Neurological soft signs in patients with schizophrenia: current knowledge and future perspectives in the post-genomics era

Abstract

Neurological soft signs (NSS), minor and subtle neurological abnormalities in sensory integration and motor performance that are not part of a properly defined neurological syndrome, have been consistently observed in patients with schizophrenia. The prevalence estimates of NSS in patients with schizophrenia have been reported to be higher than in healthy subjects. Current evidence suggests that NSS are an integral part of the disease and cannot be fully explained by the exposure to antipsychotic medication, as they are already present in treatment-naïve patients. NSS have been associated with cardinal features of the disorder such as cognitive impairment, psychopathological severity, or functional outcome. The increased prevalence of NSS and/or related motor precursors has been described at different stages of development (infancy, childhood, adolescence) in those subjects who later developed schizophrenia. Evidence from family and twin studies indicates that genetic factors play an important role in the emergence of NSS, and some authors have already suggested that such neurological anomalies are suitable endophenotypes for schizophrenia. Some genetic association studies based on a candidate gene approach have already reported the association of genetic variants with the severity of NSS. This non-systematic review describes the potential relevance of NSS 1) in the understanding of schizophrenia as a neurodevelopmental disorder, 2) as outcome predictors, 3) as biological markers during several stages of development, and 4) as a candidate (endo)phenotype for genetic analyses. Likewise, the possibilities afforded by the advances in high-throughput techniques in genomic analysis are also discussed.

• genome-wide association studies
• motor function
• sensory integration
• endophenotype

Schizophrenia is an extremely severe and disabling neuropsychiatric disorder which usually emerges during adolescence or early adulthood. It is characterised by a set of symptoms and signs that include delusions, hallucinations, paranoid ideation, disordered thought, and cognitive impairment. Alterations in drive and volition, blunted emotions, difficulties in communication, and motor abnormalities are also commonly observed in these patients. These symptoms and signs are generally clustered into positive, negative, cognitive, disorganisation, mood, and motor symptom dimensions (1, 2). At least some of these dimensions (e.g. positive, negative, disorganisation) have different levels of prominence across time and across individuals, making the course and outcome of schizophrenia heterogeneous and clinically diverse (2).

Fig. 1. Schematic summary of the current evidences of the presence of NSS or potentially related psychomotor precursors in those subjects at high risk or who eventually developed schizophrenia across the developmental span. The developmental periods in which environmental or genetic risk factors are relevant are also depicted. The period in which schizophrenia onset is more likely is also represented. CNS: central nervous system; NSS: neurological soft signs.

Regarding the aetiology of the disorder, compelling evidence supports the notion that schizophrenia is a neurodevelopmental disorder (3, 4). This model suggests that schizophrenia is the final behavioural consequence of neurodevelopmental disturbances that start long before the onset of behavioural/clinical symptoms. Such disturbances would be the result of certain combinations of genetic and environmental risk factors. As regards genetic risk factors, recent genome-wide association studies (GWAS) have just started to shed light on the polygenic architecture of this and other neuropsychiatric disorders (5, 6). However, the role of this genetic risk component in a developmental framework still needs to be ascertained. Among environmental insults potentially related with alterations in neurodevelopment, obstetric/perinatal complications and prenatal infections are perhaps the most commonly reported and replicated factors increasing the risk of schizophrenia (7–10).

Several indicators of an abnormal development already present in early childhood also support a neurodevelopmental model. Significant neuromotor impairments are frequently observed in an important proportion of schizophrenia patients many years before the full-blown clinical symptomatology appears (11). Developmental delays in the areas of motor/language development are more frequently found in patients with schizophrenia than in the general population (12–14). Likewise, lower general cognitive abilities and impaired social development have been observed in children who will later develop this disorder with respect to those who will not (11, 13) (15–19). Brain imaging has also provided evidence of structural brain disturbances present before the onset of the disease or in first-episode schizophrenia patients (20–22).

Congenital abnormalities are perhaps the features that better illustrate the effect of early developmental insults on the overall phenotype of this disorder. In this sense, minor physical anomalies (23, 24) or dermatoglyphic deviances (25, 26) have been found in these patients in a higher frequency than in the general population. Besides these features, neurological soft signs (NSS) are another marker of especial interest (27). The present non-systematic review will focus on the potential relevance of NSS 1) in the understanding of schizophrenia as a neurodevelopmental disorder, 2) as outcome predictors, 3) as biological markers during several stages of development, and 4) as a suitable phenotype for genetic analyses.

