Early and broadly defined psychosis risk mental states

Abstract

Current definitions of the prodromal (or at-risk mental state) phase of schizophrenia include attenuated and/or transient psychotic symptoms as well as a combination of different risk indicators and a recent significant deterioration in global functioning. Data accumulated to date suggest rates of conversion to frank psychosis within two years in 25 to 40% of cases supporting the validity of these criteria. However, at this late phase of illness, functional deterioration is often already pronounced, highlighting the need for earlier identification. Moreover, negative symptoms and social impairments, cognitive deficits, other non-psychotic psychopathology and/or functional decline and non-specific biological indicators, often can be detected well before the at-risk mental state as currently defined; indicating that a broad characterization of an earlier stage may be possible. Identifying specific criteria to define this group of individuals, starting from the framework of familial high risk, can help define a broader group of people, including earlier at-risk mental states, for future research. The hope is that this research will help facilitate intervention at earlier stages that may in turn minimize functional deterioration, and delay, attenuate or even prevent transition to psychosis. The disadvantages as well as the potential benefits of this approach are discussed.

Introduction

In recent years, there has been a surge of interest in studies characterizing the early phases of schizophrenia with the goals of earlier detection and intervention. It has long been recognized that the onset of overt psychosis is often preceded by a phase of variable duration, characterized by sub-threshold positive and negative symptoms and functional decline (Chapman & Chapman, 1987; Hafner et al., 1999; Sullivan, 1994). This very vulnerable stage in which changes in an individual’s functioning are becoming more pronounced leading to later risk for schizophrenia has been named an “at-risk mental state (ARMS)”, an “ultra high risk syndrome”(Yung & McGorry, 1996), a “clinical high risk syndrome”(Cannon et al., 2008) or a “prodromal risk syndrome”(Woods et al., 2009). These terms are used inter-changeably in the literature, as well as in this paper, though the term prodrome can imply inevitability of "conversion" to a psychotic disorder, while the term "at-risk" does not have this connotation. Thus, we will adopt the latter convention. Early studies of this "at-risk" mental state, or prodrome, relied upon retrospective reconstructions of information from the patient and/or other relatives after the onset of psychosis. Some general changes, during a period of months to 2 years, such as a decline in academic performance, disinterest in usual social contacts and emerging odd behaviors have been associated with this stage. However, retrospective studies (Hafner & Maurer, 2006) are biased by the knowledge that the person has schizophrenia and thus any unusual behavior in retrospect could be considered “prodromal”.

More recent conceptualizations of the prodrome have been based on prospective studies of individuals identified as being at risk for first episode psychosis (McGlashan, 2001; Yung & McGorry, 1996). Efforts have therefore been made to define diagnostic criteria for prospective identification such as comprehensive assessment of at-risk mental states (CAARMS) (Yung & McGorry, 1996) and the criteria of prodromal syndromes (COPS) (Miller et al., 2003). Observations that a substantial proportion, i.e. 25–40% of patients defined in this way develop psychotic disorders upon follow-up of up to 2.5 years (Cannon et al., 2008; Klosterkotter et al., 2001; Ruhrmann et al., 2010) support these criteria.

The Melbourne (CAARMS) and Yale (COPS) criteria are not intended to detect all prodromal patients but, rather, prodromal patients at imminent (i.e. within one year) risk of conversion to a schizophrenic psychosis. These “late” risk syndromes defined by the CAARMS and COPS may represent an accelerating trajectory that involves decline in functioning and an increase in the number, frequency, and/or severity of attenuated positive symptoms. In fact, this approach has amplified many-fold the positive predictive power of identifying a person likely to develop schizophrenia using classic familial high-risk approaches (Cannon et al., 2008). However, data suggest that this "late" prodromal phase may be preceded by an earlier period of variable duration characterized by more subtle symptoms and functional impairment that may be termed the "early" at-risk mental states. However, as well be discussed later, an alternative term such as "Broad/ early at-risk mental states" may also be appropriate, by contrast with the prodromal concept as narrowly defined currently. Figure 1 provides a schematic depiction of the premorbid, early, and late at-risk mental states and the psychotic phases that illustrates evolution of functional decline and psychopathology in those at risk for psychotic disorders. It is to be noted from Figure 1 that the narrowly defined “ “late at-risk mental states, as currently operationalized, are characterized by recency (worsening or new onset of symptoms), and higher severity/ frequency of symptoms as well as marked functional decline; by contrast, a broader group of the at-risk individuals will manifest a wider range of symptoms including cognitive deficits negative symptoms and general psychopathology, milder attenuated positive symptoms and/ or less than marked functional declines.

