Neuroscience and Biobehavioral Reviews 35 (2011) 1902–1915
Contents lists available at SciVerse ScienceDirect
Neuroscience and Biobehavioral Reviews
journal homepage: www.elsevier.com/locate/neubiorev
Review
Toward affective circuit-based preclinical models of depression: Sensitizing
dorsal PAG arousal leads to sustained suppression of positive affect in rats
Jason S. Wright ∗ , Jaak Panksepp
Center for the Study of Animal Well-being, Department of Veterinary & Comparative Anatomy, Pharmacology and Physiology, College of Veterinary Medicine,
Washington State University, Pullman, WA 99164-6520, United States
a r t i c l e
i n f o
Article history:
Received 10 May 2011
Received in revised form 23 July 2011
Accepted 3 August 2011
Keywords:
Depression
Preclinical models
Periaqueductal gray
Brain stimulation
Separation-distress
Aversion
PANIC/GRIEF circuitry
Appetitive motivation
SEEKING circuitry
50-kHz USVs
Rats
a b s t r a c t
Little is known about why clinical depression feels so bad, perhaps because optimal neural circuit-based
animal models of depression do not yet exist. Our goal here was to develop a strategy of inducing and
measuring depressive-like states in the rat using neural circuits as both the independent and major dependent variables. We hypothesized that repeated electrical stimulation of the brain (ESB) within the dorsal
periaqueductal gray (dPAG) aversion circuits would lead to a long-lasting suppression of 50 kHz ultrasonic vocalizations (USVs), a validated measure of positive social affect. Fifteen consecutive daily 10 min
sessions of intermittent PAG-ESB reduced systematically evoked 50 kHz USVs for up to 29 days following termination of ESB treatment, along with altering traditional measures of negative affect, including
behavioral agitation, sucrose intake, and decreased exploratory behavior. These findings suggest a new
affective circuit-based preclinical model of depression.
© 2011 Elsevier Ltd. All rights reserved.
Contents
1.
2.
3.
4.
5.
6.
7.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Depression modeling: brain circuit “break it and fix it” strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Background perspectives: from classic stress models to modern affect models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Background: the PANIC circuitry of the PAG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
A vocal affective indicator of SEEKING circuitry arousal—50 kHz USVs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1.
Experiment 1: dorsal PAG stimulation can abolish positive affect as monitored by 50 kHz USVs
and promotes negative affect as measured by 22 kHz USVs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2.
Experiment 2: PAG stimulation leads to a conditioned suppression of positive affect as monitored with 50 kHz USVs . . . . . . . . . . . . . . . .
5.3.
Experiment 3: chronic dorsal PAG stimulation leads to decreased exploration, altered sucrose intake,
and a persistent suppression of positive affect as indexed by 50 kHz USVs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Interpretation and discussion of results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Conclusions and future possibilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Appendix A.
Supplementary data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1902
1903
1904
1905
1906
1907
1907
1908
1910
1912
1913
1913
1913
1. Introduction
∗ Corresponding author at: Center for the Study of Animal Well-being, Department
of Veterinary & Comparative Anatomy, Pharmacology and Physiology, College of
Veterinary Medicine, Washington State University, PO Box 646520, Pullman, WA
99164-6520, United States. Tel.: +1 509 948 2203.
E-mail address: jwright@vetmed.wsu.edu (J.S. Wright).
0149-7634/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.neubiorev.2011.08.004
The clinical picture of major depression includes a variety of
affective and cognitive symptoms that are difficult to simulate in
preclinical (animal) models. It is widely believed that emotional
processes critical for understanding depression cannot be well
J.S. Wright, J. Panksepp / Neuroscience and Biobehavioral Reviews 35 (2011) 1902–1915
studied in animals. Thus, most work has been conducted modeling
various behavioral symptoms, many of which may be of debatable
validity. Putative behavioral indices, such as volume of sweet fluid
intake and swimming persistence are frequently used as measures
of depression in preclinical models, and validated largely on behavioral changes induced by approved anti-depressants (Philip et al.,
2010), which themselves have modest efficacy in treating depression (Philip et al., 2010), and lead to various undesirable side-effects
with long-term use (Pigott et al., 2010; Sinyor et al., 2010). This has
been recently emphasized by Angell (2004), an outspoken critic of
the pharmaceutical industry (e.g., see her multiple book reviews on
such topics in the New York Review of Books). Perhaps the basic science in the area would improve if we had preclinical models that
directly manipulated and monitored the relevant brain affective
systems. The premise of this paper is that it would be desirable to
have preclinical models that evaluate relevant positive and negative affective shifts as directly as possible, and our goal is to provide
evidence for one such approach.
For a thorough coverage of current animal models of depression
see: special issue of Neuroscience and Biobehavioral Reviews, edited
by Markou (2005). Intake of water laced with sucrose is commonly
used as a measure of depression, and though sucrose consumption may reflect hedonic shifts, it is not an established symptom
of human depression. For years, the forced swim test has been
shown to effectively predict the efficacy of currently used antidepressants, more recently supplemented by struggling in mice
during tail-suspension tests, but the face validity of such tests is
debatable. Our goal here was to focus on how to better achieve
optimal translational research using preclinical models of depression that not only take direct affective state shifts in animals as the
primary targets of inquiry, but also utilize induction procedures
that challenge relevant neural affective systems with direct electrical stimulation of the brain (ESB), namely direct activation of the
periaqueductal gray (PAG) in this work.
We argue that optimal preclinical progress in psychiatric
research could be achieved by directly focusing on the affective states of animals. The main presenting symptom of human
depression is sustained negative feelings that diminish eagerness to engage positively with life affirming activities. We now
know that subcortical circuits of all mammalian brains are capable of producing intense negative affective states including FEAR,
RAGE and a social-separation distress that has been labeled
the PANIC/GRIEF system (Panksepp, 1998; Panksepp and Biven,
2011; capitalizations reflect a nomenclature used to designate
primary-process emotional systems of the brain and denote
both a psychological experience and arousal of specific neuralcircuits). It is not uncommon for such sustained negative affect
to promote suicidal thoughts and actions. An estimated 864,950
individuals attempt to take their own lives each year in the
U.S. (see Substance Abuse and Mental Health Services Administration web site: http://www.samhsa.gov/). Within a given
year, 5–8% of the adult population suffers from a depressive
episode (see web sites for the National Alliance on Mental Illness:
http://www.nami.org/; Hope for Depression Research Foundation: http://hopefordepression.org), and depression is consistently
ranked in the top 10 of the world’s most debilitating disorders.
According to the DSM-IV, 10–25% of women and 5–12% of men
will suffer from a major depressive episode during their lifetime.
But what is depression? Presumably depression will eventually
be understood in terms of the up and down regulation of specific emotional-affective neural networks in the brain, at times
sufficiently intense and persistent enough to consider the final psychological states pathological. However, the scientific analysis of
depressive brain changes has yet to focus on specific cross-species
emotional networks of the brain. From the above perspective, the
development of animal models that begin to specify the therapeu-
1903
tically relevant affective brain networks involved in depression
should provide better guidance to understand what has gone
wrong in the depressed brain than very general stress procedures,
and thereby facilitate the discovery of novel treatments.
The new approach advocated here is what might be called
a “break-it and fix-it” strategy (this label is based on personal
communication with Rene Hen). The “breaking it” part will first
be considered. In order to do this coherently, animal emotional
states must be taken more seriously than has been common in
the field (Panksepp, 2011). This will entail selecting targets of
depressive-state inductions (e.g., manipulating brain emotional
systems directly to shift underlying affective neural-network variables) and development of more face-valid brain affective network
related outcome measures (utilizing affectively relevant neuralbased dependent variables). Focusing on the actual evolutionary
organization of emotional networks in the brain should allow a
more optimal understanding of what may be “broken” in depression and what may need to be “fixed.” A better understanding of
such possibilities should provide a basis for novel treatment modalities, which are needed in light of the recent STAR*D findings that
many currently used antidepressants are only marginally effective
(Philip et al., 2010; Pigott et al., 2010; Sinyor et al., 2010).
2. Depression modeling: brain circuit “break it and fix it”
strategies
Future preclinical models of depression need to take specific emotional systems of the brain more seriously as potential
endophenotypes that can guide both basic research and the development of future therapeutics (Panksepp, 2006; Sheehan et al.,
2004). Psychiatric modeling (e.g., Panksepp, 2010) demonstrates
that the most valuable translational evidence can be achieved by
understanding the various imbalances of brain emotional circuits
(Panksepp, 1998; Panksepp and Watt, 2011; Watt and Panksepp,
2009). Dynamic shifts in the operation of specific brain emotional
systems (Panksepp, 1998; Panksepp and Biven, 2011) may figure
heavily in dysphoric moods that characterize clinical depression
(also see contributions by Coenen et al., 2011, in this issue; Zellner
et al., 2011, in this issue).
There is already abundant support for these perspectives in both
human and non-human research. Neuroanatomical studies clearly
indicate that human depression is characterized by an over activity
of midline brain structures that mediate “resting state” activities in
the brain which are important for affectively relevant self-related
information processing in humans (Northoff et al., 2011, in this
issue). Homologous midline areas of many non-human animals
also exhibit the ‘footprints’ of over arousal in such brain regions
when exposed to repeated depressogenic challenges, such as variable intermittent stress procedures (Harro et al., 2011, in this issue;
Singewald, 2007). Translational analysis has highlighted that these
brain regions exhibit excessive glutamatergic activity and diminished GABAergic tone (Alcaro et al., 2010). Interestingly, the PAG is
under tonic GABAergic inhibition, and reductions of GABA tone here
promote negative affective states, partly by removing the breaks on
glutamatergic activity (Brandão et al., 2005). One might conclude
that affective imbalances within the midline subcortical and cortical networks of the brain are critically important for generating
depressive states. If so, they deserve the most intensive experimental focus in preclinical depression research.
Consider one potential model of animal depression that has
been largely neglected in the literature. Since the 1950s, it has
been known that bilateral damage to the lateral hypothalamic area
(LHA) along the trajectory of the medial forebrain bundle (MFB)
can produce a very depressed looking animal—one that does not
explore, eat, or drink normally (Teitelbaum and Epstein, 1962).
1904
J.S. Wright, J. Panksepp / Neuroscience and Biobehavioral Reviews 35 (2011) 1902–1915
Fig. 1. Successive effects of repeated injection of 0.1 g of colchicine on either side
of the far lateral hypothalamus via an electrode cannulae, on current thresholds
(60 Hz sine wave) for elicitation of forced motor responses (current and colchicine
spread probably extending to cerebral peduncles). Each curve is the average for four
animals (data from Panksepp et al., 1976).
These animals are inactive and not motivated to take care of their
bodily needs—but they can be partly rehabilitated by abundant
nursing care. They exhibit all the defining behavioral characteristics
of depression, but few have explicitly linked such classic observations with the etiology of depression, even though a reduction of
brain dopamine and reward SEEKING urges has recently been recognized as an endpoint of depressive cascades (e.g., Nestler and
Carlezon, 2006). Partial damage to the LHA/MFB may be a suitable animal model of depression. This could be further explored
and developed using reversible lesions that can be achieved with
bilateral colchicine lesions that disrupt axoplasmic flow (Panksepp
et al., 1976), and the shifting baselines of functional recovery could
be treated as measures of potential anti-depressant effects (Fig. 1).
This kind of decline in SEEKING tone could also be achieved by
a variety of other neural activity disruptors. As we will discuss,
localized electrical stimulation of the brain (ESB) applied to specific emotional networks can also lead to a depressive syndrome in
animals.
