Relationship of Posttreatment Decentering and Cognitive Reactivity to
Relapse in Major Depression
David M. Fresco
Kent State University
Zindel V. Segal, Tom Buis, and Sydney Kennedy
Centre for Addiction and Mental Health
Z. V. Segal et al. (2006) demonstrated that depressed patients treated to remission through either
antidepressant medication (ADM) or cognitive–behavioral therapy (CBT), but who evidenced mood-
linked increases in dysfunctional thinking, showed elevated rates of relapse over 18 months. The current
study sought to evaluate whether treatment response was associated with gains in decentering—the
ability to observe one’s thoughts and feelings as temporary, objective events in the mind—and whether
these gains moderated the relationship between mood-linked cognitive reactivity and relapse of major
depression. Findings revealed that CBT responders exhibited significantly greater gains in decentering
compared with ADM responders. In addition, high post acute treatment levels of decentering and low
cognitive reactivity were associated with the lowest rates of relapse in the 18-month follow-up period.
Keywords: major depression, cognitive behavior therapy, decentering, modes of mind
Major depressive disorder (MDD) is a prevalent and debilitating
national health problem with lifetime and 12-month prevalence
rates of 17% and 7%, respectively, according to the replication of
the National Comorbidity Survey (Kessler, Berglund, et al., 2005;
Kessler, Chiu, Demler, Merikangas, & Walters, 2005). Despite the
availability of treatments of established efficacy, there is a growing
consensus that the outcome for patients following acute treatment
is relatively poor (Fava, Rafanelli, Grandi, Conti, & Belluardo,
1998). Estimates suggest that approximately 50% of patients ex-
perience a recurrence within 2 years of remission following short-
term treatment (Keller, Lavori, Rice, Coryell, & Hirschfeld, 1986),
and the risk of MDD’s becoming a chronic problem increases
substantially with each additional episode experienced (Solomon
et al., 2000). Because clinical management of patients with de-
pression targets symptom reduction within the episode as the
primary goal, little attention has been paid to the development of
interventions that provide both symptomatic relief and durable
Hollon, Stewart, and Strunk (2006) reviewed evidence that
cognitive–behavioral therapy (CBT) interventions for depression
are equivalent to antidepressant medication (ADM) in reducing
acute symptoms and are twice as durable, as compared with
medication, for reducing rates of relapse or recurrence (Hollon et
al., 2006). Findings such as these have generated considerable
interest in understanding the mechanisms by which CBT produces
these benefits. Hollon et al. (2006) differentiated between two
kinds of enduring effects. First, treatment effects refer to factors
that mitigate problems that would not have gone away in the
absence of some formal intervention. From a research standpoint,
interventions can be evaluated for their treatment effects by as-
sessing the problem they putatively target. This assessment is
accomplished by demonstrating that the problem does not reoccur
or, if it reoccurs, that the frequency and/or intensity is lessened as
compared with before the delivery of the intervention. Thus,
curative interventions work by eliminating or reversing the under-
lying processes that would otherwise lead to the continuation of
the disorder. Similarly, preventive effects refer to an intervention’s
capacity to reduce risk for problems in the future—presumably by
thwarting the causal processes that could result in the initial onset
or recurrence of the disorder. The value of this distinction lies in
its suggestion that some treatment-induced changes may be rele-
vant to symptom reduction, whereas others pertain to illness mit-
igation and future risk. Although the benefits of CBT in producing
durable treatment gains for MDD are well established, isolating the
mechanisms by which CBT produces these results has proven
more elusive. For instance, reductions in negative cognitive prod-
ucts (i.e., the surface-level thoughts or dysfunctional attitudes) do
not appear to be the mechanism by which CBT produces its results,
as reductions in these thoughts have tended to covary with reduc-
tions in depression symptoms, occur in response to noncognitive
treatments such as ADM therapy and interpersonal psychotherapy,
and show little association with the durability of the treatment
gains (cf. Hollon et al., 1992; Hollon & Kriss, 1984; Imber et al.,
1990; Safran, Vallis, Segal, & Shaw, 1986; Simons, Garfield, &
Murphy, 1984). Rather, changes in cognitive structures, core
schema, or core processes are regarded as representing critical
change mechanisms in cognitive therapy (Beck, Rush, Shaw, &
Emery, 1979; Safran et al., 1986).
A potentially fruitful avenue for identifying vulnerability factors
for depression and mechanisms for durable change resides at the
confluence of cognitive science, clinical science, and affective
neuroscience (cf. Barnard & Teasdale, 1991; Brewin & Power,
1999; LeDoux, 1996; Samoilov & Goldfried, 2000; Teasdale,
1999). These models all take the view that there exists a complex
multilevel relationship between cognitive processing and emo-
David M. Fresco, Department of Psychology, Kent State University;
Zindel V. Segal, Tom Buis, and Sydney Kennedy, Centre for Addiction
and Mental Health.
Correspondence concerning this article should be addressed to David M.
Fresco, Department of Psychology, Kent State University, Kent, OH
44242. E-mail: email@example.com
Journal of Consulting and Clinical Psychology
2007, Vol. 75, No. 3, 447–455
Copyright 2007 by the American Psychological Association
tional processing and distinguish between explicit higher order
conceptual processing, which involves primarily rule-based learn-
ing, and more rapid, associational processing. Building on this
higher and lower processing typology, Teasdale further articulated
three modes of mind that have a differential association with
emotional disorders such as major depression. Of most relevance
to major depression is what Teasdale called mindless emoting,
which is associated with purely reactive, sensory-driven emoting
without focal awareness of conceptual or schematic meanings. In
this mode of mind, individuals’ thinking may be characterized as
rigid, interlocked processing patterns such as repetitive, rumina-
tive, and negatively self-focused thinking. Another important
mode of mind is conceptualizing–doing, which Teasdale charac-
terized as predominated by a focus on conceptual content and
analysis (e.g., writing a paper, planning one’s day, or solving a
logical problem). Finally, Teasdale described mindful
experiencing–being, which he characterized as times when indi-
viduals are focally aware of thoughts, internal and external sensa-
tions, and potential holistic meanings that these multiple sources of
information might convey. Of interest, Teasdale’s modes-of-mind
typology corresponds favorably to the distinctions of emotional
mind, reasonable mind, and wise mind, described by Linehan
(1993) in dialectical behavior therapy. Further, both theorists have
posited that individuals use all three modes throughout their day
and that it is commonplace to experience many switches among
modes. However, in the context of psychotherapy, helping indi-
viduals to cultivate the capacity to process the world from a
wise-mind or mindful experiencing–being perspective is regarded
as the optimal mode of mind.