Figure 1.

Schematic diagram to illustrate the early phases of psychotic disorders such as schizophrenia. Note that the type, frequency and severity of positive and negative symptoms and of the functional decline may be used to determine the transitions between early/ broadly defined and late/ narrowly defined at-risk mental states, and the psychotic episode.

Clinical features of the early at-risk mental states

The characterization of the "early" at-risk mental states has to be considered in the context of several caveats. First, it is increasingly recognized that symptoms and syndromes in developmental psychopathology can be continuous or discontinuous (Cicchetti & Rogosch, 2002); thus, depending on a mix of risk and protective factors, symptoms or disorders may progress and remain reliably present over a developmental epoch in some individuals, while for others, there may be offset or a waxing and waning course (Figure 1). Symptoms and cognitive deficits commonly used to identify prodromal patients are often transitory and may not reflect the stable core of developing psychosis (Jahshan et al., 2010; Simon & Umbricht, 2010). Secondly, the pathways to any disorder are likely to be heterogeneous. Early at-risk states may be produced by different developmental pathways, those that may emerge from familial risk, those that may emerge with trait related predispositions such as schizotypy (regardless of family history) and those that are defined mainly by emergent symptomatology (Figure 2). In the field of developmental psychopathology such an outcome is referred to as “equifinality” (Luyten et al., 2008). This view is supported by observations of the substantive variability of premorbid dysfunction, ranging from developmental delays and academic difficulties in early educational years to high achievement through schooling and into college years (Haas & Sweeney, 1992; Schmael et al., 2007). As individuals then enter an early phase of the at-risk mental state, they may have clinical features that indicate heightened risk to psychotic disorders, but do not fulfill all of the so far established “late” prodromal criteria. Third, in identifying an early at-risk syndrome, one needs to consider risk factors (e.g. family history, schizotypy) as well as the risk indicators, such as the presence of a constellation of symptoms, cognitive or neurobiological alterations and/or functional decline. Current definitions do not include aspects of the at-risk state that may not be obvious until formal testing is done, such as cognitive impairment (Seidman et al., 2010). In addition, individuals showing “non-diagnosable” clinical features may already have neurobiological alterations of an evolving disorder, and may often be at considerable distress so that they seek help from professionals; thus, efforts to characterize this population and develop criteria for their reliable identification would be worthwhile. We will briefly summarize the primary existing approaches to identifying early/ broad at- risk mental states involving the “basic symptoms” (Klosterkotter et al., 2001), or a neurodevelopmentally based approach reflecting a cluster of symptoms (Cornblatt et al., 2003).

Figure 2.

Multiple pathways to the sequential emergence of early and late at- risk mental states. Note that familial risk as well as psychometric risk (presence of schizotypal personality disorder) or clinical high risk (presence of sub-threshold symptoms and functional decline may be used to define the heterogeneous at-risk mental states.

Self-related psychotic-like symptoms

Huber and Gross (Huber, 1989), originally described subtle, often only self-experienced deficits, which were reported not only for the post-psychotic stages in which the patients were examined, but also, retrospectively, for the early course, as "basic symptoms”. These include self- perceived thought pressure and blockage, bodily perceptual alterations, disturbances in receptive and expressive speech, decreased ability to distinguish between ideas, perception and true memory, fantasy, perceptual alterations derealization and unstable ideas of reference. Klosterkotter followed up 160 patients with basic symptoms (Klosterkotter et al., 2001). Over a mean period of 9.6 years (and a minimum of 5 years), 79 (49.4%) of the 160 patients had specifically transitioned to schizophrenia. The absence of these symptoms excluded a subsequent schizophrenia with a probability of 96%, suggesting strong predictive value of these symptoms. In addition, the mean time to transition was 5.6 years (median was 4 years), well beyond the imminent scope targeted by existing methods of defining psychosis risk syndromes (e.g., COPS, CAARMS). The symptoms of high predictive value for schizophrenia were thought interference, disturbances of receptive language, and visual distortions. These findings suggest that subtle, transient psychotic-like symptoms may precede the emergence of the “late” at-risk syndrome by several years or more. It is worth noting that in Huber and Gross’s earlier work, it was documented that 37% of patients showed basic symptoms prior to the emergence of psychosis, and an additional 15% had shown the same kind of self-disturbances that spontaneously remitted after their first occurrence, but then recurred prior to progression to a first episode of psychosis (Klosterkotter et al., 2001). These symptoms have been used in recent psychosis prediction studies (Ruhrmann et al., 2010). Recently, a related phenomenological approach has emerged, focusing more theoretically on disorders of the “self” than psychotic symptoms per se, that may also have utility in early prediction when it is applied to longitudinal studies (Sass & Parnas, 2003). These approaches have in common a focus on the inner world as experienced and verbalized by the person at risk for schizophrenia. It is also to be noted that self-reported psychotic experiences are very common in the general population and may reflect developmentally mediated genetic liability to psychosis (Lataster et al., 2009).

