Anda di halaman 1dari 11

ORI GI NAL ARTI CLE

Implicit Memory and Posttraumatic Stress Symptoms


Nader Amir Amy S. Leiner Jessica Bomyea
Published online: 28 August 2008
Springer Science+Business Media, LLC 2008
Abstract Recent theories of Posttraumatic Stress
Disorder (PTSD) suggest that memory dysregulation
plays a crucial role in symptom maintenance. How-
ever, it is not clear which specic memory systems are
involved in PTSD. In this study we used a visual
implicit memory paradigm to examine memory bias in
individuals with PTSD symptoms. Three hundred
nineteen participants provided self-report measures
of PTSD, anxiety and depression symptoms. Next they
completed a visual clarity-rating task (Amir et al.
Cognition and Emotion 17(4):567583, 2003) to assess
implicit memory for three picture types (trauma-
relevant, neutral, and negative). Results revealed that
participants with PTSD symptoms had greater implicit
memory for negative and trauma-relevant pictures
compared to neutral pictures. Traumatized individuals
without PTSD symptoms showed greater implicit
memory for negative pictures relative to trauma and
neutral pictures. This pattern of results suggests that
implicit memory bias for trauma-relevant and negative
visual information may be an important factor in the
maintenance of PTSD symptoms.
Keywords Trauma PTSD Implicit memory
Introduction
Posttraumatic Stress Disorder (PTSD) is a response to a
traumatic stressor involving witnessing or experienc-
ing actual or threatened death, injury or threat to
personal integrity, as well as feelings of intense fear,
helplessness or horror associated with the event.
Symptoms of this disorder are organized into three
clusters: (a) recurrent re-experiencing of the event,
including intrusive memories, ashbacks, and night-
mares, (b) avoidance of thoughts, feelings or situations
which are reminders of the trauma, and (c) hyper-
arousal symptoms such as sleep disturbances and
hypervigilance (American Psychiatric Association
2000). Although many trauma survivors experience
these symptoms in the acute aftermath of the event,
PTSD is associated with a persistent failure to recover
from initial reactions to the trauma that may continue
up to several years thereafter (Kessler et al. 1995).
Researchers have focused on the mechanisms under-
lying these symptoms, including the role of
information processing biases during and after trau-
matic events. According to current models of PTSD,
memory dysregulation may be a key contributor to the
development and maintenance of PTSD symptoms.
More specically, at least three models of the disorder,
including Emotional Processing Theory (Foa and
N. Amir (&) J. Bomyea
San Diego State University, San Diego, CA, USA
e-mail: namir@sciences.sdsu.edu
N. Amir J. Bomyea
University of California San Diego, San Diego, CA, USA
A. S. Leiner
Atlanta VAMU, Atlanta, GA, USA
1 3
Cogn Ther Res (2010) 34:4958
DOI 10.1007/s10608-008-9211-0
Kozak 1986), Dual Representation Theory (Brewin
et al. 1996), and cognitive theory Elhers and Clarks
(2000) of PTSD, integrate components of memory
functioning into explanations of PTSD symptoms (for
a review see Dalgleish 2004). A few common features
are readily identiable in these models descriptions of
memory functioning in PTSD.
First, these models emphasize the importance of
memory processes that function outside of conscious
awareness in storing and retrieving trauma-relevant
information in PTSD. While conscious avoidance of
memories of traumatic events is considered a core
symptom for individuals with PTSD, activation of
trauma representations via automatic (i.e., capacity-
free, unconscious, involuntary; McNally 1995) mem-
ory systems may be heightened for these individuals.
This activation may lead to maladaptive, involuntarily
access to trauma-relevant memories (e.g., situationally
accessible memories in the Dual Representation
model, Brewin et al. 1996; perceptual priming in
Elhers and Clark 2000; fear network activation, Foa
and Kozak 1986). To examine automatic memory
representations researchers have used implicit memory
tasks, also referred to as priming paradigms (Schacter
1987; Schacter and Badgaiyan 2001). Implicit memory
is demonstrated when prior experiences inuence
subsequent performance without conscious awareness
of these prior experiences (Schacter 1987). While some
studies nd evidence for enhanced implicit memory
biases for trauma-specic material in PTSD(e.g., Amir
et al. 1996), others do not (Paunovic et al. 2002;
McNally and Amir 1996; Golier et al. 2003). However,
methodological differences between studies make
comparisons difcult. For example, these studies have
used different paradigms to assess implicit memory,
and targeted different populations (e.g., Holocaust
survivors, combat veterans).
The second common feature of cognitive models of
PTSD is the role of visual information in symptom
maintenance. According to Emotional Processing
Theory, stimuli in multiple modalities (such as visual
representations or bodily sensations) that match the
fear structure may serve as a trauma reminder and
subsequently activate the fear network (Foa and Kozak
1986). Similarly, a variety of stimuli modality repre-
sentations may comprise Situationally Accessible
Memories according to Dual Representation Theory
(Brewin et al. 1996). Moreover, the visual sensory
modality may be a particularly salient component of
trauma-relevant encoding and retrieval in PTSD. For
example, re-experiencing symptoms such as intrusive
memories comprise brief sensory fragments from the
time of the traumatic event and are most frequently
visual in nature (Elhers and Steil 1995; Elhers et al.