From these perspectives, it could be argued that most current
preclinical models of psychiatric disorders, though not without
merit, are not ideally suited to focus on the key neural features
of the underlying brain affective processes relevant to many psychiatric disorders, especially depression. Most current studies use
a large variety of general social and non-social stressors to induce
depressive states, and a large number of quite general outcome
measures to monitor the induced depressive states (again, for an
overall recent summary, see the Vol. 29, pp. 501–909 special issue of
this journal, edited by Markou, 2005). For instance, among the more
compelling existing models, repeated defeats in resident-intruder
aggression paradigms and sustained, unpredictable administration
of various physical stressors are used on the input (independent
variable) side, and a large number of general behavioral measures
(dependent variables such as exploration, swimming, and various
social interactions) are used to monitor the inferred status of brain
changes that may correspond to clinical depression.
Although these were excellent choices for the first generation of
preclinical modeling of depression, we would suggest that in addition to the knowledge already achieved at that level of analysis, a
shift to more specific affective approaches will be even more important for a more focused understanding that can lead to the discovery
of new and more effective pharmacotherapies. Since most popular
animal models of depression do not adequately focus on specific
brain emotional networks, the rest of this paper will focus on how
such a refined strategy could be implemented. In short, the time is
now ripe for a new generation of animal models based on existing
knowledge about the primary-process emotional network infrastructure of human and animal brains.
Accordingly, we here focus on a potential new era of preclinical
psychiatric model development—namely the use of both independent and dependent variables that manipulate and monitor specific
brain emotional systems. In this way, we may be able to more
precisely identify the types of brain systems that lead to various
forms of depressive despair and sift through their neurochemical
underpinnings for the most promising vectors for new medicinal
development. This approach may yield a more precise understanding of the underlying brain network changes that may eventually
provide better therapeutics, and may allow investigators to induce
and monitor specific brain affective changes that may guide the
development of more effective treatments for depression. Indeed,
such affective neuroscientific strategies have already yielded new
treatment strategies for depression and autism (see Burgdorf et al.,
2011, in this issue; Moskal et al., 2011, in this issue).
Thus, our theoretically derived goal in the following work was to
establish imbalanced relationships between the mammalian social
separation-distress (i.e., PANIC/GRIEF) system of the brain, specifically in the dorsal periaqueductal gray (dPAG), and to monitor acute
and chronic changes in the SEEKING system (for critical overviews
of this strategy, see Panksepp and Watt, 2011; Watt and Panksepp,
2009). Our hypothesis was that if we specifically over-activate negative separation-distress affects and related feelings for sustained
periods of time, we will produce behavioral indices of diminished
arousability of brain reward SEEKING systems (for a summary of
this system, see Alcaro and Panksepp, 2011, in this issue; Ikemoto,
2010; Panksepp and Moskal, 2008).
3. Background perspectives: from classic stress models to
modern affect models
Currently, one of the best animal models of inducing depression is the use of various “chronic unpredictable stress” paradigms,
where animals are constantly under the influence of unpredictable
stressors, such as shocks, wet cages, changing light cycles, periodic
food deprivation, and water deprivation; for reviews on chronic
stress models of depression see Willner (1997) and Anisman
and Matheson (2005). Though useful, the number of independent variables used in this model makes drawing inferences about
underlying neural causes of resulting depressed-like states and
behavior quite difficult. This is even the case when one simply uses
a single stressor such as foot shock or social defeat, since these
manipulations impact the whole brain through many input pathways. The complexity of such models increases the potential for
error in terms of sifting depression promoting brain causes vs. generalized stress correlates. Similarly, current behavioral measures
of depression in the rat, such as the forced-swim test, sucrose
intake, and varied measures of anxiety, lack the specificity of representing distinct neural-circuit functioning, making meaningful
depression-related interpretations challenging. Arguably, models
of depression that focus more directly on the most relevant underlying brain systems are needed. Of course, this is next to impossible
if we have no coherent vision of how affective processes emerge
from the brain. However, affective neuroscience approaches have
consistently indicated that wherever in the brain one evokes coherent emotional responses with ESB, those brain arousals mediate
affective states since they serve as “rewards” and “punishments” in
various learning tasks.
The linkage of depressive disorders to specific brain emotional
systems is in its infancy. Here we will focus on one model along
these lines that we are currently developing. The guiding idea is
that depression results from shifting levels of activity in two major
primary-process emotional systems: (i) sustained arousal of brain
PANIC/GRIEF (separation-distress) networks, which if sustained,
will then lead to feelings of “despair”, which (ii) partly reflects
J.S. Wright, J. Panksepp / Neuroscience and Biobehavioral Reviews 35 (2011) 1902–1915
diminishing activity in the general purpose appetitive, reward
SEEKING system, which is all too commonly labeled, rather imprecisely, as “the brain reward system” (for discussion, see Panksepp
and Moskal, 2008; Panksepp and Watt, 2011).
According to our vision of primary-process emotional systems
(Panksepp, 1998; Panksepp and Biven, 2011), it is the sustained
arousal of the PANIC/GRIEF emotional network that is most commonly activated by diminished social support which underlies the
depressogenic feelings of loss and sorrow often associated with
the loss of a loved one. In epidemiological studies, social loss is
a common precipitating cause of depression, but we hypothesize
that its final manifestation also requires a decrease in the SEEKING
urges that underlie euphoric excitement and intense exploratory
urges/motivations. If SEEKING is significantly diminished, then
this may contribute substantially to the dysphoria experienced in
depression. Thus, we believe that an optimal approach to understanding depression is to create an animal model that mimics the
affective imbalances that accompany increased PANIC/GRIEF and
decreased SEEKING feelings that may be foundational for many
forms of depression (Coenen et al., 2011, in this issue; Panksepp
and Watt, 2011; Watt and Panksepp, 2009; Zellner et al., 2011, in
this issue).
As a proof of concept, we conducted preliminary experiments where we attempted to provoke depressive states by
over-stimulating the neuroanatomical foci of the PANIC system
and evaluated whether the positive affect typically generated by
the SEEKING system was diminished. In other words, our aim
in the studies summarized here was to electrically stimulate
affects related to social separation-distress and measure resultant changes in SEEKING arousability by monitoring a particular
ultrasonic vocalization (USV) that rats emit, most commonly
during positive social interactions, which serves as a readout of
the activity within SEEKING—namely the 50 kHz USV. In past
work, this dependent measure has been validated to reflect a
general affective state that is aroused when a variety of rewarding
stimuli, especially social rewards, are anticipated and engaged
(Burgdorf et al., 2008; Knutson et al., 2002; Panksepp et al., 2002).
The mesolimbic dopamine-based SEEKING system is of critical
importance for the actual generation of the 50 kHz calls that reflect
arousal of this system (Burgdorf et al., 2007), and it is already
being used as a spontaneously manifested marker of drug desire in
rat drug addiction studies (Browning et al., 2011; for background
reviews, see Panksepp et al., 2002, 2004). Thus, we sought to create
a situation where both our major independent variable (PANIC via
PAG stimulation) and dependent variable (50 kHz USV-SEEKING
index) represented distinct, and theoretically relevant, affective
neural circuit functions within the brain.
There are many well-characterized “hot spots” within the overall PANIC network, which has been mapped in guinea pigs (Herman,
1979; Herman and Panksepp, 1981), domestic chickens (Bishop,
1984; Panksepp et al., 1988) and primates (Jürgens, 2009; Newman,
1988) that we could have chosen to stimulate. Among all these
candidates, one site stands out as potentially the best—the dorsal
PAG. It has long been known that the PAG is critical in the regulation and experience of both physical and psychological pain. In
fact, the dorsal PAG is the neural region within the PANIC network that requires the least electrical current to elicit coherent
separation distress vocalizations, as well as various other negative
emotional responses. There is also an abundance of data implicating the PAG in the generation of fear (Adamec et al., 2008;
Brandão et al., 2005, 2008; Graeff, 2004). This may be important
in the use of rats as a model species, since through selective breeding their separation-distress system may be vestigial (Panksepp,
2003a; Panksepp et al., 1992). In any event, from a primary-process
emotional perspective, the PAG can be thought of as a center of
negative affect integration and regulation (Panksepp, 1998—for
1905
overview of the basic functional anatomy of the PAG see Bandler
and Shipley, 1994; Behbehani, 1995). Because of its role in mediating several distinct negative affects (since such stimulations can
serve as punishments in learning tasks), we hypothesized that
electrical stimulation of the dorsal PAG (activation of PANIC and
probably FEAR and RAGE networks) would lead to a persistent
suppression of 50 kHz USVs—reflecting, ultimately, inhibition of
SEEKING and a diminished capacity to experience social reward
and euphoria.
4. Background: the PANIC circuitry of the PAG
Just as human infants emit panicked cries when they feel
lost as a result of being separated from social comfort, so do
many other mammalian and avian species, albeit not all. Most
notably, among the most studied species that exhibit “real”
separation-distress calls are primates (Levine and Wiener, 1988;
Suomi, 2006), dogs (Panksepp et al., 1978; Scott, 1974), guinea
pigs (Hennessy et al., 2009; Herman and Panksepp, 1978, 1981;
Pettijohn, 1979) and young chickens (Panksepp et al., 1980, 1988).
In contrast, the validity of laboratory rat and mouse models of
separation-distress remains debatable (Blumberg and Sokoloff,
2001; Panksepp, 2003a), albeit rats (but not mice) are a wonderful
species for the study of social play, which can highlight the condition of the underlying dopamine “energized” SEEKING system
(for overview, see Siviy and Panksepp, 2011, in this issue). 50 kHz
vocalizations are prevalent during play in rats, supporting the argument of their potential utility as a measure in depression studies.
Interestingly, there is high overlap between the SEEKING system
and neural regions involved in play (indeed in all the other positive social emotions, namely LUST and CARE)—most likely including
the neurochemisties involved in the actual generation of euphoric
excitement. This comes as no surprise, as play is a behavior that
consists of a cycling between anticipation and intense engagement.
It would be desirable to have a rat-sized laboratory-friendly
creature that would exhibit both real separation-calls and playfulness to evaluate the role of these social emotions in the regulation of
depressive disorders in a single species. Accordingly, we are in the
process of characterizing the social-phenotype of Octodon degus as
a potential convenient laboratory species for preclinical depression
studies. Degu pups exhibit robust separation distress vocalizations
when separated from their family and mothers. They also exhibit
physical “rough-and-tumble” play, and abundant social-affective
vocalizations, that can be used to index positive social emotional
processes (for a summary of this work, see Colonnello et al., 2011,
in this issue).
Although we anticipate future affective neuroscience work on
degus, until now, our work on the neural circuitry of depression
has been done in rats, using stimulation of PAG regions which contain separation-distress/PANIC circuitry in every species that this
response system has been mapped (i.e., chicks, guinea pigs and
primates). For present purposes, in order to evaluate whether stimulation of this brain region can evoke a depressive phenotype, we
assume that the relevant affective process still exists in rat brains
even though their spontaneous vocal expression of separation-calls
is vestigial. Still, we know that affectively negative 22 kHz USVs
can be evoked from the dorsal PAG, and the cFos neuronal activity
marker protein is expressed in this brain region when animals hear
such distress calls (Beckett et al., 1997; Sadananda et al., 2008).
