ORIGINAL ARTICLE
Mental Representations in Women With Panic Disorder
An Urban African-American Sample
John H. Porcerelli, PhD, ABPP,* Steven K. Huprich, PhD,† and Tsveti Markova, MD, FAAFM*
Abstract: Psychodynamic theories of panic disorder (PD) suggest an asso-
ciation between PD and impairments in mental representations. This study
tested this hypothesis by comparing mental representations of 25 African-
American women recruited from an urban primary care clinic with PD with
a group of 25 women without PD and matched on race, age, marital status,
education, and income. Mental representations were assessed through spon-
taneous descriptions of mothers and reliably coded with the Qualitative and
Structural Dimensions of Object Representations scale (Blatt et al., Unpub-
lished manual, 1992). Results essentially supported the hypotheses for
impaired mental representations (lower benevolence, higher ambivalence,
and lower conceptual level) for women with PD as compared with women
without PD. No differences in mental representations were found between
women with PD with major depression and women with only PD. The
findings provide preliminary support for a psychodynamic theory of PD in a
sample of African-American women.
Key Words: Mental representations, maternal representations, panic
disorder, African American, primary care.
(J Nerv Ment Dis 2010;198: 144 –149)
T he empirical study of mental representations has provided clini-
cians and researchers with valuable information about normal
and pathological personality functioning as well as various types of
Axis I pathology (Huprich and Greenberg, 2003). This paper ex-
plores the relationship between mental representations and panic
disorder (PD). More specifically, we assess the content (benevo-
lence/malevolence, punitiveness, and ambivalence) and structural
development of maternal representations in African-American
women, with and without PD, seeking care from an urban primary
care clinic. Maternal representations were chosen due to the fact that
the majority of women in our sample (n ⫽ 39, 78%) were raised in
father-absent homes.
Epidemiological community-based studies suggest that the
prevalence of PD is similar among African Americans and Cauca-
sians (Horwath et al., 1993). PD patients often suffer from comorbid
anxiety, depressive, and substance abuse disorders (Sanderson et al.,
1990; Goisman et al., 1994) as well as medical comorbidities (e.g.,
irritable bowel syndrome) (Zaubler and Katon, 1996). However,
compared with Caucasians with PD, urban-dwelling African Amer-
icans with PD tend to have a greater number of unnecessary
psychiatric hospitalizations and emergency room visits, a higher
*Department of Family Medicine and Public Health Sciences, Wayne State
University School of Medicine, Detroit, MI; and †Department of Psychology,
Eastern Michigan University, Ypsilanti, MI.
Supported by a grant from the Wayne State University Department of Family
Medicine Research Fund.
Send reprint requests to John H. Porcerelli, PhD, ABPP, Department of Family
Medicine and Public Health Sciences, Wayne State University School of
Medicine, 1101 W. University Dr, 3-North, Rochester, MI 48307. E-mail:
jporcer@med.wayne.edu.
Copyright © 2010 by Lippincott Williams & Wilkins
ISSN: 0022-3018/10/19802-0144
DOI: 10.1097/NMD.0b013e3181cc41ca
144 | www.jonmd.com The Journal of Nervous and Mental Disease • Volume 198, Number 2, February 2010
prevalence of childhood trauma, and greater psychosocial stress
(Friedman et al., 1995).
In an effort to understand the contribution of psychodynamics
to PD, Shear (Shear et al., 1993) conducted in-depth interviews of 9
patients formally diagnosed with PD (5 with comorbid major de-
pression). These patients reported “meaningful stressors preceding
panic onset that were typically linked to childhood experiences and
presented a threat to important attachments” (p 137) (Busch et al.,
2009). They also reported memories suggestive of temperamental
shyness/anxiety as children, which infant researchers have identified
as a predisposing factor in the development of anxiety disorders
(Kagan et al., 1990). Finally, these patients described their parents as
angry, frightening, critical, and controlling. Shear et al. (1993)
suggested that children with inborn inhibitions and low assertiveness
may attribute their fearful behavior as a submission to powerful and
controlling parents. However, they also noted that parents who
“restrict or frighten children might be even more aversive to the
timid child than to another child” (p 862). Regardless of the exact
origin, these experiences contribute to the child’s internal represen-
tations of the parents as powerful (punitive and abandoning) and of
the self as weak. At the same time, in the young child’s mind, his or
her self-representation may be one of being very powerful (i.e., his
or her anger can destroy the parents) and the parental representations
as weak. These representations thus pave the way for conflicts over
dependence, independence, ambivalence, and anger. Defenses such
as denial and projection are often used by these children to rid
themselves of their (feared) anger toward parents which only serves
to exacerbate their fears of separation (Busch et al., 1991). Neglect-
ful and/or overly punitive parenting can further exacerbate these
conflicts, leaving such individuals overly fearful of separation of
significant others.