Teasdale’s (1999) modes-of-mind typology also provides guid-
ance to better understand the cognitive–behavioral model of the
etiology and treatment of MDD. For instance, vulnerability for
MDD can be regarded as an individual’s tendency to slip into a
mindless emoting mode when confronted with a stressful event or
a potentially transient sad mood. By predominately processing the
world through the mindless emoting mode, one becomes more
likely to convert a transient sadness into an enduring depression, in
part by undermining healthy emotional processing. Similarly,
Teasdale cautioned that an implementation of CBT that is overly
reliant on cognitive disputation might, in turn, promote an over-
utilization of a conceptualizing–doing mode of processing to the
detriment of acquiring and utilizing mindful experiencing–being.
Thus, from this way of thinking, durable CBT for MDD is likely
to follow when an individual’s predominant mode of mind is
characterized by reductions in mindless emoting and increases in
Cognitive reactivity, defined as a change in one or more cogni-
tive indices in response to an emotion evocation challenge, in the
nomenclature of Teasdale (1999), represents a mindless emoting
mode of mind as well as a promising marker of psychological
vulnerability and increased risk to depression. In the methodology
most often used to assess cognitive reactivity, patients are asked to
listen to a piece of music designed to make them feel briefly sad,
while also recalling a personally upsetting time in their life. In an
initial study, Segal, Gemar, and Williams (1999) demonstrated that
patients with depression treated to remission with either open-label
ADM or open-label CBT evidenced different patterns of cognitive
reactivity following a negative mood induction, and that this
cognitive reactivity predicted depression relapse. Specifically,
ADM responders demonstrated increases in dysfunctional atti-
tudes, whereas CBT responders demonstrated no change in dys-
functional attitudes. In a replication study using random assign-
ment to treatment condition, Segal et al. (2006) similarly found
that ADM responders evidenced significantly more cognitive re-
activity as compared with CBT responders (d ? 0.42). Further,
treatment responders, irrespective of whether they received ADM
or CBT, who also evidenced cognitive reactivity (equivalent to an
8-or-more-point increase in dysfunctional attitudes following an
emotion evocation challenge), were at increased risk of relapse in
the subsequent 18 months as compared with participants with
minimal or no cognitive reactivity. Findings from these two studies
provide evidence that cognitive reactivity following an emotion
evocation challenge may confer vulnerability for depression re-
lapse or recurrence and that CBT may target this vulnerability
more effectively than ADM.
The current study represents a secondary analysis of Segal et al.
(2006) with the aim of identifying treatment-related factors that
might moderate the association of cognitive reactivity and relapse
in this group of patients. Decentering is described as the capacity
to take a present-focused, nonjudgmental stance in regard to
thoughts and feelings and to accept them. The construct of decen-
tering is of particular interest, as recent conceptualizations of
cognitive treatment suggest that decentering may reduce levels of
depressive rumination by teaching patients more adaptive ways of
relating to their thinking (Segal, Williams, & Teasdale, 2002).
However, decentering has also been a topic of discussion early on
in the cognitive–behavioral movement. Both Beck et al. (1979)
and Safran and Segal (1990) have described it as an important
ingredient of change in cognitive therapy. Ingram and Hollon
(1986) posited not only that decentering was important in helping
to reduce current depression symptoms but also that “the long term
effectiveness of cognitive therapy may lie in teaching patients to
initiate this process in the face of future stress” (p. 272). Decen-
tering also corresponds to the mindful experiencing–being mode
of mind within Teasdale’s (1999) typology.
Initial efforts to empirically evaluate the relationship of decen-
tering to depression and to the treatment of depression arose
through the study of metacognitive awareness, a construct related
to decentering. Metacognitive awareness refers to “the process of
experiencing negative thoughts and feelings within a decentered
perspective” (Teasdale et al., 2002, p. 276) and is typically as-
sessed with the Measure of Awareness and Coping in Autobio-
graphical Memory (MACAM; Moore, Hayhurst, & Teasdale,
1996), a semistructured clinical interview. In this study, 158 pa-
tients who had achieved partial remission of MDD following
treatment with ADM were randomized to receive ADM and clin-
ical management alone or together with cognitive therapy for 20
weeks and then receive follow-up continuation and maintenance
medication for 48 weeks. Findings indicated that lower levels of
baseline metacognitive awareness predicted earlier relapse across
both treatment groups. There was also a larger increase in meta-
FRESCO, SEGAL, BUIS, AND KENNEDY
cognitive awareness in the cognitive therapy group as compared
with the clinical management group.
Measurement of Decentering
Although the findings associated with metacognitive awareness
are promising, the MACAM is time intensive and requires clini-
cian assessment. Consequently, Fresco and colleagues (in press)
undertook a psychometric analysis of the Experiences Question-
naire (EQ), a rationally derived, self-report measure designed by
Teasdale to measure both decentering (and related processes) and
rumination in the context of mood disorders. The EQ was designed
as a relatively brief, practical measure of the primary psychother-
apy process in mindfulness based cognitive therapy (MBCT) but
had not undergone rigorous psychometric evaluation. Using ex-
ploratory and confirmatory factor analysis in two consecutive large
samples of college students, an 11-item decentering factor
emerged. This factor structure was again confirmed in a sample of
patients with remitted MDD. Findings from this latter sample
revealed that healthy control participants endorsed higher levels of
decentering as compared with MDD patients (d ? 1.3) and that
decentering was significantly and negatively correlated with levels
of clinician-assessed and self-reported symptoms of depression in
the patients with MDD.
The Present Study
The present study represents a secondary analysis of Segal et al.