Our premise here is that this brain region that provokes negative
affective vocalizations in rats may be used as a neuroanatomical
focus for the generation of affective shifts that are relevant for
depression. It is also important to note that evidence from human
brain imaging suggests that a homologous sadness-promoting
emotional system exists in this region of the human brain (Damasio
1906
J.S. Wright, J. Panksepp / Neuroscience and Biobehavioral Reviews 35 (2011) 1902–1915
Fig. 2. Self-induced sadness in humans (top) shows a similar arousal pattern as
the separation distress vocalization circuit in the guinea pig mapped with electrical stimulation of the brain (bottom). The human data according to Damasio et al.
(2000); the guinea pig data according to Herman and Panksepp (1981) and Panksepp
et al. (1988). Graph adapted from Panksepp (2003a,b).
et al., 2000). Quite strikingly, the neural substrates for human
sadness overlap substantially with the separation distress/PANIC
system of guinea pigs (for an overview, see Fig. 2) and FEAR also. In
sum, the fact that negative 22 kHz USVs are integrated in this brain
region in rats gives us confidence that relevant negative affects are
still elaborated in this species, albeit it may not be precisely the
negative social affect demonstrated in other species.
Of course, many consider the attribution of feeling states in
animals to be inappropriate anthropomorphism, but this view is
not supported by existing evidence (Panksepp, 1998, 2011). For
instance, as predicted, humans report intense feelings of negative affect, which could be characterized as fear, anxiety, agitation,
and/or impending doom when the dorsal PAG is electrically stimulated (Nashold et al., 1967; Young and Rinaldi, 1997). In a recent
study, human patients were implanted with electrodes into the rostral portion of the PAG (the same half of the PAG stimulated in the
experiments that will be described in this paper), to study the effect
on blood pressure. It was found that ESB of the dorsal regions led
to increases in blood pressure, and some of the patients reported
feelings of intense distress, as well as nausea (Green et al., 2005).
The autonomic and physiological alterations accompanying
panic attacks and those observed with electrical brain stimulation
of the dorsal PAG in both rats and humans show striking similarity
(for review/systematic comparison see: Graeff, 2004; Jenck et al.,
1995; Schenberg et al., 2001; Lovick, 2000), and there is abundant
evidence of a potential developmental link between separation
distress and depression, ever since John Bowlby emphasized the
life-long benefits of secure parent-infant relationships (Bowlby,
1988; for full summary, see Watt and Panksepp, 2009). According to
the DSM-IV, up to 65% of people with panic disorder also have major
depressive disorder. Based on symptom similarities, it has also
been proposed that panic-attacks in humans are more related to
separation-distress/PANIC feelings rather than FEAR systems that
are situated nearby in the PAG (Panksepp, 1998, 2003b). Interestingly, in about a third of individuals with both panic disorder and
major depressive disorder, the depression preceded the panic disorder (DSM-IV). Therefore, human studies of panic disorder may
aid our understanding of the functioning of the dorsal PAG and
may perhaps illuminate the nature of the etiology of depression.
Indeed, increases in gray matter within the midbrain of individu-
als with panic disorder suggest they may have enlargements of the
PAG (Protopopescu et al., 2006). Furthermore, increased cerebral
blood flow to the midbrain has been found in patients with panic
disorder in anticipation of anxiety-provoking stimuli (Boshuisen
et al., 2002). In addition, increases in glucose uptake in the midbrain are often evident in people with panic disorders (Sakai et al.,
2005). Finally, fMRI-based brain imaging has indicated that imminent threats readily “light-up” the PAG, while distal threats only
activate the amygdala (Mobbs et al., 2007).
Social defeat, an adult form of social loss, also impacts the PAG.
This model has been used to study features of the early onset
stages of depression in rats (Malatynska and Knapp, 2005), and
this stressor strongly impacts the PAG. Defeated animals exhibit
up-regulated beta-2 nicotinic acetylcholine receptor subunit transcription specifically, and potentially exclusively, in the PAG (Kroes
et al., 2007). Such up-regulation takes place in regions of the PAG
where social defeat promotes abundant 22 kHz vocalizations, calls
also observed with electrical stimulation of the dorsal PAG. Playback of these 22 kHz vocalizations promotes arousal of neurons in
the dorsal PAG (Beckett et al., 1997; Sadananda et al., 2008). Unconditioned fear/aversive stimuli, such as exposing rats to cats, also
increases neural arousal in the dorsal PAG, especially in the rostral
half (Canteras and Marina, 1999).
Most importantly, ESB of the dorsal PAG has long been shown
to sustain escape motivation conditioned place aversion in rats
(Delgado et al., 1954; Roberts and Cox, 1987; Schmitt et al., 1981,
1984). In such studies, rats show increased locomotor behavior
and agitation at low levels of PAG stimulation; at medium levels,
rats freeze, and at higher levels, rats exhibit intense arousal and
active escape behaviors. Interestingly, some of these behaviors are
inhibited in rats treated daily with 1 mg/kg of the antidepressant
fluoxetine for three weeks (Schenberg et al., 2001). Thus, abundant
data converge on the conclusion that the negative affect evoked by
PAG stimulation may have relationships to clinical depression.
5. A vocal affective indicator of SEEKING circuitry
arousal—50 kHz USVs
We will now briefly discuss the diminished arousability of the
reward SEEKING system as a second major vector in the genesis
of depression. Parenthetically, this system has long been known
as “The Brain Reward System” although this is a misnomer since
the main function of this circuitry is not to mediate the pleasure
of sensation but the eager pursuit of all resources needed for survival (Panksepp, 1981, 1982, 1998). Obviously, organisms must
seek resources to survive. Such a simple idea becomes complex
when one considers the multitude of behaviors involved in bringing about this adaptive end. Animals need many resources—food,
water, shelter, sex, to name a few—and it is clear that a single underlying neural system provides this capability and also elaborates the
affective urge to fulfill these needs. This urge is not well described
by the monolithic concept of “reward” or “pleasure” (although still
commonly used), but rather the positive excitement that can border on euphoria, of aspiring to forage for and thereby achieve life
supportive goals. The mesocorticolimbic dopamine system creates the overall arousal promoting foundation for the SEEKING
system—a generalized urge to pursue rewards—which then integrates with numerous unconditioned and conditioned drives and
environmental events, especially those that predict rewards (incentive salience states), yielding appetitive learning, which allows
organisms to become motivated toward specific ends as needs
arise. Easily monitored, quantifiable behaviors serve as readouts of
activity in the SEEKING system—from complex motor activities, to
simple measures such as invigorated sniffing (Rossi and Panksepp,
1992; Ikemoto and Panksepp, 1994) in rats, and most recently it has
been validated that socially-relevant affective vocalizations, emit-
J.S. Wright, J. Panksepp / Neuroscience and Biobehavioral Reviews 35 (2011) 1902–1915
ted in great quantities during positive social interactions, can also
serve as readouts of SEEKING. Namely 50 kHz USVs, especially of the
more abundant frequency modulated variety, are indicative of the
SEEKING system in action. Thus we have a direct behavioral way to
estimate arousal and arousability of the SEEKING system (Burgdorf
et al., 2007) during potential depressogenic manipulations such as
stimulation of the dorsal PAG.
To briefly amplify on these key issues, with half a dozen confirmatory lines of evidence: (i) 50 kHz USVs are representative of a
positive affective state in the rat (Knutson et al., 2002). (ii) Rats emit
increased rates of 50 kHz vocalizations in anticipation of electrical
stimulation to rewarding brain regions (Burgdorf et al., 2000) and
in anticipation of play (Knutson et al., 1998). (iii) Increased rates of
50 kHz USVs are also produced by rats during play and when tickled (Burgdorf and Panksepp, 2001; Panksepp and Burgdorf, 2003).
(iv) Microinjection of amphetamine into the nucleus accumbens
has been found to induce 50 kHz USVs (Burgdorf et al., 2001a,b;
Brudzynski et al., 2011), while lesions to the mesolimbic dopamine
pathway or dopamine receptor blockade attenuate them (Burgdorf
et al., 2007). (v) The number of 50 kHz vocalizations has been found
to positively correlate with conditioned place preference (Burgdorf
et al., 2007) and negatively correlate with aversion (Burgdorf et al.,
2001a,b). (vi) And perhaps most importantly, every brain region
that leads to the elicitation of 50 kHz vocalizations via electrical
brain stimulation has also supported self-stimulation, albeit, not
every region found to support self-stimulation has elicited 50 kHz
USVs (Burgdorf et al., 2007).
All this suggests that 50 kHz USVs may provide a more spontaneous and hence probably more specific index of positive SEEKING
affect than self-stimulation itself (Burgdorf et al., 2007), which is
an explicitly learned behavior. It is also clear that aversive stimuli
suppress 50 kHz USVs. For example, predatory odors, bright light,
and foot shock have all been found to decrease rates of 50 kHz USVs
(Burgdorf et al., 2001a,b; Covington and Miczek, 2003; Knutson
et al., 1999; Panksepp, 1998). Also, in rats that have been bred to
show high and low levels of 50 kHz USVs, the high-line animals
seem to be more resistant to depression than the other lines of
animals, which has already led to new ideas in antidepressant
development (see Burgdorf et al., 2011, in this issue). Finally, it
should be noted that our systematic tickling (i.e., heterospecific
1907
hand play procedure), allows us to directly assay the state of the
social-positive affect system that generates 50 kHz USVs (Panksepp
and Burgdorf, 2000, 2003; Panksepp, 2007).
5.1. Experiment 1: dorsal PAG stimulation can abolish positive
affect as monitored by 50 kHz USVs and promotes negative affect
as measured by 22 kHz USVs
We first determined the effect of dorsal PAG stimulation on positive and negative affect as measured by 50 kHz and 22 kHz USVs,
respectively. Six male adult rats were implanted with bipolar electrodes into the rostral half of the dorsal PAG and were repeatedly
habituated to a test chamber. Animals initially received a single one
second pulse of 60 Hz stimulation to the dorsal PAG every 10 s for
1 min (i.e., priming period intended to simulate conditions that are
known to induce Long Term Potentiation (LTP) in neural circuits).
They then received one pulse every 60 s for the remaining 9 min.
Animals were then left to freely vocalize for 5 min without intrusion. Two tickle sessions followed, each consisting of 1 min of 15
second-on/15-second-off tickling stimulation, followed by 4 min of
non-stimulated free vocalization. The total duration of the test was
25 min. Baseline data (i.e., same procedure without PAG stimulation collected the previous day) provided baseline control data for
behavioral changes observed on the stimulation day.
PAG stimulations completely abolished 50 kHz vocalizations
during the 10 min intermittent stimulation period, and also during
the 5 min post-stimulation observation as well as the two succeeding tickle periods (Fig. 3a). Robust 22 kHz USVs were emitted
during the brain stimulation and persisted during the 5 min poststimulation period (Fig. 3b). No 22 kHz USVs were emitted during
baseline testing (data not shown).
5.2. Experiment 2: PAG stimulation leads to a conditioned
suppression of positive affect as monitored with 50 kHz USVs
We next determined whether the suppression of 50 kHz USVs
could be contextually conditioned to verify that our results did not
represent direct motor output free of emotional value, or a simple lingering unconditioned response in the previous experiment
(i.e., in humans, affective changes provoked by brief emotional
Fig. 3. Male adult rats implanted with bipolar electrodes in the dorsal PAG received a 1 s 60 Hz sine wave pulse (current range 10–80 A) every 60 s for 10 min followed by
a 5-min post-stimulation break and two 5-min tickle periods. 50 and 22 kHz USVs were monitored during the testing session (n = 6). (A) Stimulation of the dPAG inhibits
positive affective 50 kHz USVs. Data is grouped together for the entire 25 min testing period (but the sequence of testing was the same as the right graph). (B) Stimulation of
the dPAG induces negative-affective 22 kHz USVs. Graph shows vocalizations for each of the segments of the testing period to illustrate the lasting effect of stimulation. No
22 kHz USVs were observed during baseline testing (data not shown).