NSS in schizophrenia patients

An interesting observation in an important percentage of patients with schizophrenia after a careful clinical evaluation is the presence of NSS, minor and subtle neurological abnormalities in sensory integration and motor performance (Fig. 1). NSS have been defined as neurological abnormalities that, in principle, are not part of a properly defined neurological syndrome (27). The Neurological Evaluation Scale (28) and the Cambridge Neurological Inventory (29) are the most commonly used scales to measure NSS. Clustering analyses considering the items contained in these or similar scales have shown that NSS can be grouped into several categories: integrative sensory function, motor coordination, sequencing of complex motor acts, and primitive reflexes (30). These categories should not be considered as definitive since different authors have proposed different groupings of signs depending on the analytical approaches/principles applied (31–33).

The original definition of NSS considered the fact that they cannot be easily linked to abnormalities in specific brain regions. However, advances in structural and functional neuroimaging techniques have already provided some clues regarding the neuroanatomical substrate of, at least, some of these NSS (34–39). In a recent meta-analysis, NSS have been associated with atrophy in precentral gyrus, cerebellum, inferior frontal gyrus, and thalamus (40). The same study also showed that NSS-related tasks resulted in abnormal brain activation in inferior frontal gyrus, bilateral putamen, cerebellum, and superior temporal gyrus (40).

Despite the differences in the scales and cluster definitions, both descriptive reviews and meta-analytic approaches have consistently shown that NSS are more commonly observed in patients with schizophrenia compared with healthy controls (27, 30) (41–43). Likewise, a very important fraction of these patients (73%) has been reported to show NSS aggregate scores outside the range observed in the healthy population (42). According to these studies, NSS represent one of the traits that better differentiates schizophrenia patients from healthy subjects. However, it has to be mentioned that it is not a specific feature of this disorder since NSS have also been observed in other neuropsychiatric disorders (30), although discussion exists about the specificity of some NSS clusters (32, 39).

An important confounding factor that may have influenced the aforementioned studies is antipsychotic medication. Motor side effects induced by this treatment such as tardive dyskinesia may easily be rated as NSS. As regards this issue, studies based on neuroleptic-naïve patients have shown that NSS are, to a large extent, not dependent on the exposure to antipsychotic treatment (39, 44) (45). Moreover, no associations have been found between antipsychotic medication dosage and NSS (30, 42). However, some studies have observed associations between extrapyramidal symptoms or tardive dyskinesia and NSS (44, 46). Taken together, these results indicate that NSS are an integral part of the disease rather than the result of the exposure to antipsychotic medication.

NSS and the clinical phenotype of schizophrenia

Besides the intrinsic interest of NSS as indicators of an altered neurodevelopment in patients with schizophrenia, an increasing body of evidence suggests that NSS might be an important predictor of psychopathological severity, neurocognitive impairment, and overall functional outcome in this disorder (30, 42) (43).

Specifically, studies ascertaining the relationship between psychopathology and NSS have reported that neurological signs have an important correlation with the severity of negative symptoms and the degree of disorganised behaviour (47–51). On the contrary, evidence of an influence on positive symptomatology is less conclusive (52, 53).

Cognitive impairment is increasingly being recognised as a core feature of schizophrenia. It affects several domains: working memory, executive function, verbal/visual learning, attention or speed of processing, among others (54, 55). The literature regarding the relationship between NSS and cognition suggests that specific NSS clusters are associated with specific cognitive domains (30, 42) (43). Regarding functional outcome, several studies have reported that NSS are associated with worse pre-morbid functioning, longer hospitalisation length, or a worse social functioning (56–59).

An important consideration while assuming that NSS are a schizophrenia trait is their stability over the course of the disorder. The static/dynamic status of NSS has important implications not only for their potential predictive value regarding clinical course but also for their definition as a candidate endophenotype for genetic analyses (41). Several studies have reported that some motor anomalies, which could be understood as precursors of NSS, are already present long before the onset of the disease (12, 60) (61). Along the same lines, studies in high-risk individuals and in first-episode medication-naïve patients have also reported the higher prevalence of NSS with respect to healthy controls (62–66). These and other studies provide evidence for the pre-existence of neurological disturbances in subjects at risk to develop schizophrenia and reinforce the hypothesis of schizophrenia as a neurodevelopmental disorder (see the next section for further details). These studies, however, lack the capacity to explain if once NSS are established, they remain with or without fluctuations over the course of the disorder. As regards the NSS stability over time, some studies have shown that the decrease of NSS scores across adolescence is a normal brain maturation process in healthy subjects that is to a certain degree impaired in patients with schizophrenia. However, it is not clear which NSS domains are affected as a result of a deviation from this maturation process (67, 68).

