Welcome to the Schizophrenia Resource Centre

Welcome, this website is intended for healthcare professionals in EMEA with an interest in the treatment of schizophrenia. By clicking the link below you are declaring and confirming that you are a healthcare professional

You are here

The negative syndrome of schizophrenia: Three -underlying components are better than two

Schizophrenia Research, Volume 166, Issue 1-3, August 2015, Pages 115 - 118

Summary

Aims

To analyse the underlying structure of the negative syndrome of schizophrenia as it is represented in the Brief Negative Symptom Scale.

Methods

Cross-sectional, multicentre study, employing data from 190 evaluations. Statistics: Exploratory factor analysis using the principal component analysis method.

Results

The three-component solution explained 77.4% of the total variance. Pearson correlation coefficients between components were: 1–2 = − 0.494, 1–3 = − 0.117, and 2–3 = 0.179.

Conclusion

Our solution favours a three-component structure of the negative syndrome, consisting of: external world (anhedonia and asociality), inner world (avolition and blunted affect), and alogia, with the latter only marginally related to the two former components.

Keywords: Negative syndrome, BNSS, Principal component analysis.

1. Introduction

The negative syndrome of schizophrenia, which includes social isolation, anhedonia, avolition, flat affect and alogia ( Kirkpatrick et al., 2006 .; Blanchard and Cohen, 2006 ), comprises about 20% of patients, occurs from the onset of the disorder, and lacks effective treatment. Together with cognitive symptoms, it has a significant impact on functioning, especially on the social domain (Bowie et al, 2006, Bowie et al, 2010, Harvey et al, 2012, and Rocca et al, 2014).

Previous studies on the factor structure of the negative syndrome, as it is represented in second-generation negative symptom scales ( Garcia-Portilla and Bobes, 2013 ), suggest a two-dimensional underlying structure, with one factor – expression – related to observable signs (blunted affect and alogia), and the other – experience – related to motivation, pleasure and social involvement (Horan et al, 2011, Kirkpatrick et al, 2011, Strauss et al, 2012, and Kring et al, 2013). However close examination of the factor analyses published to date raises a number of doubts that challenge the two-factor solution. Regarding the Brief Negative Symptom Scale (BNSS), four of the seven items of Factor 2 also load quite strongly on Factor 1 (> 0.45) ( Kirkpatrick et al., 2011 ), and there is no information on the correlation coefficient between these two factors that supports their independence (Kirkpatrick et al, 2011 and Strauss et al, 2012). In the case of the Clinical Assessment Interview for Negative Symptoms (CAINS), none of the papers (Horan et al, 2011 and Kring et al, 2013) provided information on the percentage of the variance explained (for the solution or for each of the factors obtained), or on the loadings of all items on each of the factors, so the reader is not able to know the adequacy of the factor structure. In addition, in the first paper ( Horan et al., 2011 ) the scree plot suggested a two- or three-factor solution, yet the choice of the two-factor solution was not clearly explained by the authors.

Finally, the most important factor of both scales (Factor 1 — expression), explained 57% of the variance in the case of the BNSS (compared with 14% explained by Factor 2 — motivation and pleasure). This data was not available for the CAINS (Horan et al, 2011, Kirkpatrick et al, 2011, Strauss et al, 2012, and Kring et al, 2013). This means that blunted affect and alogia represent the “core” of the negative syndrome, while motivation, pleasure and social involvement would be peripheral symptoms.

Bearing the above in mind, we set out to identify the underlying structure of the negative syndrome as part of the Spanish validation study of the BNSS ( Mané et al., 2014 ). Based on our previous work ( Garcia-Portilla and Bobes, 2013 ), we hypothesized that alogia would not belong to the negative but to the cognitive dimension of schizophrenia. Therefore we speculated that alogia would be an independent factor, with low correlation coefficients with the other symptoms of the negative dimension.

2. Methods

2.1. Study design

The data presented in this paper are from a cross-sectional, naturalistic, multicentre, validation study in patients with schizophrenia, conducted at 3 outpatient services in Spain (Parc de Salut Mar Barcelona, Hospital Clinic de Barcelona and Universidad de Oviedo). The data reported here are confined to the BNSS. For further clinical details on the sample please refer to Mané et al., 2014 . The study was approved by the Ethics Committee for Clinical Research of each site, and is in accordance with the 1975 Declaration of Helsinki, as revised in 1983. Written informed consent was obtained from all subjects prior to enrolment.