In an early qualitative study by Tucker (1956), perceptions
of early parenting were linked to phobic reactions (i.e., panic
attacks) in a sample of mostly women between the ages of 20 and
40 years old. In describing the dynamics of these women, Tucker
noted immaturity in their personality development combined with a
high degree of ambivalence toward their mothers. Feelings of
insecurity in childhood were brought about by observing conflict
between the parents, a lack of parental affection, overprotection and
punitiveness.
Arrindell et al. (1983) assessed early memories of parenting
behavior in patients diagnosed with social phobia, agoraphobia, and
acrophobia and compared them to a sample of nonpatient volunteers.
They found that, compared with controls, social and height phobics
reported both parents as less warm, more rejecting, and more
overprotective. They found that agoraphobics also reported both
parents as less warm but only mothers as more rejecting that
controls. These findings suggest that patients with PD may exhibit
more punitive maternal representations than patients without PD.
Several studies have been conducted comparing patients with
and without PD using a self-report measure of parental representa-
tions (Parental Bonding Instrument, Parker et al., 1979) that retro-
spectively assesses perceptions of parental care and overprotection
during childhood (Faravelli et al., 1991; Pacchierotti et al., 2002;
Silove, 1986; Wilborg and Dahl, 1997; Silove et al., 1991). These
The Journal of Nervous and Mental Disease • Volume 198, Number 2, February 2010
studies found lower parental care, higher parental protectiveness, or
both to be associated with PD. Wilborg and Dahl (1997) and Farvelli
et al. (1991) reported these differences only with PD patients with
severe (versus less severe) agoraphobia versus controls. It should be
noted that the PBI assess parental care and overprotection but does
not assess parental punitiveness.
Therefore, data obtained from psychodynamic interviews and
self-report measures suggest that patients with PD represent their
parents as less caring and more overly protective. What has not yet
been explored is whether PD patients’ current representations of
parents (i.e., object representations) would follow a similar pattern.
Shear et al. (1993) have noted that eliciting descriptions of
parents during diagnostic interviews is “likely to provide a rough
approximation of the quality of internalized object relations” (p 862)
in patients with PD. Blatt et al. (1992) developed the Qualitative and
Structural Dimensions of Object Representations scale to evaluate
the content and the structure of open-ended descriptions of parents.
Content dimensions include the degree of benevolence-malevolence,
punitiveness, and ambivalence of representations. The structural
variable (Conceptual Level - CL) assesses the development of
representations along 5 levels: sensorimotor-preoperational (others
described primarily in terms of need-gratification), concrete percep-
tual (others described in terms of their physical characteristics),
external iconic (description of the behavior and actions of others),
internal iconic (describing the feelings and mental states of others),
and conceptual (describing others as developing, autonomous, and
complex beings). In Blatt et al. (1994) cognitive-developmental/
psychodynamic model, representations become increasingly accu-
rate, articulated, and conceptually complex structures over time.
Representations begin as global, diffuse, fragmentary, and inflex-
ible. With adequate parenting they transform into increasingly
differentiated, flexible, and hierarchically organized representa-
tions. At the more adaptive realm of functioning, complex rela-
tionships of the parents exist, and individuals tend not to have as
severe psychopathology as those with more simplistic and im-
mature representations (Blatt et al., 1996; Huprich and Green-
berg, 2003; Porcerelli et al., 2006).
Based upon Shear et al.’s (1993) psychodynamic model of PD
as well as previous studies of the relationship between parental
representations and PD (Faravelli et al., 1991; Pacchierotti et al.,
2002; Silove, 1986; Wilborg and Dahl, 1997; Shear et al., 1993;
Silove et al., 1991), we hypothesize that women with PD (versus
women without PD) will have: (1) Lower maternal ratings of care
(i.e., lower Benevolence factor score), (2) Higher maternal ratings of
punitive and judgmental behavior (i.e., higher Punitive factor score),
(3) Higher ratings of maternal ambivalence, (4) Less developmen-
tally mature conceptual level representations, and (5) Lower levels
of physical functioning.
METHOD
Participants
Study participants included women from an urban primary
care Family Medicine residency training clinic in Detroit, Michigan.
Two-hundred fifty consecutive patients were asked to participate in
a larger study on women’s health, and 220 (88%) agreed to partic-
ipate. The larger study did not focus on either PD or parental
representation. All participants signed a university-approved IRB-
approved consent form. Only measures relating to PD and parental
representations are reported here.
Procedures
Patients were recruited from the clinic waiting room. In a
private area of the clinic, participants completed several health and
psychiatric (self-report) measures and provided descriptions of their
© 2010 Lippincott Williams & Wilkins
Representations in Panic Disorder
mother. A masters-level research assistant (RA) asked participants
to “Describe your mother. Describe her in a way that we can get a
picture of who she is.” Patients received an honorarium of $15 for
their participation.
Parental descriptions were written verbatim by a RA and later
transcribed for coding purposes. Another RA, a graduate student
from an A.P.A.- accredited, psychodynamically-oriented, doctoral
program in clinical psychology, underwent extensive training for
coding representations. Once an acceptable level of reliability was
obtained with practice protocols, the RA coded all 50 study proto-
cols. Twenty-five randomly chosen protocols were also coded by the
first author (J.H.P.), for assessing inter-rater reliability. Coders were
blind to demographics, diagnoses, and scores on all other rating
scales. All scoring discrepancies between the raters were rectified
through discussion for final data analysis.
