Symptom trajectories and psychosis onset in a clinical high-risk cohort: the relevance of subthreshold thought disorder

Prediction of psychosis across protocols and risk cohorts using automated language analysis

Language and speech are the primary source of data for psychiatrists to diagnose and treat mental disorders. In psychosis, the very structure of language can be disturbed, including semantic coherence (e.g., derailment and tangentiality) and syntactic complexity (e.g., concreteness). Subtle disturbances in language are evident in schizophrenia even prior to first psychosis onset, during prodromal stages. Using computer-based natural language processing analyses, we previously showed that, among English-speaking clinical (e.g., ultra) high-risk youths, baseline reduction in semantic coherence (the flow of meaning in speech) and in syntactic complexity could predict subsequent psychosis onset with high accuracy. Herein, we aimed to cross-validate these automated linguistic analytic methods in a second larger risk cohort, also English-speaking, and to discriminate speech in psychosis from normal speech. We identified an automated machine-learning speech classifier - comprising decreased semantic coherence, greater variance in that coherence, and reduced usage of possessive pronouns - that had an 83% accuracy in predicting psychosis onset (intra-protocol), a cross-validated accuracy of 79% of psychosis onset prediction in the original risk cohort (cross-protocol), and a 72% accuracy in discriminating the speech of recent-onset psychosis patients from that of healthy individuals. The classifier was highly correlated with previously identified manual linguistic predictors. Our findings support the utility and validity of automated natural language processing methods to characterize disturbances in semantics and syntax across stages of psychotic disorder. The next steps will be to apply these methods in larger risk cohorts to further test reproducibility, also in languages other than English, and identify sources of variability. This technology has the potential to improve prediction of psychosis outcome among at-risk youths and identify linguistic targets for remediation and preventive intervention. More broadly, automated linguistic analysis can be a powerful tool for diagnosis and treatment across neuropsychiatry.

Do fewer males present to clinical high-risk services for psychosis relative to first-episode services?

AIM

A decline in the rate of transition to psychosis in patients presenting with clinical high-risk has been reported in the literature. Several hypotheses have been put forward to explain this decline. In this brief report, we aimed to explore whether the demographic group presenting to clinical high-risk services differs from the 'end-point' population who present with first-episode psychosis (FEP), by focusing on gender.

METHOD

Gender distribution was compared between clinical high-risk (CHR) and FEP using data extracted from published study samples and clinical data from corresponding cohorts within the same catchment area in South London.

RESULTS

The proportion of males was significantly higher in FEP compared to CHR services in the literature describing Europe, Australia and North America and in the clinical cohort from South London.

CONCLUSION

Males are under-represented in existing CHR services in Europe, Australia and North America. This could reflect less willingness in males to seek help if experiencing low-level psychological distress and may be related to the declining transition.

Baseline demographics, clinical features and predictors of conversion among 200 individuals in a longitudinal prospective psychosis-risk cohort

BACKGROUND

DSM-5 proposes an Attenuated Psychosis Syndrome (APS) for further investigation, based upon the Attenuated Positive Symptom Syndrome (APSS) in the Structured Interview for Psychosis-Risk Syndromes (SIPS). SIPS Unusual Thought Content, Disorganized Communication and Total Disorganization scores predicted progression to psychosis in a 2015 NAPLS-2 Consortium report. We sought to independently replicate this in a large single-site high-risk cohort, and identify baseline demographic and clinical predictors beyond current APS/APSS criteria.

METHOD

We prospectively studied 200 participants meeting criteria for both the SIPS APSS and DSM-5 APS. SIPS scores, demographics, family history of psychosis, DSM Axis-I diagnoses, schizotypy, and social and role functioning were assessed at baseline, with follow-up every 3 months for 2 years.

RESULTS

The conversion rate was 30% (n = 60), or 37.7% excluding participants who were followed under 2 years. This rate was stable across time. Conversion time averaged 7.97 months for 60% who developed schizophrenia and 15.68 for other psychoses. Mean conversion age was 20.3 for males and 23.5 for females. Attenuated odd ideas and thought disorder appear to be the positive symptoms which best predict psychosis in a logistic regression. Total negative symptom score, Asian/Pacific Islander and Black/African-American race were also predictive. As no Axis-I diagnosis or schizotypy predicted conversion, the APS is supported as a distinct syndrome. In addition, cannabis use disorder did not increase risk of conversion to psychosis.

