Factor structure of the Neurological Evaluation Scale in a predominantly African American sample of patients with schizophrenia, unaffected relatives, and non-psychiatric controls

https://doi.org/10.1016/j.schres.2006.03.022Get rights and content

Abstract

Objective

This study aimed to identify latent factors of the most widely used scale to measure neurological soft signs, the Neurological Evaluation Scale (NES), in a sample of schizophrenia patients, their relatives, and non-psychiatric controls.

Method

The NES was administered following the format and instructions of the original scale [Buchanan, R.W., Heinrichs, D.W., 1989. The Neurological Evaluation Scale (NES): a structured instrument for the assessment of neurological signs in schizophrenia. Psychiatry Res. 27, 335–350]. Three factor analyses were conducted using: (1) 23 items of the scale, (2) these 23 items in a sample restricted to African American participants, and (3) 12 items previously recommended based on adequate inter-rater reliability [Sanders, R.D., Forman, S.D., Pierri, J.N., Baker, R.W., Kelley, M.E., van Kammen, D.P., Keshavan, M.S., 1998. Inter-rater reliability of the neurological examination in schizophrenia. Schizophr. Res. 29, 287–292].

Results

Exploratory factor analysis in the overall sample (n = 110) revealed three factors of interest: a “coordination/Romberg” factor (which included the fist-edge-palm test, finger-thumb opposition, rapid alternating movements, and the Romberg test), a “sensory integration” factor (which included synkinesis, extinction, stereognosis, and audio-visual integration), and an “eye movements/tandem walk/overflow movements” factor (which included convergence, gaze impersistence, tandem walk, and adventitious overflow). Minimal differences were apparent when comparing these results with those obtained from an analysis including only African American participants (n = 99). Restricting the analysis to only 12 items resulted in two factors (with poor discriminant validity and internal consistency), which showed few similarities with previously reported exploratory factor analysis results.

Conclusions

This study shows that results of empirically-derived factors of the NES vary substantially across study samples and research groups, though there appear to be at least two consistent latent variables, one related to the repetition of motor tasks and another tapping the processing of sensory stimuli. The factors demonstrated in the present study are more consistent with the original conceptually-derived subscales than prior factor analyses. Thus, until further research replicates consistent subscales of subtle neurological impairment in a variety of settings and samples, the total NES score may be more useful when comparing results across studies. Further research on latent neurological factors is needed.

Introduction

Schizophrenia is conceptualized as a neurodevelopmental psychiatric illness (Davies et al., 1998, Cannon et al., 2003, McGrath et al., 2003) affecting multiple brain functions including those in behavioral, cognitive, perceptual, and motor domains. Subtle neuromotor dysfunction appears to be present well before the onset of psychotic symptoms (Walker, 1994), and may represent a trait marker of the illness and/or of vulnerability toward developing the illness. An elevated prevalence of neurological signs in individuals with schizophrenia has been recognized for several decades (Heinrichs and Buchanan, 1988).

Neurological “soft signs” are subtle neurological abnormalities that, unlike “harder” neurological signs, are nonspecific and generally do not localize a lesion. Soft neurological signs may indicate a generalized neurodevelopmental deficit. Multiple studies have revealed an excess of soft signs among patients with schizophrenia, including those with first-episode psychosis (Sanders et al., 1994, Mohr et al., 2003, Dazzan et al., 2004, Scheffer, 2004, Chen et al., 2005), and neuroleptic-naïve or medication-free patients (Gupta et al., 1995, Shibre et al., 2002, Keshavan et al., 2003, Venkatasubramanian et al., 2003). Prevalence rates of neurological soft signs in patients with schizophrenia range from 50% to 65% compared with only about 5% in controls (Bombin et al., 2005), though estimates vary by the measurement methods used.

The presence of neurological soft signs represents a potential vulnerability marker, or intermediate phenotype, for schizophrenia (Egan et al., 2001, Bombin et al., 2005). Supporting this, several studies have found that neurological soft signs occur at higher rates among unaffected first-degree relatives of patients with schizophrenia compared to controls without a family history of the illness (Rossi et al., 1990, Griffiths et al., 1998, Ismail et al., 1998, Flyckt et al., 1999). In contrast, one study found no difference between Neurological Evaluation Scale (NES) scores in siblings of patients and well-matched controls, though this discrepant finding could be accounted for by different ascertainment strategies and exclusion criteria (Egan et al., 2001).