Measures
Mental Representations
Spontaneous descriptions of mothers were coded with the
Qualitative and Structural Dimensions of Object Representations
scale (Blatt et al., 1992). Qualitative dimensions include 13
features of parents and comprise 3 factors: (1) Benevolent-
Malevolent, (2) Punitiveness, and (3) Ambitiousness (not in-
cluded in this study). Characteristics that form the Benevolent-
Malevolent factor include: weak-strong, cold-warm, successful,
malevolent-benevolent, nurturing, affectionate, negative-positive
ideal, and constructive involvement. The weak-strong and suc-
cessful parental characteristics were not included in the data
analysis because they were unscorable in over 20% of the
protocols. Characteristics that form the Punitiveness factor in-
clude: judgmental and punitive. Maternal ambivalence ratings
were also obtained. A fourth factor, length of description (i.e.,
number of word), is used as a covariate if length significantly
correlates with qualitative or structural ratings. Maternal descrip-
tions were rated on the 11 aforementioned features using a
7-point Likert scale, except for the Ambivalent dimension which
is scored on a 5-point scale. Higher scores on the Benevolent
factors and lower scores on the Punitive factor represent healthier
functioning. CL is rated on a 9-point ordinal scale with a rating
of 1 representing developmentally immature representations and
ratings of 9 representing developmentally mature representations.
Level 1 (Sensorimotor-Preoperational) scores primarily reflect
descriptions of parents’ activity in reference to the gratification or
frustration they provide. At level 1, parents are not experienced
as having separate identities from the person providing the
description. Level 3 (Concrete-Perceptual) scores involve global,
concrete, literal descriptions of parents (e.g., what they look
like), although they are described as separate beings. Level 5
(External Iconic) scores focus on specific part properties of the
parents, often in terms of their functional activities or attributes.
Level 7 (Internal Iconic) scores involve part properties of
thoughts and feelings rather than just activities. Internal Iconic
descriptions convey aspects of the parents’ internal state. Level 9
(Conceptual) scores involve the most complex descriptions of
parents that integrate prior conceptual levels and thus combining
external and internal characteristics of parents with an apprecia-
tion of changes over time and the role of conflict in mental life.
Scores of 2, 4, 6, and 8 are used when criteria for a particular
level are not fully achieved. Numerous studies support the
reliability and validity of the qualitative factors and structural
scores (Blatt et al., 1979; Blatt et al., 1991; Blatt et al., 1996;
Blatt, 2004; Bornstein and O’Neill, 1992; Bornstein et al., 1986;
Levy et al., 1998; Porcerelli et al., 2006; Quinlan et al., 1992).
www.jonmd.com | 145
Porcerelli et al. The Journal of Nervous and Mental Disease • Volume 198, Number 2, February 2010

Physical Functioning TABLE 1. Demographics

The Medical Outcomes Study Short-Form (SF-20; Stewart et
al., 1988), a widely used measure of health status, is a 20-item Panic Disorder Matched Controls
self-report scale that assesses physical and mental health. The Variable N ⴝ 25 N ⴝ 25
physical functioning score is made up of 6 items involving overall Age, M (SD) 40.84 (8.93) 38.68 (10.28)
health, physical functioning, pain, and health perceptions. Raw Gender
scores from each scale are converted to percentage scores ranging Female (%) 25 (100%) 25 (100%)
from 1 to 100. Higher scores represent healthier functioning. For a Race (%)
review of the reliability and validity data, see Stewart et al. (1988, African American (%) 25 (100%) 25 (100%)
1989) and Stein et al. (1998).
Marital status n (%)
Married 6 (24%) 4 (16%)
Panic Disorder
Living with partner 2 (8%) 0 (0%)
The Patient Health Questionnaire (PHQ; Spitzer et al., 1999)
was used to assess PD and Major Depressive Disorder (MDD). The Single 6 (24%) 10 (40%)
PHQ is a self-report scale that was developed specifically for Separated/divorced 10 (40%) 9 (36%)
primary care practices to assess common DSM-IV (American Psy- Widowed 1 (4%) 2 (8%)
chiatric Association, 1994) disorders. The 15-item PD scale and the Income n (%)
9-item MDD scales were used for this study. Spitzer et al. (1999) $0–$ 9999 13 (52%) 10 (40%)
have reported sensitivity and specificity at 0.81 and 0.99, respec- $10,000–$19,999 8 (32%) 12 (48%)
tively, for the Panic Disorder scale, and 0.73 and 0.98, respectively, $20,000–$29,999 0 (0%) 0 (0%)
for the MDD scale. The validity of the PHQ has been supported by $30,000–$39,999 1 (4%) 0 (0%)
studies of patients in primary care settings (Spitzer et al., 1999) and $40,000–$49,999 3 (12%) 3 (12%)
obstetric-gynecologic (Spitzer et al., 2000) patients who completed Education n (%)
the PHQ, as well as in structured diagnostic telephone interviews by
Less than high school 2 (8%) 5 (20%)
a mental health professional blind to the results of the PHQ.