2002; Hackmann et al. 2004). Additionally, triggering
stimuli for these symptoms are generally perceptually
similar to the content of intrusions and, thus, are most
commonly visual. For example, the sight of sunlight
reecting on grass might trigger the visual memory of
headlights approaching immediately before a car
accident (Elhers et al. 2002; Hackmann et al. 2004).
This phenomenon suggests that exploring cognitive
processing of trauma-relevant visual information may
be particularly relevant for individuals with PTSD
symptoms. To date only one study has examined
implicit memory for visual stimuli associated with
traumatic images. Michael and Ehlers (2007) admin-
istered a visual priming task (an implicit memory
paradigm) to non-anxious individuals. These authors
showed participants a series of three pictures that
narrated a story comprising either traumatic or neutral
content. They found enhanced priming for neutral
visual stimuli that appeared before traumatic content,
indicating greater implicit memory for these stimuli.
Furthermore, the magnitude of this implicit memory
bias was associated with frequency of subsequent re-
experiencing (images, intrusive feelings, dreams).
However, in this study enhanced priming was assessed
only for neutral stimuli immediately preceding the
trauma in the narrative, and participants were not
selected for prior trauma experiences or PTSD symp-
toms. Therefore, the role of visual information specic
to traumatic stimuli remains unclear.
The third common feature of cognitive models of
PTSD is that biases in memory for trauma are not
considered a by-product of experiencing trauma or
heightened levels of psychopathology, but a result of
PTSD symptoms. Although some information pro-
cessing research has used clinical control groups to
address this issue (e.g., Stroop task in PTSD and
specic phobia groups; Bryant and Harvey 1995),
most studies examining memory functioning in
PTSD have used healthy individuals or traumatized
individuals without PTSD or anxiety symptoms as
comparison groups. Therefore, these studies cannot
examine the effects of other psychological symptoms
such as state or trait anxiety on their results. Although
researchers have noted the potential role of
50 Cogn Ther Res (2010) 34:4958
1 3
depression in memory biases within PTSD groups
(e.g., Harvey et al. 1998; Brennen et al. 2007; Moore
and Zoellner 2005), to our knowledge no study has
sought to control for level of anxiety (for reviews see
Brewin 2007; Brewin et al. 2007). Thus in the current
study two separate control groups were included to
address both trauma history and anxiety level.
Finally, the current models of PTSD postulate that
unique memory biases may be involved in the
processing of trauma-relevant material. Differentiat-
ing types of emotional stimuli when evaluating
memory disturbances in PTSDis potentially important
because other information processing studies (e.g.,
Martin et al. 1991) suggest that a processing bias for
threat-relevant material in anxiety may be due to the
high level of emotionality of the stimuli rather than to
its threatening nature. While some studies have
attempted to control for emotionality using positive
stimuli in implicit memory paradigms (e.g., Paunovic
et al. 2002), the blunted positive affect demonstrated
by many individuals with PTSD may render this
emotional control condition less effective. Therefore,
negative stimuli may be more salient and relevant for
examining the specicity of trauma related memory
bias in PTSD. To our knowledge, only one published
study (Michael et al. 2005) contrasted implicit mem-
ory effects for trauma-relevant stimuli to that of other
negative emotional stimuli in a sample of clinically
anxious adults. These researchers administered a
modied word-stem completion task comprising
trauma-relevant, general threat and neutral words to
traumatized individuals with and without PTSD. They
found enhanced memory for trauma-relevant words
but not for negative words in individuals with PTSD.
Moreover, this bias for trauma-relevant words was
related to symptom severity over time.
Informed by previous research, in the current study
we attempted to use procedures, materials, partici-
pants, and a design that would allow us to isolate
specic memory phenomena theorized to be involved
in PTSD. Participants were classied into one of four
groups based on self report of prior trauma exposure
and level of anxiety (PTSD symptom group, trauma
control group, anxious control group, and non-anxious
control group). We then administered an implicit
memory paradigm (i.e., clarity-rating task) using
visual stimuli that depicted trauma-relevant, neutral,
and negative images. In the clarity-rating task, partic-
ipants rate the clarity of images (clear: focused,
articulated, or unclear: blurry, out of focus, mumbled).
Implicit memory is demonstrated when previously
studied materials are rated as more clear than novel
materials. This paradigmhas been used successfully to
test implicit memory processes in other anxious
populations (e.g., social phobia; Amir et al. 2003).
We predicted that individuals with PTSD would
demonstrate greater implicit memory bias for trauma
pictures than negative or neutral pictures while the
comparison groups would not show differential
implicit memory for any picture types.
