November 2011  << Back  

  Can't open the PDF? Click here for help.



Factor structure of manic symptoms in adolescents

Subhash Chandra Gupta, MD, DPM

Department of Psychiatry, Peninsula College of Medicine, and Dentistry, Teignmouth, United Kingdom

Vinod Kumar Sinha, MD, DPM

Department of Psychiatry, Kasturba Medical College, Manipal, Karnataka, India

Samir Kumar Praharaj, MD, DPM

Department of Psychiatry, Kasturba Medical College, Manipal, Karnataka, India

Sachin Gandotra, MD, DPM

Consultant Psychiatrist, Mental Health Foundation, St. Maarten, Netherlands

OBJECTIVE: To identify the factor structure of manic symptoms in adolescents as assessed by the Scale for Manic States (SMS).

METHOD: Pattern of symptoms was assessed in a group of 100 adolescents with a diagnosis of manic episode as defined by the International Classification of Diseases, 10th revision - Diagnostic Criteria for Research. A principal component analysis of the broad range of psychiatric symptoms covered by the SMS was conducted.

RESULTS: Seven eigenvalues were greater than unity, and parallel analysis revealed 5 factors, whereas scree plot was inconclusive. Five-factor solution as obtained by parallel analysis was chosen, which described our data appropriately and were clinically relevant. The 5 factors were: aggressive overactivity, dysphoria, psychosis, hedonia, and thought retardation. These captured 58.14% of the total variance.

CONCLUSIONS: These 5 factors explain the clinical dimensions in adolescent mania similar to those of the adult population. Nevertheless, certain features, such as presence of psychosis along with euphoric mood and thought retardation, distinguish adolescent from adult mania.

KEYWORDS: bipolar disorder, mania, adolescents, factor analysis, principal components



Factor analysis of manic symptoms may reveal new dimensions of mania and assist in the classification of manic symptoms by identifying subgroups. Early factor analytic studies of manic symptoms were conducted by Murphy and Biegel,1 based on analysis of only 30 patients. The first factor consisted of loading from the “core” symptoms of mania. The second factor, bipolar, included a large positive loading from items identified as “paranoid-destructive” and a large negative loading from items identified as “elation-grandiose.” On the basis of this, Murphy and Biegel1 proposed 2 manic subtypes: paranoid-destructive and euphoric-grandiose.

The generation of interpretable factors requires appropriate variables be measured in the original observations. Comrey2 emphasized variables should provide representative measures of each expected factor, and omitting key variables obscures the complete factor solution. In this regard, although the Manic State Rating Scale (MSRS)3 included depressed mood, it did not include many clinically suggestive factors of mixed mania, such as lability and anxiety. Double4 examined the factorial structures of the Young Mania Rating Scale (YMRS)5 and MSRS on 81 manic patients and found a solution of 3 factors for both scales. The factor structure of the YMRS revealed a psychotic or thought disturbance factor, an overactive and aggressive factor, and a hedonism factor, but an absence of depression. In order to overcome this problem, some studies6-8 have included both mania and depression rating scales in their factor analytic studies. Cassidy et al9 overcame these shortcomings by developing their own tool, the Scale for Manic States (SMS), which identified 5 factors. The first and strongest factor represented dysphoria, with strong positive loadings for depressed mood, lability, guilt, anxiety, and suicidal thoughts and behaviors as well as a strong negative loading for euphoric mood. The other factors were psychomotor pressure, psychosis, increased hedonic functions, and irritable aggression. Since then, several factor analytic studies on manic symptoms have been conducted in the adult population.10 Gupta et al10 utilized the SMS to measure manic symptoms in adults, and reported 6 factors that were clinically relevant: psychosis, irritability/aggression, dysphoria, accelerated thought stream, hedonia, and hyperactivity.