Non-psychotic psychopathology

Non-psychotic psychopathologies, notably affective disturbances, have long been reported to be relatively early precursors of schizophrenia. Hafner and colleagues (Hafner et al., 1999), in a systematic, retrospective investigation of the schizophrenia prodrome, reported that depression was the most frequently reported and earliest symptom often preceding the onset of psychosis by 5 years or more (Hafner et al., 1999; Maurer & Hafner, 1995). They suggested that depression is not necessarily a reaction to the symptoms of schizophrenia or a consequence of medication but may be a direct expression of the disease process (Hafner et al., 1999). Prospective studies of familial high-risk relatives (Table 1) have revealed higher rates of non-psychotic psychopathology, particularly affective, anxiety (Amminger et al., 1999) as well as conduct and attention deficit disorders (Keshavan et al., 2008). It is of importance to consider including them in definitions of at-risk mental states since they are frequently help- seeking, and early intervention could potentially mitigate later, more serious psychopathology. In the Edinburgh High Risk study (focusing on people at familial high risk), pre-illness, non-specific anxiety, and affective symptomatology had greater power than most psychotic phenomena in predicting conversion to a psychotic disorder (Owens & Johnstone, 2006). Moreover, social deficits are a pervasive problem before psychotic illness begins, often observable in childhood, and have been supported by a meta-analysis of several studies (Tarbox & Pogue-Geile, 2008). The non-specific nature of general psychopathology and the relative lack of sensitivity to predict schizophrenia are clearly limitations. When used in isolation, these symptoms may precede later evolution of non-psychotic psychopathology, elevating false positive rates, and thereby causing unnecessary anxiety about later developing an illness that carries considerable social stigma. However, some non-psychotic symptoms, especially anxiety and affective disturbances, may have some value in increasing predictive value when used in combination with other features that increase risk for psychosis, such as the basic symptoms listed above, cognitive deficits, and family history of psychotic disorders.

Table 1.

Biobehavioral alterations in young relatives at familial risk for schizophrenia in prospective studies.

	Do they differ between Familiar High Risk for Schizophrenia and Healthy Controls?	Are they potential predictors for later psychosis?	Are they specific for schizophrenia?
Behavioral	Childhood behavioral disturbance, especially social difficulty and anhedonia (Tarbox & Pogue-Geile, 2008) (Glatt et al., 2006); schizotypy (Miller e al., 2002; Miller et al., 2002); childhood premorbid maladjustment (Dworkin et al., 1991)	Childhood behavior predicted later schizophrenia (Miller et al., 2002; Whyte et al., 2006); Schizotypy predicted later psychosis (Miller et al., 2002)	Childhood behavioral disturbances elevated in FHR- BP as well (Giles et al., 2007); Premorbid social competence may predict risk for SZ more than for than for affective disorder (Dworkin et al., 1991)
Cognitive/ neurological	Impaired attention, verbal memory and executive functions, and social cognition (Keshavan et al., 2010), (Seidman et al., 2006); neurological signs (Marcus et al., 1993; McNeil et al., 1993; Prasad & Keshavan, 2008), IQ (Woodberry et al., 2008)	Verbal memory and executive impairment progressed during follow-up (Bhojraj et al., 2010; Cosway et al., 2000); IQ (Woodberry et al., 2008), Predictive value of neurological signs (Fish, 1987)	Impaired response inhibition, may reflect familial predisposition to BP (Frangou et al., 2005). Neuropsychological impairment greater in FHRS compared to FHR-BP (Seidman et al., 2006; Stone et al., 2005)
Brain structure	Increased brain ventricular size (DeLisi et al., 1986); Volume reductions in amygdala and hippocampal régions (Keshavan et al., 2002; Seidman et al., 2003), superior temporal (Rajarethinam et al., 2004), prefrontal (Diwadkar et al., 2006), and caudate (Rajarethinam et al., 2007)	Medial temporal volume loss may predict psychosis (Job, 2005)	FHRS showed frontostriatal and temporal volume loss, but FHR-BP showed cingulate and ventral striatal reductions (McDonald et al., 2004)
In vivo Brain chemistry	Reduced NAA in unaffected offspring (Keshavan et al., 2009) and in unaffected siblings (Callicott et al., 1998)		NAA reductions in FHR-BP in some (Cecil et al., 2003) but not all studies (Gallelli et al., 2005)
Functional imaging studies	Variable findings but impaired PFC, parahippocampus; fMRI evidence of Hyperactivity of the default network in FHRS (MacDonald et al., 2009)	No data available	Not known