Method
Participants
Participants were 1179 undergraduate students at a the
University of Georgia who completed the Posttrau-
matic Diagnostic Scale (PDS; Foa et al. 1997), the
Beck Depression Inventory (BDI-II; Beck et al. 1996),
and the State-Trait Anxiety Inventory (STAI; Spiel-
berger et al. 1983) as part of an initial screening. The
PDS is a 49-item self-report questionnaire designed to
map onto the DSM-IV diagnostic criteria for PTSD.
The PDS has satisfactory agreement with the Struc-
tured Clinical Interview for DSM-III-R (Spitzer et al.
1990; kappa = .65, agreement = 82%, sensitiv-
ity = .89, specicity = .75; Foa et al. 1997) and has
been used to select PTSD high symptom groups in
student samples (e.g., Twamely et al. 2004). Items
assess type of trauma experienced as well as symptom
severity over the past month, and are scored on a zero to
three scale. The STAI is a 40-iteminventory with items
scored on a one to four scale; 20 items reect current
state anxiety and 20 items reect more general feelings
of trait anxiety. According to Spielberger et al. (1970),
the measure possesses adequate psychometric charac-
teristics. The BDI-II is a 21-item self report measures
assessing depressive symptomology over the previous
2 weeks. Items are multiple choice and scored on a
scale of zero to three; total scores are based on the sum
of items ranging from zero to 63. This measure has
strong psychometric properties (Beck et al. 1988). All
data were collected in large group testing sessions
(n = 50300 per session). Participants were selected
from the university-wide subject pool and received
course credit as compensation. Based on responses to
the questionnaires we created four groups: (1) PTSD
Cogn Ther Res (2010) 34:4958 51
1 3
analog group (PTSD), (2) Anxiety Control Group
(AC), (3) Trauma Control Group (TC) and (4) Non-
anxious Control Group (NAC).
Participants in the PTSD group (n = 40) reported
at least one DSM-IV-dened trauma on the PDS and
reported PTSD symptoms at levels similar to treat-
ment-seeking samples (i.e., PDS total symptoms score
[17). Participants in the AC group (n = 122) did not
report a trauma on the PDS, but endorsed anxiety
symptoms matched to the PTSD group. Participants
in the TC group (n = 70) reported at least one trauma
but reported few symptoms of PTSD (i.e., PDS \5)
and relatively low symptoms of psychopathology
(i.e., BDI \5, STAI Trait \35). Participants in the
NAC group (n = 87) did not experience a trauma and
endorsed relatively low symptoms of psychopathol-
ogy (BDI \5, STAI Trait \35). Participants who
were not eligible for one of these four groups based
on the criteria described below were not included in
the analyses.
Materials
To measure implicit memory bias we used a picture
clarity-rating task (e.g., Amir et al. 2003). We selected
36 digitized pictures from the International Affective
Picture System (IAPS; Lang et al. 1993). Picture
types were categorized as trauma-relevant, neutral or
negative based on pilot data. This pilot was conducted
on separate group of traumatized individuals with
PTSD symptoms from a group of undergraduate
students not included in the current study. To collect
this data, we printed the images used in the current
study and asked participants to identify the content
(trauma-relevant or not) and the emotional valence
using a Likert scale. Trauma-relevant pictures had
assault-related content and negative affective ratings.
Negative pictures had negative affective ratings and
no assault-related content. Neutral pictures had neu-
tral affective ratings and no assault-related content.
We presented the pictures to participants using the
Microsoft PowerPoint presentation program, a
projector, and a large projector screen. We created
two presentation les for the encoding phase: Set 1
and Set 2. Set 1 included 18 randomly selected
pictures (6 trauma-relevant, 6 negative, 6 neutral)
arranged in a random order, programmed to transition
every 5 s. Set 2 comprised the remaining 18 pictures
programmed in the same manner. The order of
presentation was counterbalanced across participants
such that half the participants from each group saw
the pictures from each set at encoding. We created
one le for the memory phase, which included all 36
selected pictures arranged in random order and
programmed to transition every 12 s. Thus, during
the memory phase all participants saw all pictures.
Procedure
The experiment was administered in 10 group
sessions. Participants rst read and signed a consent
form, completed self-report measures (PDS, BDI,
STAI), and then completed the clarity-rating task.
This task had two phases: encoding and memory
test.
During Phase 1 (encoding) participants saw either
Set 1 or Set 2 and were asked to rate the emotional
valence of each picture on a scale from 1 to 9, with
lower numbers indicating more negative valence and
higher numbers indicating more positive valence
(Amir et al. 2003). During Phase 2 (memory test), we
presented the retrieval le which consisted of 36
pictures, half of which had been presented during
encoding and half of which were new. For each
picture participants made two ratings. First, partici-
pants rated the clarity of each picture on a scale from
1 to 5. Clarity was explained to the participants as the
visual quality of the picture in terms of how blurry or
clear the picture appears. A rating of 1 indicated not
clear; 3 indicated somewhat clear; and 5 indicated
very clear. Reliability analysis for ratings in the
current sample produced Cronbachs alpha values
ranging from .73 to .80, with the exception of one set
with a value of .68. Second, participants indicated
whether the picture was old (seen during the encoding
slideshow) or new (never seen before). We included
this rating measure to assess explicit memory effects.
