz K, Essig M,
Resch F, Stieltjes B (2010) Reduced prefrontal and orbitofrontal
gray matter in female adolescents with borderline personality
disorder: Is it disorder specific? Neuroimage 49(1):114–120. doi:
10.1016/j.neuroimage.2009.07.070
Chanen AM, Velakoulis D, Carison K, Gaunson K, Wood SJ, Yuen
HP, Yucel M, Jackson HJ, McGorry PD, Pantelis C (2008)
Orbitofrontal, amygdala and hippocampal volumes in teenagers
with first-presentation borderline personality disorder. Psychiatry
Res 163(2):116–125. doi:10.1016/j.pscychresns.2007.08.007
Cohen RA, Grieve S, Hoth KF, Paul RH, Sweet L, Tate D, Gunstad J,
Stroud L, McCaffery J, Hitsman B, Niaura R, Clark CR,
McFarlane A, Bryant R, Gordon E, Williams LM (2006) Early
life stress and morphometry of the adult anterior cingulate cortex
and caudate nuclei. Biol Psychiatry 59(10):975–982. doi:
10.1016/j.biopsych.2005.12.016
Congdon E, Canli T (2008) A neurogenetic approach to impulsivity.
J Pers 76(6):1447–1484. doi:10.1111/j.1467-6494.2008.00528.x
780 Brain Struct Funct (2012) 217:767–782
123
De la Fuente J, Goldman S, Stanus E, Vizuete C, Morla
´n I, Bobes J,
Mendlewicz J (1997) Brain glucose metabolism in borderline
personality disorder. J Psychiatr Res 31(5):531–541. doi:10.1016/
s0022-3956(97)00001-0
Donegan NH, Sanislow CA, Blumberg HP, Fulbright RK, Lacadie C,
Skudlarski P, Gore JC, Olson IR, McGlashan TH, Wexler BE
(2003) Amygdala hyperreactivity in borderline personality
disorder: implications for emotional dysregulation. Biol Psychi-
atry 54(11):1284–1293. doi:10.1016/s0006-3223(03)00636-x
Driessen M, Herrmann J, Stahl K, Zwaan M, Meier S, Hill A,
Osterheider M, Petersen D (2000) Magnetic resonance imaging
volumes of the hippocampus and the amygdala in women with
borderline personality disorder and early traumatization. Arch
Gen Psychiatry 57(12):1115–1122 (pii:yoa9416)
Duman RS (2002) Pathophysiology of depression: the concept of
synaptic plasticity. Eur Psychiatry 17(Suppl 3):306–310
(pii:S0924933802006545)
Fossati A, Novella L, Donati D, Donini M, Maffei C (2002) History
of childhood attention deficit/hyperactivity disorder symptoms
and borderline personality disorder: a controlled study. Compr
Psychiatry 43(5):369–377 (pii:S0010440X02000159)
Foti ME, Geller J, Guy LS, Gunderson JG, Palmer BA, Smith LM
(2010) Borderline personality disorder: considerations for inclu-
sion in the Massachusetts parity list of ‘‘biologically-based’’
disorders. Psychiatr Q. doi:10.1007/s11126-010-9154-y
Friedel RO (2004) Dopamine dysfunction in borderline personality
disorder: a hypothesis. Neuropsychopharmacology 29(6):1029–
1039. doi:10.1038/sj.npp.1300424
Giedd JN, Blumenthal J, Jeffries NO, Castellanos FX, Liu H,
Zijdenbos A, Paus T, Evans AC, Rapoport JL (1999) Brain
development during childhood and adolescence: a longitudinal
MRI study. Nat Neurosci 2(10):861–863. doi:10.1038/13158
Goodman M, New A, Siever L (2004) Trauma, genes, and the
neurobiology of personality disorders. Ann N Y Acad Sci
1032:104–116. doi:10.1196/annals.1314.008
Hall J, Olabi B, Lawrie SM, McIntosh AM (2010) Hippocampal and
amygdala volumes in borderline personality disorder: A meta-
analysis of magnetic resonance imaging studies. Pers Ment
Health 4(3):172–179. doi:10.1002/pmh.128
Hazlett EA, New AS, Newmark R, Haznedar MM, Lo JN, Speiser LJ,
Chen AD, Mitropoulou V, Minzenberg M, Siever LJ, Buchs-
baum MS (2005) Reduced anterior and posterior cingulate gray
matter in borderline personality disorder. Biol Psychiatry
58(8):614–623. doi:10.1016/j.biopsych.2005.04.029
Herpertz SC, Dietrich TM, Wenning B, Krings T, Erberich SG,
Willmes K, Thron A, Sass H (2001) Evidence of abnormal
amygdala functioning in borderline personality disorder: a
functional MRI study. Biol Psychiatry 50(4):292–298. doi:
10.1016/s0006-3223(01)01075-7
Hoerst M, Weber-Fahr W, Tunc-Skarka N, Ruf M, Bohus M,
Schmahl C, Ende G (2010a) Correlation of glutamate levels in
the anterior cingulate cortex with self-reported impulsivity in
patients with borderline personality disorder and healthy
controls. Arch Gen Psychiatry 67(9):946–954. doi:10.1001/
archgenpsychiatry.2010.93
Hoerst M, Weber-Fahr W, Tunc-Skarka N, Ruf M, Bohus M,
Schmahl C, Ende G (2010b) Metabolic alterations in the
amygdala in borderline personality disorder: a proton magnetic
resonance spectroscopy study. Biol Psychiatry 67(5):399–405.