Cognitive deficits

Cognitive deficits are observed in youth who later develop schizophrenia, as well as in non-symptomatic children and adolescent relatives at genetic risk for schizophrenia (for reviews see Keshavan (Keshavan et al., 2010) and Seidman (Seidman et al., 2006)). The best-studied dimension of cognition in predicting schizophrenia is intelligence (Woodberry et al., 2008). Across 18 studies carried out in a variety of countries, the effect size impairment of IQ is consistently moderate (Cohen’s d = −.54). Deficits in persons at familial risk are of a similar magnitude (~ .5 SD) across studies (Bertisch et al., 2010; Hoff et al., 2005; Shedlack et al., 1997), and somewhat less severe than those with the “late” at-risk mental states in which the impairments have an approximate effect size of d = −.70 (Seidman et al., 2010). Follow-up of familial high-risk individuals during the critical years of adolescence indicates progression of some of these deficits (Bhojraj et al., 2010; Cosway et al., 2000) suggesting that the illness may begin to evolve, and cognitive decline may set in well before the individuals meet the "late" at-risk mental states criteria.

Evidence for a worsening cognitive disorder is clear, in that the effect size impairments in the first episode of schizophrenia typically average about d=1.0 (Mesholam-Gately et al., 2009). Simon (Simon et al., 2007) investigated cognitive functioning in 93 subjects who met either basic symptom or prodromal (ARMS) criteria and compared them with 43 patients with a first episode of psychosis and to 49 help-seeking patient controls. All groups performed significantly below normative expectations. Both at-risk groups performed at intermediate levels between the first-episode (FE) group and normative values. The ARMS group demonstrated intermediate performance between the FE and BS groups. Overall, auditory working memory, verbal fluency/processing speed, and verbal declarative memory were most impaired. This was also observed in studies by Eastvold and Pukrop (Eastvold et al., 2007; Pukrop & Klosterkotter, 2010). The nature of cognitive deficits may also evolve during the transition from the early to the late phases of the prodrome; Frommann (Frommann et al., 2010) observed that executive control impairment may be seen in the early prodromal phase, while additional memory dysfunction may accrue in the late phases, although there is currently very little research directly demonstrating a decline in function (Jahshan et al., 2010), More studies are needed to evaluate this question. Moreover, cognitive deficits in prodromal patients clearly underlie the functional disability that sets in early in this phase of the illness (Niendam et al., 2006). Taken together, these findings support early intervention efforts to prevent or delay the onset of the “late” prodromal and psychotic phases and thus prevent further cognitive decline. Cognitive impairment is not included as part of the current prodromal criteria though recent data suggest that some cognitive deficits (such as memory impairments) may add to value of prodromal symptoms in predicting time to conversion (Seidman et al., 2010). Further research utilizing cognitive measures as risk predictors is needed in combination with other measures to determine their utility.