There was a 3 min delay between encoding and
memory phases, during which participants were
asked to list as many U.S. states and state capitals
as possible to remove any pictures from memory.
Results
We rst compared groups on self-report measures,
including trauma type endorsement for the PTSD and
TC groups. Next, we analyzed implicit memory bias
52 Cogn Ther Res (2010) 34:4958
1 3
and explicit memory (recognition) by group and
picture type using analysis of variance. We then
correlated self report measures with implicit memory
bias and explicit memory recognition scores. Finally,
we correlated implicit memory score with explicit
memory scores within groups.
Self Report Measures
As anticipated, the AC group and PTSD group had
signicantly higher scores on measures of anxiety
and depression than the TC and NAC groups. The AC
and PTSD groups did not differ on anxiety, but the
PTSD group reported higher levels of depression. The
TC group and the NAC group did not differ
signicantly on anxiety or depression. The four
groups did not differ on age, education or gender
[ps [.2]. Table 1 presents the demographic and self-
report measures for the four groups.
Trauma Types
Table 2 presents the frequency of traumas reported
on the PDS for the PTSD and TC groups. We
conducted chi squared analyses on the percentage of
individuals in the two traumatized groups endorsing
each trauma type. These analyses revealed that the
PTSD group endorsed more of the following trauma
types: non-sexual assault by someone known by the
victim, sexual assault by someone known by the
Table 1 Demographics and questionnaire data
Group
PTSD (N = 40) AC (N = 122) TC (N = 70) NAC (N = 87)
% Female 73 72 59 70
Age 19.6 (7.4) 19.2 (1.9) 19.3 (2.1) 19.3 (1.5)
Education 13.6 (1.2) 13.7 (1.2) 13.7 (1.4) 13.8 (1.4)
PDS 26.03 (7.35)
a
N/A .89 (1.21)
b
N/A
BDI 19.9 (8.3)
a
14.5 (5.9)
b
2.1 (1.3)
c
1.9 (1.1)
c
STAI-T 51.5 (9.6)
a
49.3 (8.8)
a
25.4 (2.7)
b
25.5 (2.7)
b
STAI-ST 47.0 (11.5)
a
44.4 (10.0)
a
25.1 (4.9)
b
25.3 (5.0)
b
Note: PDS = Posttraumatic diagnostic scale (Foa et al. 1997); BDI = Beck depression inventory (Beck et al. 1996); and
STAI = Spielberger state trait anxiety inventory (T = Trait, ST = State) (Spielberger et al. 1983). We conducted Tukey follow-up
analyses to compare groups. Different superscripts refer to signicant group differences between controls, trauma victims and
individuals with PTSD symptoms. All participants were recruited from a university-wide research pool
Table 2 Percent of groups
endorsing each trauma type
Note: Individuals could
endorse two or more
traumatic experiences, thus
percentages sum to greater
than 100
PTSD TC v
2
(P); df = 1
Serious accident, re, explosion 55 57 .05 (.83)
Natural disaster 43 57 2.20 (.14)
Non-sexual assault (family member) 25 6 8.52 (.01)
Non-sexual assault (stranger) 25 13 2.63 (.11)
Sexual assault (family member) 43 4 25.00 (.01)
Sexual assault (stranger) 28 6 10.27 (.01)
Military combat or war zone 5 2 3.97 (.05)
Sexual contact under 18 with someone
5 or more years older
14 5 13.82 (.01)
Imprisonment 9 2 10.91 (.01)
Torture 10 1 15.71 (.01)
Life threatening illness 16 23 .57 (.45)
Cogn Ther Res (2010) 34:4958 53
1 3
victim, sexual assault by a stranger, military combat,
sexual contact under the age of 18 with someone 5 or
more years older, imprisonment, and torture (see
Table 2).
Clarity-Rating Task (Implicit Memory)
We computed implicit memory indices by subtracting
mean clarity ratings for new pictures from mean
clarity ratings for old pictures (Amir et al. 2003;
Perruchet and Baveux 1989) for each participant and
each picture type (trauma-relevant, negative, and
neutral). A positive index indicated that previously
seen pictures were rated as more clear than novel
pictures. Indices for each group and each picture type
are presented in Table 3.
We submitted the implicit memory indices to a 4
(Groups: PTSD, TC, AC, NAC) 9 3 (Picture type:
Trauma-relevant, Negative, Neutral) factorial analy-
sis of variance (ANOVA) with repeated measurement
on the last factor. This analysis did not reveal a main
effect of Group, F(3, 315) = 1.40, P = .24. How-
ever, there was a signicant main effect of Picture
Type, F(2, 630) = 7.22 P = .001, g
2
= .02, that was
modied by an interaction of Group 9 Picture Type,
F(6, 630) = 2.86, P\.01, g
2
= .03.