1908
J.S. Wright, J. Panksepp / Neuroscience and Biobehavioral Reviews 35 (2011) 1902–1915
Fig. 4. The suppression of 50 kHz USVs was shown to be conditioned on extinction
day 1. Graph depicts 50 kHz USVs during 11 min recording period. On stimulation
days, PAG animals (n = 10) compared to control (n = 10) received a 1 s 60 Hz sine wave
pulse every 5 s for 60 s (current range 5–20 A: below escape threshold) followed
by a single 1 s pulse every 30 s (current range 10–80 A) for 9 minutes followed by a
one minute post-ESB interval. Following the stimulation period (represented above),
animals were tickled using the protocol described in the text (data not shown). All
animals had been given two habituation days in the test chamber (without ESB) prior
to baseline testing. The suppression of 50 kHz USVs showed a conditioned effect on
extinction day 1. **p ≤ 0.01; *p ≤ 0.05.
episodes are known to last longer than the precipitating circumstances). Twenty adult female rats were implanted with bipolar
electrodes into the rostral half of the dorsal PAG. Females were
used because they are more responsive to playful tickle interventions at older ages (Panksepp and Burgdorf, 2003). Ten female rats
served as sham operated controls that did not receive PAG-ESB, and
ten served as emotional-arousal treatment animals and received
PAG-ESB. Though formally considered a contextual conditioning
experiment, the brain stimulator, which was nearby, does produce an auditory ‘click’ at the initiation of brain stimulation as well
as sham stimulation. Therefore, it is possible that this ‘click’ also
served as a cue for conditioning.
After habituation and baseline days, treatment animals received
PAG stimulation on two consecutive days followed by two consecutive days of extinction trials and a spontaneous recovery test
30 days later. The procedure consisted of a 10 min “stimulation
period” followed by two tickle periods after a 1-min post-ESB interval. As in the previous experiment, the stimulation period began
with a “priming” period during the first 60 s, though here animals received a pulse every 5 s during the priming period (again
intended to potentially induce LTP). During this priming period,
animals received a one second 60 Hz sine wave pulse (5–20 A;
below escape threshold) every 5 s for the first 60 s of the stimulation period. A single pulse (10–80 A) was then administered every
30 s for the remaining 9 min. The current threshold for stimulations
was determined by the behavioral response of the animals ∼25%
of the stimulations induced only freezing behavior, ∼25% escape
behavior, and ∼50% behavioral responses in-between.
As observed in the first experiment, 50 kHz USVs were entirely
suppressed during stimulation days. The effect was conditioned, as
50 kHz USVs were also suppressed 24 h after the stimulation on the
first extinction day, but extinguished on the second extinction test.
(Fig. 4: results shown for the stimulation period.)
Fig. 5. 50 kHz USVs were suppressed, in animals having received chronic PAGESB (n = 10 treatment; 10 controls), during the 5 min restriction period in the
familiar box prior to gaining access to the open field (for time-line of procedure,
see Supplementary Figure 1). The suppression was stable across days. **p ≤ 0.01:
unpaired two tailed t-test.
as in the above experiment, but rebalanced for prior stimulation
experience). The same 10-min stimulation protocol employed in
experiment two was again used. Animals received this stimulation
protocol every day for fifteen consecutive days. Behavioral testing was conducted during the stimulation period (always at least
2 h after the stimulation treatment of that day), and during the 30
days following the final stimulation treatment (see Supplementary
Figure 1 for timeline of stimulation/testing).
The first test used was the voluntary exploration test. During
this test, animals were placed in a familiar test box (12 (1/4)′′ H × 12
(1/4)′′ W × 19 (1/2)′′ H, that they had previously been habituated to
for one 15 min session) with a small (2 (1/8)′′ diameter circular)
door, where they were restricted during the first 5 min on testing
days (17 and 18 days after the final stimulation treatment). The
door was then opened to allow 5 min of voluntary exploration of a
novel open field (24 (1/4)′′ × 24 (1/4)′′ floor area × 12′′ high).
Dorsal PAG stimulated animals showed suppressed 50 kHz
vocalizations in the familiar box during the 5 min restriction on
both test days, demonstrating a long term suppression in 50 kHz
vocalizations that was stable across days (Fig. 5). When the door
was opened, animals that had received dorsal PAG stimulation
(dPAG animals) showed fewer investigatory nose pokes into the
open field prior to exiting than controls (Fig. 6). Nose pokes were
measured as an animal’s head emerging out of the familiar start
box (at least beyond the ears) into the open field. After exiting into
the open field, dPAG animals spent less time exploring the environment (Fig. 7). Furthermore, dPAG animals entered into and exited
5.3. Experiment 3: chronic dorsal PAG stimulation leads to
decreased exploration, altered sucrose intake, and a persistent
suppression of positive affect as indexed by 50 kHz USVs
We next determined the non-contextual and long-term consequences of chronic dorsal PAG stimulation, in order to evaluate
whether a sustained depressive type response emerged. Again,
20 female rats with dorsal PAG electrodes were used (the same
Fig. 6. When placed into a familiar box with a small hole (2 (1/8)′′ circumference)
giving access to an open field, rats having received chronic PAG stimulation performed fewer nose pokes into a novel open field during a 5 min voluntary access
period (see Section 5.3 paragraphs 1 and 2 for more info). *p ≤ 0.05: unpaired two
tailed t-test.
J.S. Wright, J. Panksepp / Neuroscience and Biobehavioral Reviews 35 (2011) 1902–1915
Fig. 7. After 5 min in the restricted familiar box, the door was opened, allowing free
access into an open field. Rats having received chronic PAG stimulation (as in Fig. 5)
spent less time in the open field on post day 17 (when the open field was novel).
*p ≤ 0.05: unpaired two tailed t-test.
Fig. 8. Animals were placed into a familiar box (having been habituated for 15 min
on previous day) with a small hole (2 (1/8)′′ circumference) giving access to a novel
open field. During the first 5 min the rats were restricted in the familiar box, after
which the door was opened and animals were allowed to freely explore for 5 min.
Animals having received chronic PAG stimulation crossed between the familiar box
and the open field more than controls during the 5 min voluntary exploration test,
suggesting increased anxiety in the open field. *p ≤ 0.05:unpaired two tailed t-test.
out of the open field more than controls (Fig. 8), and produced more
fecal boli (Fig. 9).
Sucrose tests were administered for 2 consecutive days, with a
measurement and refilling taking place every 24 h; 0.8% sucrose
Fig. 9. Animals were placed into a familiar box (habituated for 15 min on previous
day) with a small hole giving access to a novel open field. During the first 5 min the
rats were restricted in the familiar box, after which the door was opened and animals
were allowed to freely explore for 5 min. Rats having received chronic PAG stimulation excreted more fecal pellets than controls during the 5-min restriction. *p ≤ 0.05.
1909
Fig. 10. Stimulation of the dPAG modestly decreased sucrose intake during dPAG
stimulation on post ESB day 1 (for time line, see Supplementary Figure 1) but exhibited increased intake when re-tested 30 and 31 days after the final stimulation,
with no intervening sucrose testing (n = 10) compared to control (n = 10). *p ≤ 0.05;
**p ≤ 0.01: repeated measures ANOVA with Bonferroni’s post test.
was used, and spillage was negligible. At the end of the 15
days of stimulation treatment, dPAG animals showed a decreased
sucrose intake. It is important to note that post day 1 could still
be considered as part of the stimulation period, as animals had
received their last stimulation within 24 h of the beginning of
the sucrose test. Surprisingly, 30 days following stimulation, the
inverse trend was evident, with animals that had received stimulation exhibiting a sustained increase of sucrose intake compare to
controls (Fig. 10).
To monitor changes in SEEKING system reactivity, 50 kHz USVs
were monitored at the indicated time points (Fig. 11), using novel
boxes of varying size, shape and color. In this type of test, which
promotes active SEEKING and exploration of the novel environment, it is well known that animals typically exhibit investigatory
vocalizations (Schwarting et al., 2007; Wöhr et al., 2008) soon after
simply being placed into the novel environment (a different novel
box was used for each test: contact corresponding author for specific information if interested). The test lasted 10 min and all USVs
were harvested. This test was conducted both during the stimulation period (approximately 2 h after stimulation) and periodically
during the 29 days following stimulation. Tests were conducted
both during the light and dark cycle to control for diurnal effects
that might be endogenously present as well as those resulting from
the dPAG-ESB. Stimulated animals showed a persistent suppression
Fig. 11. Animals were repeatedly placed in a novel box (different box each time) for
10 min, during which USVs were recorded. Animals having received chronic dPAG
stimulation (for time line, see Supplementary Figure 1) emit fewer 50 kHz USVs
when placed in a novel environment for at least 29 days after the final day of dPAGESB treatment (n = 10) compared to control (n = 10). Each point represents a distinct
10 min test. Testing was conducted periodically, sometimes in the light and sometimes in the dark period, both during the 15 days of the stimulation treatment and
intermittently during the 29 day period following the final stimulation.
1910
J.S. Wright, J. Panksepp / Neuroscience and Biobehavioral Reviews 35 (2011) 1902–1915
of 50 kHz USVs during each of the novel box tests, including the last
one (4 weeks) after the final PAG stimulation session (Fig. 11).
6. Interpretation and discussion of results
Understanding the neural networks responsible for the regulation and generation of negative affect is surely important for
understanding depression. The separation-distress circuit (PANIC)
is a prime candidate to participate in the genesis of depression, for it
provokes a strong negative affective state that was first emphasized
by John Bowlby in the psychiatric literature (for recent summaries,
see Panksepp and Watt, 2011; Watt and Panksepp, 2009). First,
when this circuit is chronically aroused at low levels sadness may be
experienced (Damasio et al., 2000, and see Fig. 2). Second, at potentially higher levels of arousal, more intense anxiety type affective
states are produced as reflected in behavioral indices in animals,
with freezing transitioning to flight, while in humans, feelings of
doom and more intense anxious states have been reported (Mobbs
et al., 2007; Nashold et al., 1967; Panksepp et al., 2011). Finally, at
peak levels, it is possible that full-blown panic attacks are aroused
(Graeff, 2004). Therefore, multiple shades of negative affect can
potentially arise from either a single or multiple emotional systems
based on the magnitude and/or patterns of activity.
Although we have focused on the PANIC system here, it is
clear that FEAR and RAGE are also represented in the dorsal PAG
(Panksepp, 1998) and should not be discounted as contributory
causes of the effects observed. However, RAGE responses are predominately elicited from intermediate regions of the rostral-caudal
axis of the PAG in the lateral column (Depaulis et al., 1992; Bandler
and Shipley, 1994), and afferents from the central nucleus of the
amygdala (important in the genesis of FEAR) project more strongly
to caudal aspects of the dorsal PAG, though these afferents also
synapse in the rostral portions.