Negative symptoms

Negative symptoms refer to deficits that indicate a lack of one or other mental attribute as compared to “positive” symptoms that indicate new active abnormal phenomena or perceptions). Social isolation, a negative symptom, is one of the key manifestations of liability in genetic high-risk adolescents (Dworkin et al., 1994) observed in epidemiological studies of birth cohorts (Tarbox & Pogue-Geile, 2008). This may stem from social anhedonia (Cohen et al., 2006) and impaired affect perception (Eack et al., 2009; Phillips & Seidman, 2008) that characterize young people at risk for schizophrenia. Genetic liability to schizophrenia is well known to be associated with associated with negative symptoms, cognitive impairment and psychosocial dysfunction, a constellation of features collectively termed schizotaxia which may be a treatable condition (Faraone et al., 2001). Despite these observations, the focus of both the Melbourne and the Yale prodromal criteria (CAARMS, COPS) has been on emergent positive symptoms (with the exception of the genetic risk + deterioration syndrome (Miller et al., 2003; Yung & McGorry, 1996). This contrasts with the Recognition and Prevention (RAP) program (Cornblatt et al., 2003) in New York, in which a broader strategy was adopted based on a model of neurodevelopmental dysfunction. This model included the nonspecific symptoms reflecting the core of schizophrenia, i.e. cognitive deficits, affective disturbances, social isolation, and school failure (“CASIS” cluster) to define a group of clinical high-risk persons without positive symptoms (the CHR- group). CHR- group was similar to CHR+ group (those with attenuated psychotic-like symptoms) in functional deficits and conversion rates. Ongoing research will clarify whether persons who begin with these non-specific deficits will evolve into having attenuated psychotic-like symptoms, remit, or remain stably impaired. This perspective also emphasizes the key fact that disability may be significant (and requires treatment) regardless of whether the person transitions to psychosis.

Neurobiological alterations

While neuroanatomical changes have been identified in prodromal individuals, particularly those who convert to psychosis (Pantelis et al., 2003; Sun et al., 2009; Takahashi et al., 2009), accumulating evidence suggests that neuroanatomical, functional and neurochemical alterations are seen in young relatives at risk for schizophrenia well before psychotic or "late" at-risk mental states manifest (see Table 1). Brain structural alterations are highly heritable (Bertisch et al., 2010; Prasad & Keshavan, 2008). Imaging studies of those at familial high risk, such as the Edinburgh High-Risk Study (Johnstone et al., 2005), report gray matter density reductions in medial and lateral temporal lobes, as well as functional abnormalities as the at-risk relatives begin to develop symptoms. Similar findings have been reported from the Pittsburgh familial High Risk study, with alterations in brain structure in at-risk relatives (Diwadkar et al., 2006; Keshavan et al., 2002; Keshavan et al., 2005). Abnormalities in the medial prefrontal cortex and the frontal pole have also been identified in the Harvard Adolescent High Risk Study (Rosso et al., 2010). Job (Job, 2005) has shown progressive alterations in prefrontal and medial temporal regions in young familial high risk relatives during follow-up. Similarly, we (Bhojraj et al., 2010) have recently reported progressive structural brain changes in superior temporal regions in adolescent (mean age 15) high risk relatives, none of whom had met COPS criteria for the prodrome; these alterations, however, correlated with the severity of the prodromal symptoms assessed using the structured interview for prodromal symptoms (SIPS). DeLisi (DeLisi et al., 2006) and Hoptman (Hoptman et al., 2008) have reported evidence of subtle atrophy and white matter changes already present in people at genetic risk for schizophrenia. Nevertheless, these are preliminary group data results and the clinical utility of performing an MRI scan to characterize early at-risk mental states in individuals is yet to be determined.

In vivo Neurochemistry has been examined in young relatives at risk for schizophrenia using Phosphorus (³¹P) and Proton (¹H) Magnetic Resonance spectroscopy (MRS). Adolescents at high genetic risk have ³¹P MRS evidence of altered membrane phospholipid metabolism (Keshavan et al., 2003) and reduced neuronal integrity as measured by N-acetyl aspartate levels in ¹H MRS (Keshavan et al., 2009; Yoo et al., 2009). A recent study suggests altered glutamate levels in the thalamus in young persons at ultra-high risk for psychoses (Stone et al., 2009); reduced thalamic glutamate levels were correlated with medial temporal and insular gray matter volumes. This observation is consistent with the prediction that pre-psychotic phases of schizophrenia might be associated with glutamatergic alterations (Keshavan, 1999), a view that has relevance for early intervention and prevention.