(2006), who found that cognitive reactivity (a form of mindless
emoting mode of mind) in conjunction with an emotion evocation
challenge predicted relapse in the subsequent 18-month span. We
sought to evaluate three additional hypotheses. First, we predicted
that patients who achieved a positive treatment response following
random assignment to CBT would evidence significantly greater
gains in decentering (a mindful experiencing–being mode of
mind) as compared with patients with a positive treatment re-
sponse to ADM. Second, we predicted that CBT and ADM re-
sponders would show equivalent reductions in dysfunctional atti-
tudes (an index of cognitive content). Third, low cognitive
reactivity has already been shown to predict a more durable
treatment response (Segal et al., 2006). Thus, we predicted that
posttreatment levels of decentering would also show a positive
association with treatment durability, particularly in conjunction
with low levels of cognitive reactivity.
Participants for the current study were 111 patients with MDD
who accepted random assignment to either ADM (n ? 43) or CBT
(n ? 68) and completed the treatment as well as study measures.
Segal et al. (2006) screened 484 patients and, after applying
exclusion criteria, randomly assigned 301 patients to either ADM
(n ? 149) or CBT (n ? 152). Patients not invited for the acute
treatment phase of the study were disqualified for not meeting
diagnostic criteria for MDD; endorsing the presence of another
psychiatric condition judged to require more immediate treatment;
having a comorbid diagnosis of borderline, antisocial, or schizo-
typal personality disorder; having evidence of a prior poor re-
sponse to study interventions; and having scheduling difficulties.
One hundred forty-four patients who enrolled in the acute treat-
ment phase completed treatment. The attrition following initial
consent for randomization occurred prior to the onset of treatment
(n ? 69), within the first four sessions (n ? 51), or at some later
point in the course of acute treatment (n ? 37). Refusal of
randomization or disappointment with treatment assignment, par-
ticularly to ADM, was the most frequent explanation given for
Patients accepting initial randomization were approximately 37
years old (SD ? 11.5); were primarily female (62%); and had, on
average, experienced 1.37 (SD ? 0.48) lifetime episodes of major
depression, with the current episode having lasted approximately
33 weeks (SD ? 24). As reported by Segal et al. (2006), demo-
graphic characteristics were unrelated to patient attrition or treat-
ment response. CBT patients received 20 individual weekly ses-
sions over a 22- to 24-week span. ADM patients received 10 to 13
individual sessions with a study psychiatrist over a 26-week span.
Assessment of treatment response and posttreatment levels of
decentering were completed on the day of the last treatment
session. For treatment responders, the mood priming challenge
procedure occurred between 3 and 7 days following the posttreat-
(Fresco et al., in press) is an 11-item self-report instrument that
assesses the construct of decentering. Sample items include “I am
better able to accept myself as I am” and “I can observe unpleasant
feelings without being drawn into them.” In the current study, the
Decentering subscale evidenced good internal consistency at both
pretreatment (? ? .81) and posttreatment (? ? .84).
The Dysfunctional Attitude Scale (DAS; Weissman & Beck,
1978) is a 40-item self-report inventory used to assess dysfunc-
tional thinking. Participants rate the degree to which they agree or
disagree with statements that assess dysfunctional thoughts on a
7-point Likert-type scale with response options ranging from to-
tally agree to totally disagree. The DAS has previously shown
strong internal consistency (? ? .89; Weissman, 1979). In the
present study, the comparable 40-item Versions A and B were
used. Participants received a version of the DAS at the outset of the
study, following the completion of treatment as well as just before
and immediately following an emotion evocation challenge.
The Beck Depression Inventory—II (BDI–II; Beck, Steer, &
Brown, 1996) is a 21-item self-report measure that assesses the
affective, cognitive, behavioral, and somatic symptoms of depres-
sion as well as motivational components and suicidal wishes. Items
reflect a 2-week time period and are rated on a 4-point scale. In the
current study, the BDI–II was used to assess for the recurrence of
major depression post acute treatment.
view for DSM–IV, Research Version (SCID; First, Spitzer, Gib-
bon, & Williams, 1996) is a widely used semistructured interview
to determine current and lifetime diagnoses of Axis I disorders.
Ventura, Liberman, Green, Shaner, and Mintz (1998) reported
high interrater agreement for current diagnosis with an overall
weighted kappa of .82. Kappas for MDD are good (.80) to excel-
The Decentering subscale of the EQ
The Structured Clinical Inter-
SPECIAL SECTION: DECENTERING AND CBT
The Hamilton Rating Scale for Depression (HRSD; Hamilton,
1960) is a clinician-administered measure of depression symptom
severity that correlates highly with the Beck Depression Inventory
(BDI; Beck et al., 1979). The present study used the 17-item
version of the HRSD. Acute treatment response was assessed with
the HRSD using a criterion of scoring less than 8 (cf. Frank et al.,
1991; Jacobson et al., 1996; Zimmerman, Chelminski, & Poster-
nak, 2004). Using this criterion, 79% of AMD patients (34/43) and
68% of CBT patients (46/68) were classified as responders.
The Longitudinal Interview Follow-up Evaluation (LIFE; Keller
et al., 1987) is a semistructured interview used to measure partic-
ipants’ recovery from or occurrence of new depressive episodes
and the occurrence of relapse in depression. In the current study,
the LIFE interview in conjunction with phone and mail contact
with participants was used to determine relapse of MDD. Patients
who endorsed elevations in depression were brought in and eval-
uated by a clinician with the HRSD. Any patient scoring 15 or
greater on the BDI–II or 16 or greater on the HRSD-17 was
reinterviewed in a week’s time. If their scores remained in the
same range, they were assessed with the LIFE to determine
whether their level of symptomatology met criteria for MDD.
Patients were judged to have relapsed if they were given a diag-
nosis of MDD at any time during the follow-up.
treated with one of three first-line, antidepressant medications
(sertraline, 50–200 mg; paroxetine, 20–50 mg; or venlafaxine,
75–225 mg) for a period of 6 months. Choice of medication was
naturalistic and based on the treating psychiatrist’s clinical judg-
ment. Patients who failed to achieve an adequate response to the
initial ADM were allowed to discontinue it and start on a second
medication. Patients who failed both trials were removed from the
study and offered alternative care. Pharmacotherapy sessions were
20 min in duration and followed the recommendations for clinical
management developed by Fawcett, Epstein, Fiester, Elkin, and
Patients in the CBT condition received a
course of 20 individual weekly sessions over a 22- to 24-week
Patients in the ADM condition were
span according to the manualized protocol developed by Beck et
al. (1979). Treatment fidelity was assessed using the Cognitive
Therapy Scale (Vallis, Shaw, & Dobson, 1986) by applying ratings
of 18 audiotapes (early, middle, and late sessions) from 6 ran-
domly chosen patients. The mean rating for each of the two raters
across all tapes was 46.94 and 47.35, and the level of interrater
agreement (intraclass correlation coefficient) was .83.
