A core deficit in schizophrenia is problems with behavioural inhibition. Behavioural inhibition has been defined as the ability to utilise weaker sources of information or select less likely behaviour in the presence of a competing behavioural tendency (Miyake et al., 2000). Research has previously focused on tasks such as the Stroop task to measure behavioural inhibition, with schizophrenic individuals showing perturbed reaction times (RTs). However, there are significant limitations of schizophrenic samples in research. Here, we drew on schizotypal traits as measured by the Oxford-Liverpool Inventory of Feelings and Emotions (O-LIFE; Mason, Claridge & Jackson, 1995). Schizotypal traits in normal individuals are thought to form part of the continuum with schizophrenia, acknowledging it is not a binary phenotype (Ettinger et al., 2015; David, 2010). Indeed, individuals with high traits of schizotypy have been reported to similarly display maladaptive behaviours and cognitive deficits at a non-clinical level. We drew on a recently developed task from Daffron & Davis (2015), which has suggested that individuals use templates for rejection (TFR) when inhibiting attention to stimuli on the basis of high-level semantic features, compared to low-level visual features when searching using top-down target templates. As individuals with high traits of schizotypy have previously shown cognitive deficits, including problems with behavioural inhibition, the current study investigated whether such individuals would have trouble recruiting these high-level cognitive abilities, resulting in slowed ability to inhibit their visual attention in a computer-based visual search task. Consistent with previous research (e.g. Daffron & Davis, 2015), we found a significant difference between the find and ignore tasks, which provides promising reliability for this experimental paradigm. No significant difference between observers with high traits of schizotypy compared to observers with low traits of schizotypy in RTs was found, nor was a significant difference found for errors. Furthermore, error data was interpreted with caution due to assumption violations and a high accuracy rate across both groups. We suggest that future research could involve adjusting the difficulty of the task to potentially “scale up” any differences between those with high traits compared to those with low traits of schizotypy. This experimental procedure holds promise for future research.
Cognitive deficits, including inhibitory failures, have been identified as a core feature of schizophrenia since Bleuer (1950). More recent research from Miyake and colleagues (2000) placed behavioural inhibition as a key component of executive functions with behavioural inhibition representing a central subcomponent of executive functions. It has been defined by Miyake and colleagues as the ability to utilise weaker sources of information or select less likely behaviour in the presence of a competing behavioural tendency (Miyake et al., 2000). Indeed, meta-analyses of the schizophrenia literature show that executive functions per se are substantially impaired (Fioravantia et al., 2005) which also includes problems with working memory; the ability to transiently maintain and manipulate information internally (Forbes et al., 2009). In support of this research, Lewis and colleagues noted a disruption of inhibitor interneurons, leading to cortical disinhibition in schizophrenia neuropathology (Lewis et al., 2005).
An understanding of these cognitive deficits which include behavioural inhibition and working memory problems has implications for the prognosis of these individuals. Indeed, the level of cognitive functioning has been shown to be an important predictor for remission and functional outcome (Harvey et al., 1998). Obtaining a good understanding of the nature of these cognitive deficits associated with schizophrenia will likely optimise rehabilitation and intervention programs. The idea of cognitive functioning being a reliable predictor for the patients’ recovery has also led to considerations that cognitive remediation therapy can be applied (McGurk et al., 2007), improving the functional outcome for these individuals. However, there are confounds with using patients with schizophrenia in research. Namely, the role of anti-psychotic medication and discrepancies with treatment status presenting as confounding variables. For example, Starc and colleagues (2016) noted their findings be taken with caution as they could not definitively rule out the effects of anti-psychotic medication or more long-term poly-pharmacy (Starc et al., 2016). This inevitably leads to difficulties when interpreting the schizophrenia literature.
Conceptualisations of schizotypy within the normal population have been proposed as an alternative to using patients with schizophrenia in research. Schizotypy refers to a set of personality traits thought to reflect the subclinical expression of the signs and symptoms of schizophrenia (Raine, 2006) within the normal population, and is widely used to inform models of psychosis. This falls in line with the notion of schizophrenia being understood as a spectrum rather than a binary phenotype (David, 2010) due to the vast range in symptomatology and prognosis across patients. Consensus is now that schizotypy reduces to three components viz. the positive, negative and disorganised dimensions which correspond well the three-factor model of schizophrenia (Vollema & Hoijtink, 2000).
Methodological advantages of using a schizotypy sample include the absence of confounds in schizophrenia patient research such as long-term pharmacological treatment or chronicity effects. Furthermore, as schizotypy occurs within the normal population, research is able to recruit a relatively inexpensive, reliable, and available sample. Schizotypy also importantly differs from other models of schizophrenia, such as ketamine, sleep deprivation or animal models, in that it represents a better model of the neurodevelopmental aspects of schizophrenia (Ettinger et al., 2015).