Interestingly, none of the abovementioned studies addressed the specific issue of mania factors in the adolescent population. Carlson and Kashani11 and Lewinsohn et al12 suggested bipolar disorder is at least as prevalent among adolescents as it is among adults. Various researchers have conducted descriptive studies on the symptomatology of mania in childhood and adolescence.13 The latter is usually rather similar to that observed in adults, although psychotic features may be more common.14 The concurrence of mania and psychosis in children has been recognized for many years, although misdiagnosis of schizophrenia was, and remains, tentatively common.15 Manic symptoms are reported to vary with age, and older children typically present with euphoria, grandiosity, or paranoid ideation in addition to flight of ideas.16 In addition, the presence of pressured speech, overactivity, and distractibility was noted in both younger and older children. Further, mania during childhood and adolescence may also be confused with attention-deficit/hyperactivity disorder, oppositional defiant disorder, conduct disorder, and substance-related disorders.13 Hsu and Starzynski17 reported the classic manic picture is rarely seen in teenagers; rather, they often present with irritability, argumentative behavior, rapid mood oscillations, anger, annoyance, and aggressive grandiosity. Others have reported dysphoric/mixed mania as a frequent presentation in adolescence.18,19 In a cluster analytic study20 of the phenomenology of childhood and adolescent mania, 2 clusters were extracted: patients with cluster 1 were cheerful and energetic, showing more flight of ideas and less psychotic features; patients with cluster 2 were more hyperactive, and often presented with hallucinations and/or delusions as well as sexual disinhibition and reckless behavior.

There have been few studies that have examined the factor structure on symptoms of mania in adolescence. Youngstrom et al21 compared the factor structure of the YMRS between children and adolescents. They found a unifactorial structure equivalent across younger and older children. A similar unifactorial solution was reported by Frazier et al22 in the adolescent population for mania using the YMRS and the Kiddie Schedule for Affective Disorders and Schizophrenia Mania Rating Scale; a smaller second factor emerged for depression using the Children’s Depression Rating Scale-Revised. Pavuluri et al23 also reported a single large factor from the Child Mania Rating Scale in a study of 150 children with a mean age of 10.3 (SD 2.9) years. Although they had found 4 factors with eigenvalues >1 during the exploratory principal component analysis (PCA), a single-factor model fit well using confirmatory factor analysis. None of these studies addressed the issue of depressive symptoms in adolescent patients with mania, although dysphoric symptoms have been reported to occur in adults with mania. Therefore, we conducted an exploratory factor analysis of manic symptoms in adolescents using the SMS,9 which captures both the depressive symptoms and manic symptoms. Our aim was to explore the underlying factors of adolescent manic symptoms and determine whether they are similar or different than those obtained in the adult population.


This study was conducted at the Central Institute of Psychiatry, a psychiatry tertiary referral center in Ranchi, India. The study was approved by the institutional ethical committee. The study sample consisted of 100 adolescent patients (age 12 to 19) fulfilling the criteria for a manic episode according to the International Classification of Disease, 10th revision - Diagnostic Criteria for Research criteria.24 Patients were excluded if they had received a diagnosis of mixed affective episode. This is because the conceptualization of mixed mania is inconsistent due to definition variations regarding the required temporal relationship between manic and depressive symptoms.25 Written informed consent was obtained from all participants in the study.


Comprehensive Assessment of Symptoms and History26 was used to interview the study sample and gather information on 3 main sections: present state, past history of psychotic or affective illness, and lifetime history. The SMS9 a 20-item scale with good inter-rater and test-retest reliability, was used to measure the severity of mania.9 Additional information from clinical records and staff observations were incorporated into the ratings.

Factor analysis

Data were analyzed using Statistical Package for Social Sciences version 10.0 (SPSS Inc., Chicago, IL). Kaiser-Meyer-Olkin measure of sampling adequacy27 was .58 and Barlett’s test of sphericity was significant (P < .001). Exploratory factor analysis was performed to determine factors within the sample. The item “guilt” was excluded from factor analysis because the entire sample gave a rating of “0” on this item. Three criteria for retaining the number of components were considered: Kaiser’s criterion to retain eigenvalues greater than unity,27 Cattell’s scree plot inspection for the point of inflexion,28 and Horn’s parallel analysis.29 Parallel analysis was performed with criterion values replicating a randomly generated data set obtained using the regression equations developed by Keeling.30 Only factor loadings with an absolute value >.4 was interpreted.31 Orthogonal rotation (varimax rotation with Kaiser normalization) was initially done, and revealed dual loading >.4 on more than 1 factor. Oblique rotation (oblimin rotation with Kaiser normalization, delta at 0) was done, and revealed a more interpretable solution. The pattern matrix was interpreted, as it had no complex variables and a simpler structure.


