Subgenual Prefrontal Cortex Activity Predicts
Individual Differences in Hypothalamic-Pituitary-
Adrenal Activity Across Different Contexts
Allison L. Jahn, Andrew S. Fox, Heather C. Abercrombie, Steven E. Shelton, Terrence R. Oakes,
Richard J. Davidson, and Ned H. Kalin
Background: Hypothalamic-pituitary-adrenal (HPA) system activation is adaptive in response to stress, and HPA dysregulation occurs in
stress-related psychopathology. It is important to understand the mechanisms that modulate HPA output, yet few studies have addressed
the neural circuitry associated with HPA regulation in primates and humans. Using high-resolution F-18-fluorodeoxyglucose positron
emission tomography (FDG-PET) in rhesus monkeys, we assessed the relation between individual differences in brain activity and HPA
function across multiple contexts that varied in stressfulness.
Methods: Using a logical AND conjunctions analysis, we assessed cortisol and brain metabolic activity with FDG-PET in 35 adolescent
archival data set. In this data set, brain metabolic activity and cortisol were assessed in 17 adolescent male rhesus monkeys that were
exposed to three stress-related contexts.
Results: Results from the two studies revealed that subgenual prefrontal cortex (PFC) metabolism (Brodmann’s area 25/24) consistently
individuals who consistently show heightened cortisol and may be at risk for stress-related HPA dysregulation.
Key Words: Cortisol, HPA, monkey, PET, stress, subgenual PFC
primary mechanism through which emotions influence chronic
disease. Persistently increased cortisol occurs in some individuals
with depression (1,2) as well as in youth at risk for developing
stress-related psychopathology (3). Moreover, in various chronic
illnesses, increased cortisol is associated with poorer medical
outcomes (4,5). Although rodent studies have examined brain
mechanisms involved in HPA regulation, little is known about
neural circuitry related to individual differences in primate and
human cortisol concentrations.
Evidence from rodent studies points to medial prefrontal
cortex (mPFC) as being important in HPA regulation and regu-
lation of autonomic, behavioral, and emotional responses asso-
ciated with stress (6–8). Of particular relevance to these func-
tions are efferents that project from ventral PFC to subcortical and
brain stem structures (9–11). Data from rodent studies indicate
that dorsomedial PFC regions (prelimbic cortex) are involved in
HPA inhibition, whereas ventral regions (infralimbic cortex) are
involved in HPA activation (12,13). Subgenual PFC (Brodmann’s
area [BA] 25/24) is the primate and human cortical region most
analogous to the rat infralimbic cortex (7).
Subgenual PFC is of particular interest in conceptualizing the
relation between HPA overactivity and stress-related psychopa-
ortisol, a glucocorticoid released from the adrenal cortex,
is critical for survival, reflects activation of the hypotha-
lamic-pituitary-adrenal (HPA) system and represents a
thology, because this region has been the site of both structural
and functional alterations in some depressed patients, as well as
a successful target for treatment of severe depression (14–17). In
a study in treatment-resistant depressives, deep brain stimulation
decreased subgenual PFC hypermetabolism resulting in a marked
part of an “integrated pathway” for regulation of HPA is hypothe-
sized to account partially for involvement of this region in depres-
To study brain circuitry associated with individual differences
in HPA regulation in primates, we used data from a study aimed
at establishing the relation between anxious temperament and
regional brain metabolic activity (20,21). The availability of
cortisol data allowed us to examine the relation between cortisol
and metabolic activity in monkeys exposed to different contexts.
Rhesus monkeys are an ideal species to provide insights into
human HPA regulation because they share similarities with
humans in HPA function, brain structure (e.g., prefrontal cortical
development), and behavior (10,22). Rhesus monkeys are
among the best species to investigate brain mechanisms under-
lying emotion regulation and psychopathology because multiple
scans can be repeatedly obtained in the same individual during
exposure to ethologically relevant settings (23).
Plasma cortisol and metabolic brain activity were assessed
repeatedly in the same monkeys in response to four situations
varying in their degree of stressfulness. Two of the conditions
(separation from cagemate and introduction of a human in-
truder) elicit different threat-related adaptive responses, whereas
the other two conditions involved assessment in the monkeys’
home cages (24). Logical AND conjunction analysis (see Nichols
et al. 2005 ) was used to find clusters of regional metabolic
activity that correlated with cortisol concentrations across all four
conditions. This logical AND conjunction analysis identifies
regions where metabolism and cortisol are significantly corre-
atry (HCA, SES, RJD, NHK) and Waisman, Laboratory for Brain Imaging
and Behavior (ALJ, ASF, TRO, RJD), University of Wisconsin—Madison.