The NES, which standardizes the administration and scoring of neurological impairment (Buchanan and Heinrichs, 1989), is the most widely used instrument for assessing neurological deficits in schizophrenia. The original subscales of the NES were derived theoretically. That is, items were initially grouped conceptually based on similarities between tasks or brain regions thought to underlie the functions being tested (Table 1). Such groupings of items allowed for the computation of subscale scores. The designated subscales include Sensory Integration, Motor Coordination, and Sequencing of Complex Motor Tasks (Buchanan and Heinrichs, 1989). Other theoretically-derived groupings, based on neuroanatomical considerations, include Cerebellar, Frontal, and Parietal subscales (Egan et al., 2001).

Factor analysis is an empirical technique for the reduction of data to underlying coherent subsets or factors. Thus, this method is used to find latent variables among many observed variables by grouping together those with similar characteristics. As shown in Table 2, several studies, from at least three research groups, have reported NES subscales based on exploratory factor analysis (Malla et al., 1997, Sanders et al., 2000, Keshavan et al., 2003, Emsley et al., 2005, Goldstein et al., 2005). Additionally, one study has reported subscales based on a confirmatory factor analysis that included 95 unmedicated patients (Sanders et al., 2005). Although the NES was constructed with several conceptual factors of hypothetical significance, this factorial structure has never been confirmed because empirical factors found in prior studies differ from these theoretical subscales (Krebs et al., 2000), and prior exploratory factor analyses have yielded generally inconsistent results. Further research is needed into the underlying structure of the neurological soft signs construct, especially given emerging evidence that such subtle neurological abnormalities may serve as a vulnerability marker for schizophrenia that is measured easily, noninvasively, and at low cost.

The objective of the current study was to perform an exploratory factor analysis in a sample of patients with schizophrenia and related disorders, their healthy relatives, and non-psychiatric controls. Because neurological soft signs are increasingly recognized as a potential vulnerability marker, a variety of studies use the NES in samples comprised of these three subgroups. Yet, to date, published factor analytic studies of the NES have included only patients (and less often, healthy controls). Furthermore, this study assessed participants from an urban, predominantly African American population. Previous studies have suggested that race/ethnicity may have an impact on the prevalence of neurological soft signs (Bombin et al., 2005), though most prior factor analytic studies have relied upon samples composed predominantly of Caucasian patients. Three factor analyses were conducted using: (1) 23 items of the scale, (2) these 23 items in a sample restricted to only African American participants, and (3) a subset of 12 items previously recommended as most reliable (Sanders et al., 1998). Comparisons of neurological soft signs, as well as minor physical anomalies, across the three groups (patients, relatives, and controls) will be presented in another report.

Section snippets

Setting and sample

The study took place at a large, urban, university-affiliated, public-sector health system in the southeastern United States. This health system provides medical and psychiatric services for a predominantly uninsured, low-income, African American population. Patients and their healthy relatives were drawn from an adult community mental health clinic (where patients are typically seen monthly by a social worker and quarterly by a psychiatrist) and an adult outpatient intensive day treatment

Results

Basic demographic data for the overall sample of 110 participants are shown in Table 3. Just over half of participants (57.3%) were female, and 90% self-identified as African American. Nearly half (49.1%) of participants were single/never married, and only 30% were employed. The mean age of participants was 42.9 ± 13.4 years, and the mean number of years of education completed was 12.4 ± 2.2.

Table 4 displays statistically significant correlations between the 23 NES items. In general, item–item

Discussion

As is the case with prior exploratory factor analyses of the NES, there is some difficulty in identifying what the factors reported here represent conceptually. The statistically-derived factors in this study share some commonalities with theoretical subscales and with factors reported previously, but prominent differences are also apparent.

The first factor in this sample, which is termed here “coordination/Romberg”, included three items related to repetitive movements of the hands or fingers

Acknowledgements

This research was supported by a “Young Minds in Psychiatry” award to the first author from the American Psychiatric Institute for Research and Education/AstraZeneca. The authors gratefully acknowledge data collection assistance provided by Andrea Cunningham, Michelle Esterberg, Will Fleiss, LaTasha McKenzie Mack, Laura Pence, and Luis Quintero.

References (31)

  • R.D. Sanders et al.

    Confirmatory factor analysis of the neurological evaluation scale in unmedicated schizophrenia

    Psychiatry Res.