To followup this interaction, we conducted one way
ANOVAs within each group. These analyses revealed
that the AC group and the NAC group did not respond
differently to different picture types, ps [.3. How-
ever, the PTSD group had signicantly different
implicit memory indices for different picture types,
F(2, 78) = 3.37, P\.05, g
2
= .08. Negative pictures,
t(39) = 2.03 P\.05, and trauma-relevant pictures,
t(39) = 2.85 P\.01, resulted in greater implicit
memory than neutral pictures. The TC group also had
differing indices for different picture types, F(2,
138) = 9.60, P\.001, g
2
= .12, with greater implicit
memory for negative than neutral pictures,
t(69) = 3.75, P\.01, and negative than trauma-
relevant pictures, t(69) = 3.74, P\.01.
Recognition Task (Explicit Memory)
We also examined participants explicit memory for
each picture type. We submitted participants per-
centage of correct picture recognition to a 4 (Groups:
PTSD, TC, AC, NAC) 9 3 (Picture type: Trauma-
relevant, Negative, Neutral) ANOVA with repeated
measurement on the last factor. These analyses did not
reveal a main effect for Group, F(3, 315) = 1.67,
P = .17, or an interaction of Group 9 Picture Type,
F(6, 630) = .95, P = .46. However, there was a main
effect of Picture Type, F(2, 630) = 37.44, P\.001,
g
2
= .11. Explicit memory for negative pictures,
t(318) = 6.76, P\.001, and trauma-relevant pic-
tures, t(318) = 7.22, P\.001, was better than for
neutral pictures. Explicit memory for trauma-relevant
and negative pictures was not signicantly different,
P[.9. Accuracy (percent correct) for explicit mem-
ory ratings for each group and each picture type are
presented in Table 4.
Correlational Analyses
We rst computed correlations between self-report
measures and memory scores. Memory scores were
not signicantly correlated with STAI-S, STAI-T, or
BDI scores (ps [.15). However, for individuals
completing the PDS (PTSD and TC groups) there
was a signicant relationship between PDS and
implicit memory bias for trauma pictures, r = .22,
Table 3 Mean and standard deviations for implicit memory
bias across picture type
Group
PTSD AC TC NAC
Picture type
Negative .23 (.80) .18 (.46) .26 (.44) .04 (.49)
Trauma-relevant .21 (.54) .11 (.43) -.00 (.42) .08 (.34)
Neutral -.01 (.56) .11 (.41) .02 (.34) .05 (.37)
Note: PTSD = PTSD symptom group; AC = Anxiety control
group; TC = Trauma control group; NAC = Non-anxious
control group
Table 4 Mean and standard deviations for percent accuracy
on explicit memory test across picture type
Group
PTSD AC TC NAC
Picture type
Negative 97 (8) 95 (14) 96 (12) 97 (11)
Trauma-relevant 96 (9) 95 (13) 97 (7) 97 (7)
Neutral 86 (16) 86 (16) 92 (11) 89 (13)
Note: PTSD = PTSD symptom group; AC = Anxiety control
group; TC = Trauma control group; NAC = Non-anxious
control group
54 Cogn Ther Res (2010) 34:4958
1 3
P\.05, as well as explicit memory recognition for
neutral pictures r = .19, P\.05.
Next, we examined the relationship between
implicit and explicit memory within the groups. For
trauma-relevant pictures, results revealed that in the
PTSD group explicit memory for trauma-relevant
pictures was negatively correlated with implicit mem-
ory for trauma-relevant pictures, r = -.64, P\.001,
negative pictures, r = -.45, P\.01, and neutral
pictures, r = -.45, P\.01. These relationships were
not signicant in the TC group. For negative pictures,
explicit memory was negatively correlated with
implicit memory for negative pictures in the PTSD
group, r = -.36, P\.05, and the TCgroup, r = -.26,
P\.05. No other correlations were signicant,
ps [.1.
Discussion
Our results suggest that individuals with PTSD symp-
toms show enhanced implicit memory for trauma-
relevant and negative images relative to neutral
images. Individuals who experienced a trauma but
were asymptomatic demonstrated enhanced implicit
memory for negative pictures relative to trauma-
relevant and neutral pictures. Thus, experiencing a
traumawith or without PTSDsymptomswas asso-
ciated with enhanced implicit memory for negative
pictures relative to neutral pictures, while PTSD
symptoms were also associated with an implicit
memory bias for trauma pictures relative to neutral
pictures. Individuals without a history of trauma,
regardless of current level of anxiety, did not demon-
strate differential implicit memory between picture
types. This nding is consistent with prior literature
examining implicit memory in individuals high in trait
anxiety (for a review see Russo et al. 1999). Consid-
ered together, our results are consistent with prior
research suggesting that trauma-relevant information
is preferentially activated and processed in individuals
with PTSD symptoms at an automatic level (i.e., Amir
et al. 1996).
Our results regarding explicit memory bias were
consistent with research demonstrating superior recall
and recognition biases for emotional words in unse-
lected participants (e.g., Ochsner 2000). This nding
may be due to unique properties of emotional infor-
mation that enable more elaborate encoding (e.g.,
physiological and evaluative responses, speeded pro-
cessing, increased rehearsal; Ochsner 2000). However,
explicit memory for trauma pictures was not enhanced
for individuals in the PTSD group relative to control
groups. Although other studies have found explicit
memory bias for trauma-relevant information in indi-
viduals with PTSD (e.g., Golier et al. 2003; Paunovic
et al. 2002), methodological differences between the
studies may have contributed to this inconsistency. For
example, explicit memory for visual material used in
the current study was very high relative to the verbal
material used in other studies. This may have made the
current paradigm less sensitive to detecting explicit
memory bias differences between picture types.