Address correspondence to Ned H. Kalin, M.D., 6001 Research Park Boule-
vard, Madison, WI 53719; E-mail: firstname.lastname@example.org.
Received Apr 13, 2009; revised Jul 6, 2009; accepted Jul 22, 2009.
BIOL PSYCHIATRY 2010;67:175–181
© 2010 Society of Biological Psychiatry
lated during separation from cagemate and introduction of a
human intruder and home-cage alone and home-cage with
cagemate; thus, excluding regions that do not show significant
correlations across all conditions. To confirm these results, we
used the same analytic strategy in a different data set of previ-
ously studied animals in which F-18-fluorodeoxyglucose posi-
tron emission tomography (FDG-PET) and cortisol were col-
lected. On the basis of rodent data suggesting that mPFC acts as
a relay for limbic-cortical pathways and integrates various as-
pects of stress responses, we hypothesized that individual differ-
ences in mPFC activity, specifically, the subgenual PFC, would
predict individual differences in cortisol concentrations across
Methods and Materials
Animals and Procedures. Subjects included 35 adolescent
rhesus monkeys (Macaca mulatta; mean [M] age ? standard
deviation; 2.61 [.56]; range: 1.83–3.96; male: 12, M weight: 4.06
[.77] kg; n ? 36, one animal excluded for refusing to leave its
cage). Animals were pair-housed at the Wisconsin National
Primate Center and Harlow Primate Laboratory and maintained
on 12-hour light–dark cycle (lights on 6 AM). All experimental
procedures were approved by the University of Wisconsin
Institutional Animal Care and Use Committee (IACUC).
Animals were exposed to four conditions before FDG-PET
scanning, with each condition occurring at least 1 week apart.
Because the bulk of FDG uptake occurs within 30 minutes after
injection (see Rilling et al. 2001 ), we designed conditions
lasting 30 minutes. In the home cagemate (H-CM) condition,
animals remained in their home cage with their partner for 30
minutes. In the home alone (H-ALN) condition, animals re-
mained alone in their home cage for 30 minutes. In the no eye
contact condition (NEC), animals were placed in a test cage while
a human entered the room for 10 minutes, presented his profile,
and avoided eye contact with the monkey while standing 2.5
miles away. The human left the test room for 5 minutes,
reentered for 5 minutes, and left again for 5 minutes, then
reentered for 5 minutes. This procedure was used to avoid
habituation. In the alone condition (ALN), animals were placed
in a foreign cage and left alone for 30 minutes. Exposure to each
condition was counterbalanced: eight animals received H-ALN
first, nine animals received H-CM, nine received ALN, and nine
received NEC first. For each animal, all conditions were run at
approximately the same time of day (minimum difference in
scanning times within an individual: 4 minutes; maximum differ-
ence in scanning times: 2 hours 48 minutes). Portions of this data
set have been previously described (20,21).
Cortisol. Blood was sampled by femoral venipuncture at the
end of each condition, before the FDG-PET scan. Samples were
obtained between 8:45 AM and 4:15 PM.
Basal cortisol was evaluated twice at intervals of at least 1
week. Baseline samples were collected in the morning between
8:00 AM and 9:30 AM.
Plasma was immediately extracted from blood by centrifuga-
tion at 4°C and frozen at ?70°C until assayed. Cortisol was
measured in plasma using enzyme immunoassay (Diagnostic
Systems Laboratories, Webster, Texas) with intraassay variability
of 6.1% and interassay variability of 6.3%. The detection limit was
.125 ng. For analysis, time of day was regressed on cortisol values
to remove variance associated with time of day.
Magnetic Resonance Imaging Acquisition. Structural mag-
netic resonance images (MRI) were used to aid in structural
normalization of PET data. See Additional Methods in Supple-
ment 1 for more information.