doi:10.1016/j.biopsych.2009.09.030
Insel TR (2010) Faulty circuits. Sci Am 302(4):44–51
Irle E, Lange C, Sachsse U (2005) Reduced size and abnormal
asymmetry of parietal cortex in women with borderline person-
ality disorder. Biol Psychiatry 57(2):173–182. doi:10.1016/
j.biopsych.2004.10.004
Jacob GA, Gutz L, Bader K, Lieb K, Tuscher O, Stahl C (2010)
Impulsivity in borderline personality disorder: impairment in
self-report measures, but not behavioral inhibition. Psychopa-
thology 43(3):180–188. doi:10.1159/000304174
Juengling FD, Schmahl C, Hesslinger B, Ebert D, Bremner JD,
Gostomzyk J, Bohus M, Lieb K (2003) Positron emission
tomography in female patients with borderline personality disor-
der. J Psychiatr Res 37(2):109–115 (pii:S0022395602000845)
Jung RE, Yeo RA, Love TM, Petropoulos H, Sibbitt WL Jr, Brooks
WM (2002) Biochemical markers of mood: a proton magnetic
resonance spectroscopy study of normal human brain. Biol
Psychiatry 51(3):224–229 (pii:S0006322301012240)
Karl A, Schaefer M, Malta LS, Dorfel D, Rohleder N, Werner A
(2006) A meta-analysis of structural brain abnormalities in
PTSD. Neurosci Biobehav Rev 30(7):1004–1031. doi:10.1016/
j.neubiorev.2006.03.004
Kaufman J, Charney D (2001) Effects of early stress on brain
structure and function: implications for understanding the
relationship between child maltreatment and depression. Dev
Psychopathol 13(3):451–471
Koenigsberg HW, Fan J, Ochsner KN, Liu X, Guise KG, Pizzarello S,
Dorantes C, Guerreri S, Tecuta L, Goodman M, New A, Siever
LJ (2009a) Neural correlates of the use of psychological
distancing to regulate responses to negative social cues: a study
of patients with borderline personality disorder. Biol Psychiatry
66(9):854–863. doi:10.1016/j.biopsych.2009.06.010
Koenigsberg HW, Siever LJ, Lee H, Pizzarello S, New AS, Goodman
M, Cheng H, Flory J, Prohovnik I (2009b) Neural correlates of
emotion processing in borderline personality disorder. Psychiatry
Res 172(3):192–199. doi:10.1016/j.pscychresns.2008.07.010
Macqueen G, Frodl T (2010) The hippocampus in major depression:
evidence for the convergence of the bench and bedside in
psychiatric research? Mol Psychiatry 1-13. doi:10.1038/mp.
2010.80
Mathew SJ, Shungu DC, Mao X, Smith EL, Perera GM, Kegeles LS,
Perera T, Lisanby SH, Rosenblum LA, Gorman JM, Coplan JD
(2003) A magnetic resonance spectroscopic imaging study of
adult nonhuman primates exposed to early-life stressors. Biol
Psychiatry 54(7):727–735 (pii:S0006322303000040)
Mechelli A, Price CJ, Friston KJ, Ashburner J (2005) Voxel-based
morphometry of the human brain: methods and applications.
Current Medical Imaging Reviews 1:105–113
Minzenberg MJ, Fan J, New AS, Tang CY, Siever LJ (2007a) Fronto-
limbic dysfunction in response to facial emotion in borderline
personality disorder: an event-related fMRI study. Psychiatry
Res 155(3):231–243. doi:10.1016/j.pscychresns.2007.03.006
Minzenberg MJ, Fan J, New AS, Tang CY, Siever LJ (2007b) Fronto-
limbic dysfunction in response to facial emotion in borderline
personality disorder: an event-related fMRI study. Psychiatry
Res Neuroimaging 155(3):231–243. doi:10.1016/j.pscychresns.