The sample characteristics have been summarized in TABLE 1. There were 100 patients, comprised of 56 males and 44 females. The mean age of the sample was 16.9 (SD 1.87) and the mean total SMS score was 34.59 (SD 6.87). There was a family history of affective illness in 43 patients, of non-affective psychosis in 9 patients, and substance dependence in 5 patients. Sixty-two patients were in their first manic episode.


Sample characteristics (N = 100)

  Mean SD
Age 16.90 1.87
Education years 7.71 2.90
Past episodes (total) 1.61 1.05
Manic episodes 1.11 1.03
Depressive episodes 0.50 0.86
SMS total score 34.59 8.29
  Motor activity 3.60 1.22
  Decreased sleep 3.67 1.06
  Pressured speech 2.97 1.14
  Racing thoughts/disturbed concentration 2.05 1.40
  Mood lability 0.77 1.20
  Euphoric mood 2.69 1.13
  Depressed mood 0.36 0.81
  Suicide 0.15 0.54
  Psychosis 3.13 1.45
  Paranoia 1.12 1.49
  Grandiosity 2.79 1.59
  Lack of insight 3.05 0.82
  Increased contact 1.76 1.21
  Increased sexuality 0.96 1.30
  Humor 0.37 0.53
  Anxiety 0.71 1.02
  Irritability 2.66 1.08
  Aggression 1.55 1.09
  Dress 0.23 0.53
Entire population scored “0” on item “guilt,” which has been omitted.
SD: standard deviation; SMS: Scale for Manic State.
Principal component analysis

The PCA resulted in an initial 7 factors with eigenvalues greater than unity.27 The 7 factors captured 70.19% of the variance. The scree plot was not useful as there was no clear inflexion after the first factor (FIGURE 1). Parallel analysis using Keeling’s regression equation30 showed only the first 5 eigenvalues exceeded the criterion values for a randomly generated data matrix of the same size (100 patients × 19 items) (TABLE 2). The 5 factors accounted for 58.14% of the variance. This 5-factor solution appeared clinically most appropriate for our data, and none of the items loaded on >2 components. TABLE 3 shows the rotated (oblimin: delta = 0) component matrix for a 5-factor solution with factor loadings >.4.31

FIGURE 1: Cattell’s scree plot of extracted components


Comparison of eigenvalues from PCA and the corresponding criterion values obtained from parallel analysis

Component number Actual eigenvalue from PCA Criterion value from parallel analysisa Decision
1 3.81 1.85 Accept
2 2.32 1.73 Accept
3 1.93 1.61 Accept
4 1.66 1.50 Accept
5 1.41 1.39 Accept
6 1.20 129 Reject
aParallel analysis using Keeling’s regression equation.
PCA: principal components analysis.

Table 3

Factor structure of SMS (PCA with oblimin rotation) (N = 100)

Manic symptoms Factor loading (>.4)
Factor 1
Aggressive overactivity
Factor 2
Factor 3
Factor 4
Factor 5
Thought retardation
Aggression .866        
Irritability .859        
Motor activity .596        
Lack of insight .560        
Depressed mood   .852      
Mood lability   .839      
Anxiety   .671      
Psychosis     .885    
Grandiosity     .825    
Euphoric mood     .634    
Paranoia     .442    
Dress       .687  
Increased sexuality       .673  
Increased contact       .451  
Racing thought/disturbed concentration         -.939
Pressured speech         -.882
Eigenvalues 3.81 2.32 1.93 1.66 1.41
Variance %a 20.07 11.75 10.14 8.76 7.42
aUnrotated (when components are correlated, sums of squared loadings cannot be added to obtain a total variance).
PCA: principal component analysis; SMS: Scale for Manic States.