PET Data Acquisition and Pre-Processing. Immediately be-
fore each 30-minute condition, animals were injected with ?10
mCi of FDG through a 19-gauge intravenous catheter in the
saphenous vein. After the 30-minute session, blood was drawn
for cortisol samples, and then monkeys were anesthetized with
15 mg/kg ketamine followed by approximately 1.5% isoflurane
gas. The animal’s head was positioned in a stereotaxic apparatus
to maintain head positions across conditions in a P4 microPET
scanner (Concorde Microsystems, Knoxville, Tennessee). PET
resolution was 2-mm full width at half maximum (FWHM;
volumetric resolution of approximately 2 mm3). To facilitate
across-subject comparisons, PET images were transformed into a
standard space as defined by Paxinos, Huang, and Toga, 2000
We corrected for potential differences in PET scan intensity by
applying a global scale factor to each scan. This ensured each
PET image volume had the same mean-intensity value. Using
standard methods, global scale factors were determined by
adjusting the mean-intensity based on a whole brain region of
interest (30). To compute true regional cerebral metabolic rate of
glucose (rCMRglu) values, quantitative arterial blood radioactivity
measurements are required; however, obtaining these measures
would have interfered with the naturalistic (awake) paradigm.
Instead, we performed intensity normalization to obtain a proxy
for rCMRgluto facilitate intersubject comparisons.
PET images were smoothed using a 4 mm FWHM Gaussian
kernel to accommodate small differences in locations of func-
tional regions across subjects (e.g., due to differences in gyral
features) and increase signal-to-noise ratio (31).
Gray-matter probability (GMP) maps were created for each
subject and used as a regressor in tests of FDG uptake to control
for anatomical differences (32,33). See Supplement 1 for addi-
Logical AND Conjunction Analysis. Correlations between
cortisol and metabolism were performed within each condition
using “fmristat” (34,35). Correlations were performed controlling
for age, time of day, and anatomic differences by using the GMP
map as a voxelwise covariate in our functional analyses as
described in Oakes et al. 2007 (32).
To identify brain regions consistently correlated with cortisol
levels across contexts, we performed a logical AND conjunction
analysis using a minimum statistic (25). This analysis identified
brain regions in which metabolism was significantly correlated
with cortisol levels in each condition (p ? .05, two-tailed
uncorrected). We combined the statistical parametric maps from
each condition and obtained a map revealing regions where
metabolism and cortisol were significantly correlated during the
NEC condition and the ALN condition and the H-ALN condition
and the H-CM condition. Thus, clusters represent brain regions
where metabolism and cortisol are correlated across all four
Follow-up analyses were performed by extracting mean val-
ues for each cluster of interest within each condition. Metabolic
activity in the threat conditions was calculated by averaging the
mean value within each region of interest for ALN and NEC
conditions. Metabolic activity in the home conditions was calcu-
lated by averaging the mean glucose metabolism value within
each region of interest for H-ALN and H-CM. Difference scores
were created for each subject by subtracting the mean metabo-
176 BIOL PSYCHIATRY 2010;67:175–181
A.L. Jahn et al.
lism value for home conditions from the mean metabolism value
for threat conditions for each cluster of interest.
Animals and Procedures. Data include 17 adolescent male
rhesus monkeys (M age: 2.99 [.48]; range: 2.18–3.57; M weight:
4.78 [.88] kg; n ? 22, five animals excluded from analysis due to
problems with alignment in the scanner and missing cortisol
data). Housing procedures were identical to those presented in
Experiment 1. Experimental procedures were approved by the
University of Wisconsin IACUC.
Animals were exposed to three 30-minute conditions: ALN,
NEC, and stare (ST). In the ST condition, animals were placed in
a test cage while an experimenter entered the room and looked
directly at the animal in 10-minute increments from 2.5 miles
away with a 5-minute break in which the human left the testing
room. Breaks were designed to minimize habituation, thus
extending the heightened sensitively to the intruder over the
30-minute FDG uptake. The NEC and ALN conditions were
identical to conditions from Experiment 1. Six animals received
ALN first, seven animals received ST first, and four animals
received NEC first. Each condition was not performed more than
once per week. Data from Experiment 2 have been described in
Fox et al. (2005) (36) and Kalin et al. 2005 (23).
Cortisol. Collection, storage, and processing were identical
to Experiment 1.
MRI Acquisition. Structural images were obtained for 6 of
the 17 monkeys. Data collection was identical to Experiment 1.
PET Data Acquisition, Processing, and Analysis. Data col-
lection procedures and analysis of brain metabolic activity were
identical to Experiment 1 (with the exception of GMP as a
covariate, which was not used). Because anatomic data were
obtained for only 6 of the 17 monkeys, procedures for registra-
tion and transformation to a standard space (27) differed from
GMP maps could not be created because of the absent anatomic
Logical AND Conjunction Analysis. Correlations between
cortisol and brain activity were performed within each condition
using “fmristat” (34,35), and we statistically controlled for age
and time of day. We again performed a logical AND conjunction
analysis (25) using a minimum statistic (p ? .05, one-tailed
uncorrected) to determine brain regions that were related to
cortisol across contexts. We combined the statistical parametric
maps from each condition and obtained a map revealing regions
in which metabolism and cortisol were significantly correlated
during the NEC condition, the ST condition, and the ALN
condition. Clusters represent brain regions where metabolism
and cortisol are correlated across all three conditions.