Emotion Evocation Challenge
Patients listened through headphones to a piece of music pre-
sented on a CD player while following instructions to recall a time
in their lives when they felt sad. The piece of music was “Russia
Under the Mongolian Yoke,” composed by Sergei Prokofiev
(1934/1987). This piece was remastered at half speed and runs for
approximately 8 min. This piece, played at half speed, has been
shown to be very effective in inducing a negative or depressed
mood (e.g., Fresco, Heimberg, Abramowitz, & Bertram, 2006;
Segal et al., 1999, 2006). A meta-analysis of musical mood induc-
tion procedures found a good effect size (Cohen’s f ? .56) for sad
mood inductions including, but not limited to, this Prokofiev piece
played at half speed (Westermann, Spies, Stahl, & Hesse, 1996).
The first set of analyses examined patients randomly assigned to
ADM or CBT on baseline measures of depression symptoms,
decentering, and dysfunctional attitudes. Findings indicated that
the two treatment groups did not differ significantly on baseline
measures of self-reported depression, F(1, 110) ? 0.29, ns;
clinician-assessed depression, F(1, 110) ? 0.39, ns; decentering,
F(1, 110) ? 0.03, ns; or dysfunctional attitudes, F(1, 110) ? 1.00,
ns. Sample means and standard deviations are presented in
Cognitive Change as a Function of Treatment Response
The next set of analyses evaluated the degree of change on
cognitive measures as a function of treatment condition and re-
Means and Standard Deviations for Depression Symptom and Cognitive Measures as Assessed Among Treatment Responders at
Pretreatment and Post Acute Treatment
(n ? 68):
(n ? 43):
(n ? 46)
(n ? 34)
MSD M SDM SDMSD MSD M SD
Depression Inventory—II; HRSD ? Hamilton Rating Scale for Depression; Decentering ? Experiences Questionnaire—Decentering subscale.
CBT ? cognitive–behavioral therapy; ADM ? antidepressant medication; Pre ? pretreatment; Post ? post acute treatment; BDI–II ? Beck
FRESCO, SEGAL, BUIS, AND KENNEDY
sponse status. A series of 2 (response status: responder, nonre-
sponder) ? 2 (time: pretreatment, post acute treatment) repeated
measures analyses of variance with decentering and dysfunctional
attitudes as the dependent measures were conducted separately for
ADM and CBT patients. The findings were evaluated using custom-
ary standards of significance (e.g., p ? .05) as well as Cohen’s (1988)
effect size conventions for f (small ? .10, medium ? .25, and large ?
.40) and d (small ? 0.20, medium ? 0.50, and large ? 0.80).
Our first hypothesis posited that CBT responders
with ADM responders. Findings for ADM patients revealed a signif-
icant main effect for time, F(1, 42) ? 11.74, p ? .002, f ? .56; a
nonsignificant main effect for response status, F(1, 42) ? 1.22, ns,
f ? .18; and a nonsignificant Response Status ? Time interaction,
F(1, 42) ? 1.58, ns, f ? .21. CBT patients evidenced a significant
main effect for time, F(1, 67) ? 51.47, p ? .001, f ? .85, and a
nonsignificant main effect for response status, F(1, 67) ? 3.05, p ?
.08, f ? .21. However, these main effects were qualified by a signif-
icant Response Status ? Time interaction, F(1, 67) ? 8.05, p ? .006,
f ? .34. Examination of the group means presented in Table 1,
coupled with this interaction effect, revealed that CBT treatment
responders evidenced significantly larger gains in decentering as
compared with CBT nonresponders.
Our second hypothesis posited that
CBT responders and ADM responders would evidence equivalent
drops in dysfunctional attitudes. Findings for ADM patients re-
vealed a significant main effect for time, F(1, 42) ? 14.97, p ?
.001, f ? .60; a nonsignificant main effect for response status, F(1,
42) ? 2.19, ns, f ? .22; and a nonsignificant Response Status ?
Time interaction, F(1, 42) ? 0.73, ns, f ? .14. Similarly, CBT
patients evidenced a significant main effect for time, F(1, 67) ?
26.68, p ? .001, f ? .64; a nonsignificant main effect for response
status, F(1, 67) ? 0.56, ns, f ? .10; and a nonsignificant Response
Status ? Time interaction, F(1, 67) ? 1.86, ns, f ? .17. Exami-
nation of the group means presented in Table 1, coupled with this
nonsignificant interaction effect, indicated that ADM and CBT
patients experienced drops in dysfunctional attitudes over the
course of acute treatment that occurred irrespective of treatment
Cognitive Change Among Treatment Responders
responders to assess whether posttreatment levels of decentering or
dysfunctional attitudes differed as a function of treatment condition
while controlling for the corresponding baseline measure of either
decentering or dysfunctional attitudes. For decentering, findings re-
vealed a significant main effect for treatment condition, F(1, 77) ?
6.45, p ? .013, d ? 0.58, indicating that CBT responders evidenced
significantly larger gains in decentering as compared with ADM
responders when we controlled for baseline levels of decentering. By
contrast, CBT responders and ADM responders did not differ signif-
icantly from one another on posttreatment levels of dysfunctional
attitudes, F(1, 77) ? 1.49, ns, d ? 0.28.
Association of Cognitive Reactivity and Decentering With
Segal et al. (2006) reported that 99 patients who completed and
responded to either ADM or CBT entered the emotion evocation
and 18-month follow-up phase of the study. The findings indicated
that the emotion evocation challenge successfully induced short-
lived sadness and that some patients who had recovered from
depression evidenced cognitive reactivity that, in turn, predicted
likelihood of relapse (see Segal et al., 2006, for details).