Some individual longitudinal studies (69) and a recent meta-analysis (70) have provided interesting evidence of NSS scores decreasing in parallel with remission of psychopathological symptoms, although not reaching the scores typically observed in healthy subjects, suggesting the putative role of NSS as disease progression markers. It should be noted, however, that some follow-up studies have observed an important degree of stability in NSS over the course of schizophrenia (71).

Neurological abnormalities in childhood and adolescence

As already mentioned in the previous section, a significant amount of evidence indicates that 1) NSS are present before the onset of the disease and 2) a decrease in NSS scores during adolescence is a normal process that may be altered in schizophrenia patients (Fig. 1). From a neurodevelopmental perspective, the close relationship between an altered development and NSS has been shown by several studies reporting that very early developmental insults are associated with increased NSS scores in the postnatal period (59, 72–75).

Prospective studies spanning infancy through adolescence have reported compelling evidence of the association of neurodevelopmental abnormalities/delays with schizophrenia (76). In a prospective study based on the British 1946 Birth Cohort, patients with schizophrenia achieved motor development milestones later than controls (13). The milestone with the most striking differences between cases and controls was walking (13). Data from the Northern Finland 1966 Birth Cohort pointed in a similar direction, showing that a later attainment of motor development milestones such as standing or walking was associated with a higher risk of schizophrenia in adulthood (77). The analysis of the 1972–1973 Cohort from Dunedin, New Zealand, reported that children who eventually developed a schizophreniform disorder showed a poorer motor function over the several assessments they underwent during childhood (11). Rosso et al., also in a prospective approach based on the 1959–1966 Boston NCPP Cohort, observed anomalies in motor coordination at 7 years of age and unusual movements at ages 4 and 7 in children who later developed schizophrenia (78). Cannon et al., by using a case-control design nested within the 1951–1960 Helsinki Cohort, observed a motor coordination deficit in those children who developed schizophrenia in adulthood (79).

In this context, studies based on the analysis of home movies of children who later developed schizophrenia are also of special interest. Walker et al. showed that during the first 2 years of life, motor skills and neuromotor functioning had important differences between children who later developed schizophrenia and those who did not (12). Subsequent studies, based on a similar design (videotaped behaviours), also reported these deficits on general neuromotor functioning at older ages in children who later suffered the disorder (80).

Another methodological approach that has reported interesting results is the assessment of neuromotor impairments in high-risk children (81, 82). High-risk children are defined as those with relatives (mainly parents) with schizophrenia/psychosis diagnoses. Taking into account the outstanding heritability estimates of the disorder (83) and the evidence from previous family studies (84), these children are potential carriers of a substantial genetic risk to develop schizophrenia. According to the results of the Swedish high-risk study, already on the third or fourth day of life, neurological deviations can be observed in the offspring of mothers diagnosed with schizophrenia (85).

The combination of a sample characterised by a high genetic risk burden together with a prospective design is theoretically optimal to understand how high-risk subjects behave regarding neuromotor function, cognition, social adjustment during childhood–adolescence, or their response upon exposure to environmental risk factors such as obstetric complications or drug use. Likewise, such a design has the potential to identify indicators/markers of risk long before the onset of the disorder. The first notable approach using this design was carried out in the context of the New York Infant Study (86). The author showed that 7 out of 12 high-risk children had a neurointegrative defect characterised by a delay of motor development (together with other developmental features). These seven infants received a diagnosis of schizophrenia, or schizotypal or paranoid personality disorder (1:5:1) in adulthood. Several of these types of studies indicated the existence of an altered psychomotor development during the first 2 years of life (86–90). It should be noted, however, that two of the studies observed a reduction of these abnormalities with age (90, 91). Such a resolution with age during infancy is difficult to interpret taking into account the disturbances in coordination and fine motor control commonly observed in high-risk individuals during childhood and adolescence (63, 86) (88, 92). For example, the New York High-Risk Project reported, after neurobehavioural measuring, deficits in gross motor skills in 75% of the children between 7 and 12 years of age who were diagnosed with schizophrenia-related psychosis in adulthood (63). Very similar results were obtained by the Swedish high-risk study after analysing 6-year-old children as regards gross and fine neuromotor deviations (93). Mild incoordination and awkwardness while performing rapidly alternating movements at age 7 were especially frequent in the offspring of patients with schizophrenia in the Boston NCPP Study (94). With regards to school age and adolescence periods, the Jerusalem Infant Development Study reported poor motor coordination, poor right–left orientation, poor balance, and motor overflow in the offspring of patients with a schizophrenia diagnosis (95). The same study reported that poor neurobehavioural functioning was two times more frequent in the adolescent offspring of patients with this disorder than in the offspring of control parents, with male offsprings showing a poorer functioning (96).