CONCLUSIONS

NAPLS SIPS findings were replicated while controlling for clinical and demographic factors, strongly supporting the validity of the SIPS APSS and DSM-5 APS diagnosis.

Formal thought disorder in people at ultra-high risk of psychosis

BACKGROUND

Formal thought disorder is a cardinal feature of psychosis. However, the extent to which formal thought disorder is evident in ultra-high-risk individuals and whether it is linked to the progression to psychosis remains unclear.

AIMS

Examine the severity of formal thought disorder in ultra-high-risk participants and its association with future psychosis.

METHOD

The Thought and Language Index (TLI) was used to assess 24 ultra-high-risk participants, 16 people with first-episode psychosis and 13 healthy controls. Ultra-high-risk individuals were followed up for a mean duration of 7 years (s.d.=1.5) to determine the relationship between formal thought disorder at baseline and transition to psychosis.

RESULTS

TLI scores were significantly greater in the ultra-high-risk group compared with the healthy control group (effect size (ES)=1.2), but lower than in people with first-episode psychosis (ES=0.8). Total and negative TLI scores were higher in ultra-high-risk individuals who developed psychosis, but this was not significant. Combining negative TLI scores with attenuated psychotic symptoms and basic symptoms predicted transition to psychosis (P=0.04; ES=1.04).

CONCLUSIONS

TLI is beneficial in evaluating formal thought disorder in ultra-high-risk participants, and complements existing instruments for the evaluation of psychopathology in this group.

DECLARATION OF INTERESTS

None.

COPYRIGHT AND USAGE

© The Royal College of Psychiatrists 2017. This is an open access article distributed under the terms of the Creative Commons Non-Commercial, No Derivatives (CC BY-NC-ND) license.

Prediction of transition to psychosis in patients with a clinical high risk for psychosis: a systematic review of methodology and reporting

BACKGROUND

To enhance indicated prevention in patients with a clinical high risk (CHR) for psychosis, recent research efforts have been increasingly directed towards estimating the risk of developing psychosis on an individual level using multivariable clinical prediction models. The aim of this study was to systematically review the methodological quality and reporting of studies developing or validating such models.

METHOD

A systematic literature search was carried out (up to 14 March 2016) to find all studies that developed or validated a clinical prediction model predicting the transition to psychosis in CHR patients. Data were extracted using a comprehensive item list which was based on current methodological recommendations.

RESULTS

A total of 91 studies met the inclusion criteria. None of the retrieved studies performed a true external validation of an existing model. Only three studies (3.5%) had an event per variable ratio of at least 10, which is the recommended minimum to avoid overfitting. Internal validation was performed in only 14 studies (15%) and seven of these used biased internal validation strategies. Other frequently observed modeling approaches not recommended by methodologists included univariable screening of candidate predictors, stepwise variable selection, categorization of continuous variables, and poor handling and reporting of missing data.

CONCLUSIONS

Our systematic review revealed that poor methods and reporting are widespread in prediction of psychosis research. Since most studies relied on small sample sizes, did not perform internal or external cross-validation, and used poor model development strategies, most published models are probably overfitted and their reported predictive accuracy is likely to be overoptimistic.

Improving Prognostic Accuracy in Subjects at Clinical High Risk for Psychosis: Systematic Review of Predictive Models and Meta-analytical Sequential Testing Simulation