    (2005)
  • R.E. Scheffer

    Abnormal neurological signs at the onset of psychosis

    Schizophr. Res.

    (2004)
  • A.H. Yazici et al.

    Neurological soft signs in schizophrenic patients and their nonpsychotic siblings

    Schizophr. Res.

    (2002)
  • C. Arango et al.

    Neurological signs and the heterogeneity of schizophrenia

    Am. J. Psychiatry

    (2000)
  • I. Bombin et al.

    Significance and meaning of neurological signs in schizophrenia: two decades later

    Schizophr. Bull.

    (2005)
  • Cited by (26)

    • Neurological soft signs in familial and sporadic schizophrenia

      2019, Psychiatry Research
      Citation Excerpt :

      Concerning the scoring of lateralizing items, some researchers used the mean score while others kept the highest side score (Keshavan et al., 2003; Malla et al., 1997; Sanders et al., 2000, 2005). We chose to use the sum of both (Compton et al., 2006). Two additional scores were used, the sum of all items concerning the left and the sum concerning the right side of the body.

    • Motor dysfunction as research domain in the period preceding manifest schizophrenia: A systematic review

      2018, Neuroscience and Biobehavioral Reviews
      Citation Excerpt :

      A total of 46 studies that met the inclusion criteria were identified. Among these studies, 13 reported unaffected first-degree relatives (Kinney et al., 1986; Rossi et al., 1990; Woods et al., 1987; Galindo et al., 2016; Bollini et al., 2007; Solanki et al., 2012; Chan et al., 2016; Gourion et al., 2003,2004; Compton et al., 2007; Gabalda et al., 2008; Flyckt et al., 1999; Compton et al., 2006), 1 reported familial schizophrenia relatives (Griffiths et al., 1998), 12 reported children/offsprings (Erlenmeyer-Kimling and Cornblatt, 1984; Erlenmeyer-Kimling et al., 1989; Marcus et al., 1985b,1985c; Marcus et al., 1985a; McNeil et al., 1993; Rieder and Nichols, 1979; Prasad et al., 2009; Niemi et al., 2005; Schubert and McNeil, 2004; Schubert et al., 2005; Sugranyes et al., 2017; Marcus et al., 1993), 7 reported co-twins (Picchioni et al., 2006; Sanders et al., 2006; Xu et al., 2016; Cantor-Graae et al., 1994a; Cantor-Graae et al., 1994b; Kelly et al., 2004; Niethammer et al., 2000), and 13 reported siblings (Ismail et al., 1998; Egan et al., 2001; Hyde et al., 2007; Cantor-Graae et al., 2000; Chen et al., 2000; Aksoy-Poyraz et al., 2011; Yazici et al., 2002; Mechri et al., 2010,2009; Cuesta et al., 2017; Ismail et al., 2001; Caldani et al., 2017a,2017b) of schizophrenia patients. We also identified one systematic review (Wolff and O’driscoll, 1999) and one meta-analysis that found significant NSS scores difference between first-degree relatives and both schizophrenia patients (mean effect size: 0.97) and HP (mean effect size: 0.97) (Chan et al., 2010b).

    • Neurological soft signs: Effects of trait schizotypy, psychological distress and auditory hallucination predisposition

      2017, Schizophrenia Research: Cognition
      Citation Excerpt :

      Consequently, adopting cluster analysis in combination with a more robust measure of NSS may highlight differences attributable to the dimensions of schizotypy. The current study is using the Neurological Evaluation Scale (NES; Buchanan and Heinrichs, 1989): one of the most widely used measures of NSS within the psychosis literature (e.g. Compton et al., 2006; Chan et al., 2010a; Sewell et al., 2010). Therefore the research from this study can be more easily compared with existing research in the field.

    • The confirmatory factor structure of neurological soft signs in nigerians with first episode schizophrenia

      2015, Neuroscience Letters
      Citation Excerpt :

      Factor analysis is a standard technique for reducing data sets to underlying consistent sub-sets, which can then be used to find principal variables among many observed variables. Out of a dozen reported factor analyses of NSS in schizophrenia [1–12], only one [6] had been based on a sample from the African continent. That exploratory factor analysis (EFA) was based on a sample composed of 21% white, 71% mixed and 8% Black Africans with first episode schizophrenia.

    View all citing articles on Scopus
    View full text