2.2. Subjects

Twenty participants recruited from the schizophrenia research programmes of the three sites were enrolled based on the following inclusion criteria (1) age ≥ 18 years; and (2) diagnosis of schizophrenia (made by registered consultant psychiatrists using DSM-IV criteria). Exclusion criteria were (1) IQ below 80; and (2) neurological disorders or substance dependence except tobacco.

2.3. Procedure and measures

We followed the original BNSS validation methodology ( Kirkpatrick et al., 2011 ). All patients were interviewed at baseline and ten were also interviewed one week later for the test–retest reliability analysis. All BNSS interviews were videotaped and rated independently by all the raters in the study [AM and DB (Parc de Salut Mar, Barcelona, Spain), CGR and GS (Hospital Clinic, Barcelona, Spain), EFE (University of Cambridge, UK), and PGP and LGA (Universidad de Oviedo, Spain)]. All raters are qualified psychiatrists or psychologists, all of whom participated in the BNSS adaptation and validation procedures. In order to improve inter-rater reliability a pilot study was conducted with two patients using the same inclusion criteria. The patients were interviewed with the BNSS-Sp. The interviews were videotaped by AM and were later assessed by all participating raters. Ratings were discussed to ensure that the BNSS-Sp criteria were understood. For further details see Mané et al. (2014) .

Here we report the data from the 190 evaluations obtained from the baseline visit (20 patients rated by 7 raters) and the re-test visit (10 patients rated by 5 raters).

2.4. Statistical analysis

The statistical analysis was performed using SPSS v.17.0. The statistical significance was two-tailed and was set at 0.05. An exploratory factor analysis (EFA) using the principal component analysis (PCA) method with oblimin rotation was used to explore the component structure of 12 of the 13 BNSS items. We decided not to include item 4 “Distress” as its communalities were 0.377 (for the three-component solution) and 0.357 (for the two-component solution), thus indicating that this item has a poor fit with regard to the other items in its component. This decision was also supported by the theoretical composition of the negative syndrome. However when we repeated the same analyses including this item we obtained very similar results ( Table 3 ). We decided to use the oblimin rotation method as it provides a better representation of clinical reality, characterized by relationships between the different dimensions of the negative syndrome ( Horan et al., 2011 ). Due to criticism of the Kaiser rule (i.e. retention of principal components with eigenvalues above 1) we also used the scree plot to determine the appropriate number of components to extract. We then performed a PCA analysis with oblimin rotation, so as to extract the selected components of the 12 items of the BNSS. Only items with a loading ≥ 0.40 were considered to load on a component. The component correlation matrix was also obtained. The internal consistency of the BNSS components was calculated using Cronbach's alpha coefficient at the item level. Finally we calculated Pearson correlation coefficients (r) to determine the strength of the relationships between components.

Table 1 BNSS rotated three-component structure and component loadings.



BNSS items
Component loadings
Component 1

External world
Component 2

Inner world
Component 3

Alogia
1. Pleasure: intensity 0.919 0.460  
2. Pleasure: frequency 0.873 0.392
3. Expected pleasure: intensity

5. Asociality: behaviour
0.796

0.780
0.348

0.431
6. Asociality: internal experience 0.728 0.361
7. Avolition: behaviour 0.661 0.746
8. Avolition: internal experience 0.634 0.766
9. Blunted affect: facial expression 0.383 0.912
10. Blunted affect: vocal expression 0.335 0.885
11. Blunted affect: expressive gestures 0.490 0.881 0.356
12. Alogia: quantity of speech 0.550 0.619 0.770
13. Alogia: spontaneous elaboration 0.580 0.664 0.742
Eigenvalues

Percentage of total variance
6.575

54.791
1.686

14.052
1.031

8.589

Table 2 BNSS rotated two-component structure and component loadings.