Experiment 1—Relation Between Cortisol and Regional
Cortisol. Cortisol values, adjusted for time of day, revealed
varying degrees of HPA activation in relation to the different test
conditions, F(4,136) ? 37.84, p ? .0001 (see Figure 1A). Cortisol
values (adjusted for time of day) were greater in the threat
conditions (ALN and NEC) compared with the home-cage con-
ditions (H-CM and H-ALN), t(34) ? 4.79, p ? .001. Cortisol
values (adjusted for time of day) were greater in the home-cage
conditions compared with baseline, t(34) ? 8.68, p ? .001. Tests
on raw cortisol values also indicate significant differences across
conditions (see Figure 1 and Supplement 1 for statistics).
ical AND conjunction analysis (25) revealed that metabolism in
subgenual PFC and part of pregenual PFC (BA 25/24) was
consistently correlated with individual differences in cortisol
(p ? .05; Figure 2, Table 1). A separate prefrontal cluster in which
metabolism correlated with cortisol across conditions was iden-
tified in left orbitofrontal cortex (OFC, BA 13; Figure 2A). A
region of basal forebrain including the right bed nucleus of the
stria terminalis (BNST), ventral striatum, and shell and core of the
nucleus accumbens (NAC) was the only other region in which
metabolism positively correlated with cortisol across the four
conditions (which we refer to as BNST/NAC; Figure 2B). There
were no regions in which metabolism was consistently nega-
tively correlated with cortisol.
We examined the correlation between glucose metabolism
during the four test conditions with baseline cortisol sampled in
the morning on nontesting days. Individual differences in base-
line cortisol concentrations were significantly positively corre-
lated with glucose metabolism in the subgenual PFC cluster
during the ALN, NEC, and H-ALN conditions, rs(33) ? .35; ps ?
.05; OFC during the ALN and NEC conditions, rs(33) ? .40; ps ?
.05; and BNST/NAC only during the ALN condition, r (33) ? .40;
p ? .05.
Figure 1. For Experiment 1, the two threat conditions included alone in a
test cage and no eye contact. The two home-cage conditions included
home alone and home with cagemate. (A) Both threat and home-cage
conditions produced a greater cortisol response compared with baseline
(BSL) morning cortisol after controlling for time of day, ts(34) ? 8.68, ps ?
.001. Differences between conditions were also significant when using raw
cortisol levels, F(4,136) ? 73.97, p ? .001; all ts(34) ? 3.10, ps ? .005.
Graphically, raw cortisol values are presented. (B) Threat conditions pro-
duced higher relative cerebral glucose metabolism in the subgenual pre-
frontal cortex (PFC) than home conditions, t(34) ? 3.10, p ? .005. Error bars
represent standard error of the mean. *p ? .005.
A.L. Jahn et al.
BIOL PSYCHIATRY 2010;67:175–181 177
Condition-Related Metabolism. We compared metabolism
in the threat conditions versus home-cage conditions for brain
regions identified in the logical and conjunction analysis. Com-
paring across conditions, metabolism in subgenual PFC, OFC,
and BNST/NAC (see Figure 2A for the location of these regions)
revealed a similar pattern to that of cortisol. Regional metabolism
during the threat conditions (NEC and ALN) was significantly
higher than metabolism in the home-cage conditions (H-CM and
H-ALN) in subgenual PFC, t(34) ? 3.10, p ? .005; Figure 1B;
OFC, t(34) ? 2.96, p ? .01; and BNST/NAC, t(34) ? 4.82, p ?
Follow-up correlational analyses examined the relations among
subgenual PFC, OFC, and BNST/NAC metabolism. Metabolic
activity was significantly correlated between subgenual PFC and
OFC in all four conditions (rs ? .47?.76, p ? .01). Metabolic
activity was significantly correlated between subgenual PFC and
BNST/NAC in all four conditions (rs ? .52?.65, p ? .01).
Metabolic activity was significantly correlated between OFC
and BNST/NAC in one of the four conditions (H-CM: r  ?
.45, p ? .01).