Another aim of the current study was to examine the relation-
ship of posttreatment decentering to cognitive reactivity and re-
lapse. Specifically, our third hypothesis predicted that a combina-
tion of high posttreatment decentering and low cognitive reactivity
would be associated with the most durable treatment response
during the follow-up period. This prediction lends itself to a
moderation analysis, where the interaction of a known predictor of
relapse (cognitive reactivity) is examined in relation to a putative
moderator of that main effect relationship. Kraemer, Stice, Kazdin,
Offord, and Kupfer (2001) recently provided a cogent statement
that offers clarity on methodological and statistical issues in eval-
uating both moderators and mediators. A thorough review of this
report is beyond the scope of the current study, but given the
relative statistical independence between decentering and cogni-
tive reactivity (r ? –.11) and the design of data acquisition, the
following analysis qualifies by Kraemer et al.’s standards as an
appropriate test of moderation.1To accomplish this aim, a Cox
proportional-hazards survival analysis was computed to estimate
the impact of posttreatment decentering and cognitive reactivity on
time until relapse among acute treatment responders. Treatment
condition (ADM or CBT), posttreatment depression symptoms (to
control for posttreatment depression levels), posttreatment decen-
tering, and cognitive reactivity were entered at the first step as
main effects. The nested two-way interactions composed of treat-
ment condition, posttreatment decentering, and cognitive reactivity
were entered at Step 2, followed by the three-way interaction of
these variables. The overall model resulted in a significant drop in
the survival function value, ?2(8, N ? 77) ? 21.28, p ? .006. In
the full model, cognitive reactivity (B ? –0.66, SE ? 0.22),
Wald(1) ? 8.55, p ? .003, and the three-way interaction of
treatment condition, posttreatment decentering, and cognitive re-
activity were significant (B ? 0.24, SE ? 0.10), Wald(1) ? 4.86,
p ? .028. This interaction was interpreted by solving the Cox
regression model for treatment condition (ADM ? 0; CBT ? 1) at
either 1 standard deviation above or 1 standard deviation below the
sample mean on posttreatment decentering as well as the definition
of a marked increase (DAS change ? 8) versus minimal or no
cognitive reactivity (DAS change ? 8). Solving the regression
equation in this manner estimated the proportional hazards for
each treatment condition as a function of low decentering–high
cognitive reactivity, low decentering–minimal cognitive reactivity,
high decentering–high cognitive reactivity, and high decentering–
minimal cognitive reactivity. As seen in Figures 1 and 2, findings
revealed that a combination of high decentering and low cognitive
reactivity was associated with the lowest predicted rates of relapse
1As pointed out by an anonymous reviewer, there may be theoretically
interesting mediation relationships between decentering and cognitive re-
activity. However, by Kraemer et al.’s (2001) standards, the methodology
of this secondary analysis is disqualified by virtue of the fact that the
putative mediator (decentering) was not assessed prior to the independent
variable (cognitive reactivity) and the two variables are not significantly
SPECIAL SECTION: DECENTERING AND CBT
(e.g., highest rates of survival) particularly among CBT re-
sponders. By contrast, and unexpectedly, among CBT respond-
ers with high decentering and high cognitive reactivity, the
predicted survival rate was markedly diminished. Similarly,
among CBT responders with low decentering, the relationship
between cognitive reactivity and relapse evidenced an opposite
relationship such that the combination of low decentering and
high cognitive reactivity was associated with a more durable
treatment response as compared with low decentering and low
as a function of high and low decentering and high and low cognitive reactivity.
Survival analysis showing the predict rates of survival among antidepressant medication responders
sponders as a function of high and low decentering and high and low cognitive reactivity.
Survival analysis showing the predict rates of survival among cognitive–behavioral therapy re-
FRESCO, SEGAL, BUIS, AND KENNEDY
In an attempt to better understand the relationship of decentering
to relapse, one additional survival analysis was conducted, this
time examining the relationship of high decentering and low
decentering at the sample average of cognitive reactivity. Among
CBT responders, the predicted cumulative survival associated with
high decentering was .73, whereas the predicted cumulative sur-
vival for CBT responders with low decentering was .66. For ADM
responders, the predicted cumulative survival associated with high
decentering was .55, whereas the predicted cumulative survival
associated with low decentering was .46.
The main focus of the current study was to investigate the relation-
ship of decentering, which is defined as the ability to observe one’s
thoughts and feelings as temporary, objective events in the mind, to
both recovery from major depression and protection against relapse.
We can now report two main findings with respect to decentering:
First, gains in decentering are more pronounced among CBT treat-
ment responders as compared with ADM responders. Second, high
posttreatment decentering appears to offer some protection against
relapse of MDD, particularly for patients who recovered in CBT.
We examined the relationship of treatment modality (ADM or
CBT) and response status to two self-report cognitive measures:
the DAS, a measure of dysfunctional attitudes, and the EQ, a
measure of decentering. Similar to other studies (Hollon et al.,
1992; Simons et al., 1984), findings indicated that patients, irre-
spective of treatment condition and response status, evidenced
drops in dysfunctional attitudes. However, treatment responders in
the CBT condition endorsed significantly larger gains in decenter-
ing compared with CBT nonresponders, with this finding corre-
sponding to a medium effect. Similarly, CBT responders evi-
denced significantly larger gains in decentering compared with
ADM responders, a finding that also corresponded to a medium
effect. Thus, these findings indicate that gains in decentering
demonstrate some specificity to treatment modality.
This first set of findings from the current study also corresponds
favorably to findings reported by Teasdale et al. (2002), despite
some important methodological differences. In the Teasdale et al.
(2002) study, patients with depression who had achieved a partial
remission of their depression with ADM were randomly assigned
to 20 weeks of continued medication management (CM) or 20
weeks of CM plus 16 sessions of CBT. Metacognitive awareness,
the process of experiencing negative thoughts and feelings within
a decentered perspective (as assessed with the MACAM), in-
creased significantly for patients who received the CBT augmen-
tation of CM. The gains in metacognitive awareness enjoyed by
the CM-enriched-with-CBT group held even after controlling for
relapse status in the follow-up period and in concurrent levels of
depression. Thus, in both studies, gains in metacognitive aware-
ness or decentering appear to be a by-product of receiving CBT.
