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.

Psychosocial stress is included in most etiologic models of schizophrenia, frequently as a precipitating factor for psychosis in vulnerable individuals. Nonetheless, the stress-diathesis model has not been tested prospectively in prodromal patients as a predictor of psychosis. The biological effects of stress are mediated by the hypothalamic-pituitary-adrenal (HPA) axis, which governs the release of steroids, including cortisol. The past few decades have witnessed an increased understanding of the neural effects of stress and cortisol, including both normal and abnormal diatheses. As few biological markers have been evaluated as risk factors for psychosis in prodromal patients, the HPA axis and its interaction with intervening life events are apt candidates for study. In this article, we review the HPA axis and its neural effects, present a model for how stress might precipitate psychosis in vulnerable individuals, review the empirical evidence of a link between stress and schizophrenia symptoms, and propose a research design and appropriate statistical models to test the stress-diathesis model for psychosis onset in prodromal patients.

The frontal cortex, basal ganglia, amygdala, and hippocampal formation—all brain areas implicated in schizophrenia¹—are subdivided into functionally distinct subregions (Figure 1). Each subregion is unique in its molecular expression profile² so that select regions are differentially vulnerable to mechanisms of disease.³ Nevertheless, pinpointing regions differentially targeted by schizophrenia is difficult. By the time patients are clinically diagnosed with schizophrenia, multiple brain sites are affected,⁴ making the task of dissociating primary from secondary sites of dysfunction challenging. Further confounding the problem is the absence of pathognomonic histological markers or florid neurodegeneration.⁴ Thus, because schizophrenia is primarily a disease of neuronal dysfunction, not neuronal cell death, cross-sectional comparisons of postmortem samples do not necessarily reveal patterns of anatomical progression.

Schizophrenia has been linked with intrauterine exposure to maternal stress due to bereavement, famine and major disasters. Recent evidence suggests that human vulnerability may be greatest in the first trimester of gestation and rodent experiments suggest sex specificity. We aimed to describe the consequence of an acute maternal stress, through a follow-up of offspring whose mothers were pregnant during the Arab-Israeli war of 1967. A priori, we focused on gestational month and offspring's sex. Method In a pilot study linking birth records to Israel's Psychiatric Registry, we analyzed data from a cohort of 88,829 born in Jerusalem in 1964–76. Proportional hazards models were used to estimate the relative risk (RR) of schizophrenia, according to month of birth, gender and other variables, while controlling for father's age and other potential confounders. Other causes of hospitalized psychiatric morbidity (grouped together) were analyzed for comparison. Results There was a raised incidence of schizophrenia for those who were in the second month of fetal life in June 1967 (RR = 2.3, 1.1–4.7), seen more in females (4.3, 1.7–10.7) than in males (1.2, 0.4–3.8). Results were not explained by secular or seasonal variations, altered birth weight or gestational age. For other conditions, RRs were increased in offspring who had been in the third month of fetal life in June 1967 (2.5, 1.2–5.2), also seen more in females (3.6, 1.3–9.7) than males (1.8, 0.6–5.2). Conclusion These findings add to a growing literature, in experimental animals and humans, attributing long term consequences for offspring of maternal gestational stress. They suggest both a sex-specificity and a relatively short gestational time-window for gestational effects on vulnerability to schizophrenia.