Individual Differences between Threat and Home-Cage
Conditions. To assess further the extent to which individual
differences in metabolic activity predict cortisol concentrations,
we computed difference scores for each subject between threat
and home-cage cortisol levels, as well as difference scores
between threat and home-cage regional metabolism. We corre-
lated difference scores for metabolism for each brain region of
interest and with difference scores for cortisol. This analysis
demonstrated that the difference in cortisol levels between threat
and home-cage conditions was significantly correlated with the
difference in metabolism in the subgenual PFC, r(33) ? .53, p ?
.001, Figure S1A in Supplement 1; OFC, r (33) ? .38, p ? .05,
Figure S1B in Supplement 1, and BNST/NAC, r(33) ? .55, p ?
.001, Figure S1C in Supplement 1.
Experiment 2—Analysis of Archival Data
Cortisol Data. Cortisol levels (in ?g/dL) following each of
the stressful conditions were greater than the baseline concen-
trations assessed on a different day (M: 41.94 [11.61]; F [3,48] ?
39.09, p ? .001) and did not significantly differ across ALN (M:
5.1 184.108.40.206 3.9
(µg / dL)
r = .42
5.14.8 4.5 4.2
(µg / dL)
r = .41
(µg / dL)
r = .51
(µg / dL)
r = .42
conditions. The solid pink areas represent the brain regions of overlap in which cortisol and glucose metabolism are correlated across all four conditions as
identified using the logical and conjunction analysis. Lower images: isolated image of brain regions of overlap (solid pink). (A) Anterior sagittal and coronal
slices depicting subgenual prefrontal cortex (PFC) and the left orbitofrontal cortex (OFC) in which cortisol correlates with glucose metabolism across all four
subgenual PFC and cortisol in each of the four conditions: threat conditions—alone (ALN) and no eye contact (NEC); home-cage conditions—home alone
(H-ALN) and home with cagemate (H-CM).
Table 1. Experiment 1: Brain Regions that Consistently Predict Cortisol
Responses Across Contexts
Local MaximaHemisphere Volume, mm3
Subgenual PFC BA 24/25
orbitofrontal cortex; PFC, prefrontal cortex.
178 BIOL PSYCHIATRY 2010;67:175–181
A.L. Jahn et al.
75.79 [18.10]), NEC (M: 74.10 [11.78]), and ST (M: 75.39 [16.98])
conditions ts ? .54, ps ? .60).
Consistent Correlations between Metabolism and Cortisol. To
test further the reliability of the relation between cortisol and PFC
activity, we used similar methods to examine an archival data set.
On the basis of the prior analyses, we hypothesized that across
each of the three contexts, metabolic activity in the subgenual
PFC, OFC, and BNST/NAC regions would predict individual
differences in cortisol concentrations in response to the condi-
tions. A one-tailed p ? .05 threshold for the predicted regions
was used based on the a priori directional hypothesis. Consistent
with the earlier results, across the three conditions, positive
correlations were found between cortisol levels and metabolic
rate in a cluster overlapping with the subgenual PFC (BA 25/24;
local maxima, x: ?2, y: 11, z: 6; cluster volume ? 65 mm3). In
contrast to the initial study, significant correlations between
cortisol and metabolic activity were not found in the BNST/NAC
or OFC regions (see Table S1 in Supplement 1 for all clusters).
We did not observe significant correlations between baseline
plasma cortisol levels and subgenual PFC glucose metabolism
in any of the conditions in this archival data set, rs (15) ? .31,
ps ? .22.
This study provides novel data suggesting that in primates
subgenual PFC is a brain area that is critically related to HPA
output. By using a logical AND conjunction analysis (25), we
were able to identify brain regions in which glucose metabolic
rate significantly predicts cortisol in each condition and thus is
consistently related to individual differences across multiple
contexts. This analytic strategy allowed us to pinpoint brain
regions that may act as traitlike neural regulators. The data
demonstrate that individual differences in subgenual PFC meta-
bolic rate predict cortisol levels across different contexts that vary
in their degree of stressfulness. The data also show that both
subgenual PFC metabolic activity and cortisol concentrations
during threat conditions were greater than in the less stressful
home conditions. Furthermore, analysis of the archival data
confirmed the context-independent relation between subgenual
PFC activity and cortisol concentrations. Taken together, these
data suggest that activity in the subgenual PFC has a traitlike
relationship with individual variation in cortisol levels that re-
mains stable across threatening and nonthreatening contexts.