Factor 1 was the strongest and had the highest positive loadings from aggression, irritability, motor activity, and lack of insight, and represents aggressive overactivity. Factor 2 included depressed mood, mood lability, and anxiety, which seemingly represented dysphoria. Factor 3 included psychosis, grandiosity, euphoric mood, and paranoia, and appeared to represent psychosis. Factor 4 included dress, increased sexuality, and increased contact, which seemed to represent hedonia. Factor 5 had negative loadings from racing thoughts/disturbed concentrations and pressured speech, and appeared to represent thought retardation. FIGURE 2 shows the distribution of factor scores. The Kolmogorov-Smirnov test (with Lilliefors significance correction) showed factors 1 and 3 to have normal distribution and the rest to have non-normal distributions (P < .05).

FIGURE 2: Distribution of factor scores


The PCA of manic symptoms using the SMS revealed 5 to 7 factors depending on the method of retaining factors employed. It has been demonstrated that Kaiser’s criterion tends to overestimate the number of components, and parallel analysis has been found to be conservative for estimating the number of factors.32 In our study, 5 factors obtained through parallel analysis appear to be clinically most relevant and described various aspects of manic syndrome consistent with the adult literature. The rotated factors accounted for a high proportion of the variance (58.14%) and had a simple structure. Our findings are in contrast to other studies21-23 that have reported unifactorial solutions of manic symptoms in children. It is possible mania in adolescents presents similar to mania in adults, as seen in studies examining depression in adolescents,33 whereas studies in younger children show a different symptom profile than that of adolescents. More plausible reasons could be the factors derived reflect the variables included in the analysis and SMS9 contains several items covering depressive symptoms.

The first factor obtained in this study represented aggressive overactivity, similar to the second factor of Double,4 which includes increased motor activity, irritability, and disruptive-aggressive behavior. This factor might reflect the paranoid-destructive factor of Murphy and Beigel.1 This is in contrast to a previous study by Cassidy et al,9 in which irritability, aggression, and lack of insight had loaded on a factor separate from psychomotor acceleration.

Factor 2 represented a dysphoric mood factor similar to the first factor of Cassidy et al.9 This factor has positive loadings from depressed mood, mood lability, and anxiety, which are symptoms that have been hypothesized to characterize dysphoric mania.6 In contrast to bimodal distribution of this factor in adults,9 which suggests 2 distinct subtypes of mania (euphoric and dysphoric), adolescent mania was predominantly characterized by dysphoric mood consistent with the previous literature.18,19 Because patients with a clinical diagnosis of mixed episode were excluded from the study, this finding suggests dysphoria features are present even in pure manic episodes in adolescents. This factor was evident in studies by Rossi et al,7 González-Pinto et al,8 and Cassidy et al,9 but not in studies by Murphy and Biegel1 and Double,4 as the latter did not include depressive symptoms in their assessment scales. The current evidence suggests dysphoric features should be included in the clinical diagnosis of adolescent mania.

Factor 3 represented psychosis, with the highest loadings from psychosis, grandiosity, euphoric mood, and paranoia. This factor was similar to factor 3 of Cassidy et al,9 which used the SMS; factor 4 of Dilsaver et al,34 which consisted of delusions, suspiciousness, and irritability; and factor 1 of Double,4 which included insight, language thought disorder, and items from the YMRS. This factor does not have a clear counterpart in the analysis of Murphy and Biegel1 and Rossi et al,7 possibly because the scale they used did not include defined items for psychosis. In our sample, euphoric mood had positive loadings with this factor, which suggests in adolescence, psychotic features and euphoric mood appear together and are inherent with manic symptoms. In contrast, factor analysis of mania in adults separates psychosis as an independent factor from euphoric mood, suggesting 2 subtypes of mania (with or without psychotic symptoms).9,10

Factor 4 represented hedonia, which is identical to factor 4 of Cassidy et al,9 although grandiosity and euphoric mood did not appear to be related to this factor in our study. This factor is reminiscent of the euphoric-grandiose symptom cluster observed by Beigel and Murphy1 and factor 3 of Double.4 As suggested by Cassidy et al,9 dress, increased contact, and hypersexuality can be conceptualized as alternative expressions of elevated mood and self-esteem.