2007.03.006
Minzenberg MJ, Fan J, New AS, Tang CY, Siever LJ (2008)
Frontolimbic structural changes in borderline personality disor-
der. J Psychiatr Res 42(9):727–733. doi:10.1016/j.jpsychires.
2007.07.015
New AS, Hazlett EA, Buchsbaum MS, Goodman M, Mitelman SA,
Newmark R, Trisdorfer R, Haznedar MM, Koenigsberg HW,
Flory J, Siever LJ (2007) Amygdala-prefrontal disconnection in
borderline personality disorder. Neuropsychopharmacology
32(7):1629–1640. doi:10.1038/sj.npp.1301283
New AS, Hazlett EA, Newmark RE, Zhang J, Triebwasser J,
Meyerson D, Lazarus S, Trisdorfer R, Goldstein KE, Goodman
M, Koenigsberg HW, Flory JD, Siever LJ, Buchsbaum MS
(2009) Laboratory induced aggression: a positron emission
tomography study of aggressive individuals with borderline
Brain Struct Funct (2012) 217:767–782 781
123
personality disorder. Biol Psychiatry 66(12):1107–1114. doi:
10.1016/j.biopsych.2009.07.015
Nunes PM, Wenzel A, Borges KT, Porto CR, Caminha RM, de
Oliveira IR (2009) Volumes of the hippocampus and amygdala
in patients with borderline personality disorder: a meta-analysis.
J Pers Disord 23(4):333–345. doi:10.1521/pedi.2009.23.4.333
Paus T (2005) Mapping brain maturation and cognitive development
during adolescence. Trends Cogn Sci 9(2):60–68. doi:10.1016/
j.tics.2004.12.008
Rusch N, van Elst LT, Ludaescher P, Wilke M, Huppertz HJ, Thiel T,
Schmahl C, Bohus M, Lieb K, Hesslinger B, Hennig J, Ebert D
(2003) A voxel-based morphometric MRI study in female
patients with borderline personality disorder. Neuroimage
20(1):385–392 (pii:S1053811903002970)
Salavert J, Gasol M, Vieta E, Cervantes A, Trampal C, Gispert JD
(2011) Fronto-limbic dysfunction in borderline personality disor-
der: a 18F-FDG positron emission tomography study. J Affect
Disord 131(1–3):260–267. doi:10.1016/j.jad.2011.01.001
Schmahl C, Bremner JD (2006) Neuroimaging in borderline person-
ality disorder. J Psychiatr Res 40(5):419–427. doi:10.1016/j.
jpsychires.2005.08.011
Schmahl CG, Vermetten E, Elzinga BM, Bremner JD (2003) Magnetic
resonance imaging of hippocampal and amygdala volume in
women with childhood abuse and borderline personality disorder.
Psychiatry Res 122(3):193–198 (pii:S0925492703000234)
Schmahl C, Berne K, Krause A, Kleindienst N, Valerius G, Vermetten
E, Bohus M (2009) Hippocampus and amygdala volumes in
patients with borderline personality disorder with or without
posttraumatic stress disorder. J Psychiatry Neurosci 34(4):
289–295
Silbersweig D, Clarkin JF, Goldstein M, Kernberg OF, Tuescher O,
Levy KN, Brendel G, Pan H, Beutel M, Pavony MT, Epstein J,
Lenzenweger MF, Thomas KM, Posner MI, Stern E (2007) Failure
of frontolimbic inhibitory function in the context of negative
emotion in borderline personality disorder. Am J Psychiatry
164(12):1832–1841. doi:10.1176/appi.ajp.2007.06010126
Soloff P, Nutche J, Goradia D, Diwadkar V (2008) Structural brain
abnormalities in borderline personality disorder: a voxel-based
morphometry study. Psychiatry Res 164(3):223–236. doi:
10.1016/j.pscychresns.2008.02.003
Steele H, Siever L (2010) An attachment perspective on borderline
personality disorder: advances in gene-environment consider-
ations. Curr Psychiatry Rep 12(1):61–67. doi:10.1007/s11920-
009-0091-0
Takahashi T, Chanen AM, Wood SJ, Walterfang M, Harding IH,
Yucel M, Nakamura K, McGorry PD, Suzuki M, Velakoulis D,
Pantelis C (2009a) Midline brain structures in teenagers with
first-presentation borderline personality disorder. Prog Neuro-
psychopharmacol Biol Psychiatry 33(5):842–846. doi:10.1016/
j.pnpbp.2009.03.035
Takahashi T, Chanen AM, Wood SJ, Yucel M, Tanino R, Suzuki M,
Velakoulis D, Pantelis C, McGorry PD (2009b) Insular cortex
volume and impulsivity in teenagers with first-presentation
borderline personality disorder. Prog Neuropsychopharmacol
Biol Psychiatry 33(8):1395–1400. doi:10.1016/j.pnpbp.2009.