Discriminating subjects at clinical high risk (CHR) for psychosis who will develop psychosis from those who will not is a prerequisite for preventive treatments. However, it is not yet possible to make any personalized prediction of psychosis onset relying only on the initial clinical baseline assessment. Here, we first present a systematic review of prognostic accuracy parameters of predictive modeling studies using clinical, biological, neurocognitive, environmental, and combinations of predictors. In a second step, we performed statistical simulations to test different probabilistic sequential 3-stage testing strategies aimed at improving prognostic accuracy on top of the clinical baseline assessment. The systematic review revealed that the best environmental predictive model yielded a modest positive predictive value (PPV) (63%). Conversely, the best predictive models in other domains (clinical, biological, neurocognitive, and combined models) yielded PPVs of above 82%. Using only data from validated models, 3-stage simulations showed that the highest PPV was achieved by sequentially using a combined (clinical + electroencephalography), then structural magnetic resonance imaging and then a blood markers model. Specifically, PPV was estimated to be 98% (number needed to treat, NNT = 2) for an individual with 3 positive sequential tests, 71%-82% (NNT = 3) with 2 positive tests, 12%-21% (NNT = 11-18) with 1 positive test, and 1% (NNT = 219) for an individual with no positive tests. This work suggests that sequentially testing CHR subjects with predictive models across multiple domains may substantially improve psychosis prediction following the initial CHR assessment. Multistage sequential testing may allow individual risk stratification of CHR individuals and optimize the prediction of psychosis.

Insights into psychosis risk from leukocyte microRNA expression

Dysregulation of immune system functions has been implicated in schizophrenia, suggesting that immune cells may be involved in the development of the disorder. With the goal of a biomarker assay for psychosis risk, we performed small RNA sequencing on RNA isolated from circulating immune cells. We compared baseline microRNA (miRNA) expression for persons who were unaffected (n=27) or who, over a subsequent 2-year period, were at clinical high risk but did not progress to psychosis (n=37), or were at high risk and did progress to psychosis (n=30). A greedy algorithm process led to selection of five miRNAs that when summed with +1 weights distinguished progressed from nonprogressed subjects with an area under the receiver operating characteristic curve of 0.86. Of the five, miR-941 is human-specific with incompletely understood functions, but the other four are prominent in multiple immune system pathways. Three of those four are downregulated in progressed vs. nonprogressed subjects (with weight -1 in a classifier function that increases with risk); all three have also been independently reported as downregulated in monocytes from schizophrenia patients vs. unaffected subjects. Importantly, these findings passed stringent randomization tests that minimized the risk of conclusions arising by chance. Regarding miRNA-miRNA correlations over the three groups, progressed subjects were found to have much weaker miRNA orchestration than nonprogressed or unaffected subjects. If independently verified, the leukocytic miRNA biomarker assay might improve accuracy of psychosis high-risk assessments and eventually help rationalize preventative intervention decisions.

Characterizing psychosis risk traits in Africa: A longitudinal study of Kenyan adolescents

The schizophrenia prodrome has not been extensively studied in Africa. Identification of prodromal behavioral symptoms holds promise for early intervention and prevention of disorder onset. Our goal was to investigate schizophrenia risk traits in Kenyan adolescents and identify predictors of psychosis progression. 135 high-risk (HR) and 142 low-risk (LR) adolescents were identified from among secondary school students in Machakos, Kenya, using the structured interview of psychosis-risk syndromes (SIPS) and the Washington early recognition center affectivity and psychosis (WERCAP) screen. Clinical characteristics were compared across groups, and participants followed longitudinally over 0-, 4-, 7-, 14- and 20-months. Potential predictors of psychosis conversion and severity change were studied using multiple regression analyses. More psychiatric comorbidities and increased psychosocial stress were observed in HR compared to LR participants. HR participants also had worse attention and better abstraction. The psychosis conversion rate was 3.8%, with only disorganized communication severity at baseline predicting conversion (p=0.007). Decreasing psychotic symptom severity over the study period was observed in both HR and LR participants. ADHD, bipolar disorder, and major depression diagnoses, as well as poor occupational functioning and avolition were factors relating to lesser improvement in psychosis severity. Our results indicate that psychopathology and disability occur at relatively high rates in Kenyan HR adolescents. Few psychosis conversions may reflect an inadequate time to conversion, warranting longer follow-up studies to clarify risk predictors. Identifying disorganized communication and other risk factors could be useful for developing preventive strategies for HR youth in Kenya.

Emotional Experiences Predict the Conversion of Individuals with Attenuated Psychosis Syndrome to Psychosis: A 6-Month Follow up Study

The present study explored the conversion rate in individuals with Attenuated Psychosis Syndrome (APS) and potential predictor for transition in mainland China. Sixty-three participants identified as APS were followed up 6 months later. The results showed that 17% of individuals with APS converted to full-blown psychosis. The converters exhibited significantly poorer emotional experience and expression than the non-converters at baseline. A further binary logistic regression analysis showed that emotional experience could predict the transition (Wald = 4.18, p = 0.041, 95% CI = 1.04~6.82). The present study suggests an important role of emotional processing in the prediction of the development of full-blown psychosis.

