A machine learning approach to predicting psychosis using semantic density and latent content analysis

Computational analysis of linguistic features in speech samples of first-episode bipolar disorder and psychosis

BACKGROUND

In recent years, automated analyses using novel NLP methods have been used to investigate language abnormalities in schizophrenia. In contrast, only a few studies used automated language analyses in bipolar disorder. To our knowledge, no previous research compared automated language characteristics of first-episode psychosis (FEP) and bipolar disorder (FEBD) using NLP methods.

METHODS

Our study included 53 FEP, 40 FEBD and 50 healthy control participants who are native Turkish speakers. Speech samples of the participants in the Thematic Apperception Test (TAT) underwent automated generic and part-of-speech analyses, as well as sentence-level semantic similarity analysis based on SBERT.

RESULTS

Both FEBD and FEP were associated with the use of shorter sentences and increased sentence-level semantic similarity but less semantic alignment with the TAT pictures. FEP also demonstrated reduced verbosity and syntactic complexity. FEP differed from FEBD in reduced verbosity, decreased first-person singular pronouns, fewer conjunctions, increased semantic similarity as well as shorter sentence and word length. The mean classification accuracy was 82.45 % in FEP vs HC, 71.1 % in FEBD vs HC, and 73 % in FEP vs FEBD. After Bonferroni correction, the severity of negative symptoms in FEP was associated with reduced verbal output and increased 5th percentile of semantic similarity.

LIMITATIONS

The main limitation of this study was the cross-sectional nature.

CONCLUSION

Our findings demonstrate that both patient groups showed language abnormalities, which were more severe and widespread in FEP compared to FEBD. Our results suggest that NLP methods reveal transdiagnostic linguistic abnormalities in FEP and FEBD.

Automated linguistic analysis in youth at clinical high risk for psychosis

Identifying individuals at clinical high risk for psychosis (CHRP) is crucial for preventing psychosis and improving the prognosis for schizophrenia. Individuals at CHR-P may exhibit mild forms of formal thought disorder (FTD), making it possible to identify them using natural language processing (NLP) methods. In this study, speech samples of 62 CHR-P individuals and 45 healthy controls (HCs) were elicited using Thematic Apperception Test images. The evaluation involved various NLP measures such as semantic similarity, generic, and part-of-speech (POS) features. The CHR-P group demonstrated higher sentence-level semantic similarity and reduced mean image-to-text similarity. Regarding generic analysis, they demonstrated reduced verbosity and produced shorter sentences with shorter words. The POS analysis revealed a decrease in the utilization of adverbs, conjunctions, and first-person singular pronouns, alongside an increase in the utilization of adjectives in the CHR-P group compared to HC. In addition, we developed a machine-learning model based on 30 NLP-derived features to distinguish between the CHR-P and HC groups. The model demonstrated an accuracy of 79.6 % and an AUC-ROC of 0.86. Overall, these findings suggest that automated language analysis of speech could provide valuable information for characterizing FTD during the clinical high-risk phase and has the potential to be applied objectively for early intervention for psychosis.

Detecting schizophrenia, bipolar disorder, psychosis vulnerability and major depressive disorder from 5 minutes of online-collected speech

Background

Psychosis poses substantial social and healthcare burdens. The analysis of speech is a promising approach for the diagnosis and monitoring of psychosis, capturing symptoms like thought disorder and flattened affect. Recent advancements in Natural Language Processing (NLP) methodologies enable the automated extraction of informative speech features, which has been leveraged for early psychosis detection and assessment of symptomology. However, critical gaps persist, including the absence of standardized sample collection protocols, small sample sizes, and a lack of multi-illness classification, limiting clinical applicability. Our study aimed to (1) identify an optimal assessment approach for the online and remote collection of speech, in the context of assessing the psychosis spectrum and evaluate whether a fully automated, speech-based machine learning (ML) pipeline can discriminate among different conditions on the schizophrenia-bipolar spectrum (SSD-BD-SPE), help-seeking comparison subjects (MDD), and healthy controls (HC) at varying layers of analysis and diagnostic complexity.

Methods

We adopted online data collection methods to collect 20 minutes of speech and demographic information from individuals. Participants were categorized as "healthy" help-seekers (HC), having a schizophrenia-spectrum disorder (SSD), bipolar disorder (BD), major depressive disorder (MDD), or being on the psychosis spectrum with sub-clinical psychotic experiences (SPE). SPE status was determined based on self-reported clinical diagnosis and responses to the PHQ-8 and PQ-16 screening questionnaires, while other diagnoses were determined based on self-report from participants. Linguistic and paralinguistic features were extracted and ensemble learning algorithms (e.g., XGBoost) were used to train models. A 70%-30% train-test split and 30-fold cross-validation was used to validate the model performance.

Results

The final analysis sample included 1140 individuals and 22,650 minutes of speech. Using 5-minutes of speech, our model could discriminate between HC and those with a serious mental illness (SSD or BD) with 86% accuracy (AUC = 0.91, Recall = 0.7, Precision = 0.98). Furthermore, our model could discern among HC, SPE, BD and SSD groups with 86% accuracy (F1 macro = 0.855, Recall Macro = 0.86, Precision Macro = 0.86). Finally, in a 5-class discrimination task including individuals with MDD, our model had 76% accuracy (F1 macro = 0.757, Recall Macro = 0.758, Precision Macro = 0.766).