What are the potential implications of the current
results for models of PTSD? First, our results are in
line with the hypothesis that cognitive patterns
associated with PTSD symptom maintenance occur
in memory systems operating outside of conscious
awareness. However, our data suggest that implicit
and explicit memory systems might interact for some
stimuli in individuals who have experienced trauma.
Specically, there was a high negative correlation
between explicit recognition and implicit memory
bias for trauma-related information in individuals
with PTSD symptoms. Research suggests that indi-
viduals who develop symptoms of PTSD after a
traumatic event commonly engage in maladaptive
cognitive strategies such as suppression (Elhers and
Clark 2000). Although avoiding explicit recollections
of unwanted thoughts may temporarily block con-
scious retrieval, it may also result in a rebound,
increasing activation of these same thoughts (for a
review see Wenzlaff and Wegner 2000). Conscious
avoidance of trauma-relevant memories by individu-
als with PTSD may lead to a paradoxical rebound in
retrieval of trauma-specic information in the form of
intrusive symptoms (Shipherd and Beck 2005; Steil
and Ehlers 2000).
The relationship between avoidance and automatic
retrieval in PTSD may partly explain the current
results. That is, repetitively accessing trauma-relevant
material at an automatic level might be reected in
implicit memory bias for such material. Conversely, in
the absence of PTSD symptom development or with
recovery from such symptoms (corresponding with
fewer intrusive symptoms), trauma-relevant informa-
tion may have diminished priority for preferential
automatic processing and thus less propensity to be
Cogn Ther Res (2010) 34:4958 55
1 3
remembered implicitly by individuals without PTSD.
Given the relationship between explicit recall and
implicit memory bias found in the current study, one
potential avenue for future studies would be to examine
the relationship between such memory biases and
specic symptom clusters, particularly avoidance
symptoms. If explicit avoidance is associated with
enhanced rebound processing at the automatic level,
individuals endorsing such symptoms would likely
exhibit the strongest effect on a measure of implicit
memory.
Our results also provide support for current models
emphasizing the specicity of information processing
biases for trauma-relevant visual information in
individuals with PTSD symptoms. To our knowledge,
this is the rst study to include a matched group of
generally anxious participants in addition to non-
anxious and trauma control groups within a visual
implicit memory paradigm. Enhanced implicit mem-
ory for trauma-relevant pictures relative to neutral
pictures was unique to traumatized individuals with
PTSD symptoms. The differential pattern of implicit
memory biases for different picture types demon-
strated by these groups suggests that the memory
mechanisms involved in PTSD symptoms may be
readily differentiated from those of general distress
and trauma history. This in turn lends support to the
diagnostic specicity of PTSD. Furthermore, the
implicit nature of this paradigm makes the procedure
less susceptible to demand characteristics. Thus it
may be possible to use this measure in conjunction
with self-report and clinician-assessed symptoms to
differentiate individuals presenting with general dis-
tress, trauma history, and individuals presenting
symptoms specically as a result of PTSD. Informa-
tion garnered from this type of paradigm may be also
be utilized clinically to assess symptom change
without relying exclusively on self report. In addition,
this type of information processing bias could
potentially be targeted in future treatment programs
for PTSD. Other cognitive bias training programs are
currently being utilized therapeutically to modify
attention biases associated with Generalized Anxiety
Disorder (see Amir et al. in press).
Contrary to a prior study using negative stimuli
(e.g., Michael et al. 2005), both the PTSD and TC
groups also demonstrated preferential implicit pro-
cessing of negative information relative to neutral
information. One difference between the study by
Michael et al. and the current study is the use of
visual versus verbal information. Visual information
may be more sensitive to implicit memory effects
because it is strongly perceptual in nature. The
current models of PTSD do not stipulate what
information processing bias exists for non-traumatic,
negatively valenced information. Experience of
trauma likely results in salient memories associated
with negative feelings of fear, helplessness or horror
even in the absence of pathological response. As
such, individuals with salient trauma experiences
may have greater consolidation, elaboration, and
retrieval of memories associated with negative expe-
riences. This heightened processing may be reected
in implicit memory bias for this type of material,
even for individuals who do not subsequently develop
symptoms. Future studies should examine the role of
negative visual information in implicit memory
paradigms for these groups in order to better integrate
such ndings to the information processing associ-
ated with experiencing trauma.