BNSS items
Component loadings
Component 1 Component 2
1. Pleasure: intensity

2. Pleasure: frequency

3. Expected pleasure: intensity

5. Asociality: behaviour

6. Asociality: internal experience

7. Avolition: behaviour

8. Avolition: internal experience

9. Blunted affect: facial expression

10. Blunted affect: vocal expression

11. Blunted affect: expressive gestures

12. Alogia: quantity of speech

13. Alogia: spontaneous elaboration
0.920

0.872

0.794

0.781

0.727

0.669

0.641

0.388

0.342

0.494

0.547

0.577
0.465

0.419

0.411

0.426

0.399

0.651

0.692

0.913

0.865

0.905

0.767

0.803
Eigenvalues

Percentage of total variance
6.575

54.791
1.686

14.052

Table 3 Factor analysis results of the BNSS without and with item 4 “Distress”.

  Without item 4 “Distress” With item 4 “Distress”
  2 components 3 components 2 components 3 components
KMO 0.866 0.873
Bartlett test of sphericity Chi2 = 1948.588, p < 0.001 Chi2 = 2012.145, p < 0.001
% of total variance explained 68.843 77.433 65.896 73.876
Components' items (Cronbach's alpha) C1: 1–3, 5–7 (0.888)

C2: 8–13 (0.912)
C1: 1–3, 5–6 (0.883)

C2: 7–11 (0.902)

C3: 12–13 (0.950)
C1: 1–7 (0.884)

C2: 8–13 (0.912)
C1: 1–6 (0.877)

C2: 7–11 (0.902)

C3: 12–13 (0.950)
r between components C1–C2 = − 0.522 C1–C2 = − 0.494

C1–C3 = − 0.117

C2–C3 = 0.179
C1–C2 = − 0.535 C1–C2 = − 0.512

C1–C3 = 0.081

C2–C3 = − 0.130

C: component; KMO: Kaiser–Meyer–Olkin statistic.

3. Results

Subjects' mean age was 37.3 (11.7), 70% were male, and the mean length of illness was 11.6 years (10.4). Mean BNSS total score was 23.7 (15.4). Mean scores on each of the subscales were Anhedonia [5.0 (4.5)], Distress [1.1 (1.5)], Asociality [3.8 (2.5)], Avolition [4.1 (2.9)], Blunted affect [6.5 (5.0)], and Alogia [3.3 (3.5)].

3.1. Exploratory factor analysis and principal component analysis

An EFA analysis was performed on 12 of the 13 BNSS items. The KMO (Kaiser–Meyer–Olkin) sampling adequacy statistic (0.866) and the Bartlett test of sphericity (Chi2 = 1948.588, p < 0.001) indicated that EFA was a valid option for exploring the BNSS structure. The three-components explained 77.4% of the total variance and their internal consistency was good (Cronbach's alpha of 0.883, 0.902, and 0.950 respectively). Table 1 displays the rotated component structure with the component loadings for the three-component solution. A moderate correlation was found between components 1 and 2 (r = − 0.494), while components 1 and 3 (− 0.117) and 2 and 3 (0.179) exhibited a low correlation.

Although there were three components with eigenvalues above 1, the scree plot ( Fig. 1 ) also made it possible to retain a two-factor solution for further investigation. Therefore a PCA was applied using oblimin rotation in order to extract the selected BNSS components. The two extracted components in the PCA explained 68.8% of the total variance, and their internal consistency was good (Cronbach's alpha of 0.897 and 0.915 respectively). A moderate correlation was found between the two components (r = − 0.522, p < 0.05). Table 2 depicts the rotated component structure for the two-component solution.

gr1

Fig. 1 BNSS scree plot.

4. Discussion

The aim of this study was to identify the underlying structure of the negative syndrome as it is represented in the BNSS, in patients with schizophrenia. Unlike previous studies on the dimensional structure of the BNSS (Kirkpatrick et al, 2011 and Strauss et al, 2012), we decided not to include item 4 “Distress”, given its lower communality.