Previous research (e.g. Claridge, 1997) has made distinctions between quasi-dimensional and fully dimensional models of schizotypy. The fully dimensional model emerged out of personality theory and regards psychotic traits as no different from other individual differences that potentially have either healthy or unhealthy outcomes. Current opinion (e.g. Mason & Claridge, 2006) is weighted in favour of the fully dimensional model. Our research supports this perspective.
The Oxford-Liverpool Inventory of Feelings and Emotions (O-LIFE) was introduced in 1995 as a four-scale questionnaire measuring psychosis-proneness, principally schizotypy (Mason, Claridge & Jackson, 1995). Its items were chosen to make it suitable for tapping psychotic characteristics in healthy individuals. O-LIFE has been used in a variety of studies across many research domains, establishing its construct validity as a measure of schizotypal traits. A recent large quantitative genetic analysis has also established convincing heritability for the O-LIFE scales (Liney et al., 2003). O-LIFE was designed to have four scales; Unusual Experiences (UnEx), Cognitive Disorganisation (CogDis), Introvertive Anhedonia (IntAn), and Impulsive Nonconformity (ImpNon). All four scales have established good test-retest reliability at around .70 (Burch et al. 1988).
The UnEx scale contains items describing perceptual aberrations, magical thinking and hallucinations. UnEx was of particular interest for our research questions, as it is phenomenologically related to positive symptomatology in schizophrenia. Individuals with positive symptomatology report greater problems of inhibition and suppression of thoughts. Characterising this in schizotypy samples would therefore create the best chance of a comprehensive spectrum between schizotypy traits and schizophrenia in terms of the common deficits seen with behavioural inhibition in schizophrenia research.
The scale CogDis taps aspects of poor attention and concentration as well as poor decision-making and social anxiety and can be seen to reflect thought disorder and disorganised aspects of psychosis. IntAn contains items that describe a lack of enjoyment from social and physical sources of pleasure, as well as avoidance of intimacy and can be seen to reflect weakened forms of negative symptoms (Mason, Linney & Claridge, 2005). The fourth scale, ImpNon, was not of particular importance in terms of our research interests, based on recent debate from Pickering (2004). Pickering argued that ImpNon is not a true schizotypal factor because it does not reflect cognitions or behaviours found in schizophrenia. Cochrane, Petch and Pickering (2010), in a review of the O-LIFE questionnaire, also questioned the acceptability of the ImpNon scale, as scores were not elevated in samples of individuals with schizophrenia as would be expected if schizotypy overlaps with schizophrenia, suggesting limited validity of this scale. We thus thought it would be acceptable to limit our categorisation of individuals to low and high traits of schizotypy based primarily on UnEx, which inevitably also yielded equivalently high or low CogDis, and IntAn scores.
Attentional disturbance has been emphasised as the “primary expression of the schizophrenic brain” (Heinrichs, 2005). Furthermore, it has been proposed that schizophrenia may compromise the efficiency and integrity of neural circuits underlying attentional selection which may well be expressed in widespread cognitive deficits (Ungar et al., 2010). Undoubtedly, the phenomenological similarity of schizotypy with schizophrenia has implications for our understanding of schizophrenia. High traits of schizotypy, as measured by O-LIFE, implicate cognitive deficits in behavioural inhibition and selective and sustained attention (Ettinger et al., 2015) parallel to the deficits observed in schizophrenia patients. This translates into maladaptive behaviours, such as smoking and drug use. These individuals also suffer lower social, educational, and professional levels of functions and high levels of distress (Cohen & Davis, 2009).
A thorough description of high schizotypy will also help address with more confidence the question of the existence and nature of an overlap between schizotypy and schizophrenia (Nelson et al., 2013). Knowledge of different patterns of cognition could inform specific understanding of the distribution of schizophrenia spectrum phenotypes across the population, without resorting to, as Ettinger and colleagues term it, “an arbitrary reference of a spectrum” between healthy individuals and individuals with schizophrenia (Ettinger et al., 2015). Such research is particularly important as it will aid the development of intervention strategies for at risk individuals with high schizotypy. Although these individuals rarely develop a clinical diagnosis of schizophrenia, it is important for this minority that do, and will likely extend to therapeutic interventions for the schizophrenia phenotype.
As noted, individuals with high levels of schizotypy have been shown to display sustained attention deficits. For example, impaired early visual processing has been associated with overall elevated schizotypal traits including abnormal P1 (Koychev et al., 2010) and depth perception (Barbato et al., 2012). This early visual processing dysfunction has also been proposed to cascade to all stages of information processing. For example, abnormal P1 presents higher-function encoding of stimuli, leading to potential working memory problems (Koychev et al., 2010) which fall in line with the working memory problems observed in schizophrenia patients (Forbes et al., 2009). These visual abnormalities have also been implicated with the level of cognitive disorganisation (Cappe et al., 2012) and reported positive symptoms (Richardson & Gruzelier, 1994). This has been supported by functional imaging research (fMRI), with individuals that report high schizotypy scores on the O-LIFE scale (Mason, Linney & Claridge, 2005) showing lower blood-oxygen level dependent (BOLD) signal than controls with low O-LIFE scores in occipital areas known to be associated with early sensory and attentional processing.