Factor 5 represented thought retardation, with high negative loadings from racing thoughts/disturbed concentrations and pressured speech. Increased motor/psychomotor activity as a cardinal symptom of mania has been emphasized by various early researchers.35 While Cassidy et al9 found a factor of combined psychic and motor activity, these symptoms were delineated on 2 separate factors (ie, factors 5 and 1) in our study. The separation of these 2 components also provides evidence of different subtypes of mania, in which psychic and motor activity may be dissociated. The prominence of thought retardation in adolescent mania (in contrast to thought acceleration in adult studies) highlights the prominence of mixed-like episodes in these patients. Although changes in mood are the hallmark of mania in the current nosological criteria, they are little supported in factor analytic studies. The relative weight of psychomotor activity level as opposed to mood in diagnosis of mania may need reconsideration.


In this factor analysis, the obtained factors for clinical dimensions of mania in adolescents were similar to that of adult manic patients, in contrast to previous studies that have found single-factor solutions.21-23 Nevertheless, there are certain differences, such as presence of psychosis along with euphoric mood and thought retardation, that distinguish adolescent mania from that of the adult population. Dysphoria appeared as a separate factor in our sample, which is consistent with the recent studies in adults. This again substantiates the observation that non-inclusion of depressive symptoms in previous studies might have led to exclusion of this factor. The factors generated in this study describe useful research and clinical dimensions for the study of mania. In addition to mood, they capture psychosis, hedonia, and psychic and motor activity.

One limitation of the study is small sample size, ie, the number of cases was >10 times the number of items in the scale used for the study as recommended by Streiner.36 Nevertheless, these factors allow the characterization of individual patients on these dimensions, which may extend our ability to study the relationship between mania features and biologic markers, prognosis, and treatment response. Further studies are warranted to clarify the underlying factors in adolescent mania through exploratory and confirmatory factor analyses.

DISCLOSURES: Dr. Gupta is a speaker for AstraZeneca, Bristol-Myers Squibb, Pfizer Inc., Janssen-Cilag, and Eli Lilly and Company. Drs. Praharaj, Gandotra, and Sinha report no financial relationships with any company whose products are mentioned in this article or with manufacturers of competing products.