All these possibilities will need to be disentangled by future
research, perhaps through the use of optogenetic procedures that
should allow more distinct stimulation of particular affective
systems. Such a procedure will also allow us to better localize
stimulation to the dorsomedial and dorsolateral columns of the
rostral PAG, as it is possible that, though electrodes were localized to the dorsomedial and dorsolateral columns (since they
were situated diagonally such that stimulation would arouse both
columns), electrical stimulation in the experiments reported here
might have also influenced more lateral columns and deep layers
of the superior colliculi, as well as other tissue surrounding the
dorsal PAG. Still, we would note that at present it is the separationdistress PANIC network that is more likely to figure in the type
of “psychic pain” which may help engender shame and guilt,
which are common in depression. Although some of the particularities above are speculative, until more refined functional studies
can be done contrasting potentially different influences of different primary-process emotional systems, the overall premise that
separation-distress responses arise from a distinct brain network
that may be appropriately called the PANIC circuit, has been supported by both animal and human studies, and it is clear that the
region of the dorsal PAG that received electrical stimulation here is
involved in negative affect.
In sum, the results from experiment 1 demonstrate that stimulation of the dorsal PAG inhibits the 50 kHz vocalizations that
are indicative of an excited SEEKING state, although an absence of
50 kHz USVs obviously does not imply a total absence of activity
within the SEEKING system. Behavioral analysis alone does not
provide enough resolution to adjudicate on the underlying neurophysiological details and will require the implementation of many
other approaches for clarity (e.g., electrophysiological ones such as
in Mu et al., 2011). In any event, the absence of vocalizations during
the 5-min period following dPAG stimulation suggests that the
aroused affective states persist well beyond the actual stimulation.
To probe the sensitivity of the SEEKING system we performed a
standardized assay of its reactivity to stimuli that provoke positive
social affect, namely our standard tickling test (more scientifically
known as “heterospecific hand play” that strongly promotes 50 kHz
USVs) (Panksepp and Burgdorf, 2000, 2003). Tickling administered
at 5 and 10 min after the final stimulation (in experiment 1) seemed
unable to elicit essentially any 50 kHz USVs, suggesting that dorsal
PAG stimulation creates a lasting state that resists the induction of
positive social affect. Indeed, playful excitement might be considered the antithesis of depression, and the fact that dPAG stimulation
can inhibit such excitement for at least 10 min after stimulation
suggests that dPAG stimulation, and/or the anticipation of further
stimulation, exerts a sustained powerful inhibition on the SEEKING system. The neurophysiological consequences will need to be
characterized with in vivo voltammetry for changes in dopamine
activity and even more refined measures for other types of potentially relevant mesolimbic neuronal activity (e.g., Mu et al., 2011).
It is also noteworthy that dPAG stimulation in these experiments
evoked persistent 22 kHz USVs during intervals between successive PAG stimulation—alarm vocalizations that are also commonly
emitted in the presence of cat odor. In the present experiment,
without any external threat, successive stimulations of the dPAG
led to vigorous 22 kHz vocalization emission in most rats, with
males typically exhibiting stronger responses than females. Following a single stimulation pulse a male rat might emit continuous
intermittent 22 kHz USVs for more than 10 min. In comparison,
22 kHz USVs in response to cat odor/presence are long and drawn
out, averaging more than 1 s per vocalization. After a single dorsal
PAG stimulation, such calls are emitted; however, repeated stimulations lead to shorter 22 kHz USVs, perhaps reflecting a different
quality of the evoked negative affect.
The results from experiment two demonstrate that our dependent variable, the 50 kHz USV, may exhibit both unconditioned
and conditioned reductions in response to repeated stimulation
of the dorsal PAG. Traditionally, conditioned place preferences
and avoidances have been used as the gold standard to evaluate
affective learning processes, but it is clear that emotional vocalizations can also be used (Burgdorf et al., 2001a,b). Such learning
in the present situation is suggested by the data, but not formally
evaluated; it is possible that our series of LTP-type stimulations
set in motion larger neurological/psychological processes outside
the realm of learning, such as a sustained shift in the sensitivity
of underlying brain affective systems. However, it is interesting
that on the 2nd extinction trial it was the ‘flat’ 50 kHz USVs,
which do not reflect dopamine-based euphoric arousal (Burgdorf
et al., 2007), that rebounded (data not shown), whereas the
frequency modulated vocalizations that are more responsive to
dopamine activity remained suppressed. One possibility is that flat
50 kHz calls represent a more cognitive type of call that animals use
to sample the social environment, and this type of state may also
be an initial state related to extinction learning, perhaps indicating
that animals “realize” no negative event is about to take place. If this
is true, then perhaps glutamatergic input into the nucleus accumbens from the many learning centers is involved. At the same time,
the continued suppression of frequency modulated 50 kHz calls
might reflect either a sustained suppression of euphoric excitement
related specifically to SEEKING, and/or reflect continued arousal of
negative affective substrates that are more sensitive to other types
of past social-emotional events. Yet, the sustained reductions of
positive 50 kHz USVs in a novel (unconditioned) environment up
to a month after the termination of chronic dPAG stimulation (Exp.
5.3) suggests a sustained shifting of temperamental-affective state
following acute stimulation is possible. With regard to modeling
depression, this long-term suppression of positive affect following
J.S. Wright, J. Panksepp / Neuroscience and Biobehavioral Reviews 35 (2011) 1902–1915
chronic dPAG stimulation is the most important finding of the
present work. Of course, a great deal of further work is needed to
evaluate the influence of each of the relevant variables, including
the durations and patterns of stimulation, ESB intensities, the
range of brain region within and outside the PAG able to provoke
such effect, as well as a variety of other parameters.
The results from the third experiment also help support the
relevance of chronic dPAG stimulation to a sustained depressive phenotype by the demonstration of decreased exploration,
increased anxiety, and decreased positive affect for a remarkably
long period following the final brain stimulation. This appears
to reflect an actual shift in affective phenotype. Perhaps importantly, in all of the present studies, the dPAG-ESB was administered
in a special stimulation chamber (outside the home cage), thus
allowing the home cage of animals to serve as a “safety zone.” In
future studies, it would be of interest to determine whether the
administration of the PAG-ESB within the home cage environment
would magnify the effects reported here, especially the degree and
quality of social interactions within and outside the home environments. If the currently observed depressive symptoms are more
severe when induced in the home environment, it may suggest
that such a paradigm might even more closely resemble depression in humans, since humans experiencing depression rarely have
such “safe zones,” except perhaps in the domain of formal therapy
and/or social interaction with close friends and family.
Thus, it will be important to evaluate how dPAG-ESB in both
new as well as home environments influence affective expressions
in each environment. We do know that our type of procedure is sufficiently strong for the apparent temperamental shifts to be readily
observed in new environments. For instance, the voluntary environmental exploration test (Figs. 5–9) allowed the use of repeated
test days to analyze the difference between dPAG-ESB animals
and controls in their qualitative engagement with new situations.
Though these tests were conducted more than two weeks after
the final dPAG-ESB sessions, 50 kHz USVs were still suppressed
in the PAG stimulated group across days, demonstrating the sustained power of the depressogenic cascade that had been set in
motion with a total of only 450 s of actual PAG-ESB. The finding
that dPAG animals showed fewer exploratory nose-pokes in this
test is, of course, open to multiple interpretations. One interpretation is that fewer nose pokes represent a decreased investigative
impulse, consistent with the 50 kHz vocalization data. A second
interpretation is that fewer nose pokes into a novel environment
before entering it reflect increased “caution,” or rather is perhaps
representative of an impulsive tendency to not thoroughly investigate the safety of a new environment prior to entering it. Of
course, both of these may be secondary effects of an elevated form
of anxiety.
Our interpretation of increased negative affect in animals having received dorsal PAG stimulation is supported by their apparent
increased eagerness to return to a familiar location during exploration of an open field, which in our work was represented by more
enters/exits between the familiar “start” box and the novel open
field, perhaps reflecting an increase in general anxiousness, which
is also consistent with the finding that dPAG animals produced
more fecal boli than controls. The insecure behavior displayed by
dPAG rats, i.e., fewer nose pokes into a novel environment before
entering; more enter/exits between a familiar box and a novel open
field, resembles the behavioral consequences reported in bilateral
olfactory bulbectomized (OBx) models of depression in rats. For
example, it has been reported that OBx rats show increased locomotor activity in a well-lit open field (Klein and Brown, 1969) and
increased enter/exits into the open arms of an elevated plus maze
(Song et al., 1996). A direct comparison of hedonic profiles between
OBx and dPAG-ESB rats, especially with 50 kHz USVs will be needed
to resolve such issues.
1911
Fig. 12. dPAG animals decreased in body weight by the 15th stimulation day (n = 10)
compared to control (n = 10). *p ≤ 0.05: two way ANOVA with Bonferroni post test.
They gradually regained most of their lost weight during the subsequent month of
testing.
Traditional measures may be less informative. For instance,
sucrose intake has traditionally been used as a measure of hedonic
response in preclinical animal models of depression. However, this
remains a controversial measure of anhedonia since it lacks clinical face validity (i.e., do depressed people consume fewer sweets?).
Toward the end of the 15 days of stimulation, dPAG animals showed
a significant decrease in sucrose intake, consistent with much of the
rodent literature employing various stress models of depression.
We initially observed the predicted effect. However, 30 days after
the final stimulation, dPAG animals showed a significant increase
in sucrose intake compared to controls. This later finding is consistent with depressogenic early life isolation in Octodon degus
(see Colonnello et al., 2011, in this issue). Perhaps the increased
sucrose intake is demonstrative of a self-medication response, as
highlighted by the capacity of sweets to release brain dopamine as
well as opioids (Hernandez and Hoebel, 1988; Avena et al., 2008).
From this perspective, we would also hypothesize that animals having received chronic dPAG stimulation will show a smaller
dopamine response in the nucleus accumbens in response to both
appetitive and consummatory sucrose testing. In other words,
perhaps dPAG animals require more exposure to sweet tastes to
experience hedonically “normal” levels of dopamine response. The
initially decreased sucrose intake of dPAG animals may represent
a more acute stress response from the PAG-ESB, but may also be
reflective of overall weight changes in these animals (dPAG-ESB
animals lost weight). However, by one month post-stimulation,
the weights and daily weight changes had stabilized (Fig. 12) in
both experimental and control animals, so the increased sucrose
intake then likely represents what is perhaps the more sustained
neurological-temperamental change that had consolidated as a
result of the earlier PAG-ESB experiences. In contrast, in the novel
box experiments (Fig. 11), which promote a SEEKING response in
rats, 50 kHz USVs were consistently suppressed both during stimulation days and 29 days following stimulation. This suggests that
the positive hedonics of gustatory treats and those that promote
50 kHz USVs reflect different underlying brain processes.
Obviously, without further brain analyses, there are multiple
possibilities as to what has been altered by chronic dPAG stimulation. For example, the separation-distress/PANIC system, either
within the PAG or in other parts of the network, might be sensitized
and up-regulated. Conversely, SEEKING substrates might be chronically desensitized from the stimulation, or both. Further research
is needed to determine these neurological alterations. To this end,
we plan to evaluate how dorsal PAG stimulation induces neuronal
activity markers such as fos expression to determine potential “hot
spots” within the brain that need to be further investigated. Abundant work still needs to be done with this new affective circuit level
1912
J.S. Wright, J. Panksepp / Neuroscience and Biobehavioral Reviews 35 (2011) 1902–1915
approach to preclinical modeling of depression, including evaluating various traditional markers of the depressive phenotype such
as chronic elevations of adrenal steroid production and changes in
brain growth factors, to name just a few. However, all other animal
models that do not directly address the relevant affective changes
may be even more remote from clinical reality.