In primates, the subgenual PFC is linked to many brain
regions that are involved in HPA regulation as well as visceral
and emotional responses to stress (amygdala, BNST, hypothala-
mus, and periaqueductal gray) (11,37,38). Furthermore, using a
similar analytic strategy in the same animals used in Experiment
1, we recently identified a network consisting of many of these
regions (amygdala, hippocampus, and periaqueductal gray) that
is predictive of individual differences in anxious temperament
(20). In rodent studies, it has been demonstrated that during
stress the infralimbic cortex (IL; the region in rodents that is most
analogous to subgenual PFC in primates) (7) provides input to
the basolateral amygdala (39). On the basis of rodent data, the IL
is thought to influence behavior during extinction learning
through indirect input (projections to intercalated amygdala
neurons) to the central nucleus of the amygdala (40,41). Other
studies show that lesions to the infralimbic cortex suppress
stress-related corticosterone responses (13,42), elevate basal
levels of corticosterone (42), prolong extinction learning to
fearful stimuli (43), and alter social behavior (42). The observed
relationship, across contexts, between the subgenual PFC meta-
bolic rate and cortisol is consistent with the idea that this region
is critically involved in the regulation of the HPA axis and the
stress response. In conjunction with other primate and rodent
data, our findings suggest that the primate subgenual PFC might
represent the core of an integrated pathway for both initiation
and regulation of a cortisol response.
Data from our primary study also suggest that individual
differences in activity in the OFC as well as an area encompass-
ing the BNST/NAC are associated with HPA output. Although
metabolic activity between the BNST/NAC and OFC was corre-
lated only in one of the four conditions, metabolic activity in the
subgenual PFC was correlated with BNST/NAC and OFC across
all four conditions. This further suggests that the subgenual PFC
is part of a network of structures whose activity is related to the
HPA axis. The BNST/NAC has dense projections from the mPFC
which includes BA 25 (10,44). Importantly, the BNST/NAC have
direct connections to the paraventricular nucleus of the hypo-
thalamus (9,23,45). This region of the basal forebrain is thought
to be part of a network that integrates endocrine, autonomic, and
emotional states (46). The OFC also plays a role in regulating
emotion and anxiety and is bidirectionally linked to the amyg-
dala (47). Although the relations between cortisol and metabo-
lism in these regions were evident in Experiment 1, they were not
replicated in the analysis of the archival data and thus should be
regarded with caution.
Although our findings demonstrate that greater subgenual PFC
activation is related to greater cortisol secretion, the 30-minute time
frame of FDG uptake (see, e.g., Rilling et al. 2001 ), along with
our use of a single cortisol sample per condition, makes it difficult
to interpret the specific mechanisms involved in this relationship.
Because subgenual PFC provides input to the hypothalamus, amyg-
dala, and BNST/NAC and because it is also implicated in negative
feedback of the HPA system (12), it is likely that the positive relation
between subgenual PFC activity, and cortisol reflects the combina-
tion of stimulatory and inhibitory processes within the HPA axis.
Because of the correlational nature of our data, the possibility exists
that cortisol and subgenual regional metabolism show a simulta-
neous increase during these contexts but share no mechanistic
relationship with one another. However, rodent data provide clear
evidence for a mechanistic relationship between infralimbic cortex
and HPA regulation, which suggests that our correlational findings
may identify in primates a region that is mechanistically related to
variation in HPA output. In addition, it is notable that the primate
PFC contains a wealth of glucocorticoid receptors (48), and the
distribution of receptors in the PFC is affected by stress (22), further
suggesting a functional relationship between these systems. The
possibility also exists that the relation we observed between sub-
genual PFC activity and cortisol could be due to the effects of
cortisol on brain metabolic activity. Future work should continue to
address the causal nature of this relationship and provide data with
increased sensitivity to detect time profiles of these responses.
Another potential limitation is our inability to confidently
address sex-related differences in these samples. Our sample
from Experiment 2 contained only male monkeys, and our
sample for Experiment 1 contained an insufficient number of
males (n ? 12) and females (n ? 23) to address sex-related
It has been suggested that depression is partly due to alter-
ations in stress-related systems (16). Studies in some depressed
patients demonstrate altered regulation of the HPA axis (49).
Other studies point to increased subgenual PFC activity associ-
ated with the pathophysiology of depression (18), and recent
A.L. Jahn et al.
BIOL PSYCHIATRY 2010;67:175–181 179
human neuroimaging studies support the involvement of the
subgenual PFC in HPA regulation during stressful contexts (50).
The findings presented in this study provide a link between
individual differences in subgenual PFC and HPA activity and
may be important in understanding the basis of psychopathology
related to dysregulation of the stress response. Future work
should examine whether individuals with elevations in sub-
genual PFC activity across multiple contexts are at risk for
stress-related HPA dysregulation and psychopathology.