  1. Murphy DL, Biegel A. Depression elation, and lithium carbonate responses in manic patient subgroups. Arch Gen Psychiatry. 1974;31:643–648.
  2. Comrey AL. Common methodological problems in factors analytic studies. J Consult Clin Psychol. 1978;46:648–649.
  3. Beigel A, Murphy DL, Bunney WE Jr. The Manic State Rating Scale. Scale construction reliability, and validity. Arch Gen Psychiatry. 1971;25:256–262.
  4. Double DB. The factor structure of manic rating scales. J Affect Disord. 1990;18:113–119.
  5. Young RC, Biggs JT, Ziegler VE, et al. A rating scale for mania: reliability, sensitivity and validity. Br J Psychiatry. 1978;133:429–435.
  6. Post RM, Rubinow DR, Uhde TW, et al. Dysphoric mania. Clinical and biological correlates. Arch Gen Psychiatry. 1989;46:353–358.
  7. Rossi A, Daneluzzo E, Arduini L, et al. A factor analysis of signs and symptoms of the manic episode with Bech-Rafaelsen Mania and Melancholia Scales. J Affect Disord. 2001;64:267–270.
  8. González-Pinto A, Ballesteros J, Aldama A, et al. Principal components of mania. J Affect Disord. 2003;76:95–102.
  9. Cassidy F, Forest K, Murry E, et al. A factor analysis of the signs and symptoms of mania. Arch Gen Psychiatry. 1998;55:27–32.
  10. Gupta SC, Sinha VK, Praharaj SK, et al. Factor analysis of manic symptoms. Aust N Z J Psychiatry. 2009;43:1141–1146.
  11. Carlson GA, Kashani JH. Manic symptoms in a non-referred adolescent population. J Affect Disord. 1988;15:219–226.
  12. Lewinsohn PM, Klein DN, Seeley JR. Bipolar disorders in a community sample of older adolescents: prevalence phenomenology, comorbidity, and course. J Am Acad Child Adolesc Psychiatry. 1995;34:454–463.
  13. Geller B, Luby J. Child and adolescent bipolar disorder: a review of the past 10 years. J Am Acad Child Adolesc Psychiatry. 1997;36:1168–1176.
  14. Ballenger JC, Reus VI, Post RM. The “atypical” clinical picture of adolescent mania. Am J Psychiatry. 1982;139:602–606.
  15. Weller EB, Weller RA. Assessing depression in prepubertal children. Hillside J Clin Psychiatry. 1986;8:193–201.
  16. Carlson GA. Bipolar disorder in childhood and adolescence. In: Cantwell WP Carlson GA, eds. Affective disorders in childhood and adolescence: an update. New York, NY: SP Medical and Scientific Books; 1983.
  17. Hsu LK, Starzynski JM. Mania in adolescence. J Clin Psychiatry. 1986;47:596–599.
  18. Krasa NR, Tolbert HA. Adolescent bipolar disorder: a nine-year experience. J Affect Disord. 1994;30:175–184.
  19. Geller B, Sun K, Zimerman B, et al. Complex and rapid-cycling in bipolar children and adolescents: a preliminary study. J Affect Disord. 1995;34:259–268.
  20. Kothari S. Phenomenology of childhood and adolescent mania: a clinical study [MD in Psychiatry thesis]. Ranchi India: Ranchi University; 1998.
  21. Youngstrom EA, Danielson CK, Findling RL, et al. Factor structure of the Young Mania Rating Scale for use with youths ages 5 to 17 years. J Clin Child Adolesc Psychol. 2002;31:567–572.
  22. Frazier TW, Demeter CA, Youngstrom EA, et al. Evaluation and comparison of psychometric instruments for pediatric bipolar spectrum disorders in four age groups. J Child Adolesc Psychopharmacol. 2007;17:853–866.
  23. Pavuluri MN, Henry DB, Devineni B, et al. Child mania rating scale: development, reliability, and validity. J Am Acad Child Adolesc Psychiatry. 2006;45:550–560.
  24.  World Health Organization. The ICD-10 classification of mental and behavioral disorders: diagnostic criteria for research. Geneva Switzerland: World Health Organization; 1993.
  25. McElroy SL, Keck PE Jr, Pope HG Jr, et al. Clinical and research implications of the diagnosis of dysphoric or mixed mania or hypomania. Am J Psychiatry. 1992;149:1633–1644.
  26. Andreason NC. Comprehensive Assessment of symptoms and History (CASH). Iowa City IA: The University of Iowa; 1985.
  27. Kaiser HF. A second generation little jiffy. Psychometrika. 1970;35:401–415.
  28. Cattell RB. The scree test for number of factors. Multivariate Behavioral Research. 1966;1:245–276.
  29. Horn JL. A rationale and test for the number of factors in factor analysis. Psychometrika. 1965;30:179–185.
  30. Keeling KB. A regression equation for determining the dimensionality of data. Multivariate Behavioral Research. 2000;35:457–468.
  31. Stevens JP. Applied multivariate statistics for the social sciences. Hillsdale NJ: Lawrence Erlbaum Associates; 1992.
  32. Zwick WR, Velicer WF. Comparison of five rules for determining the number of components to retain. Psychol Bull. 1986;99:432–442.
  33. Guo Y, Nilsson ME, Heiligenstein J, et al. An exploratory factor analysis of the Children’s Depression Rating Scale-Revised. J Child Adolesc Psychopharmacol. 2006;16:482–491.
  34. Dilsaver SC, Chen YR, Shoaib AM, et al. Phenomenology of mania: evidence for distinct depressed, dysphoric, and euphoric presentations. Am J Psychiatry. 1999;156:426–430.
  35. Taylor MA, Abrams R. The phenomenology of mania. A new look at some old patients. Arch Gen Psychiatry. 1973;29:520–522.
  36. Streiner DL. Figuring out factors: the use and misuse of factor analysis. Can J Psychiatry. 1994;39:135–140.

CORRESPONDENCE: Samir Kumar Praharaj, MD, DPM, Assistant Professor, Department of Psychiatry, Kasturba Medical College, Manipal, Karnataka, India – 576104, E-MAIL: