BackgroundPrevious studies have suggested that physical activity mayhave antidepressant

and/or anti-anxiety effects.AimsTo examine the bidirectional relationship between

physicalactivity and common mental disorders and establish theimportance of context, type
and intensity of activityundertaken.MethodA clinical examination of 40 401 residents of
Norway wasundertaken. Participants answered questions relating to thefrequency and
intensity of both leisure-time and workplaceactivity. Depression and anxiety were measured
using theHospital Anxiety and Depression Scale (HADS). Biological andsocial data were
also collected.ResultsThere was an inverse relationship between the amount ofleisure-time
physical activity and case-level symptoms ofdepression. This cross-sectional association was
only presentwith leisure-time (as opposed to workplace) activity and wasnot dependent on the
intensity of activities undertaken.Higher levels of social support and social engagement
wereimportant in explaining the relationship between leisureactivity and depression.
Biological changes such asalterations to parasympathetic vagal tone (resting pulse)
andchanges to metabolic markers had a less important role.ConclusionsIndividuals who
engage in regular leisure-time activity of anyintensity are less likely to have symptoms of
depression. Thecontext and social benefits of exercise are important inexplaining this
relationship.Declaration of interestNone

inhabitants of the Nord-Trøndelag County aged 20–89 years(n= 92 936) were invited to a
clinical examination as part of ageneral health screening. The overall objectives, methods
andcontents of the HUNT-2 study have been described in detailelsewhere.24Measurement of
physical activityAll participants were asked how often they engaged in both lightand intense
leisure-time physical activity. Light activity wasdefined as an activity that did not lead to
being sweaty or outof breath, while intense activity was any leisure-time activitythat did
result in sweating or breathlessness. Participants weregiven four options for each question:
none, less than 1 h per week,1–2 h per week or 3 h or more per week. They were also
askedabout how physically active they were at work. Participants weregiven four options in
relation to their work activity: mostlysedentary, required to walk a lot, walk and lift a lot, or
intensephysical work. A number of examples were given to help defineeach of these
categories.Assessment of depression and anxietyAll participants were asked to complete the
Hospital Anxietyand Depression Scale (HADS).25The HADS is a self-reportquestionnaire
comprising 14 four-point Likert-scaled itemscovering anxiety and depression over the past 2
weeks. It wasdesigned to avoid false-positive cases among individuals withsomatic illness. It
contains no questions on somatic symptoms,sleep or appetite disturbance, focusing instead on
the psychologicaland cognitive symptoms relevant to the two disorders. A cut-offscore of 8 in
each subscale has been found to be optimal for casefinding, with sensitivity and specificity
estimates of about 0.80.26Mediating and confounding variablesA wide range of potential
confounding or mediating factors werealso considered. Based on existing knowledge, a
conceptualhierarchical framework was constructed to demonstrate how eachof these factors
may interact with both levels of activity and risk ofcommon mental disorder.27This is shown
in Fig. 1.Socioeconomic factorsInformation on participants’ age, gender and marital status
wasobtained from the Norwegian National Population Registry.Participants were asked to
record their highest completededucation level, with responses coded on a three-point
ordinalscale of compulsory (primary), secondary school and university.An index of social
class based on individuals’ current occupationwas calculated according to the International
Erikson–Goldthorpe–Portocarero classification.28Participants were askedif they smoked
cigarettes and were requested to complete theCAGE questionnaire (a four-question screening
instrument foralcohol problems).29Participants were also asked how many of Physical
activity and common mental disorderstheir immediate relatives (mother, father, sibling or
child) had amental illness.Physical illness/disabilityMany somatic diseases limit the amount
of physical activity that ispossible, while also independently increasing the risk of
commonmental disorders.30Therefore they are potentially an importantconfounder in the
relationship between physical activity andcommon mental disorders. The number of somatic
diagnosesreported by each participant (from the list of angina pectoris,myocardial infarction,
asthma, cancer, diabetes, epilepsy, hyper-tension, arthritis, osteoporosis, respiratory disease,
stroke andthyroid disease) was considered as a continuous variable inregression models.
Participants were also asked to rate their levelof functional impairment owing to any physical
diseases (on afour-point scale: none, slightly, moderately or severely). They werealso asked if
their movement was limited (for any reason), and ifso, to what degree.Metabolic markersA
trained nurse obtained measurements of participants’ height,weight, hip and waist
circumference. This allowed participants’body mass index (BMI) and waist–hip ratio to be
calculated. BothBMI and waist–hip ratio have been associated with depression,31suggesting
that obesity may be an important mediator in anyrelationship between physical activity and
common mentaldisorders. Non-fasting blood samples were also taken and analysedto
measure total cholesterol, triglyceride and glucose levels. Eachof these is known to alter
according to levels of physical activityand to be associated with common mental
disorders.30,32,33Measurement of autonomic nervous system activityClinically trained
nurses or technicians measured participants’blood pressure and heart rate on three occasions.
The firstmeasurement was taken after the participant had been seated for2 min, with
subsequent measurements occurring at 1 minintervals. The mean value of the second and
third pulsemeasurement was considered to be the resting heart rate. Theexact mechanism for
exercise-induced bradycardia has been thesubject of some debate, although there is
reasonable evidence tosuggest it is due to increased parasympathetic vagal tone.20Theuse of
resting heart rate as a measure of autonomic nervoussystem activity is further supported by
studies showing a closeassociation between resting pulse and more validated measuresof
autonomic function such as respiratory sinus arrhythmia andbaroreflex sensitivity.34Level of
social supportThe extent of the individual’s social network and support wasassessed via two
separate questions. Participants were asked‘How many good friends do you have?’ with the
instruction toonly count those who they could confidentially talk to and whowould help them
when they were in need. They were also askedhow often they took part in social activities.
The estimatednumber of friends was considered as a continuous variable, andthe frequency of
social activities was recorded on a four-pointscale (never or only a few times a year, once a
week, once or twicea month, or more than once a week).