The main methodological difference between the current study and
the Teasdale et al. (2002) study was that the current study enrolled
participants who met criteria for MDD and randomly assigned them
to either ADM or CBT to treat their depression. By contrast, the
Teasdale et al. (2002) sample consisted of patients with partially
remitted depression who were randomly assigned to ADM with CM
with and without CBT. Still, in both studies, gains in either metacog-
nitive awareness or decentering occurred after receiving CBT.
The second focus of the current study was to examine the
relationship of decentering to the durability of treatment gains
following acute treatment response with either ADM or CBT.
Segal et al. (2006), from which the current study emerged, reported
that participants in CBT, compared with ADM, evidenced less
cognitive reactivity in response to a laboratory emotion evocation
challenge, and the cognitive reactivity, irrespective of treatment
modality, was associated with earlier relapse in the subsequent 18
months. Findings from the current study complement the earlier
results of Segal et al. (2006) by showing that posttreatment levels
of decentering in combination with low cognitive reactivity were
associated with the most durable treatment response, particularly
among CBT responders. The interaction of posttreatment decen-
tering, treatment modality, and cognitive reactivity significantly
improved prediction of relapse after controlling for posttreatment
levels of depressive symptoms. Overall, the results indicate that
high decentering as assessed with the EQ provides protection
against relapse in conjunction with low cognitive reactivity. Future
researchers may wish to further examine the additive and interac-
tive relationships between decentering and cognitive reactivity to
further elucidate their contribution to risk for relapse. The other
unexpected finding was associated with CBT responders with low
posttreatment decentering who seemed to have a more durable
treatment response in association with high cognitive reactivity.
One possible explanation for this finding is that low decentering
may assess a general inattention to one’s thoughts and feelings and
that it thus represents a proxy for distraction or perhaps an atten-
tional control strategy that draws the mind away from self-focused
attention (cf. Wells, 1990; Wells & Matthews, 1994). Similarly,
the finding of a fragile recovery associated with low decentering
and low cognitive reactivity among CBT patients seems compat-
ible with evolutionary accounts of depression such as the emotion
context insensitivity hypothesis (Rottenberg, Gross, & Gotlib,
2005), which posits that depression represents a general psycho-
logical inertia characterized by a lack of emotional reactivity for
both negative and positive emotional situations. In the short run,
this reduced reactivity spares the organism from emotional up-
heaval and actions that might be counterproductive, but in the long
run, this inactivity comes with costs, including greater depression
severity, longer episodes of depression, and lower levels of psy-
chosocial functioning (Rottenberg, Kasch, Gross, & Gotlib, 2002).
Perhaps the poor durability associated with low decentering and
low cognitive reactivity is a proxy for the lack of emotional
reactivity proposed by Rottenberg and colleagues to promote
longer and more severe depressive episodes. This question remains
an important area of future research.
Teasdale et al. (2002) found that high levels of metacognitive
awareness, assessed prior to a patient’s enrollment in CM or CM
enriched with CBT, improved the prediction of relapse even after
controlling for concurrent levels of depressive symptoms. The effects
of ADM on metacognitive awareness were not reported. By contrast,
in the current study, CBT but not ADM resulted in gains in decen-
tering, and the relationship of decentering to relapse was most pro-
nounced in CBT patients, particularly in conjunction with low cog-
Using the rubric for investigating the enduring effects of treatments
proposed by Hollon et al. (2006), these findings offer initial evidence
simply palliative. Both ADM and CBT were equivalent in their
SPECIAL SECTION: DECENTERING AND CBT
success in treating the symptoms of the current major depressive
episode, and similar numbers of patients in each condition enjoyed a
response such that they no longer met criteria for MDD following the
end of acute treatment. However, decentering contributed to the
durability of CBT, particularly for patients who also evidenced low
cognitive reactivity. Although these findings are necessarily tentative,
they offer evidence that decentering corresponds to the Hollon et al.
(2006) designation of a curative agent within CBT.
Findings from the current study, although encouraging, must be
interpreted in the context of some important limitations. First, the
current study represents a secondary analysis of another study. The
hypotheses proposed in the current study were not central to the
earlier investigation, and thus, methodological decisions made in
the original study, such as the schedule for the administration of
measures such as decentering, prevent us from drawing any con-
clusions of a causal nature regarding the role of decentering in the
process of recovery from MDD. Rather, the findings show that
gains in decentering covary with recovery from MDD. Future
research is needed that assesses decentering during the course of
treatment and prior to the determination of acute treatment re-
sponse before statements about its role as a mechanism of change
can be evaluated (cf. DeRubeis & Feeley, 1990; Hollon, DeRubeis,
& Evans, 1987; Jacobson et al., 1996). Findings of the current
study represent just the second statement to attest to the psycho-
metric properties and predictive validity of the Decentering scale
of the EQ (Fresco et al., in press). Although the present results are
encouraging, the accumulation of more research evaluating the
performance and assessment of decentering of the EQ is needed.
Three additional questions and areas of future research arise
from the findings of the current study. First, the Decentering scale
of the EQ represents a new candidate measure assessing a meta-
cognitive factor and mindful experiencing–being mode of mind
associated with recovery and relapse prevention. However, a fruit-
ful area of future research will be to evaluate how the EQ relates
to other measures of metacognitive abilities. One measure, already
discussed, is the MACAM (Moore et al., 1996; Teasdale et al.,
2002). Another promising measure of metacognitive ability is the
frequency with which individuals endorse extreme response cate-
gories, positive or negative, on self-report measures of cognition,
such as the Attributional Style Questionnaire (Peterson et al.,
1982) and the DAS (Weissman & Beck, 1978). Teasdale and
colleagues (2001) found that higher rates of extreme responding
predicted relapse and differential response to cognitive therapy for
depression. Similarly, Petersen et al. (2007) found that reductions
in extreme responding predicted full depression remission over 8
weeks of treatment with ADM. Furthermore, during the mainte-
nance phase, patients randomly assigned to ADM-only mainte-
nance evidenced significantly greater increases in extreme re-
sponding compared with patients in ADM maintenance
augmented with CBT. Second, Teasdale’s (1999) modes-of-
mind model posits that patterns of cognitive rigidity such as
cognitive reactivity and depressive rumination confer risk for
occurrence–recurrence of MDD. An important follow-up to the
current study will be to simultaneously evaluate the relationship
of cognitive reactivity, depressive rumination, and decentering
in the context of recovery from MDD and in the emotion
evocation challenge methodology. A particular question that
remains is whether gains in decentering simply reflect drops in
depressive rumination or whether gains in decentering represent
the acquisition of a metacognitive ability to counteract its
effects. Third, findings from the current study show that a
traditional implementation of CBT for depression (Beck et al.,
1979) produced gains in decentering. The version of CBT used
in the current study did not include any of the modifications and
additions found in MBCT (Segal, Williams, & Teasdale, 2002),
even though one of the current investigators contributed to the
MBCT protocol. A fruitful area of future research will be to see
whether MBCT protects against MDD relapse via gains in a
mindful experiencing–being mode of mind while decreasing a
mindless emoting mode of mind.
Barnard, P. J., & Teasdale, J. D. (1991). Interacting cognitive subsystems:
A systemic approach to cognitive–affective interaction and change.
Cognition and Emotion, 5, 1–39.
Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive
therapy of depression. New York: Guilford Press.
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the BDI–II.
San Antonio, TX: Psychological Corporation.
Brewin, C. R., & Power, M. (1999). Integrating psychological therapies:
Processes of meaning transformation. British Journal of Medical Psy-
chology, 72, 143–157.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences
(2nd ed.). San Diego, CA: Academic Press.
DeRubeis, R. J., & Feeley, M. (1990). Determinants of change in cognitive
therapy for depression. Cognitive Therapy and Research, 14, 469–482.
Fava, G. A., Rafanelli, C., Grandi, S., Conti, S., & Belluardo, P. (1998).
Prevention of recurrent depression with cognitive behavioral therapy:
Preliminary findings. Archives of General Psychiatry, 55, 816–820.
Fawcett, J., Epstein, P., Fiester, S. J., Elkin, I., & Autry, J. H. (1987).
Clinical management: Imipramine/placebo administration manual:
NIMH Treatment of Depression Collaborative Research Program. Psy-
chopharmacology Bulletin, 23, 309–324.
First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1996).
Structured clinical interview for DSM–IV Axis I disorders—Patient
edition (SCID–I/P, Version 2.0). New York: New York State Psychiatric
Institute, Biometrics Research Department.
Frank, E., Prien, R. F., Jarrett, R. B., Keller, M. B., Kupfer, D. J., Lavori,
P. W., Rush, A. J., & Weissman, M. M. (1991). Conceptualization and
rationale for consensus definitions of terms in major depressive disorder.
Archives of General Psychiatry, 48, 851–855.
Fresco, D. M., Heimberg, R. G., Abramowitz, A., & Bertram, T. L. (2006).
The effect of a negative mood priming challenge on dysfunctional
attitudes, explanatory style, and explanatory flexibility. British Journal
of Clinical Psychology, 45, 167–183.
Fresco, D. M., Moore, M. T., van Dulmen, M., Segal, Z. V., Teasdale,
J. D., Ma, H., & Williams, J. M. G. (in press). Initial psychometric
properties of the Experiences Questionnaire: A self-report survey of
decentering. Behavior Therapy.
Hamilton, M. (1960). A rating scale for depression. Journal of Neurology,
Neurosurgery, and Psychiatry, 23, 56–62.
Hollon, S. D., DeRubeis, R. J., & Evans, M. D. (1987). Causal mediation
of change in treatment for depression: Discriminating between nonspeci-
ficity and noncausality. Psychological Bulletin, 102, 139–149.
Hollon, S. D., DeRubeis, R. J., Evans, M. D., Wiemer, M. J., Garvey, M. J.,
FRESCO, SEGAL, BUIS, AND KENNEDY
Grove, W. M., & Tuason, V. B. (1992). Cognitive therapy and pharma-
cotherapy for depression: Singly and in combination. Archives of Gen-
eral Psychiatry, 49, 774–781.
Hollon, S. D., & Kriss, M. R. (1984). Cognitive factors in clinical research
and practice. Clinical Psychology Review, 4, 35–76.
Hollon, S. D., Stewart, M. O., & Strunk, D. (2006). Enduring effects for
cognitive behavior therapy in the treatment of depression and anxiety.
Annual Review of Psychology, 57, 285–315.
Imber, S. D., Pilkonis, P. A., Sotsky, S. M., Elkin, I., Watkins, J. T., Collins,
J. F., et al. (1990). Mode-specific effects among three treatments for de-
pression. Journal of Consulting and Clinical Psychology, 58, 352–359.
Ingram, R. E., & Hollon, S. D. (1986). Cognitive therapy for depression
from an information processing perspective. In R. E. Ingram (Ed.),
Information processing approaches to clinical psychology (pp. 259–
281). San Diego, CA: Academic Press.
Jacobson, N. S., Dobson, K. S., Truax, P. A., Addis, M. E., Koerner, K.,
Gollan, J. K., et al. (1996). A component analysis of cognitive–
behavioral treatment for depression. Journal of Consulting and Clinical
Psychology, 64, 295–304.
Keller, M. B., Lavori, P. W., Friedman, B., Nielsen, E., Endicott, J.,
McDonald-Scott, P., & Andreasen, N. C. (1987). The longitudinal in-
terval follow-up evaluation: A comprehensive method for assessing
outcome in prospective longitudinal studies. Archives of General Psy-
chiatry, 44, 540–548.
Keller, M. B., Lavori, P. W., Rice, J., Coryell, W., & Hirschfeld, R. M.
(1986). The persistent risk of chronicity in recurrent episodes of non-
bipolar major depressive disorder: A prospective follow-up. American
Journal of Psychiatry, 143, 24–28.
Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., &
Walters, E. E. (2005). Lifetime prevalence and age-of-onset distributions
of DSM–IV disorders in the National Comorbidity Survey Replication.
Archives of General Psychology, 62, 593–602.
Kessler, R. C., Chiu, W. T., Demler, O., Merikangas, K. R., & Walters,
E. E. (2005). Prevalence, severity, and comorbidity of 12-month
DSM–IV disorders in the National Comorbidity Survey Replication.