Some high school 11 (44%) 7 (28%)
Inclusion Criteria and Matching Procedures High school graduate 7 (28%) 7 (28%)
Women were included in the PD group with and without Some college 5 (20%) 5 (20%)
MDD. Participants were considered for the comparison group if they College graduate 0 (0%) 1 (4%)
did not meet PHQ criteria for PD and MDD, and could be matched Employment n (%)
on race, gender, age (⫾4 years), marital status, education, and Full-time 3 (12%) 8 (32%)
income with a member of the PD group. Each PD group member Part-time 3 (12%) 4 (16%)
was matched with one comparison group member from the larger Unemployed 19 (76%) 13 (52%)
women’s health study sample. When more than one participant was Diagnosis n (%)
eligible for matching to a PD group member, one was randomly Panic disorder 25 (100%) 0 (0%)
chosen for inclusion into the study. Major depressive disorder 13 (52%) 0 (0%)

RESULTS
Sample
The demographics of the sample by group are listed in
Table 1.
Diagnostic Group and Matched Comparison Group
A total of 25 women met PHQ criteria for PD, with 13 of 25
(52%) also meeting PHQ criteria for MDD. The 25 women in the PD
group were matched according to race, gender, age (⫾4 years),
marital status (96% exact match), education (92% exact match), and
income (96% exact match).
Inter-Rater Agreement
Raters agreed 95% of the time on whether a maternal descrip-
tion was codable for each characteristic. Discrepancies were re-
solved through discussion. Inter-rater agreement was calculated
using the intraclass correlation coefficient (ICC). Shrout and Fleiss
(1979) provided guidelines for interpreting ICC values: poor ⫽
⬍0.40, fair ranges from 0.40 to 0.59, good ranged from 0.60 to 0.74,
and excellent is ⬎0.74. ICCs for the qualitative dimension scales
ranged from 0.72 to 0.88 (good to excellent) for descriptions of the
mother. ICCs for the structural dimension ratings (conceptual level)
were 0.83 (excellent) for descriptions of mother.
Correlations Between Length of Descriptions,
Factor Scores, Ambivalence and CL
If the length of the maternal descriptions is significantly
correlated with any of the representation scores, Blatt et al. (1992)
recommend that it be statistically held constant. The length of
maternal descriptions were not significantly correlated with the
Benevolence-Malevolence factor, r ⫽ ⫺0.24, p ⫽ 0.09; Punitive
factor, r ⫽ ⫺0.07, p ⫽ 0.61; Ambivalence, r ⫽ 0.23, p ⫽ 0.10; or
CL, r ⫽ 0.14, p ⫽ 0.32. Hence, no statistical control of length was
needed.
Intercorrelations
Correlations between the maternal representation factors, am-
bivalence, and CL scores were: Benevolence-Malevolence factor
and Punitive factor, r ⫽ ⫺0.35, p ⫽ 0.01; Benevolence factor and
Ambivalence, r ⫽ ⫺0.64, p ⬍ 0.001; Benevolence and CL, r ⫽
0.35, p ⫽ 0.01; Punitive and Ambivalence, r ⫽ 0.10, p ⫽ 0.47;
Punitive and CL, r ⫽ ⫺0.08, p ⫽ 0.59; Ambivalence and CL, r ⫽
⫺0.12, p ⫽ 0.41. All correlations were in the expected direction.
PD Group (n ⴝ 25) Versus Matched Comparison
Group (n ⴝ 25)
A MANOVA was performed with the PD and matched
comparison groups serving as independent variables and the mater-
nal Benevolence-Malevolence factor, Punitive factor, Ambivalence

146 | www.jonmd.com © 2010 Lippincott Williams & Wilkins

The Journal of Nervous and Mental Disease • Volume 198, Number 2, February 2010 Representations in Panic Disorder

and CL scores serving as dependent variables. Since the MANOVA
result was significant (Wilks’ lamba (4, 45) ⫽ 3.04, p ⫽ 0.027),
follow-up ANOVA tests were performed and are shown in Table 2.
The Benevolence-Malevolence factor score was significantly differ-
ent between the groups (p ⫽ 0.01) with the CL score showing a
trend toward significance (p ⫽ 0.07). The effect size for group
differences for the Benevolence-Malevolence factor was in the
medium-to-large effect range (d ⫽ 0.73); for the CL factor, it was
in the medium effect range (d ⫽ 0.52); and for the Ambivalence
score, it was in the small-to-moderate effect range (d ⫽ 0.44)
according to Cohen’s (1988) criteria. Table 2 also reports the
F-values, significance levels and effect sizes for the individual
Benevolence-Malevolence and Punitive characteristics that make
up each of the factor scores.