Declining transition rates to psychotic disorder in "ultra-high risk" clients: Investigation of a dilution effect

During recent years, a decrease has been noted in the rate of transition of ultra-high risk (UHR) clients to a psychotic disorder. Although important to the concept of the at-risk mental state, the reasons for this decline remain largely unknown. We investigated the possibility of a 'dilution effect' in contributing to the decline, i.e. if later UHR cohorts present with less severe clinical intake characteristics than earlier cohorts. Firstly, clinical intake characteristics of a large UHR sample (n=397) were compared across baseline year epochs (1995-2006). Characteristics showing significant differences were included in a Cox-regression to examine if they could explain the decline in transition rates. Secondly, because later cohorts show lower transition rates, 'more stringent' UHR-criteria were retrospectively applied to these cohorts (post-2000, n=219), investigating if this resulted in a higher transition rate. Results indicated that earlier cohorts presented with (1) a larger array of attenuated psychotic symptoms, (2) higher ratings on conceptual disorganization (formal thought disorder) and (3) a higher proportion of individuals with trait risk factor (all P<.001). However, these factors could not fully account for the decline in transition rates. Applying more stringent UHR-criteria to the post-2000-subsample did not substantially change the rate of transition. Our study suggests that later UHR cohorts presented with different clinical intake characteristics than earlier cohorts. While this may have contributed to the observed decrease in transition rates to psychosis, it does not appear to fully account for this decline, suggesting other factors have also impacted on transition rates over time.

Severity of thought disorder predicts psychosis in persons at clinical high-risk

BACKGROUND

Improving predictive accuracy is of paramount importance for early detection and prevention of psychosis. We sought a symptom severity classifier that would improve psychosis risk prediction.

METHODS

Subjects were from two cohorts of the North American Prodrome Longitudinal Study. All subjects met Criteria of Psychosis-Risk States. In Cohort-1 (n=296) we developed a classifier that included those items of the Scale of Psychosis-Risk Symptoms that best distinguished subjects who converted to psychosis from nonconverters, with performance initially validated by randomization tests in Cohort-1. Cohort-2 (n=592) served as an independent test set.

RESULTS

We derived 2-Item and 4-Item subscales. Both included unusual thought content and suspiciousness; the latter added reduced ideational richness and difficulties with focus/concentration. The Concordance Index (C-Index), a measure of discrimination, was similar for each subscale across cohorts (4-Item subscale Cohort-2: 0.71, 95% CI=[0.64, 0.77], Cohort-1: 0.74, 95% CI=[0.69, 0.80]; 2-Item subscale Cohort-2: 0.68, 95% CI=[0.3, 0.76], Cohort-1: 0.72, 95% CI=[0.66-0.79]). The 4-Item performed better than the 2-Item subscale in 742/1000 random selections of 80% subsets of Cohort-2 subjects (p-value=1.3E-55). Subscale calibration between cohorts was proportional (higher scores/lower survival), but absolute conversion risk predicted from Cohort-1 was higher than that observed in Cohort-2, reflecting the cohorts' differences in 2-year conversion rates (Cohort-2: 0.16, 95% CI=[0.13, 0.19]; Cohort-1: 0.30, 95% CI=[0.24, 0.36]).

CONCLUSION

Severity of unusual thought content, suspiciousness, reduced ideational richness, and difficulty with focus/concentration informed psychosis risk prediction. Scales based on these symptoms may have utility in research and, assuming further validation, eventual clinical applications.

Twelve-month psychosis-predictive value of the ultra-high risk criteria in children and adolescents

OBJECTIVE

The validity of current ultra-high risk (UHR) criteria is under-examined in help-seeking minors, particularly, in children below the age of 12 years. Thus, the present study investigated predictors of one-year outcome in children and adolescents (CAD) with UHR status.