Archives of General Psychology, 62, 617–627.
Kraemer, H. C., Stice, E., Kazdin, A., Offord, D., & Kupfer, D. (2001).
How do risk factors work together? Mediators, moderators, and inde-
pendent, overlapping, and proxy risk factors. American Journal of Psy-
chiatry, 158, 848–856.
LeDoux, J. (1996). The emotional brain: The mysterious underpinnings of
emotional life. New York: Simon & Schuster.
Linehan, M. M. (1993). Cognitive–behavioral treatment of borderline
personality disorder. New York: Guilford.
Moore, R. G., Hayhurst, H., & Teasdale, J. D. (1996). Measure of Aware-
ness and Coping in Autobiographical Memory: Instructions for admin-
istering and coding. Unpublished manuscript, Department of Psychiatry,
University of Cambridge, Cambridge, England.
Petersen, T., Feldman, G., Harley, R., Fresco, D. M., Graves, L., Holmes,
L., et al. (2007). Extreme response style in recurrent and/or chronically
depressed patients—Change with acute antidepressant treatment and
stability during continuation treatment. Journal of Consulting and Clin-
ical Psychology, 75, 145–153.
Peterson, C., Semmel, A., von Baeyer, C., Abramson, L. Y., Metalsky,
G. I., & Seligman, M. E. P. (1982). The Attributional Style Question-
naire. Cognitive Therapy and Research, 6, 287–300.
Prokofiev, S. (1987). Russia under the Mongolian yoke [Alexander
Nevsky]. Cleveland, OH: Telarc. (Original work published 1934)
Rottenberg, J., Gross, J. J., & Gotlib, I. H. (2005). Emotion context
insensitivity in major depressive disorder. Journal of Abnormal Psychol-
ogy, 114, 627–639.
Rottenberg, J., Kasch, K. L., Gross, J. J., & Gotlib, I. H. (2002). Sadness
and amusement reactivity differentially predict concurrent and prospec-
tive functioning in major depressive disorder. Emotion, 2, 135–146.
Safran, J. D., & Segal, Z. V. (1990). Interpersonal process in cognitive
therapy. New York: Basic Books.
Safran, J. D., Vallis, T. M., Segal, Z. V., & Shaw, B. F. (1986). Assessment
of core cognitive processes in cognitive therapy. Cognitive Therapy and
Research, 10, 509–526.
Samoilov, A., & Goldfried, M. R. (2000). Role of emotion in cognitive–
behavior therapy. Clinical Psychology: Science and Practice, 7, 373–385.
Segal, Z. V., Gemar, M., & Williams, S. (1999). Differential cognitive
response to a mood challenge following successful cognitive therapy or
pharmacotherapy for unipolar depression. Journal of Abnormal Psychol-
ogy, 108, 3–10.
Segal, Z. V., Kennedy, S., Gemar, M., Hood, K., Pedersen, R., & Buis, T.
(2006). Cognitive reactivity to sad mood provocation and the prediction
of depressive relapse. Archives of General Psychiatry, 63, 749–755.
Segal, Z. V., Williams, J. M. G., & Teasdale, J. D. (2002). Mindfulness-
based cognitive therapy for depression: A new approach to preventing
relapse. New York: Guilford Press.
Simons, A. D., Garfield, S. L., & Murphy, G. E. (1984). The process of
change in cognitive therapy and pharmacotherapy for depression:
Changes in mood and cognition. Archives of General Psychiatry, 41,
Solomon, D. A., Keller, M. B., Leon, A. C., Mueller, T. I., Lavori, P. W.,
Shea, M. T., et al. (2000). Multiple recurrences of major depressive
disorder. American Journal of Psychiatry, 157, 229–233.
Teasdale, J. D. (1999). Emotional processing, three modes of mind, and the
prevention of relapse in depression. Behaviour Research and Therapy,
Teasdale, J. D., Moore, R. G., Hayhurst, H., Pope, M., Williams, S., &
Segal, Z. V. (2002). Metacognitive awareness and prevention of relapse
in depression: Empirical evidence. Journal of Consulting and Clinical
Psychology, 70, 275–287.
Teasdale, J. D., Scott, J., Moore, R. G., Hayhurst, H., Pope, M., & Paykel,
E. S. (2001). How does cognitive therapy prevent relapse in residual
depression? Evidence from a controlled trial. Journal of Consulting and
Clinical Psychology, 69, 347–357.
Vallis, T. M., Shaw, B. F., & Dobson, K. S. (1986). The Cognitive Therapy
Scale: Psychometric properties. Journal of Consulting and Clinical
Psychology, 54, 381–385.
Ventura, J., Liberman, R. P., Green, M. F., Shaner, A., & Mintz, J. (1998).
Training and quality assurance with the Structured Clinical Interview for
DSM–IV (SCID–I/P). Psychiatry Research, 79, 163–173.
Weissman, A. N. (1979). The Dysfunctional Attitude Scale: A validation
study. Dissertation Abstracts International, 40, 1389–1390.
Weissman, A. N., & Beck, A. T. (1978, March). Development and vali-
dation of the Dysfunctional Attitude Scale: A preliminary investigation.
Paper presented at the meeting of the American Educational Research
Association, Toronto, Ontario, Canada.
Wells, A. (1990). Panic disorder in association with relaxation-induced
anxiety: An attentional training approach to treatment. Behavior Ther-
apy, 21, 273–280.
Wells, A., & Matthews, G. (1994). Attention and emotion: A clinical
perspective. Hove, England: Erlbaum.
Westermann, R., Spies, K., Stahl, G., & Hesse, F. W. (1996). Relative
effectiveness and validity of mood induction procedures: A meta-
analysis. European Journal of Social Psychology, 26, 557–580.
Zimmerman, M., Chelminski, I., & Posternak, M. (2004). A review of
studies of the Hamilton Depression Rating Scale in healthy controls:
Implications for the definition of remission in treatment studies of
depression. Journal of Nervous and Mental Disease, 192, 595–601.
Received August 18, 2006
Revision received February 2, 2007
Accepted February 12, 2007 ?
SPECIAL SECTION: DECENTERING AND CBT