Our study has limitations. For example, our
participants did not receive diagnostic interviews to
assess for PTSD status, and the sample consisted of a
primarily female students. Because of the limited
sample size and exclusive reliance on self-report, it is
difcult to generalize our ndings to clinical samples
with PTSD. However, the PDS has been used to
select PTSD diagnostic groups in student samples
(e.g., Twamely et al. 2004) and demonstrates high
sensitivity and specicity (Foa et al. 1997). Further-
more, our materials were not specic to the trauma
experienced by each individual participant, and
groups differed on trauma types experienced. Idio-
graphic selection of materials is important given the
diverse nature of experiences that may lead to
psychological distress. When examining implicit
memory only in individuals who had experienced
assault-relevant traumas we found the same pattern of
implicit and explicit memory mean scores, although
group differences did not reach statistical signicance
due to the small sample size. Future research should
address this limitation by matching participant trau-
mas to stimuli used within the study paradigm.
Additionally, although the clarity-rating paradigm
used in the current study has been used in prior
literature to assess implicit memory, further research
is needed to conrm the reliability and validity of this
paradigm in anxious samples, particularly when
56 Cogn Ther Res (2010) 34:4958
1 3
delivered in a group setting. Finally, our anxious
control sample was not matched to the PTSD
symptoms group on other psychological variables
(e.g., depression). Therefore, we cannot speak to the
specicity of the results in PTSD as opposed to
depression. However, because the magnitude of the
difference between the AC and PTSD groups was
small on our measure of depression (6 points) and
both groups were below the clinical range, it is
unlikely that depression was solely responsible for
the results.
In summary, consistent with models of PTSD we
found that implicit memory bias in traumatized
individuals with and without PTSD may be informa-
tive regarding the nature of information processing
bias in these individuals. These results suggest that
differential implicit memory effects may play a role
in the maintenance of PTSD symptoms and may
differentiate individuals with and without PTSD
symptoms after experiencing trauma. Further
research might look at whether such biases predict
subsequent development of PTSD symptoms and
how such effects change as a result of treatment.
References
American Psychiatric Association (2000). Diagnostic and
statistical manual of mental disorders (4th ed., text revi-
sion). Washington, DC: Author.
Amir, N., Beard, C., Burns, M., & Bomyea, J. Attention modi-
cation program for individuals with Generalized Anxiety
Disorder. Journal of Abnormal Psychology, in press.
Amir, N., Bower, E., Briks, J., & Freshman, M. (2003).
Implicit memory for negative and positive social infor-
mation in individuals with and without social anxiety.
Cognition and Emotion, 17(4), 567583. doi:10.1080/02
699930302300.
Amir, N., McNally, R. J., & Wiegartz, P. S. (1996). Implicit
memory bias for threat in posttraumatic stress disorder.
Cognitive Therapy and Research, 20, 625635. doi:
10.1007/BF02227965.
Beck, A. T., Steer, R. A., & Brown, G. (1996). Beck depression
inventory manual-II. San Antonio, TX: Psychological
Corporation.
Beck, A. T., Steer, R. A., & Garbin, M. G. (1988). Psychometric
properties of the beck depression inventory: Twenty-ve
years of evaluation. Clinical Psychology Review, 8(1), 77
100. doi:10.1016/0272-7358(88)90050-5.
Brennen, T., Dybdahl, R., & Kapidzic, A. (2007). Trauma-
related and neutral false memories in war-induced post-
traumatic stress disorder. Consciousness and Cognition,
16, 877885. doi:10.1016/j.concog.2006.06.012.
Brewin, C. R. (2007). Autobiographical memory for trauma:
Update on four controversies. Memory (Hove, England),
15(3), 227248. doi:10.1080/09658210701256423.
Brewin, C. R., Dalgleish, T., & Joseph, S. (1996). A dual
representation theory of posttraumatic stress disorder.
Psychological Review, 103, 670686. doi:10.1037/0033-
295X.103.4.670.
Brewin, C., Kleiner, J., Vasterling, J., & Field, A. (2007).
Memory for emotionally neutral information in posttrau-
matic stress disorder: A meta-analytic investigation.
Journal of Abnormal Psychology, 116, 448463. doi:
10.1037/0021-843X.116.3.448.
Bryant, R. A., & Harvey, A. G. (1995). Processing threatening
information in posttraumatic stress disorder. Journal of
Abnormal Psychology, 104(3), 537. doi:10.1037/0021-
843X.104.3.537541.
Dalgleish, T. (2004). Cognitive approaches to posttraumatic
stress disorder: The evolution of multirepresentational
theorizing. Psychological Bulletin, 130(2), 228260.
Elhers, A., & Clark, D. M. (2000). A cognitive model of
posttraumatic stress disorder. Behaviour Research and
Therapy, 38, 319345.
Elhers, A., Hackmann, A., Steil, R., Clohessy, S., Wenninger,
K., & Winter, H. (2002). The nature of intrusive memories
after trauma: The warning signal hypothesis. Behavior
Research & Therapy, 40, 995.
Elhers, A., & Steil, R. (1995). Maintenance of intrusive memo-
ries in posttraumatic stress disorder: A cognitive approach.
Behavioral and Cognitive Psychotherapy, 23, 217249.
Foa, E. B., Cashman, L., Jaycox, L., & Perry, K. (1997). The
validation of a self-report measure of posttraumatic stress
disorder: The posttraumatic diagnostic scale. Psycholog-
ical Assessment, 9, 445451.