METHOD

Thirty-five children and adolescents (age 9-17 years) meeting UHR criteria according to the Structured Interview for Psychosis-Risk Syndromes were followed-up for 12 months. Regression analyses were employed to detect baseline predictors of conversion to psychosis and of outcome of non-converters (remission and persistence of UHR versus conversion).

RESULTS

At one-year follow-up, 20% of patients had developed schizophrenia, 25.7% had remitted from their UHR status that, consequently, had persisted in 54.3%. No patient had fully remitted from mental disorders, even if UHR status was not maintained. Conversion was best predicted by any transient psychotic symptom and a disorganized communication score. No prediction model for outcome beyond conversion was identified.

CONCLUSIONS

Our findings provide the first evidence for the predictive utility of UHR criteria in CAD in terms of brief intermittent psychotic symptoms (BIPS) when accompanied by signs of cognitive impairment, i.e. disorganized communication. However, because attenuated psychotic symptoms (APS) related to thought content and perception were indicative of non-conversion at 1-year follow-up, their use in early detection of psychosis in CAD needs further study. Overall, the need for more in-depth studies into developmental peculiarities in the early detection and treatment of psychoses with an onset of illness in childhood and early adolescence was further highlighted.

Emotion recognition deficits as predictors of transition in individuals at clinical high risk for schizophrenia: a neurodevelopmental perspective

BACKGROUND

Schizophrenia is characterized by profound and disabling deficits in the ability to recognize emotion in facial expression and tone of voice. Although these deficits are well documented in established schizophrenia using recently validated tasks, their predictive utility in at-risk populations has not been formally evaluated.

METHOD

The Penn Emotion Recognition and Discrimination tasks, and recently developed measures of auditory emotion recognition, were administered to 49 clinical high-risk subjects prospectively followed for 2 years for schizophrenia outcome, and 31 healthy controls, and a developmental cohort of 43 individuals aged 7-26 years. Deficit in emotion recognition in at-risk subjects was compared with deficit in established schizophrenia, and with normal neurocognitive growth curves from childhood to early adulthood.

RESULTS

Deficits in emotion recognition significantly distinguished at-risk patients who transitioned to schizophrenia. By contrast, more general neurocognitive measures, such as attention vigilance or processing speed, were non-predictive. The best classification model for schizophrenia onset included both face emotion processing and negative symptoms, with accuracy of 96%, and area under the receiver-operating characteristic curve of 0.99. In a parallel developmental study, emotion recognition abilities were found to reach maturity prior to traditional age of risk for schizophrenia, suggesting they may serve as objective markers of early developmental insult.

CONCLUSIONS

Profound deficits in emotion recognition exist in at-risk patients prior to schizophrenia onset. They may serve as an index of early developmental insult, and represent an effective target for early identification and remediation. Future studies investigating emotion recognition deficits at both mechanistic and predictive levels are strongly encouraged.

Automated analysis of free speech predicts psychosis onset in high-risk youths

BACKGROUND/OBJECTIVES

Psychiatry lacks the objective clinical tests routinely used in other specializations. Novel computerized methods to characterize complex behaviors such as speech could be used to identify and predict psychiatric illness in individuals.

AIMS

In this proof-of-principle study, our aim was to test automated speech analyses combined with Machine Learning to predict later psychosis onset in youths at clinical high-risk (CHR) for psychosis.

METHODS

Thirty-four CHR youths (11 females) had baseline interviews and were assessed quarterly for up to 2.5 years; five transitioned to psychosis. Using automated analysis, transcripts of interviews were evaluated for semantic and syntactic features predicting later psychosis onset. Speech features were fed into a convex hull classification algorithm with leave-one-subject-out cross-validation to assess their predictive value for psychosis outcome. The canonical correlation between the speech features and prodromal symptom ratings was computed.

RESULTS

Derived speech features included a Latent Semantic Analysis measure of semantic coherence and two syntactic markers of speech complexity: maximum phrase length and use of determiners (e.g., which). These speech features predicted later psychosis development with 100% accuracy, outperforming classification from clinical interviews. Speech features were significantly correlated with prodromal symptoms.

CONCLUSIONS

Findings support the utility of automated speech analysis to measure subtle, clinically relevant mental state changes in emergent psychosis. Recent developments in computer science, including natural language processing, could provide the foundation for future development of objective clinical tests for psychiatry.