In summary, this study used multiple contexts, varying in their
metabolism and HPA activity in primates. Heightened subgenual
PFC glucose metabolism consistently predicted higher cortisol lev-
els across contexts that varied in their degree of stressfulness. The
findings point to a role for the primate subgenual PFC as a possible
traitlike neural regulator of HPA activity. These data linking sub-
genual PFC to HPA function may be particularly relevant to under-
related psychopathology in which both subgenual function and
HPA function can be dysregulated.
We are grateful to H. Van Valkenberg, T. Johnson, K. Meyer,
E. Zao, R. Hoks, and the staff at the Harlow Center for Biological
Psychology and the Wisconsin National Primate Research Center
(Grant No. RR000167), for their technical support. This work
was supported by Grant Nos. MH46729, MH69315, MH84051,
the HealthEmotions Research Institute, Meriter Hospital, a Na-
awarded to A. Jahn, and a National Institute of Mental Health
K08 award (No. K08 MH074715) to H. Abercrombie.
Allison Jahn, Andrew Fox, Dr. Heather Abercrombie, Dr.
Terry Oakes, and Dr. Richard Davidson reported no biomedical
financial interests or potential conflicts of interest. Dr. Shelton
reports owning stock options in General Electric, Corp. (GE). Dr.
Kalin reports serving on the scientific advisory board for Astra-
Zeneca, Bristol-Myers-Squibb, CeNeRx Biopharma, Corcept Ther-
apeutics, Cyberonics, Elsevier, Eli Lilly and Company, Forest Labo-
ratories, General Electric, Corp. (GE Healthcare), GlaxoSmithKline,
Jazz Pharmaceutical, Neuronetics, Novartis, Otsuka, American
Pharmaceuticals, Takeda International, and Wyeth Pharma-
ceutical. He reports stock options in Corcept Therapeutics and
CeNeRx Biopharma. He is the owner of Promoter Neurosciences,
LLC. He holds a patent for the promoter sequences for cortico-
tropin-releasing factor CRF2alpha and for a method of identify-
ing agents that alter the activity of the promoter sequences: USA.
Patent issued on 7/4/06; patent #7,071,323, divisional patent
applied for on 9/26/05; patent application #11,234,916. He
holds a patent for the promoter sequences for urocortin II and the
use thereof: USA Patent issued on 08/08/06; patent #7,087,385.
He holds a patent for the promoter sequences for corticotropin-
releasing factor binding protein and use thereof: USA Patent
issued on 10/17/06; patent #7,122,650. He holds a patent for the
method for reducing CRF receptor mRNA: USA Patent applied for
on 7/22/04, patent application #20,050,042,212.
Supplementary material cited in this article is available
1. Burke HM, Davis MC, Otte C, Mohr DC (2005): Depression and cortisol
in depression: Implications for therapy. J Affect Disord 62:77–91.
3. Mannie ZN, Harmer CJ, Cowen PJ (2007): Increased waking salivary
4. Leserman J, Petitto JM, Golden RN, Gaynes BN, Gu H, Perkins DO, et al.
and cortisol on progression to AIDS. Am J Psychiatry 157:1221–1228.
5. Brown ES, Varghese FP, McEwen BS (2004): Association of depression
with medical illness: Does cortisol play a role? Biol Psychiatry 55:1–9.
6. Price JL, Carmichael ST, Drevets WC (1996): Networks related to the
orbital and medial prefrontal cortex; a substrate for emotional behav-
7. Quirk GJ, Beer JS (2006): Prefrontal involvement in the regulation of
emotion: Convergence of rat and human studies. Curr Opin Neurobiol
8. Schneider M, Koch M (2005): Behavioral and morphological alterations
rats. Exp Neurol 195:185–198.
9. Herman JP, Figueiredo H, Mueller NK, Ulrich-Lai Y, Ostrander MM, Choi
DC, et al. (2003): Central mechanisms of stress integration: Hierarchical
circuitry controlling hypothalamo-pituitary-adrenocortical responsive-
ness. Front Neuroendocrinol 24:151–180.
(subgenual cortex) of the macaque monkey. J Comp Neurol 421:172–
12. Diorio D, Viau V, Meaney MJ (1993): The role of the medial prefrontal
cortex (cingulate gyrus) in the regulation of hypothalamic-pituitary-
adrenal responses to stress. J Neurosci 13:3839–3847.
13. Sullivan RM, Gratton A (1999): Lateralized effects of medial prefrontal
cortex lesions on neuroendocrine and autonomic stress responses in
rats. J Neurosci 19:2834–2840.