Foa, E. B., & Kozak, M. J. (1986). Emotional processing of
fear: Exposure to corrective information. Psychological
Bulletin, 99, 2035.
Golier, J., Yehuda, R., Lupien, S., & Harvey, P. (2003).
Memory for trauma-related information in Holocaust
survivors with PTSD. Psychiatry Research, 121, 133143.
Hackmann, A., Ehlers, A., Speckens, A., & Clark, D. (2004).
Characteristics and content of intrusive memories in
PTSD and their changes with treatment. Journal of
Traumatic Stress, 17, 231240.
Harvey, A. G., Bryant, R. A., & Dan, S. T. (1998). Autobio-
graphical memory in acute stress disorder. Journal of
Consulting and Clinical Psychology, 66(3), 500506.
Kessler, R. C., Sonnega, A., Bromet, E., Hughes, M., & Nel-
son, C. B. (1995). Posttraumatic stress disorder in the
national comorbidity survey. Archives of General Psy-
chiatry, 52, 10481063.
Lang, P., Greenwald, M., Bradley, M., & Hamm, A. (1993).
Looking at pictures: Affective, facial, visceral, and
behavioral reactions. Psychophysiology, 30, 261273.
Martin, M., Williams, R. M., &Clark, D. M. (1991). Does anxiety
lead to selective processing of threat-related information?
Behaviour Research and Therapy, 29, 147160.
McNally, R. J. (1995). Automaticity and the anxiety disorders.
Behavior Research & Therapy, 33(7), 747754.
McNally, R. J., & Amir, N. (1996). Perceptual implicit mem-
ory for trauma-related information in post-traumatic stress
disorder. Cognition and Emotion, 10(5), 551556.
Cogn Ther Res (2010) 34:4958 57
1 3
Michael, T., & Ehlers, A. (2007). Enhanced perceptual priming
for neutral stimuli occurring in traumatic context: Two
experimental investigations. Behavior Research & Ther-
apy, 45, 341358.
Michael, T., Elhers, A., & Halligan, S. L. (2005). Enhanced
priming for trauma-related material in posttraumatic stress
disorder. Emotion, 5(1), 103112.
Moore, S. A., & Zoellner, L. A. (2005). Overgeneral autobio-
graphical memory and traumatic events: An evaluative
review. Psychological Bulletin, 133(3), 419437.
Ochsner, K. R. (2000). Are affective events richly remembered
or simply familiar? The experience and process of recog-
nizing feelings past. Journal of Experimental Psychology:
General, 129(2), 242261.
Paunovic, N., Lundh, L., & Ost, L. (2002). Attentional and
memory bias for emotional information in crime victims
with acute posttraumatic stress disorder (PTSD). Journal
of Anxiety Disorders, 16, 675692.
Perruchet, P., & Baveux, P. (1989). Correlational analyses of
implicit and explicit memory performance. Memory and
Cognition, 17(1), 7786.
Russo, R., Fox, E., & Bowles, R. J. (1999). On the status of
implicit memory bias in anxiety. Cognition and Emotion,
13(4), 435456.
Schacter, D. L. (1987). Implicit memory: History and current
status. Journal of Experimental Psychology: Learning.
Memory and Cognition, 13(3), 501518.
Schacter, D. L., & Badgaiyan, R. D. (2001). Neuroimaging of
priming: New perspectives on implicit and explicit
memory. Current Directions in Psychological Science,
10(1), 14.
Shipherd, J., & Beck, G. J. (2005). The role of thought sup-
pression in posttraumatic stress disorder. Behavior
Therapy, 36(3), 277287.
Spielberger, C. D., Gorsuch, R. L., & Lushene, R. (1970).
Manual for the state-trait anxiety inventory (STAI). Palo
Alto, CA: Consulting Psychology Press.
Spielberger, C. D., Gorsuch, R. L., Lushene, R., Vagg, P. R., &
Jacobs, G. A. (1983). Manual for the state-trait anxiety
inventory. Palo Alto, CA: Consulting Psychologist Press.
Spitzer, R. L., Williams, J. B. W., Gibbon, M., & First, M. B.
(1990). Users guide for the structured clinical interview
for DSM-III-R: SCID. Washington, DC: American Psy-
chiatric Association.
Steil, R., & Ehlers, A. (2000). Dysfunctional meaning of
posttraumatic intrusions in chronic PTSD. Behavior
Research and Therapy, 38, 537558.
Twamely, E. W., Hami, S., & Stein, M. B. (2004). Neuro-
psychological function in college students with and
without posttraumatic stress disorder. Psychiatry
Research, 126, 265274.
Wenzlaff, R. M., & Wegner, D. M. (2000). Thought suppres-
sion. Annual Review of Psychology, 51, 5991.
58 Cogn Ther Res (2010) 34:4958
1 3
Copyright of Cognitive Therapy & Research is the property of Springer Science & Business Media B.V. and its
content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for individual use.

Anda mungkin juga menyukai