Similar to previous studies (Horan et al, 2011, Kirkpatrick et al, 2011, Strauss et al, 2012, and Kring et al, 2013) we were able to obtain a two-component solution, even though the three-component showed a better fit, as shown in the results section. Our results challenge the pre-established dimensions of experience and expression, given that our second component mixes experience (avolition) with expression (blunted affect) and separates the two symptoms of the expression dimension. We believe that our solution is not problematic, as there is no psychopathological or pathophysiological basis for justifying the two-dimensional structure maintained to date.

We believe the three-factor solution represents the theoretical model and nature of the negative syndrome of schizophrenia more accurately. When examining the three-component solution, we observed that the first component, “external world”, consisted of anhedonia and asociality, the second, “inner world”, included avolition and blunted affect, and the third, “alogia”, was represented only by alogia. Contrary to the previous two-factor solutions, the “core” of the negative syndrome in our analyses is represented by pleasure and social involvement (which explained 54.8% of the total variance), and this, we believe, provides a better reflection of the nature of the negative syndrome. In addition, the correlation coefficients among the three components indicate that the alogia component is barely related to the other two components, thus supporting the hypothesis of it belonging to the cognitive rather than the negative domain ( Garcia-Portilla and Bobes, 2013 ). Furthermore, most of the mechanisms proposed to explain this sign are related to executive processes associated with the frontal lobe. Berenbaum et al. (2008) found that alogia was associated with poor performance on cognitive tests of planning, verbal fluency and concentration/attention in patients with schizophrenia.

The small sample size of our study is a limitation that needs to be taken into consideration, since it could have decreased the power of the study. However as in the original BNSS validation, we enhanced the power of the study by increasing the number of raters and by using raters from different settings with different professional backgrounds. Furthermore a pre-validation exercise was conducted to ensure that BNSS criteria were understood and so as to increase inter-rater reliability (videotapes from two pilot patients were rated by all raters; the ratings were then discussed among raters for consensus).

These results have important clinical and research implications. On the one hand, there is a need for the negative syndrome to be appropriately delimitated, since this syndrome, along with cognitive deficits, has proven to be the main determinant of real-world functioning of patients with schizophrenia. This should be confirmed if alogia is eventually removed from the negative syndrome. On the other hand, the changes proposed in this paper on the internal structure of the negative syndrome may lead to new outcome measures with which to assess the efficacy of interventions targeting negative symptoms.

In conclusion, our solution favours a three-component structure of the negative syndrome: external world (anhedonia and asociality), inner world (avolition and blunted affect), and alogia, with the alogia component only marginally related to the former components. These results raise questions concerning the two-dimensional structure of the negative syndrome and regarding the inclusion of alogia, and suggest the need for further research on the structure of the negative syndrome.

Role of funding source

There is not any funding source.

Contributors

MP Garcia-Portilla, A Mané, E Fernández-Egea and J Bobes conceptualized the project and designed the tasks.

MP Garcia-Portilla, L Garcia-Alvarez, and J. Bobes wrote the first draft of the manuscript.

MP Garcia-Portilla, L Garcia-Alvarez, and J. Bobes analyzed the data.

All the authors reviewed the manuscript and contributed to its final version.

Conflict of interest

The authors declare that they have no conflicts of interests concerning this article.

Acknowledgements

We would like to thank Sharon Grevet for her assistance with the English.