Functional imaging studies have also suggested alterations in the non-psychotic relatives at risk for schizophrenia. Li (Li et al., 2007) using fMRI found language processing deficits in people at genetic high-risk for schizophrenia. Whalley (Whalley et al., 2007) has found an association between altered temporal activity and psychotic symptoms in subjects at high genetic risk of Schizophrenia. Neural substrates underlying working memory and executive function (Barbour et al., 2010; Keshavan et al., 2002; Keshavan et al., 2002; Seidman et al., 2006), and declarative memory (Thermenos et al., 2007) have been found to be abnormal in persons at familial high risk. Whitfield-Gabrieli (Whitfield-Gabrieli et al., 2009) has shown aberrant activity of default state networks in unaffected relatives at risk for schizophrenia; these alterations correlated with schizotypal psychopathology, pointing to the possibility that these functional changes may reflect early markers of pre-psychotic psychopathology.

Taken together, progressive alterations in brain structure, chemistry and function may begin insidiously in at-risk individuals who have not yet developed psychosis or the “late” prodrome, highlighting the importance of systematic studies of the “early” at-risk population. However, how these observations can be standardized to be used widely for early detection needs further investigation.

Toward operationalizing the “early” at-risk mental states

Developing operational definitions for the early at-risk mental states is of both clinical and scientific importance. First, early at-risk mental states might lead to help seeking but may have a different pathway to care. Platz (Platz et al., 2006) observed that while patients with psychotic symptoms more often contacted mental health professionals, patients with insidious and more non-specific features, including those with early features (as assessed by basic symptoms), more frequently contacted general practitioners. Second, as reviewed earlier, there is some evidence of neurobiological progression of the illness related pathophysiology in high-risk relatives most of whom had not developed “late” at-risk criteria. Thus, early treatment may be needed at this stage before progression occurs and the inevitability of developing schizophrenia or other psychotic disorders can be deterred. Third, the earliest stages of psychotic disorders [or psychosis] remain to be characterized for in-depth research, while extant efforts focus on fully symptomatic individuals who are already demonstrating imminent risk. In this sense, the natural history of onset of schizophrenia and related disorders is not fully indexed by our current methods and still not defined well enough to prevent progression.

Two groups have attempted to define the early at-risk mental states in the prior literature. The German Research Network on Schizophrenia (Bechdolf et al., 2005) has proposed a definition of "early initial prodromal states” (EIPS) based on the presence of a) basic symptoms (at least one in the previous 3 months), and/or b) functional decline (>30 points in the past year in the global assessment of functioning scale) and the presence of a risk factor, such as a having a first degree family history of schizophrenia or schizophrenia spectrum disorder. However, these criteria at least partially overlap with the Melbourne and Yale definitions of the "late" prodromal criteria. McGorry has proposed including the early prodrome (stage 1a) in his scheme of staging the early course of schizophrenia (McGorry et al., 2007; Tandon et al., 2009). The early prodrome was defined here as mild or non-specific symptoms (including neurocognitive deficits) of psychosis or severe mood disorder, but no operational criteria were proposed. It is therefore important to develop a more comprehensive set of criteria to fully capture the variations in presentations of the early at-risk mental states.

As stated earlier, the current prodromal criteria (COPS, CAARMS) may only capture the more severe and later aspects of the prodrome as it evolves toward psychosis. Specifically, the Brief Intermittent Psychotic symptom syndrome (BIPS) criteria (Miller et al., 2003; Yung & McGorry, 1996) excludes those with brief but infrequent (< 1x/month) or brief (duration shorter than several minutes or non-recent (onset before the previous 3 months) psychotic symptoms; the Yale attenuated positive symptom syndrome (APS) criteria excludes those with a) mild positive symptoms (e.g. those scoring mild, or 2 on the scale of prodromal symptoms, or SOPS scale) and b) those who may have transient symptoms before the month prior to evaluation, and c) those with no increase in symptoms over the previous year. The Global Assessment of Functioning scale (GAF) changes may not be linear and may dynamically change over time- so restricting the change to “last month” might miss some "true" at risk cases whose functioning dramatically waxes and wanes in relation to everyday stress. Additionally, the current GAF score declines (>/= 30% compared to one year ago) may be too conservative, and may miss individuals with more moderate but still clinically significant functional declines and those whose functional declines may be more gradual (e.g., achieving 30% over two years but not the past year).