With regard to the physical health status of the groups,
women with PD had lower health status scores than women
without PD, (M ⫽ 44.79, SD ⫽ 19.32 vs. M ⫽ 55.57, SD ⫽
18.26, F关1, 48兴 ⫽ 4.12, p ⫽ 0.048).
Since 13 of the 25 patients with PD also met criteria for
MDD, a second (exploratory) MANOVA was run to assess whether
the maternal factor scores, Ambivalence, and CL ratings were more
pathological in the PD group with MDD as compared with the
PD-only group. Results indicated that there were no significant
differences between the PD with MDD group and the PD-only group
on any of the dependent variables: Wilks’ lamba (4, 20) ⫽ 0.70,
p ⫽ 0.60.
DISCUSSION
This study assessed the maternal representations of a sample
of urban-dwelling African-American women with and without PD.
Drawing on previous studies using interview (Shear et al., 1993) and
self-report methodologies (Faravelli et al., 1991; Pacchierotti et al.,
2002; Silove, 1986; Wilborg and Dahl, 1997; Shear et al., 1993;
Silove et al., 1991), we assessed the mental representations via
spontaneous descriptions of mothers and coded them using the
Qualitative and Structural Dimensions of Object Representations
scale (Blatt et al., 1992). This coding system not only allowed us to
assess caring and punitive/controlling representations but also am-
bivalent feelings toward mother/caretakers as well as a structural
scale of mental representations (i.e., an overall level of psycholog-
ical maturity).
Our first hypothesis was strongly supported. Maternal Benev-
olence-Malevolence factor scores were significantly lower (i.e., less
caring) in women with PD than in the women without PD. Women
with PD experience their mother in adult life as less affectionate,
warm or nurturing, less constructively involved, and more malevo-
lent. These findings are consistent with studies using the Parental
Bonding Instrument that showed that mothers and fathers of men
and women with PD were experienced as less caring (Faravelli et al.,
1991; Pacchierotti et al., 2002; Silove, 1986; Wilborg and Dahl,
1997; Silove et al., 1991). The findings also support the psychody-
namic theory of panic. These negative maternal characteristics can
exacerbate anger and ambivalence and thus contribute to the cycle of
anger (toward mother) 3 separation fears 3 greater anger toward
mother, etc.
Our second hypothesis was not supported. No differences
were found between women with and without PD on the maternal
Punitive factor. The lack of differences between the groups may
be due in part to a characteristic of single-parent families.
Mothers who are caring for children without the presence and
support of a partner may resort to being more punitive and
judgmental as a means of discipline. However, a punitive-critical
stance may also be a function of maternal distress. This seems
consistent with the literature, in which an association has been
found between low socio-economic status and distress (Mi-
rowsky and Ross, 1989) as well as increased distress in single-
parent mothers versus partnered mothers, especially those who
are financially challenged (Butterworth, 2004; Crosier et al.,
2007; Wang, 2004).
The third hypothesis was partially supported. Although ma-
ternal ambivalence was not significantly different between the PD
and matched comparison groups, the effect size approached a
moderate level, d ⫽ 0.44, suggesting that women in the PD group
tended to have higher levels of ambivalence toward their mothers as
opposed to the women without PD. A moderate level of ambivalence
is not surprising in women with PD, especially when one takes into

TABLE 2. Means, Standard Deviations, F-Values, and Effect Sizes of Maternal Representations Scores in
Women With and Without Panic Disorder
Panic Disorder Controls
N ⴝ 25 N ⴝ 25
M (SD) M (SD) df F p Cohen’s d
Mental representations
Benevolence/malevolence factor 3.65 (1.58) 4.67 (1.19) 1, 48 6.60 0.013 0.73*
Punitive factor 3.72 (1.55) 3.54 (1.43) 1, 48 0.09 0.771 0.08
Ambivalence 3.34 (1.56) 2.64 (1.63) 1, 48 2.41 0.127 0.44
Conceptual level 4.16 (1.31) 4.92 (1.60) 1, 48 3.36 0.073 0.52*
Maternal characteristics
Affectionate 3.62 (1.41) 4.76 (1.36) 1, 48 8.46 0.005 0.82**
Benevolence-malevolence 3.92 (1.65) 4.84 (1.27) 1, 48 5.39 0.025 0.62*
Warm-cold 4.04 (1.48) 4.92 (1.32) 1, 48 4.93 0.031 0.63*
Constructive involvement 3.44 (1.72) 4.46 (1.24) 1, 48 5.80 0.020 0.68*
Nurturant 3.22 (1.89) 4.56 (1.41) 1, 48 8.06 0.007 0.80**
Positive-negative ideal 3.66 (1.83) 4.46 (1.46) 1, 48 2.92 0.094 0.48
Punitive 3.80 (1.55) 3.54 (1.43) 1, 48 0.38 0.541 0.17
Judgemental 3.80 (1.26) 3.62 (1.30) 1, 48 0.25 0.621 0.14
Effect size (Cohen’s d): small ⫽ 0.20, medium ⫽ 0.50, and large ⫽ 0.80 (Cohen, 1988).
*Moderate effect size.