Taken together, these results indicate that slow neuromotor development is one of the main known risk factors in the development of schizophrenia (97). Although some of the motor abnormalities described in these studies are not strictly NSS according to standard definitions, they could be understood as the first indicators of a neurological disturbance in these children.

Overlap between general neurological abnormalities and NSS

Although evidence suggests a certain continuum of anomalies starting with a very early neuromotor dysfunction eventually leading to subtle neurological anomalies in adult patients with schizophrenia, it has not been firmly proved. Moreover, a considerable fraction of patients with schizophrenia show motor signs not considered as NSS such as abnormal involuntary movements, parkinsonism, catatonia, or psychomotor slowing (98, 99). These reports suggest that, at least regarding motor functions, NSS would represent one of the expressions of an aberrant motor system rather than a final developmental stage of a generalised motor dysfunction.

Among those studies measuring NSS with suitable instruments, the Jerusalem Infant Development Study explored the stability of NSS over time, showing a long-term stability from the childhood to adolescent assessments regarding motor coordination, motor overflow, and sensory-perceptual signs (95). In the context of the same prospective study, Hans et al. observed that 40% of the high-risk adolescents with a poor neurobehavioural functioning already had a poor functioning in infancy and school age (96). Interestingly, a cohort study from Sweden reported that the total score for neurological abnormalities at 22 years in high-risk subjects correlated positively with the score at 6 years of age (100). However, the same study failed to find such a correlation between the score at 22 years and the score registered during early infancy (100).

Data provided by these studies suggest that 1) NSS in high-risk subjects present a certain degree of stability during childhood–adolescence and 2) the link between very early neuromotor abnormalities and NSS in adult patients is not straightforward, especially considering that fluctuations in these neurological functions are expected to happen as brain maturation progresses. Understanding whether this heterogeneous group of neurological disturbances across the developmental span have a similar origin or, on the contrary, represent different biological processes with phenotypical resemblance might help elucidate their interest as predictors of disease onset/course at different stages of development. Likewise, it might lead to the definition of suitable phenotypes for genetic analysis.

Evidence for a genetic component in NSS

The first family studies taking into consideration NSS already described their increased frequency in non-affected relatives of patients with schizophrenia compared with healthy controls (101, 102). Most of the subsequent family studies have reported that healthy relatives of patients with schizophrenia diagnosis have higher NSS scores than healthy independent controls and lower NSS scores than their relatives with schizophrenia (31, 100) (103–115). In the review by Bombin et al., motor signs are suggested to be the cluster of signs with a larger overlap with schizophrenia genetic vulnerability (30). These results have been confirmed using meta-analytic techniques (116, 117). These findings suggest that NSS are, to a certain extent, genetically determined.

In this context, NSS heritability (h²) studies are of special interest in order to ascertain the amount of phenotypic variance of NSS that can be explained by genetic factors (118). Few studies have attempted to estimate the heritability of NSS (31, 119) (120). One of them reported important h² values (between 0.53 and 0.99) for measures of neurological motor signs while the other NSS clusters barely provided any outstanding heritability estimate (120). Convergent evidence from family-based and heritability studies, therefore, suggest that NSS clusters closely related to motor function would be the ones better suited for genetic analyses. However, further heritability studies are warranted in order to obtain better estimates of the heritability of not only the total NSS but also the different clusters of signs.

According to the evidence mentioned in the previous sections, NSS 1) behave as a trait marker within the schizophrenia phenotype, 2) have a close relationship with some of the cardinal features of the disorder, 3) may have an important predictive value regarding functional outcomes, and 4) are closely related to developmental events leading to the disease onset. Therefore, understanding how genetic factors determine the emergence of NSS might provide clues to specific genetic factors that contribute to the risk of schizophrenia in the context of an altered neurodevelopment. In this sense, the potential interest of NSS as candidate endophenotypes in schizophrenia has been thoroughly discussed in a review by Chan and Gottesman (41). These authors concluded that NSS meet most of the criteria required for an endophenotype to be useful, although they acknowledged that further research is required in order to confirm their potential as suitable phenotypes for genetic analyses.