References

  • Berenbaum et al., 2008 H. Berenbaum, J.G. Kerns, L.L. Vernon, J.J. Gomez. Cognitive correlates of schizophrenia signs and symptoms: I. Verbal communication disturbances. Psychiatry Res.. 2008;159:147-156 Crossref
  • Blanchard and Cohen, 2006 J.J. Blanchard, A.S. Cohen. The structure of negative symptoms within schizophrenia: implications for assessment. Schizophr. Bull.. 2006;32:238-245 Crossref
  • Bowie et al., 2006 C.R. Bowie, A. Reichenberg, T.L. Patterson, R.K. Heaton, P.D. Harvey. Determinants of real-world functional performance in schizophrenia subjects: correlations with cognition, functional capacity, and symptoms. Am. J. Psychiatry. 2006;163:418-425 Crossref
  • Bowie et al., 2010 C.R. Bowie, C. Depp, J.A. McGrath, P. Wolyniec, B.T. Mausbach, M.H. Thornquist, J. Luke, T.L. Patterson, P.D. Harvey, A.E. Pulver. Prediction of real-world functional disability in chronic mental disorders: a comparison of schizophrenia and bipolar disorder. Am. J. Psychiatry. 2010;167:1116-1124 Crossref
  • Garcia-Portilla and Bobes, 2013 M.P. Garcia-Portilla, J. Bobes. The new challenge in identifying the negative syndrome in schizophrenia. Rev. Psiquiatr. Salud Ment. (Barc.). 2013;6:141-143
  • Harvey et al., 2012 P.D. Harvey, R.K. Heaton, W.T. Carpenter Jr., M.F. Green, J.M. Gold, M. Schoenbaum. Functional impairment in people with schizophrenia: focus on employability and eligibility for disability compensation. Schizophr. Res.. 2012;140:1-8 Crossref
  • Horan et al., 2011 W.P. Horan, A.M. Kring, R.E. Gur, S.P. Reise, J.J. Blanchard. Development and psychometric validation of the clinical assessment interview for negative symptoms (CAINS). Schizophr. Res.. 2011;132:140-145 Crossref
  • Kirkpatrick et al., 2006 B. Kirkpatrick, W. Fenton, W.T. Carpenter, S.T. Marder. The NIMH-MATRICS consensus statement on negative symptoms. Schizophr. Bull.. 2006;32:296-303
  • Kirkpatrick et al., 2011 B. Kirkpatrick, G.P. Strauss, L. Nguyen, B.A. Fischer, D.G. Daniel, A. Cienfuegos, S.R. Marder. The brief negative symptom scale: psychometric properties. Schizophr. Bull.. 2011;37:300-305 Crossref
  • Kring et al., 2013 A.M. Kring, R.E. Gur, J.J. Blanchard, W.P. Horan, S.P. Reise. The clinical assessment interview for negative symptoms (CAINS): final development and validation. Am. J. Psychiatry. 2013;170:165-172 Crossref
  • Mané et al., 2014 A. Mané, C. García-Rizo, M.P. Garcia-Portilla, D. Bergé, G. Sugranyes, L. Garcia-Alvarez, M. Bernardo, J. Bobes, E. Fernandez-Egea. Spanish adaptation and validation of the Brief Negative Symptoms Scale. Compr. Psychiatry. 2014;55:1726-1729
  • Rocca et al., 2014 P. Rocca, C. Montemagni, S. Zappia, R. Piterà, M. Sigaudo, F. Bogetto. Negative symptoms and everyday functioning in schizophrenia: a cross-sectional study in a real world-setting. Psychiatry Res.. 2014;218:284-289 Crossref
  • Strauss et al., 2012 G.P. Strauss, L.E. Hong, J.M. Gold, R.W. Buchan, R.P. McMahon, W.R. Keller, B.A. Fischer, L.T. Catalano, A.J. Culbreth, W.T. Carpenter, B. Kirkpatrick. Factor structure of the brief negative symptom scale. Schizophr. Res.. 2012;142:96-98 Crossref

Footnotes

a Department of Psychiatry, University of Oviedo, Spain

b Centro de Investigación Biomédica en Red, Área de Salud Mental (CIBERSAM), Spain

c Institut de Neuropsiquiatria i Adiccions, Parc de Salut Mar, Barcelona, Spain

d Fundació IMIM, Barcelona, Spain

e Barcelona Clinic Schizophrenia Unit, Neuroscience Institute, Hospital Clinic of Barcelona, Spain

f Child and Adolescent Psychiatric Unit, Neuroscience Institute, Hospital Clinic of Barcelona, Spain

g Institut d'Investigacions Biomediques August Pi i Sunyer (IDIBAPS), Barcelona, Spain

h Department of Psychiatry and Clinical Psychobiology, University of Barcelona, Barcelona, Spain

i Department of Psychiatry, Behavioural and Clinical Neuroscience Institute, University of Cambridge, United Kingdom

j Cambridgeshire and Peterborough NHS Foundation Trust, United Kingdom

Corresponding author at: Department of Psychiatry, University of Oviedo, Julian Claveria 6, 33006 Oviedo, Spain. Tel.: + 34 985104219; fax: + 34 985103553.