**Large effect size.

© 2010 Lippincott Williams & Wilkins www.jonmd.com | 147

Porcerelli et al. The Journal of Nervous and Mental Disease • Volume 198, Number 2, February 2010
account that these women described their mothers as cold, affec-
tionless, and malevolent.
The fourth hypothesis was also partially supported. Differ-
ences between women with PD and those without PD on the
Conceptual Level factor showed a trend toward significance (p ⫽
0.07) and reached a moderate effect size (d ⫽ 0.52). This suggests
a deficit in representation development in women with PD. Women
in the PD group, on average, displayed representations that were less
cognitively and affectively complex (i.e., more literal, global and
concrete), with an emphasis on physical part-properties of the
mother, than women without PD. Part-properties in the psychoana-
lytic literature suggests that others are represented in need-gratifying
ways without an integration of mutual aspects of relatedness. The
average representation level of the women without PD also focused
on part-properties of the mother but the descriptions were more
specific (versus nonspecific and global), suggesting a greater appre-
ciation of the mother’s characteristics beyond just the gratification
that she can provide. Producing representations that are less com-
plex, rich, and specific may leave individuals vulnerable to feelings
of anger and frustration which can contribute to fears of (real or
imagined) attachment disruptions that can contribute to the onset of
anxiety and panic.
The fifth hypothesis was supported. Women in the PD group
reported a lower level of physical functioning than the women
without PD. These findings support the findings of Zaubler and
Katon (1996) as well as more recent studies by Marshall et al. (2008)
and Kinley et al. (2009).
Panic attacks are associated with overwhelming fear and
anxiety suggesting an inability to self-soothe or bring to mind any
representation that reduces distress and suffering. In this sense, it is
not surprising that less mature, maternal representations are associ-
ated with PD, given the vital role that many mothers play in helping
the child develop the capacity to self-regulate (Fonagy et al., 2002;
Huprich, 2009; Schore, 2002), develop a sense of safety in the world
around them, (Bartholomew et al., 2001; Sullivan, 1953; Winnicott,
1965), and to develop a coherent sense of self that is not susceptible
to fragmentation (Kohut, 1971; Kohut and Wolf, 1978). Although
fathers and other caregivers help to play an important role in these
psychological achievements, in our sample of women, the presence
of PD was associated with less benevolent and developmentally
mature maternal representations as well as higher levels of ambiv-
alence toward their mothers.
The following is description of a panic attack from a young
African-American woman who meets the criteria for PD and was
seen in the primary care clinic. The interaction illustrates several of
the psychodynamic features of PD–namely, conflicts over autonomy
and anger, and several of the negative maternal representations
reported in this study.
Sarah, a 24-year-old African-American woman, lives with her
mother and 6-month-old daughter. She complained to her family
physician about her anxiety and recent panic attack. She stated that
the felt “claustrophobic” one evening after going to bed. Her heart
began to pound, and she had difficulty breathing. “I thought I was
dying, having a heart attack!” When asked about possible precipi-
tating events from that day, Sarah said she had a “bad fight” with her
mother that evening about disagreements over parenting her infant
daughter. Sarah stated that her mother is “always involved in my
business . . . telling me I don’t take care of my baby right. She never
lets me have my own life . . .. even after I turned 18 . . . even after
I had my baby.” Sarah later added that if she does not take her
mother’s advice, she “makes you feel like you really hurt her.”
One surprising finding is the lack of differences in represen-
tations scores between the women meeting criteria for both PD and
major depression and the women diagnosed with only PD. While it
148 | www.jonmd.com
is known that MDD is associated with problematic parental repre-
sentations (Blatt et al., 1979), it does not appears that MDD and PD
additively are associated with poorer maternal representations. How-
ever, this interpretation is offered tentatively awaiting additional
research comparing these groups involving larger sample sizes.
Our study has limitations. First, though the homogenous
sample of African-American women provides more opportunities to
generalize to African-American women as a group, there clearly are
concerns about how generalizable the results are to a more diverse
sample. Second, the measure used to assess PD was a self-report
instrument. It is widely known that self-report measures are limited
in their reliability and validity, though Spitzer et al. (1999) did report
good sensitivity and specificity values for the PHQ in detecting PD.
Third, our study is a preliminary investigation of the association
between maternal representations and PD. There are many potential
mediating and moderating variables that could affect the associa-
tions found in this study. Future research on this topic should
consider third variables and other potential confounds to further
elucidate the way in which maternal representations are associated
with PD. Fourth, we had fairly modest sample sizes, which likely
contributed to the lack of statistical significance on some of the
analyses. Our effect sizes suggest that increased sample sizes would
likely affect significance. Lastly, further research should examine
maternal and paternal representations across both culture and gender
in PD patients to determine further if the same patterns exist
cross-culturally and in both sexes.
CONCLUSIONS
Maternal representations of urban-dwelling African-Ameri-
can women with PD, as assessed with the Qualitative and Structural
Dimensions of Object Representations scale, exhibit lower levels of
parental care, higher levels of ambivalence, and structurally less
mature representations than women without PD. Our study did not
support higher levels of punitiveness in representations of women
with PD. With regard to the finding of lower parental care, our
results are consistent with previous studies that used a different
method of assessment (self-report) and ethnically-different samples.
REFERENCES
American Psychiatric Association (1994) Diagnostic and Statistical Manual of
Mental Disorders (4th ed). Washington (DC): American Psychiatric Associa-
tion.
Arrindell WA, Emmelkamp PM, Monsma A, Brilman E (1983) The role of
perceived parental rearing practices in the aetiology of phobic disorders: A
controlled study. Br J Psychiatry. 143:183–187.
Bartholomew K, Kwong MH, Hart SD (2001) Attachment. In WJ Livesley, (Ed),
Handbook of Personality Disorders. (pp 197–227). New York (NY): Guilford.
Blatt SJ (1974) Levels of object representations in anaclitic and introjective
depression. Psychoanal Study Child. 29:107–157.
Blatt SJ (2004) Experiences of depression: Theoretical, clinical and research
perspectives. Washington (DC): American Psychological Association.
Blatt SJ, Chevron ES, Quinlan DM, Schaffer CE, Wein S (1992) The Assessment
of Qualitative and Structural Dimensions of Object Representations. Unpub-
lished manual. New Haven (CT): Yale University.
Blatt SJ, Stayner DA, Auerbach JS, Bahrends RS (1996) Change in object and self
representations in long-term, intensive, inpatient treatment of seriously dis-
turbed adolescents and young adults. Psychiatry. 59:82–107.
Blatt SJ, Wein SJ, Chevron ES, Quinlan DM (1979) Parental representations and
depression in normal young adults. J Abnorm Psychol. 88:388 –397.
Blatt SJ, Wiseman H, Prince-Gibson E, Gatt C (1991) Object representations and
change in clinical functioning. Psychotherapy. 28:273–283.
Bornstein RF, Galley DJ, Leone DR (1986) Parental representations and orality.
J Pers Assess. 50:80 – 89.
Bornstein RF, O’Neill RM (1992) Parental perceptions and psychopathology.
J Nerv Ment Dis. 180:475– 483.
Busch FN, Cooper AM, Klerman GL, Penzer RJ, Shapiro T, Shear MK (1991)
© 2010 Lippincott Williams & Wilkins
The Journal of Nervous and Mental Disease • Volume 198, Number 2, February 2010
Neurophysiological, cognitive-behavioral and psychodynamic approaches to
panic disorder: Toward an integration. Psychoanal Inq. 11:316 –332.
Busch FN, Milrod BL, Sandberg LS (2009) A study demonstrating efficacy of
psychoanalytic psychotherapy for panic disorder: Implication for psychoana-
lytic research, theory and practice. J Am Psychoanal Assoc. 57:131–148.
Butterworth P (2004) The prevalence of psychiatric disorders among lone moth-
ers: Association with physical and sexual violence. Br J Psychiatry. 184:21–27.
Cohen J (1988) Statistical Power Analysis for the Behavioral Sciences. (2nd ed).
Hillsdale (NY): Erlbaum Press.
Crosier T, Butterworth P, Rodgers B (2007) Mental health problems among single
and partnered mothers: The role of financial hardship and social support. Soc
Psychiatry Psychiatr Epidemiol. 42:6 –13.
Faravelli C, Panichi C, Pallanti S, Paterniti S, Grecu LM (1991) Perception of
early parenting in panic and agoraphobia. Acta Psychiatr Scand. 84:6 – 8.
Fonagy P, Gergely G, Jurist E, Target M (2002) Affect Regulation, Mentalization
and the Development of the Self. New York (NY): Other Press.
Friedman S, Paradis CM, Hatch M (1995) Characteristics of African American
and White patients with panic disorder and agoraphobia. Hosp Community
Psychiatry. 45:798 – 803.
Goisman RM, Warshaw MG, Peterson LG, Rogers MP, Cuneo P, Hunt MF,
Tomlin-Albanese JM, Kazin A, Gollan JK, Epstein Kaye T, Reich JH, Keller
MB, et al. (1994) Panic, agoraphobia and panic disorder with agoraphobia. Data
from a multicenter anxiety disorders study. J Nerv Ment Dis. 182:72–79.
Horwath E, Johnson J, Hornig CD (1993) Epidemiology of panic disorder in
African Americans. Am J Psychiatry. 150:465– 469.
Huprich SK (2009) Psychodynamic Therapy: Conceptual and Empirical Foun-
dations. New York (NY): Taylor and Francis.
Huprich SK, Greeberg RP (2003) Advances in the assessment of object relations
in the 1990s. Clin Psychol Rev. 23:665– 698.
Kagan J, Reznick JS, Snidman N, Johnson MO, Gibbons G, Gerston M, Bieder-
man J, Rosenbaum JF (1990) Origins of panic disorder. In JC Ballanger, (Ed),
Neurobiology of Panic Disorder. (pp 71– 87). New York (NY): Allan R. Liss.