Molecular genetics and NSS

Despite the interest of NSS in a genetic context, few studies have attempted to associate them with specific genetic variants in genes of interest. The scarce literature based on candidate-gene approaches, however, has reported some results of interest. The first of these studies analysed a single nucleotide polymorphism (SNP) in the serotonin 2-A receptor gene (HTR2A, chr13q14.2) in a sample of Han Chinese population (121). This study found a trend for association between this genetic variant (T102C) and lower NSS scores related to motor coordination in schizophrenia. Interestingly, very similar results have been obtained in a more recent study based on a Caucasian sample (122). In this study, the T102C polymorphism of the HTR2A gene was also associated with a lower degree of NSS related to motor coordination in a sample of schizophrenia patients. Another study based on the catechol-O-methyltransferase gene (COMT, chr22q11.21) analysed the influence on NSS of another SNP which modifies the structure and function of the resulting protein (Val¹⁵⁸Met) (123). This study, based on a Caucasian population, found an association of Val¹⁵⁸Met polymorphism with cognitive and motor deficits in patients with schizophrenia. Another study based on general healthy population reported the effect apolipoprotein E gene (APOE, chr19q13.32) variants have on NSS (124). This study showed the association of APOE4 variant with NSS and its interest to identify subjects at risk of cognitive decline in later life.

NSS in the post-genomics era: current and future challenges

Genetic studies of complex traits like psychiatric phenotypes have experienced a revolution in the past few years as a result of the technological advance in high-throughput techniques for genomic analysis (5). In this regard GWAS have significantly contributed to the understanding of the genetic architecture of many complex traits (see GWAS Catalog www.ebi.ac.uk/gwas/).

Neuropsychiatric disorders are not an exception, and GWAS have currently identified more than 100 loci associated with an increased risk of schizophrenia (6). Moreover, these studies have shed light on the polygenic nature of this disorder in which, as in many other complex traits such as height (125), body mass index (126), or multiple sclerosis (127), the contribution of each common genetic variant to the phenotype is small but the global effect of many of these variants accounts for an important fraction of the phenotype.

To our knowledge, only one study has analysed the role of specific genetic variants identified by GWAS on NSS. In this study, a genetic variant (rs1344706) in the ZNF804A gene (chr2q32.1) has been found to be associated with NSS, although results are difficult to interpret since the schizophrenia risk allele was associated with the presence of fewer NSS (128). Such an approach can be useful to unmask the role of common genetic variants in both the risk of schizophrenia and the emergence of these neurological anomalies.

The possibility of carrying out genome-wide association studies with NSS as the target phenotype might pave the way for a better understanding of the genetic architecture of NSS. Since small genetic effects are expected despite the use of such a potential endophenotype, large samples comprehensively characterised for their neurological profile, ideally using the same instruments for the assessment, would be needed in order to gain enough statistical power. Likewise, this approach should include not only patients with schizophrenia but also healthy relatives and individuals from the general population in order to determine if the potential genetic associations are diagnosis specific or, on the contrary, such genetic effects influence the expression of NSS regardless of the diagnostic status. Moreover, equally assessed subjects at genetic risk (high-risk subjects) at different developmental times should also be included in order to ascertain the role of genetic variants in premorbid stages of the disease. Last but not least, the potential interactions between a genetic risk background with environmental factors acting in early developmental stages such as, for example, obstetric complications or prenatal infections should also be considered.

If the difficulties of collecting a sample with these characteristics could be overcome, GWAS analyses using NSS as the target phenotype might help elucidating the role of genetic factors intimately related to the processes leading to an altered neurodevelopment in patients with schizophrenia. This could consequently provide a better understanding of the aetiology of schizophrenia in the context of a neurodevelopmental model, despite the fact that NSS cannot be considered as predictors of the risk of schizophrenia due to the lack of specificity. Likewise, it could enhance the identification of genetic factors closely related to disease features involving psychopathology, cognition, or functional outcome, finally finding paths from genes to the clinical subphenotypes that have been elusive so far.

Conflict of interest and funding

The authors report no conflict of interest. This research was funded by the Deutsche Forschungsgemeinschaft (DFG) grants: Klinische Forschergruppe (KFO) 241: ‘Genotype-phenotype relationships and neurobiology of the longitudinal course of psychosis’ TP1 (SCHU 1603/5-1) and PsyCourse Project 1: ‘Complex clinical, neurobiological, and molecular signatures of the longitudinal course of psychosis: leveraging comprehensive phenotyping, novel machine learning, and (epi)genomic approaches’ (SCHU 1603/7-1). Thomas G. Schulze was supported by the Lisa-Oehler-Foundation. This work was also supported by the Spanish Ministry of Economy and Competitivity, Instituto de Salud Carlos III (PI15/01420) – co-financed by Fondo Europeo de Desarrollo Regional (FEDER) ‘Una manera de hacer Europa’. Thanks to CIBERSAM and to Comissionat per a Universitats i Recerca del DIUE (2014SGR1636).