7. Conclusions and future possibilities
According to the DSM-IV, a major depressive episode must last
at least two weeks and include one or two core symptoms and a
subset of additional symptoms, for a total of 5 symptoms. A full set
of criteria has yet to be achieved by any preclinical model, but the
current approach is one of the few that seems to be resolving some
of the cardinal emotional symptoms. Indeed, the core symptoms
of depression are affective: first, a depressed mood; and second, a
decreased interest or ability to feel pleasure. Chronic dorsal PAG
stimulation in these experiments clearly induced such affective
changes, with effects that were sustained for more than two weeks.
In our estimation, the persistent and long term suppression of positive affect, as indicated by diminished 50 kHz USVs, may be the
gold-standard “direct” preclinical indicator of a depressed mood in
rat models, thus potentially fulfilling the first of the two core symptoms of a depressive episode. Dorsal PAG stimulated animals also
exhibited behaviorally apparent decreased interest in their environments, as indicated by their decreased exploration of a novel
open field, thus potentially fulfilling the second core symptom of a
major depressive episode.
To meet criteria for a major depressive episode, a person must
also exhibit at least 3 of the following 4 symptoms, according to
the DSM-IV, depending on whether both or only one core symptom is present: (i) greater than 5% body weight change in a month
(either increase or decrease); (ii) changes in sleep (either insomnia or hypersomnia); “psychomotor agitation or retardation”; (iii)
fatigue; decreased concentration and/or indecisiveness; inappropriate and/or (iv) persistent feelings of guilt; finally, feelings and
thoughts of death. Of course the last symptom cannot be preclinically modeled; issues such as feelings of “guilt” and “thoughts about
suicide”–tertiary-processes of the mind, cannot be probed in animal models. However, we anticipate that rats that have received
chronic dorsal PAG stimulation will exhibit the other symptoms.
Indeed, it was clear that our animals exhibited more than a 5%
decrease in body weight during the 15 day stimulation period,
and that this decreased weight persisted following the stimulation.
However, this trait was less stable and showed increased variability
as time progressed following stimulation compared to the suppression of 50 kHz vocalizations. Dorsal PAG stimulated animals
also showed more enter/exits between a familiar box and a novel
open field, which may be indicative of an anxious, indecisive state
of mind, but obviously further research is needed to classify and
delineate these consequences. However, the important point here
is that shifts in specific affective states can now be targeted as both
independent and dependent variables in preclinical modeling of
depression.
This initial attempt at direct affective modeling of depression
in rats suggests that chronic stimulation of the rostral half of the
dorsal PAG could be developed into the first affective circuit-based
preclinical approach to understanding depression in a commonly
available laboratory animal. Such results encourage us to move
toward other new, direct brain-based animal models of depression.
There are, of course many other possibilities along these lines, and
for that reason we presented data with a reversible lateral hypothalamic lesion approach (Fig. 1), which, along with other direct neural
challenges, could be used to directly simulate reductions in SEEKING urges that characterize depression, as lesions there are known
to markedly reduce the frequency modulated 50 kHz USVs that
are premier indicators of depression-relevant positive affect in rats
(Burgdorf et al., 2007) as well as many of the other symptoms of
depression.
Currently, the only other direct brain-based model is the
robust depressive syndrome induced by olfactory bulbectomy (for
overview, see Song and Leonard, 2005), but this model was a
serendipitous discovery, as opposed to a theoretically developed
strategy focused on manipulating and monitoring the underlying affective states of the brain. We realize that the discussion
of affectively experienced states in non-human animals remains
a controversial and contentious topic in animal brain research, but
we think that we have reached a time where the weight of evidence strongly supports the existence of such psychological states
in the brains of many other creatures (Panksepp, 2011). Our contention is that the more we invest in such lines of research, the
more rapidly we will solve the dilemma of affective disorders in
humans. And perhaps, from the knowledge of the underlying shifts
in neuroaffective substrates, we will be able to develop not only a
better understanding of depression but eventually better therapies
(Burgdorf et al., 2011, in this issue).
In closing, although this set of studies only points the way
toward a new class of preclinical models of depression, one must
wonder why such approaches remain to be developed. Our suspicion is that this is due to several factors, especially the necessary
perspective that in order to pursue such lines of thought one
has to take the emotional feelings of animals seriously—a difficult
assumption for many, as such a premise is perhaps unable to be
deductively demonstrated. It has to be inferred inductively from
convergent evidence, and new predictions.
The premise that non-human animals have affective experiences that can be scientifically studied has not been either a popular
or even widely recognized aspect of animal models of psychiatric
disorders, yet functional and structural deformations in the brain
mechanisms that mediate various types of affective experiences,
woven from underlying neural circuitry changes, are probably the
way many psychiatric disorders arise. It is surely no secret that
most investigators in behavioral neuroscience do not feel comfortable talking about the feelings of the animals they study. Let us
consider these two issues briefly, to weigh the potential benefits of
making the above assumption, versus the potential risks.
The issue of whether animals have any form of consciousness—
defined here as internal subjective experience (i.e., phenomenal
consciousness or qualia)—that can be scientifically studied—is a
very controversial topic. Our view is on the liberal fringe of educated
views, since we think the evidence itself now robustly supports the
contention that the subjective experience of others, including nonhuman animals, can be scientifically investigated and understood
with the aid of modern neuroscience (for a review of this issue,
see Panksepp, 2011). With respect to brain emotional systems, it is
basically a law of nature that wherever one applies ESB and evokes
coherent emotional behaviors, animals treat those shifted brain
states as rewards and punishments in learning situations, such as
conditioned place preferences and aversions. Since animals do not
exhibit any clear propositional-symbolic communication, we obviously must rely on their behaviors to read their minds. However,
this “reading” is not just a matter of belief in science, but a matter of whether a view generates a rich set of testable/verifiable
predictions.
Thus we are in an unusual intellectual era where the weight
of evidence supports the unpopular view that animals do have
emotional feelings while the weight of educated opinion supports
the opposite traditional Cartesian view. Although few would disagree that new approaches to modeling psychiatric disorders in
non-humans are needed, it still remains generally unacceptable
in neuroscience circles to make the leap to subjective inferences,
J.S. Wright, J. Panksepp / Neuroscience and Biobehavioral Reviews 35 (2011) 1902–1915
and to base preclinical psychiatric models on a direct confrontation with animal emotions. This is not at all a unique instance in
the history of science. There are abundant examples during the
past four centuries since the start of “The Enlightenment” with a
similar flavor, and though this alone is not sufficient to promote
acceptance of new world views, it certainly suggests that sometimes the most obvious phenomena continue to be ignored during
periods when most investigators feel we are making great advances
using standard approaches. However, no systematic brain research,
in human or other animals, has yet yielded a brand new type of
anti-depressant, despite half a century of efforts.
In any event, to utilize neural circuit functions within animal
brains as a foundation for understanding the evolutionary sources
of human emotional feelings is a basic evolutionary premise of
cross-species affective neuroscience (Panksepp, 1998). Again, the
key evidential base of the argument is that restricted stimulation
of the subcortical neural circuits that control emotional-instinctual
behaviors can routinely serve as “rewards” and “punishments”
in the learned control of behavior. This is the gold-standard of
evidence that suggests that in order to make progress on many psychiatric issues, we simply must take the emotional feelings of other
animals seriously, for they seem to arise from ancient subcortical
brain regions that control instinctual emotional-action sequences
in all mammalian brains.
Of course, the attribution of emotional feelings to animals
raises a host of ethical issues which we will not address here (see
Panksepp, 2011). Still, this suggests that investigators are wise to
consider psychological distress issues in other animals. Accordingly, for our part, one of our research goals is to see if we can
induce the temperamental changes described in this paper by
applying PAG-ESB to animals that are anesthetized with agents
such as Versted, etc. Since we think as mechanistically as most
neuroscientists, our belief is that we should be able to simulate
some of the emotional effects described in this paper in animals
even when stressors are applied when they are not fully conscious. The empirical evidence is suggestive for such possibilities,
as we have been able to characterize key attributes of the SEEKING
system—such as ESB evoked sniffing—in anesthetized animals. For
instance, the thresholds for the induction of sniffing in anesthetized
rats is slightly higher but strictly correlated (r’s about .9 and above)
with self-stimulation of the same electrodes in fully awake animals (Rossi and Panksepp, 1992). Likewise, we have mapped the
circuitry for separation distress calls (Fig. 2) in anesthetized guinea
pigs (Herman, 1979; Herman and Panksepp, 1981) and chickens
(Bishop, 1984; Panksepp et al., 1988).
Granting that we are able to promote a depressive phenotype by
stimulating the dorsal PAG in anesthetized rats, this finding may be
of great benefit for understanding the underlying neural circuitry,
not only from a practical but also a scientific point of view. A big
problem in the field is that global external stressors that are used to
evoke depressive phenotypes in animals quite simply have massive
effects on the brain (see Harro et al., 2011, in this issue), and the
sifting of the most important brain changes is a major challenge,
not fully accomplished yet.
Considering its global stressful nature, it is to be expected that
dPAG ESB will also have massive brain effects in waking animals, including those arising from fear of the next stimulation,
but perhaps such brain changes are more limited in the brain than
states evoked by massive peripheral stressors (but this remains to
be demonstrated). However, we anticipate that ESB under anesthesia may help delimit the critical brain systems even further.
And with ongoing spectacular developments in optogenetic brain
stimulation procedures, more focal activation of the most relevant brain systems could be achieved, and thereby candidate
circuits most relevant to depression research could be more easily
characterized.
1913
In conclusion, we feel one key aim for the next generation of
preclinical models of depression should be to get at the various
positive and negative brain affective systems as directly as possible.
The challenges in achieving this are not only scientific, but cultural.
However, it is our prediction that when we start to more widely utilize such models, substantive progress will be made more rapidly,
revealing more precisely the specific brain regions and processes
that need to be influenced. If so, we can also anticipate that with
a better understanding of the underlying affective circuit imbalances, new treatment vectors will be identified and implemented.
For instance, it is now known that modulation of glutamatergic
and cholinergic drives in midbrain regions may figure in depressive cascades (e.g., Burgdorf et al., 2011, in this issue, and Kroes
et al., 2007) not only in the PAG, but also nearby affective processing regions of the brainstem (see Normansell and Panksepp, 2011,
in this issue). In other words, a variety of positive and negative feelings are stitched into primordial brain regions in ways that have
barely been functionally envisioned by modern neuroscience.
There are other new lines of research related to depression that
may also lead to new treatment options, such as those that surround
the hypothesis that altered/hyperactive glia-related cytokines and
other pro-inflammatory immune-molecules are involved in psychiatric disorders (see Hennessy et al., 2009; Arakawa et al., 2011,
in this issue). This line of research might not be independent of the
dorsal PAG, as a relatively high glia/neuron ratio exists there (Beitz,
1985). Indeed, this, and many other issues need to be addressed in
order to understand depression. We are optimistic that with the
gains in molecular biology, neuroscience, and behavioral research
during the last century, we are approaching an era of developing
novel treatments for depression and other psychiatric and neurological diseases based on our emerging understanding of the shared
primary-process emotional processes of mammalian brains.
Acknowledgement
This work was supported by Hope for Depression Research
Foundation. We thank Sheri Six for her assistance with this work.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.neubiorev.2011.08.004.
References
Adamec, R., Holmes, A., Blundell, J., 2008. Vulnerability to lasting anxiogenic effects
of brief exposure to predator stimuli: sex, serotonin and other factors-relevance
to PTSD. Neurosci. Biobehav. Rev. 32, 1287–1292.