07.017
Tebartz van Elst L, Thiel T, Hesslinger B, Lieb K, Bohus M, Hennig
J, Ebert D (2001) Subtle prefrontal neuropathology in a pilot
magnetic resonance spectroscopy study in patients with
borderline personality disorder. J Neuropsychiatry Clin Neurosci
13(4):511–514
Tebartz van Elst L, Hesslinger B, Thiel T, Geiger E, Haegele K,
Lemieux L, Lieb K, Bohus M, Hennig J, Ebert D (2003)
Frontolimbic brain abnormalities in patients with borderline
personality disorder: a volumetric magnetic resonance imaging
study. Biol Psychiatry 54(2):163–171 (pii:S0006322302017432)
Tebartz van Elst L, Ludaescher P, Thiel T, Buchert M, Hesslinger B,
Bohus M, Rusch N, Hennig J, Ebert D, Lieb K (2007) Evidence
of disturbed amygdalar energy metabolism in patients with
borderline personality disorder. Neurosci Lett 417(1):36–41. doi:
10.1016/j.neulet.2007.02.071
Vollm BA, Zhao L, Richardson P, Clark L, Deakin JF, Williams S,
Dolan MC (2009) A voxel-based morphometric MRI study in
men with borderline personality disorder: preliminary findings.
Crim Behav Ment Health 19(1):64–72. doi:10.1002/cbm.716
Weniger G, Lange C, Sachsse U, Irle E (2009) Reduced amygdala and
hippocampus size in trauma-exposed women with borderline
personality disorder and without posttraumatic stress disorder.
J Psychiatry Neurosci 34(5):383–388
Whittle S, Chanen AM, Fornito A, McGorry PD, Pantelis C, Yucel M
(2009) Anterior cingulate volume in adolescents with first-
presentation borderline personality disorder. Psychiatry Res
172(2):155–160. doi:10.1016/j.pscychresns.2008.12.004
Whitwell JL (2009) Voxel-based morphometry: an automated tech-
nique for assessing structural changes in the brain. J Neurosci
29(31):9661–9664. doi:10.1523/JNEUROSCI.2160-09.2009
Williamson A (2007) Using self-report measures in neurobehavioural
toxicology: can they be trusted? Neurotoxicology 28(2):227–234.
doi:10.1016/j.neuro.2006.03.009
Wingenfeld K, Spitzer C, Rullkotter N, Lowe B (2010) Borderline
personality disorder: hypothalamus pituitary adrenal axis and
findings from neuroimaging studies. Psychoneuroendocrinology
35(1):154–170. doi:10.1016/j.psyneuen.2009.09.014
Wiseman RM, Saxby BK, Burton EJ, Barber R, Ford GA, O’Brien JT
(2004) Hippocampal atrophy, whole brain volume, and white
matter lesions in older hypertensive subjects. Neurology 63(10):
1892–1897
Zanarini MC, Williams AA, Lewis RE, Reich RB, Vera SC, Marino
MF, Levin A, Yong L, Frankenburg FR (1997) Reported
pathological childhood experiences associated with the develop-
ment of borderline personality disorder. Am J Psychiatry 154(8):
1101–1106
Zanarini MC, Yong L, Frankenburg FR, Hennen J, Reich DB, Marino
MF, Vujanovic AA (2002) Severity of reported childhood sexual
abuse and its relationship to severity of borderline psychopa-
thology and psychosocial impairment among borderline inpa-
tients. J Nerv Ment Dis 190(6):381–387
Zetzsche T, Frodl T, Preuss UW, Schmitt G, Seifert D, Leinsinger G,
Born C, Reiser M, Moller HJ, Meisenzahl EM (2006) Amygdala
volume and depressive symptoms in patients with borderline
personality disorder. Biol Psychiatry 60(3):302–310. doi:
10.1016/j.biopsych.2005.11.020
Zetzsche T, Preuss UW, Frodl T, Schmitt G, Seifert D, Munchhausen
E, Tabrizi S, Leinsinger G, Born C, Reiser M, Moller HJ,
Meisenzahl EM (2007) Hippocampal volume reduction and
history of aggressive behaviour in patients with borderline
personality disorder. Psychiatry Res 154(2):157–170. doi:
10.1016/j.pscychresns.2006.05.010
782 Brain Struct Funct (2012) 217:767–782
123