Initiated as part of the ongoing deliberation about the nosological structure of DSM, this
review aims to evaluatewhether the anxiety disorders share features of responding that define
them and make them distinct from depressivedisorders, and/or that differentiate fear disorders
from anxious-misery disorders. The review covers symptom self-reportas well as on-line
indices of behavioral, physiological, cognitive, and neural responding in the presence of
aversive stim-uli. The data indicate that the anxiety disorders share self-reported symptoms of
anxiety and fear; heightened anxietyand fear responding to cues that signal threat, cues that
signal no threat, cues that formerly signaled threat, and contextsassociated with threat;
elevated stress reactivity to aversive stimuli; attentional biases to threat-relevant stimuli
andthreat-based appraisals of ambiguous stimuli; and elevated amygdala responses to
threat-relevant stimuli. Some differ-ences exist among anxiety disorders, and between anxiety
disorders and depressive disorders. However, the differencesare not fully consistent with
proposed subdivisions of fear disorders vs. anxious misery disorders, and comparative datain
large part are lacking. Given the high rates of co-morbidity, advances in our understanding of
the features of re-sponding that are shared across vs. unique to anxiety and depressive
disorders will require dimensional approaches. Insummary, the extant data help to define the
features of responding that are shared across anxiety disorders, but are in-sufficient to justify
revisions to the DSM nosology at this time.

In summary (Table 3), the findings from contextconditioning suggest that anxiety disorders
thathave been tested (PTSD and PD) are characterizedby hypersensitivity to contexts in
which threat will/may be delivered. However, the degree to whichthis hyper-sensitivity
discriminates anxiety disor-ders from mood disorders, or discriminates fear dis-orders from
other anxiety disorders, is not yetknown.Summary of explicit threat cue and
contextconditioning and implications for DSM.Whereas elevated sensitivity to explicit threat
cuesand safety cues (including cues during extinction)may be characteristic of anxiety
disorders, it is un-known whether it is more characteristic of someanxiety disorders vs. other
anxiety disorders, andwhether it is exclusive to anxiety disorders relativeto depression.
Similarly, extant data suggest thatcertain anxiety disorders are characterized by
hy-persensitivity to contexts in which threat will/maybe delivered, but the degree to which
this hypersen-sitivity discriminates anxiety disorders from otherdisorders, or discriminates
fear disorders from anx-ious-misery disorders is not known. Thus, the lit-erature pertaining to
aversive conditioning helps todefine features of responding that are common toanxiety
disorders, but is insufficient to justify revi-sions to the organizational structure of the
DSMnosology.ANXIETY DISORDERS AND STRESS REACTIVITYAnother extensive
body of research measures psy-chophysiological response to a variety of differentaversive
stimuli, without testing thelearningof anassociation with neutral stimuli. Such “stress
reac-tivity” measurement illuminates differences be-tween anxious groups and controls, and
betweenanxiety disorders. Anticipatory baseline respondingin studies of “stress reactivity”
can be viewed as anindex of contextual anxiety. The acute responsemeasured during actual
delivery of stressors can beviewed as an UR (when using generic stressors, suchas shock), or
a response to personally relevant stim-uli (when using disorder-specific stressors, such
astrauma reminders for PTSD). Stress reactivity par-adigms typically measure skin
conductance, heartrate, respiration, muscle tension, or startle reflex toindex the
somato-visceral response component offear and anxiety. The most thorough evaluationwould
entail assessment of generic stressors and dis-order-specific stressors across different anxiety
dis-orders and mood disorders. However, instead, al-most every study pertains to a single
anxietydisorder vs. healthy controls, and usually tests ei-ther generic or disorder-specific
stressors. Results

are summarized in Table 5. Notably, this reviewdoes not cover pharmacological probes that
exertcentral nervous system effects, such as meta-chloro-phenylpiperazine or yohimbine, but
rather is re-stricted to assessment of reactivity to inherentlyaversive stimuli that may occur in
the natural envi-ronment.Combining the psycho-physiological resultsfrom the studies of
individual anxiety disorders andthe few studies that address more than one anxietydisorder
(Table 5), it may be concluded that PTSDand PD show elevations in baseline anticipation
ofgeneric stressors, (58, 59) and perhaps more so inPTSD than PD, (60) but do not show
elevations inacute response to generic stressors relative to con-trols (58, 59). Also,
individuals with PD show astronger baseline response (e.g., (61)) but not astronger
physiological acute response to disorder-specific stressors (e.g., (62, 63)), although there
areoccasional exceptions (e.g., (64)). The disorder-specific stressors for PD have consisted
mostly ofcarbon dioxide inhalation paradigms.10,11The ma-jority of PTSD individuals
additionally exhibitstronger acute physiological response to trauma re-minders compared to
controls (e.g., (69)] althoughnot always (e.g., (70, 71)). SOP (nongeneralized)and SP may
possess stronger baseline and acutephysiological response to disorder-specific
stimulicompared to healthy controls (e.g., (72–75)), al-though baseline and acute response
have not beenevaluated to highly aversive generic stressors such asshock, and the findings
remain tentative. Evidencepertaining to GAD and OCD also is limited byinsufficiently
aversive generic stressors. Availableevidence provides an inconsistent picture
regardingbaseline differences between GAD and healthycontrols (76, 77) and
fornon-anxiousindividuals tosometimes have stronger autonomic responses toacute generic
stressors than individuals with GAD(e.g., (77)), which may be due to tonic inhibitoryeffects
in GAD (e.g., (78)). Individuals with OCDdid not differ from controls during baseline record-