A frequently applied paradigm considered to assess behavioural inhibition is the Stroop task (Stroop, 1935). The Stroop task is a well-studied measure of selective attention and has been used across many patient groups. RT scores have been interpreted as an inverse indicator of behavioural inhibition. Individuals with deficits in inhibition should result in an increased interference. We theorise that individuals with high schizotypy traits should fall into this category. Indeed, Wapner and Krus (1960) were the first to show that patients with schizophrenia needed over-proportionally more time in the interference condition than controls. Meta-analyses (e.g. Westerhausen et al., 2011, 2013) have also indicated that the reported deficits in executive functioning found in schizophrenia are at least partly due to a reduced ability of cognitive inhibition.
One approach to the Stroop paradigm would be to focus on individual differences, for example, how different symptoms of schizophrenia or subgroups of individuals (i.e. low versus high traits of schizotypy) affect performance. For example, Buchanan and colleagues (1994) have addressed the questions on how schizophrenia patients with positive versus negative symptoms perform in the Stroop task. A systematic analysis of the effect of these variables would provide valuable information about how much the impairment in cognitive inhibition can be considered part of the spectrum from schizotypy to schizophrenia (Westerhausen et al., 2011). Such work will not only elucidate the cognitive processing patterns of individuals with high expressions of schizotypy traits, but will also inform aetiological theories of schizophrenia.
The cognitive deficits that have been explored in schizophrenia and schizotypy research implicate high-level cognitive abilities. The Stroop task, on the other hand, is based on low level visual features i.e. colour, making it a less desirable task to assess the high-level cognitive problems thought to be implicated in individuals high in traits of schizotypy. Higher-level behavioural inhibition in healthy individuals was investigated by Daffron and Davis (2015, 2016) in a recently developed paradigm designed to provide a more direct and intuitive measure of high-level inhibition. This has presented a promising alternative to previous paradigms such as the Stroop task. The work from Daffron &Davis focused on a key debate in current literature, concerning the ability of individuals with high and low traits of schizotypy to suppression attention to some categories of objects. Many real-world search tasks involve searching for a category of objects, rather than for a specific target image. The researchers therefore took a different approach to visual search, using photorealistic objects against natural backgrounds, increasing ecological validity. This stemmed from recent findings that search is less efficient when two or more distinct categories of targets may be present (Cunningham & Wolfe, 2014). This paradigm relied on the assumption of a “two-category cost”; that, if two types of target objects have overlapping features that distinguish them as useful from other objects (e.g. locks and keys), the observer could use a template that specified this overlap between the two categories and search could be as efficient as when searching for one category only (Daffron & Davis, 2015).
Researchers (Daffron & Davis, 2015) focus on a poorly researched area of visual search; search that proceeds when a target’s properties are not known. They proposed this occurs by rejecting the stimuli known to be irrelevant non-targets, termed templates for rejection (TFR; Woodman & Luck, 2007). Researchers found that when the two non-targets were drawn from physically dissimilar, but semantically related, categories (e.g. locks and keys), no two-template costs were observed. The absence of the two-category cost was theorised to be because some semantic features were presumably shared when the two possible categories of non-targets were semantically related (e.g. locks and keys). Consequent research using the two-template cost in top-down search paradigms when the target item is specified has found however, that semantic knowledge did not influence the guidance of attention. Observers encountered a two-template cost. This fell in line with models of top-down search, such as Wolfe’s Guided Search Model (Wolfe, 2007; Wolfe et al., 2011). This model posits that target templates bias attention toward targets based on visual features (Wolfe et al., 2011) such that sets of visual features guide search, while semantic features, as a secondary ability to be recruited, can contribute to the speed of recognition, if necessary for objects that are more complex.
This paradigm presented an ideal opportunity to establish any deficits in high-level processing in individuals with high traits of schizotypy compared to individuals low on traits of schizotypy. Because TFR recruits higher cognitive processes like semantic knowledge, we hypothesised that observers with high traits of schizotypy, compared to those with low traits of schizotypy, would find this more difficult and display slowed RTs. This two-template cost paradigm across top-down and bottom-up visual attention seemed more appropriate than previous research that has relied on paradigms such as the Stroop task, as TFR allows individuals to recruit these high-level cognitive skills, such as behavioural inhibition, compared to low-level visual features such as colour.