14. Drevets WC (1999): Prefrontal cortical-amygdalar metabolism in major
(2001): Changes in regional brain glucose metabolism measured with
positron emission tomography after paroxetine treatment of major
et al. (2004): Functional but not structural subgenual prefrontal cortex
abnormalities in melancholia. Mol Psychiatry, September 325:393–405.
a new nosology with therapeutic implications. Biol Psychiatry 61:729–
anatomy of melancholia. Annu Rev Med 49:341–361.
20. Fox AS, Shelton SE, Oakes TR, Davidson RJ, Kalin NH (2008): Trait-like
mates. PLoS ONE 3:e2570.
serotonin transporter genotype is associated with intermediate brain
frontal cortex. Psychoneuroendocrinology 33:360–367.
associated with the expression and contextual regulation of anxiety in
primates. Biol Psychiatry 58:796–804.
24. Kalin NH, Shelton SE (1989): Defensive behaviors in infant rhesus mon-
keys: Environmental cues and neurochemical regulation. Science 243:
tion inference with the minimum statistic. Neuroimage 25:653–660.
26. Rilling JK, Winslow JT, O’Brien D, Gutman DA, Hoffman JM, Kilts CD
27. Paxinos G, Huang XF, Toga AW (2000): The Rhesus Monkey Brain in Ste-
reotaxic Coordinates. San Diego, CA: Academic Press.
180 BIOL PSYCHIATRY 2010;67:175–181
A.L. Jahn et al.
28. Jenkinson M, Smith S (2001): A global optimisation method for robust Download full-text
affine registration of brain images. Med Image Anal 5:143–156.
29. Woods RP, Grafton ST, Watson JD, Sicotte NL, Mazziotta JC (1998): Au-
The influence of biological and technical factors on the variability of
global and regional brain metabolism of 2-[18F]fluoro-2-deoxy-D-glu-
activation in PET images. Hum Brain Mapp 4:74–16.
32. Oakes TR, Fox AS, Johnstone T, Chung MK, Kalin N, Davidson RJ (2007):
Integrating VBM into the general linear model with voxelwise anatom-
ical covariates. Neuroimage 34:500–508.
33. Zhang Y, Brady M, Smith S (2001): Segmentation of brain MR images
through a hidden Markov random field model and the expectation–
A general statistical analysis for fMRI data. Neuroimage 15:1–15.
multivariate random field theory. Neuroimage 23(Suppl 1):S189–S195.
Calling for help is independently modulated by brain systems underly-
37. Critchley HD (2005): Neural mechanisms of autonomic, affective, and
and prelimbic areas of the medial prefrontal cortex in the Japanese
39. Maroun M (2006): Stress reverses plasticity in the pathway projecting
from the ventromedial prefrontal cortex to the basolateral amygdala.
40. Quirk GJ, Likhtik E, Pelletier JG, Pare D (2003): Stimulation of medial
prefrontal cortex decreases the responsiveness of central amygdala
output neurons. J Neurosci 23:8800–8807.
42. Rangel A, Gonzalez LE, Villarroel V, Hernandez L (2003): Anxiolysis fol-
lowed by anxiogenesis relates to coping and corticosterone after me-
dial prefrontal cortical damage in rats. Brain Res 992:96–103.
43. Morgan MA, Romanski LM, LeDoux JE (1993): Extinction of emotional
learning: Contribution of medial prefrontal cortex. Neurosci Lett 163:
44. Haber SN, McFarland NR (1999): The concept of the ventral striatum in
45. Davis M (2006): Neural systems involved in fear and anxiety measured
with fear-potentiated startle. Am Psychol 61:741–756.
neuropsychiatric disorders. Neuroscience 76:957–1006.
tal cortex in mediating anxious temperament. Biol Psychiatry 62:1134–
48. Sanchez MM, Young LJ, Plotsky PM, Insel TR (2000): Distribution of
corticosteroid receptors in the rhesus brain: Relative absence of glu-
cocorticoid receptors in the hippocampal formation. J Neurosci 20:
49. Young EA, Kotun J, Haskett RF, Grunhaus L, Greden JF, Watson SJ, et al.
methasone in depression. Arch Gen Psychiatry 50:395–403.
50. Kern S, Oakes TR, Stone CK, McAuliff EM, Kirschbaum C, Davidson RJ
(2008): Glucose metabolic changes in the prefrontal cortex are associ-
A.L. Jahn et al.
BIOL PSYCHIATRY 2010;67:175–181 181