Given these considerations, we have proposed (Table 2) a set of research criteria that would capture the “early” or broader at risk states and distinguish it from the “late” prodromal phase. Our approach to these criteria involves the use of a trait (risk) plus state approach; we have focused on the familial risk since genetic factors are among the best established risk factors for schizophrenia. We have focused on these criteria to allow for the more subtle positive symptoms, definitive negative symptoms, general psychopathology to be incorporated, and to capture the more moderate or protracted functional decline. Additionally, the cognitive deficits, which frequently precede the prodrome as currently defined, are included. Since these criteria will represent an at-risk state more distal to psychosis onset than the COPS and CAARMS definitions, and may include some milder manifestations of the at-risk mental states, it is likely that rates of conversion to psychosis will be lower, at least when individuals are studied prospectively over short periods (e.g., 1–2.5 years). However, conversion to psychosis may not be the only relevant clinical endpoint, since functional decline and symptomatic distress clearly precede psychosis and are themselves in need of active intervention (Woods et al., 2010). In addition, understanding progression from the early/broadly defined prodrome to later/ more narrowly defined phases as defined by COPS or CAARMS represents a relevant outcome. Overall, we hope that the proposed criteria will serve a useful purpose for more fully elucidating the neurobiology and clinical course characteristics of the early prodromal phase.

Table 2.

Proposed at risk mental state criteria by contrast with a representative “late” set of criteria (COPS, Miller et al 2003). Early/broad at-risk mental state criteria are met if (1) is present plus any two of the items 2–6. COPS criteria are fulfilled if 1 and 6 (genetic risk- deterioration syndrome), 2 (attenuated positive syndrome or 3 are present (brief intermittent psychotic symptoms). GAF= Global Assessment of Functioning scale. SOPS= scale of positive symptoms.

	Early/ Broad at-risk mental states: proposed operational criteria.	Current “Late” prodromal criteria (COPS)
1. Genetic risk	Having a first degree relative with a psychotic disorder	First degree relative with psychosis or schizotypal PD in patient
2. Attenuated psychotic-like (positive) symptoms	One or more of the attenuated SOPS Positive or disorganized items1 scoring mild, moderate or severe but not at a psychotic level; these may include one or more basic symptoms (Klosterkotter et al 2001) AND Symptoms beginning or increasing one or more points within the past 3 yrs AND Symptoms occurring < once per wk but at least once a month for the past year	One or more of the SOPS positive items scoring moderate to severe AND Symptoms beginning or increasing one or more points within the past year AND Symptoms occurring >once per wk for last month
3. Brief Intermittent Psychotic symptoms	Not considered	One or more of the five SOPS positive items rated at a psychotic level AND Symptoms beginning in the past 3 mos AND Symptoms occurring currently at least several minutes per day at least once per mo
4. Attenuated negative / general psychopathology	Two or more of the SOPS negative/ general symptoms rated at least moderate in severity	Not considered
5. Cognitive impairment	Evidence of executive cognitive impairment (at least 1.0 S.D. deficit relative to age-expected norms on at least 30% of the measures (i.e., 3/10 tests) (Doyle et al., 2000)	Not considered
6. Functional decline	GAF decline > 15 but < 30% over the last 2 years, sustained for > 1 mo	GAF decline > 30% over the last mo vs 1 yr ago

¹. SOPS positive symptoms include unusual thought content/delusional Ideas, suspiciousness/persecutory Ideas, grandiosity, perceptual Abnormalities/hallucinations, and disorganized communication (Miller et al., 2003).

Defining the features of the early and broader range of at-risk mental states is fraught with a number of limitations, that have been well addressed in the clinical high risk literature; these include ethical concerns about labeling (Cornblatt et al., 2001; Filakovic et al., 2007), lower rates of transition to psychoses, high false positive rates (Haroun et al., 2006) and the danger of inappropriately treating something not in need of the specific types of treatment more appropriate for psychotic conditions. For these reasons, the proposed framework is designed to facilitate research, and is far from being considered as part of a diagnostic scheme for clinical use at this time. The term “early prodrome” may imply inevitable progression to a disorder, which clearly does not happen, and therefore an alternative term such as “Early/ Broad at- risk states” (E-BARS) may be more appropriate. Nevertheless, we believe that defining and refining this concept may pave the way toward detection and intervention efforts even earlier than those currently being contemplated. The reliability, validity, and psychometric properties of the E-BARS criteria need to be further investigated.