Alcaro, A., Panksepp, J., Witczak, J., Hayes, D.J., Northoff, G., 2010. Is
subcortical–cortical midline activity in depression mediated by glutamate and
GABA? A cross-species translational approach. Neurosci. Biobehav. Rev. 34,
592–605.
Alcaro, A., Panksepp, J., 2011. The SEEKING mind: primal neuro-affective substrates
for appetitive incentive states and their pathological dynamics in addictions and
depression. Neurosci. Biobehav. Rev. 35, 1805–1820.
Angell, M., 2004. The Truth About the Drug Companies: How They Deceive Us and
What to do About it. Random House, New York.
Anisman, H., Matheson, K., 2005. Stress, depression, and anhedonia: caveats concerning animal models. Neurosci. Biobehav. Rev. 29, 525–546.
Arakawa, H., Cruz, S., Deak. T. 2011. From models to mechanisms: Odorant
communication as a key determinant of social behavior in rodents during illnessassociated states. Neurosci Biobehav Rev. this issue, in press.
Avena, N.M., Rada, P., Hoebel, B.G., 2008. Evidence for sugar addiction: behavioral
and neurochemical effects of intermittent, excessive sugar intake. Neurosci.
Biobehav. Rev. 32, 20–39.
Bandler, R., Shipley, M., 1994. Columnar organization in the midbrain periaqueductal
gray—modules for emotional expression? Trends Neurosci. 17, 379–389.
Beckett, S.R.G., Duxon, M.S., Aspley, S., Marsden, C.A., 1997. Central C-Fos expression
following 20 kHz/ultrasound induced defense behavior in the rat. Brain Res. Bull.
42, 421–426.
Behbehani, M.M., 1995. Functional-characteristics of the midbrain periaqueductal
gray. Prog. Neurobiol. 46, 575–605.
1914
J.S. Wright, J. Panksepp / Neuroscience and Biobehavioral Reviews 35 (2011) 1902–1915
Bishop, P., 1984. Brain and opiate modulation of avian affective vocalizations.
Unpublished Ph.D. Dissertation, Bowling Green State Univ., Bowling Green, OH.
Beitz, A.J., 1985. The midbrain periaqueductal gray in the rat. I. Nuclear volume, cell
number, density, orientation, and regional subdivision. J. Comp. Neurol. 237,
445–459.
Blumberg, M.S., Sokoloff, G., 2001. Do infant rats cry? Psychol. Rev. 108, 83–95.
Boshuisen, M.L., Ter Horst, G.J., Paans, A.M.J., Reinders, A.A.T.S., Den Boer, J.A., 2002.
rCBF differences between panic disorder patients and control subjects during
anticipatory anxiety and rest. Biol. Psychiat. 52, 126–135.
Bowlby, J., 1988. A Secure Base: Parent Child Attachment and Healthy Human Development. New York, Basic Books.
Brandão, M.L., Borelli, K.G., Nobre, M.J., Santos, J.M., Albrechet-Souza, L., Oliveira,
A.R., Martinez, R.C., 2005. Gabaergic regulation of the neural organization of fear in the midbrain tectum. Neurosci. Biobehav. Rev. 29, 1299–
1311.
Brandão, M.L., Zanoveli, J.M., Ruiz-Martinez, R.C., Oliveira, L.C., Landeira-Fernandez,
J., 2008. Different patterns of freezing behavior organized in the periaqueductal
gray of rats: association with different types of anxiety. Behav. Brain. Res. 188,
1–13.
Browning, J.R., Browning, D.A., Maxwell, A.O., Dong, Y., Jansen, H., Panksepp, J.,
Sorg, B.A., 2011. Positive affective vocalizations during cocaine and sucrose
self-administration: a model for spontaneous drug desire in rats. Neuropharmacology 61, 268–275.
Brudzynski, S.M., Silkstone, M., Komadoski, M., Scullion, K., Duffus, S., Burgdorf,
J., Kroes, R.A., Moskal, J.R., Panksepp, J., 2011. Effects of intraaccumbens
amphetamine on production of 50 kHz vocalizations in three lines of selectively
bred Long-Evans rats. Behav. Brain Res. 217, 32–40.
Burgdorf, J., Knutson, B., Panksepp, J., 2000. Anticipation of rewarding electrical
brain stimulation evokes ultrasonic vocalization in rats. Behav. Neurosci. 114,
320–327.
Burgdorf, J., Knutson, B., Panksepp, J., Ikemoto, S., 2001a. Nucleus accumbens
amphetamine microinjections unconditionally elicit 50-kHz ultra sonic vocalizations in rats. Behav. Neurosci. 115, 940–944.
Burgdorf, J., Knutson, B., Panksepp, J., Shippenberg, T., 2001b. Evaluation of rat ultrasonic vocalizations as predictors of the conditioned aversive effects of drugs.
Psychopharmacology 155, 35–42.
Burgdorf, J., Panksepp, J., Moskal., J.R. 2011. Frequency-modulated 50kHz ultrasonic
vocalizations a tool for uncovering the molecular substrates of positive affect.
Neurosci Biobehav Rev. this issue, in press.
Burgdorf, J., Panksepp, J., 2001. Tickling induces reward in adolescent rats. Physiol.
Behav. 72, 167–173.
Burgdorf, J., Wood, P.L., Kroes, R.A., Moskal, J.R., Panksepp, J., 2007. Neurobiology of
50-kHz ultrasonic vocalizations in rats: electrode mapping, lesion, and pharmacology studies. Behav. Brain Res. 182, 274–283.
Burgdorf, J., Kroes, R.A., Moskal, J.R., Pfaus, J.G., Brudzynski, S.M., Panksepp, J.,
2008. Ultrasonic vocalizations of rats (Rattus norvegicus) during mating, play,
and aggression: behavioral concomitants, relationship to reward and selfadministration of playback. J. Comp. Psychol. 122, 357–367.
Canteras, N.S., Marina, G., 1999. Fos-like immunoreactivity in the periaqueductal
gray of rats exposed to a natural predator. NeuroReport 10, 413–418.
Coenen, V.A., Schlaepfer, T.E., Maedler, B., Panksepp, J., 2011. Cross-species affective
functions of the medial forebrain bundle-Implications for the treatment of affective pain and depression in humans. Neurosci. Biobehav. Rev., 2010 December
22 [Epub].
Colonnello, V., Iacobucci, P., Fuchs, T., Newberry, R.C., Panksepp, J., 2011. Octodon
degus. A useful animal model for social-affective neuroscience research: basic
description of separation distress, social attachments and play. Neurosci. Biobehav. Rev. 35, 1854–1863.
Covington III, H.E., Miczek, K.A., 2003. Vocalizations during withdrawal from opiates
and cocaine: possible expressions of affective distress. Eur. J. Pharmacol. 467,
1–13.
Damasio, A.R., Grabowski, T.J., Bechara, A., Damasio, H., et al., 2000. Subcortical
and cortical brain activity during the feeling of self-generated emotions. Nat.
Neurosci. 3, 1049–1056.
Delgado, J.M., Robert, W., Miller, N.E., 1954. Learning motivated by electrical stimulation of the brain stem. Am. J. Physiol. 179, 581–593.
Depaulis, A., Keay, K.A., Bandler, R., 1992. Longitudinal neuronal organization of
defensive reactions in the midbrain periaqueductal gray region of the rat. Exp.
Brain Res. 90, 307–318.
Graeff, F.G., 2004. Serotonin, the periaqueductal gray and panic. Neurosci. Biobehav.
Rev. 28, 239–259.
Green, A.L., et al., 2005. Deep brain stimulation can regulate arterial blood pressure
in awake humans. NeuroReport 16, 1741–1745.
Harro, J., Kanarik, M., Matrov, D., Panksepp, J. 2011. Mapping patterns of depressionrelated brain regions with cytochrome oxidase histochemistry: Relevance of
animal affective systems to human disorders, with a focus on resilience to
adverse events. Neurosci Biobehav Rev. this issue, in press.
Hennessy, M.B., Schiml-Webb, P.A., Deak, T., 2009. Separation, sickness, and depression: a new perspective on an old animal model. Curr. Dir. Psychol. Sci. 18,
227–231.
Herman, B.H., 1979. An exploration of brain social attachment substrates in guinea
pigs. Unpublished Ph.D. Dissertation, Bowling Green State Univ., Bowling Green,
OH.
Herman, B.H., Panksepp, J., 1978. Effects of morphine and naloxone on separation
distress and approach attachment: Evidence for opiate mediation of social affect.
Pharmacol Biochem Behav. 9, 213–220.
Herman, B.H., Panksepp, J., 1981. Ascending endorphin inhibition of distress vocalization. Science 211, 1060–1062.
Hernandez, L., Hoebel, B.G., 1988. Feeding and hypothalamic stimulation increase
dopamine turnover in the accumbens. Physiol. Behav. 44, 599–606.
Ikemoto, S., 2010. Brain reward circuitry beyond the mesolimbic dopamine theory: a neurobiological theory. Neurosci. Biobehav. Rev., doi:10.1016/j.neubiorev.
2010.02.001, February 10. Epub ahead of print.
Ikemoto, S., Panksepp, J., 1994. The relationship between self-stimulation and sniffing in rats: does a common brain system mediate these behaviors? Behav. Brain
Res. 61, 143–162.
Jenck, F., Moreau, J.L., Martin, J.R., 1995. Dorsal periaqueductal gray-induced aversion as a simulation of panic anxiety: elements of face and predictive validity.
Psychiat. Res. 57, 181–191.
Jürgens, U., 2009. The neural control of vocalization in mammals: a review. J. Voice
23, 1–10.
Klein, D., Brown, T.S., 1969. Exploratory behavior and spontaneous alternation in
blind and anosmic rats. J. Comp. Physiol. Psychol. 68, 107–110.
Knutson, B., Burgdorf, J., Panksepp, J., 1998. Anticipation of play elicits highfrequency ultrasonic vocalizations in young rats. J. Comp. Psychol. 112, 65–73.
Knutson, B., Burgdorf, J., Panksepp, J., 1999. High-frequency ultrasonic vocalizations
index conditioned pharmacological reward in rats. Physiol. Behav. 66, 639–643.
Knutson, B., Burgdorf, J., Panksepp, J., 2002. Ultrasonic vocalizations as indices of
affective states in rats. Psychol. Bull. 128, 961–977.
Kroes, R.A., Burgdorf, J., Otto, N.J., Panksepp, J., Moskal, J.R., 2007. Social defeat, a
paradigm of depression in rats that elicits 22-kHz vocalizations, preferentially
activates the cholinergic signaling pathway in the periaqueductal gray. Behav.
Brain Res. 182, 290–300.
Levine, S., Wiener, S.G., 1988. Psychoendocrine aspects of mother–infant relationships in nonhuman primates. Psychoneuroendocrinology 13, 143–154.
Lovick, T.A., 2000. Panic disorder—a malfunction of multiple transmitter control systems within the midbrain periaqueductal gray matter? Neuroscientist 6, 48–59.
Malatynska, E., Knapp, R.J., 2005. Dominant-submissive behavior as models of mania
and depression. Neurosci. Biobehav. Rev. 29, 715–737.
Markou, A. (Ed.), 2005. Special issue: animal models of depression and antidepressant activity. Neurosci. Biobehav. Rev. 29, 501–909.
Moskal, J. R., Burgdorf, J., Kroes, R.A., Brudzynski, S.M., Panksepp, J. 2011. A novel
NMDA receptor glycine-site partial agonist, GLYX-13, has therapeutic potential
for the treatment of autism. Neurosci Biobehav Rev. this issue, in press.