Kinley DJ, Cox BJ, Clara I, Goodwin RD, Sareen J (2009) Panic attacks and their
relation to psychological and physical functioning in Canadians: Results from
a nationally representative sample. Can J Psychiatry. 54:113–122.
Kohut H (1971) The Analysis of the Self. New York (NY): International Univer-
sities Press.
Kohut H, Wolfe ES (1978) The disorders of the self and their treatment: An
outline. Int J Psychoanal. 59:413– 425.
Levy KN, Blatt SJ, Shaver PR (1998) Attachment styles and parental represen-
tations. J Pers Soc Psychol. 74:407– 419.
Marshall EC, Zvolensky MJ, Sachs-Ericsson N, Schmidt NB, Bernstein A (2008)
Panic attacks and physical health problems in a representative sample: Singular
and interactive associations with psychological problems and interpersonal and
physical disability. J Anxiety Disord. 22:78 – 87.
Mirowsky J, Ross CE (1989) Social Causes of Psychological Distress. (2nd ed).
New York (NY): Aldine de Gruyter.
Pacchierotti C, Bossini F, Castrogiovanni A, Pierachinni L, Soreca I,
Castrogiavinnin P (2002) Attachment and panic disorder. Psychopathology.
35:347–354.
© 2010 Lippincott Williams & Wilkins
Representations in Panic Disorder
Parker G, Tupling H, Brown LB (1979) A parental bonding instrument. Br J Psy-
chiatry. 134:138 –147.
Porcerelli JH, Huprich SK, Binienda J, Karana D (2006) Object representations
and their relationship to psychopathology and physical health status in African
American women in primary care. J Nerv Ment Dis. 194:838 – 844.
Quinlan DM, Blatt SJ, Chevron ES, Wein SJ (1992) The analysis of descriptions
of parents: Identification of a more differentiated factor structure. J Pers Assess.
59:340 –351.
Sanderson WC, Di Nardo PA, Rapee RH, Barlow DH (1990) Syndrome comor-
bidity in patients diagnosed with a DSM-III-R anxiety disorder. J Abnorm
Psychol. 90:308 –312.
Schore AN (2002) Advances in neuropsychoanalysis, attachment theory and
trauma research: Implications for self psychology. Psychoanal Inq. 22:433–
484.
Shear MK, Cooper AM, Klerman GL, Busch FN, Shapiro T (1993) A psychody-
namic model of panic disorder. Am J Psychiatry. 150:850 – 866.
Shrout PE, Fleiss JL (1979) Intraclass correlation: Uses in assessing rater reli-
ability. Psychol Bull. 86:420 – 428.
Silove D (1986) Perceived parental characteristics and reports of early parental
deprivation in agoraphobic patients. Aust N Z J Psychiatry. 20:365–369.
Silove D, Parker G, Hadzi-Pavlovic D, Manicavasagar V, Blaszczynski A (1991)
Parental representations of patients with panic disorder and generalized anxiety
disorder. Br J Psychiatry. 159:835– 841.
Spitzer RL, Kroenke K, Williams JBW (1999) Validation and utility of a
self-report version of PRIME-MD. JAMA. 282:1737–1744.
Spitzer RL, Williams JB, Kroenke K, Hornyak R, McMurray J (2000) Validity
and utility of the PRIME-MD patient health questionnaire in assessment of
3000 obstetric-gynecologic patients: The PRIME-MD patient health question-
naire obstetrics-gynecologic study. Am J Obstet Gynecol. 183:759 –769.
Stewart AL, Greenfield S, Hays RD, Wells R, Rogers WH, Berry SD, McGlynn
EA, Ware JE (1989) Functional status and well-being of patients with chronic
conditions: Results from the medical outcomes study. JAMA. 262:907–913.
Stewart AL, Hays RD, Ware JE (1988) The MOS short-form general health
survey: Reliability and validity in a patient population. Med Care. 26:724 –732.
Stein MD, Mulvey KP, Plough A, Samet JH (1998) The functioning and
well-being of persons who seek treatment for drug and alcohol use. J Subst
Abuse. 10:75– 84.
Sullivan HS (1953) The Interpersonal Theory of Psychiatry. New York (NY):
Norton.
Tucker WI (1956) Diagnosis and treatment of the phobic reaction. Am J Psychi-
atry. 112:825– 830.
Wang JL (2004) The difference between single and married mothers in the
12-month prevalence of major depressive syndrome, associated factors and
mental health service utilization. Soc Psychiatry Psychiatr Epidemiol.
39:26 –32.
Wilborg IM, Dahl AA (1997) The recollection of parental rearing styles in
patients with panic disorder. Acta Psychiatr Scand. 96:58 – 63.
Winnicott DW (1965) The Maturational Process and the Facilitating Environ-
ment. New York (NY): International Universities Press.
Zaubler TS, Katon W (1996) Panic disorder and medical comorbidity: A review
of the medical and psychiatric literature. Bull Menninger Clin. 60(2 suppl
A):12–38.
www.jonmd.com | 149