Mobbs, D., Petrovic, P., Marchant, J.L., et al., 2007. When fear is near: threat imminence elicits prefrontal-periaqueductal gray shifts in humans. Science 317,
1079–1083.
Mu, P., Neumann, P.A., Panksepp, J., Schlüter, O.M., Dong, Y., 2011. Exposure to
cocaine alters dynorphin-mediated regulation of excitatory synaptic transmission in nucleus accumbens neurons. Biol. Psychiat. 69, 228–235.
Nashold, B.S., Wilson, W.P., Slaughter, D.G., 1967. Sensations evoked by stimulation
in the midbrain of man. J. Neurosurg. 30, 14–24.
Newman, J.D. (Ed.), 1988. The Physiological Control of Mammalian Vocalizations.
Plenum, New York.
Nestler, E.J., Carlezon Jr., W.A., 2006. The mesolimbic dopamine reward circuit in
depression. Biol. Psychiat. 59, 1151–1159.
Normansell, L., Panksepp, J. 2011. Glutamatergic modulation of separation distress:
Profound emotional effects of excitatory amino acids in chicks. Neurosci Biobehav Rev. this issue, in press.
Northoff, G., Wiebking, C., Feinberg, T., Panksepp, J., 2011. The ‘resting-state
hypothesis’ of major depressive disorder—a translational subcortical–cortical
framework for a system disorder. Neurosci. Biobehav. Rev., 2010 December 28
[Epub ahead of print].
Panksepp, J., 1981. Hypothalamic integration of behavior: rewards, punishments,
and related psychobiological process. In: Morgane, P.J., Panksepp, J. (Eds.),
Handbook of the Hypothalamus, vol. 3. Part A. Behavioral Studies of the Hypothalamus. Marcel Dekker, New York, pp. 289–487.
Panksepp, J., 1982. Toward a general psychobiological theory of emotions. Behav.
Brain Sci. 5, 407–467.
Panksepp, J., 1998. Affective Neuroscience: The Foundations of Human and Animal
Emotions. Oxford University Press, New York.
Panksepp, J., 2003a. Can anthropomorphic analyses of “separation cries” in other
animals inform us about the emotional nature of social loss in humans? Psychol.
Rev. 110, 376–388.
Panksepp, J., 2003b. Trennungsschmerz als mogliche ursache für panikattacken—
neuropsychologische Uberlegungen und Befunde. Personlichkeitsstorung: Theorie und therapie 7, 245–251.
Panksepp, J., 2006. Emotional endophenotypes in evolutionary psychiatry. Prog
Neuro-Psychopharm Biol Psychiat. 30, 774–784.
Panksepp, J., 2007. Neuroevolutionary sources of laughter and social joy: modeling
primal human laughter in laboratory rats. Behav. Brain Res. 182, 231–244.
Panksepp, J., 2010. Affective neuroscience of the emotional BrainMind: evolutionary perspectives and implications for understanding depression. Dialog Clin
Neurosci. 12, 533–545.
Panksepp, J., 2011. Toward a cross-species neuroscientific understanding of the
affective mind: do animals have emotional feelings? Am. J. Primatol. 73,
545–561.
Panksepp, J., Biven, L., 2011. The Archaeology of Mind: Neuroevolutionary Origins
of Human Emotion. Norton, W. W. & Company, Inc., New York.
Panksepp, J., Fuchs, T., Iacabucci, P., 2011. The basic neuroscience of emotional experiences in mammals: The Case of subcortical FEAR circuitry and implications for
clinical anxiety. App Animal Behav Sci. 129, 1–17.
J.S. Wright, J. Panksepp / Neuroscience and Biobehavioral Reviews 35 (2011) 1902–1915
Panksepp, J., Burgdorf, J., 2000. 50k-Hz chirping (laughter?) in response to conditioned and unconditioned tickle-induced reward in rats: effects of social housing
and genetic variables. Behav. Brain Res. 115, 25–38.
Panksepp, J., Burgdorf, J., 2003. “Laughing” rats and the evolutionary antecedents of
human joy? Physiol. Behav. 79, 533–547.
Panksepp, J., Herman, B., Conner, R., Bishop, P., Scott, J.P., 1978. The biology of social
attachments: Opiates alleviate separation distress. Biol Psychiat. 13, 607–618.
Panksepp, J., Knutson, B., Burgdorf, J., 2002. The role of emotional brain systems in
addictions: a neuro-evolutionary perspective. Addiction 97, 459–469.
Panksepp, J., Newman, J.D., Insel, T.R., 1992. Critical conceptual issues in the analysis
of separation distress systems of the brain. In: Strongman, K.T. (Ed.), International Review of Studies on Emotion, vol. 2. John Wiley & Sons, pp. 51–72.
Panksepp, J., Nocjar, C., Burgdorf, J., Panksepp, J.B., Huber, R., 2004. The role of emotional systems in addiction: a neuroethological perspective. In: Bevins, R.A.,
Bardo, M.T. (Eds.), 50th Nebraska Symposium on Motivation: Motivational Factors in the Etiology of Drug Abuse. Nebraska, Lincoln, pp. 85–126.
Panksepp, J., Meeker, R., Bean, N.J., 1980. The neurochemical control of crying. Pharmacol. Biochem. Behav. 12, 437–443.
Panksepp, J., Meeker, R., Reilly, P., Vilberg, T., 1976. Reversible CNS lesions and
disruption of self-stimulation by inhibition of axoplasmic flow. In: Wauquiex,
A., Rolls, E.R. (Eds.), Brain Stimulation Reward. North Holland Publishing Co.,
Amsterdam, pp. 118–120.
Panksepp, J., Moskal, J. 2008. Dopamine and SEEKING: Subcortical reward systems
and appetitive urges. In: A. Elliot, A. (Ed.) Handbook of approach and avoidance
motivation, Taylor & Francis Group, LLC, New York, pp. 67-87.
Panksepp, J., Normansell, L.A., Herman, B., Bishop, P., Crepeau, L., 1988. Neural and
neurochemical control of the separation distress call. In: Newman, J.D. (Ed.),
The Physiological Control of Mammalian Vocalizations. Plenum, New York, pp.
263–300.
Panksepp, J., Watt, J., 2011. Why does depression hurt? Ancestral primary-process
separation-distress (PANIC) and diminished brain reward (SEEKING) processes
in the genesis of depressive affect. Psychiatry 74, 5–14.
Pettijohn, T.F., 1979. Attachment and separation distress in the guinea pig. Dev.
Psychobiol. 12, 73–81.
Pigott, H.E., Leventhal, A.M., Alter, G.S., Boren, J.J., 2010. Efficacy and effectiveness
of antidepressants: current status of research. Psychother. Psychosomat. 79,
267–279.
Philip, N.S., Carpenter, L.L., Tyrka, A.R., Price, L.H., 2010. Pharmacologic approaches
to treatment resistant depression: a re-examination for the modern era. Expert
Opin. Pharmacother. 11, 709–722.
Protopopescu, X., Pan, H., Tuescher, O., et al., 2006. Increased brainstem volume in
panic disorder: a voxel-based morphometric study. NeuroReport 17, 361–363.
Roberts, V.J., Cox, V.C., 1987. Active avoidance conditioning with dorsal central gray
stimulation in a place preference paradigm. Psychobiology 15, 167–170.
Rossi III, J., Panksepp, J., 1992. Analysis of the relationships between self-stimulation
sniffing and brain-stimulation sniffing. Physiol. Behav. 51, 805–813.
Sadananda, M., Wohr, M., Schwarting, R.K., 2008. Playback of 22-kHz and 50-kHz
ultrasonic vocalizations induce differential c-fos expression in rat brain. Neurosci. Lett. 435, 17–23.
1915
Sakai, Y., Kumano, H., Nishikawa, M., et al., 2005. Cerebral glucose metabolism associated with a fear network in panic disorder. NeuroReport 16, 927–931.
Schenberg, L.C., Bittencourt, A.S., Sudre, E.C.M., Vargas, L.C., 2001. Modeling panic
attacks. Neurosci. Biobehav. Rev. 25, 647–659.
Schmitt, P., Di Scala, G., Jenck, F., Sandner, G., 1984. Periventricular structures, elaboration of aversive effects and processing of sensory information. In: Modulation
of Sensorimotor Activity During Alterations in Behavioral States. Alan R. Liss,
Inc., New York, pp. 393–414.
Schmitt, P., Sandner, G., Karli, P., 1981. Escape and approach induced by brain stimulation: a parametric analysis. Behav. Brain Res. 2, 49–79.
Schwarting, R.K.W., Jegan, N., Wöhr, M., 2007. Situational factors, conditions and
individual variables which can determine ultrasonic vocalizations in male adult
Wistar rats. Behav. Brain Res. 182, 208–222.
Scott, J.P.,1974. Effects of psychotropic drugs on separation distress in dogs. In:
Proceedings of the Collegium International Neuropsychopharmacologicum.
Excerpta Medica International Congress Series, no. 359, pp. 735–745.
Sheehan, T.P., Chambrs, R., Russell, D.S., 2004. Regulation of affect by the lateral
septum: implications for neuropsychiatry. Brain Res. Rev. 46, 71–117.
Singewald, N., 2007. Altered brain activity processing in high-anxiety rodents
revealed by challenge paradigms and functional mapping. Neurosci. Biobehav.
Rev. 31, 18–40.
Sinyor, M., Schaffer, A., Levitt, A., 2010. The sequenced treatment alternatives to
relieve depression (STAR*D) trial: a review. Can. J. Psychiat. 55, 126–135.
Siviy, S., Panksepp, J., 2011. In search of the physiological substrates for social playfulness in mammalian brains. Neurosci. Biobehav. Rev. 35, 1821–1830.
Song, C., Earley, B., Leonard, B.E., 1996. The effects of central administration of
neuropeptide Y on behavior, neurotransmitter, and immune functions in the
olfactory bulbectomized rat model of depression. Brain Behav. Immunol. 10,
1–16.
Song, C., Leonard, B.E., 2005. The olfactory bulbectomised rat as a model of depression. Neurosci. Biobehav. Rev. 29, 627–647.
Suomi, S.J., 2006. Risk, resilience, and gene × environment interactions in rhesus
monkeys. Ann. N. Y. Acad. Sci. 1094, 52–62.
Teitelbaum, P., Epstein, A.N., 1962. The lateral hypothalamic syndrome: recovery of
feeding and drinking after lateral hypothalamic lesions. Psychol. Rev. 69, 74–90.
Watt, D.F., Panksepp, J., 2009. Depression: an evolutionarily conserved mechanism
to terminate separation-distress? A review of aminergic, peptidergic, and neural
network perspectives. Neuropsychoanalysis 11, 5–104.
Willner, P., 1997. Validity, reliability and utility of the chronic mild stress model of
depression: a 10-year review and evaluation. Psychopharmacology (Berl.) 134,
319–329.
Wöhr, M., Houx, B., Schwarting, R.K.W., Spruijt, B., 2008. Effects of experience and
context on 50-kHz vocalizations in rats. Physiol. Behav. 93, 766–776.
Young, R.F., Rinaldi, P.C., 1997. Brain stimulation. In: North, R.B., Levy, R.M. (Eds.),
Neurosurgical Management of Pain. Springer, New York, pp. 288–290.
Zellner, M.R., Watt, D.F., Solms, M., Panksepp, J., 2011. Affective neuroscientific and
neuropsychoanalytic approaches to two intractable psychiatric problems: why
depression feels so bad and what addicts really want. Neurosci. Biobehav. Rev.,
January 15 [Epub ahead of print].