Conclusion

Our ML pipeline demonstrated disorder-specific learning, achieving excellent or good accuracy across several classification tasks. We demonstrated that the screening of mental disorders is possible via a fully automated, remote speech assessment pipeline. We tested our model on relatively high number conditions (5 classes) in the literature and in a stratified sample of psychosis spectrum, including HC, SPE, SSD and BD (4 classes). We tested our model on a large sample (N = 1150) and demonstrated best-in-class accuracy with remotely collected speech data in the psychosis spectrum, however, further clinical validation is needed to test the reliability of model performance.

Large language models in psychiatry: Opportunities and challenges

The ability of Large Language Models (LLMs) to analyze and respond to freely written text is causing increasing excitement in the field of psychiatry; the application of such models presents unique opportunities and challenges for psychiatric applications. This review article seeks to offer a comprehensive overview of LLMs in psychiatry, their model architecture, potential use cases, and clinical considerations. LLM frameworks such as ChatGPT/GPT-4 are trained on huge amounts of text data that are sometimes fine-tuned for specific tasks. This opens up a wide range of possible psychiatric applications, such as accurately predicting individual patient risk factors for specific disorders, engaging in therapeutic intervention, and analyzing therapeutic material, to name a few. However, adoption in the psychiatric setting presents many challenges, including inherent limitations and biases in LLMs, concerns about explainability and privacy, and the potential damage resulting from produced misinformation. This review covers potential opportunities and limitations and highlights potential considerations when these models are applied in a real-world psychiatric context.

Natural Language Processing Applied to Spontaneous Recall of Famous Faces Reveals Memory Dysfunction in Temporal Lobe Epilepsy Patients

Objective and Background

Epilepsy patients rank memory problems as their most significant cognitive comorbidity. Current clinical assessments are laborious to administer and score and may not always detect subtle memory decline. The Famous Faces Task (FF) has robustly demonstrated that left temporal lobe epilepsy (LTLE) patients remember fewer names and biographical details compared to right TLE (RTLE) patients and healthy controls (HCs). We adapted the FF task to capture subjects' entire spontaneous spoken recall, then scored responses using manual and natural language processing (NLP) methods. We expected to replicate previous group level differences using spontaneous speech and semi-automated analysis.

Methods

Seventy-three (N=73) adults (28 LTLE, 18 RTLE, and 27 HCs) were included in a case-control prospective study design. Twenty FF in politics, sports, and entertainment (active 2008-2017) were shown to subjects, who were asked if they could recognize and spontaneously recall as much biographical detail as possible. We created human-generated and automatically-generated keyword dictionaries for each celebrity, based on a randomly selected training set of half of the HC transcripts. To control for speech output, we measured the speech duration, total word count and content word count for the FF task and a Cookie Theft Control Task (CTT), in which subjects were merely asked to describe a visual scene. Subjects' responses to FF and CTT tasks were recorded, transcribed, and analyzed in a blinded manner with a combination of manual and automated NLP approaches.

Results

Famous face recognition accuracy was similar between groups. LTLE patients recalled fewer biographical details compared to HCs and RTLEs using both the gold-standard human-generated dictionary (24%±12% vs. 31%±12% and 30%±12%, p=0.007) and the automated dictionary (24%±12% vs. 31%±12% and 32%±13%, p=0.007). There were no group level differences in speech duration, total word count, or content word count for either the FF and CTT to explain difference in recall performance. There was a positive, statistically significant relationship between MOCA score and FF recall performance as scored by the human-generated (ρ= .327, p= .029) and automatically-generated dictionaries (ρ= .422, p= .004) for TLE subjects, but not HCs, an effect that was driven by LTLE subjects.

Discussion

LTLE patients remember fewer details of famous people than HCs or RTLE patients, as discovered by NLP analysis of spontaneous recall. Decreased biographical memory was not due to decreased speech output and correlated with lower MOCA scores. NLP analysis of spontaneous recall can detect memory dysfunction in clinical populations in a semi-automated, objective, and sensitive manner.

Natural Language Processing and Schizophrenia: A Scoping Review of Uses and Challenges

(1) Background: Approximately 1% of the global population is affected by schizophrenia, a disorder marked by cognitive deficits, delusions, hallucinations, and language issues. It is associated with genetic, neurological, and environmental factors, and linked to dopaminergic hyperactivity and neurotransmitter imbalances. Recent research reveals that patients exhibit significant language impairments, such as reduced verbal output and fluency. Advances in machine learning and natural language processing show potential for early diagnosis and personalized treatments, but additional research is required for the practical application and interpretation of such technology. The objective of this study is to explore the applications of natural language processing in patients diagnosed with schizophrenia. (2) Methods: A scoping review was conducted across multiple electronic databases, including Medline, PubMed, Embase, and PsycInfo. The search strategy utilized a combination of text words and subject headings, focusing on schizophrenia and natural language processing. Systematically extracted information included authors, population, primary uses of the natural language processing algorithms, main outcomes, and limitations. The quality of the identified studies was assessed. (3) Results: A total of 516 eligible articles were identified, from which 478 studies were excluded based on the first analysis of titles and abstracts. Of the remaining 38 studies, 18 were selected as part of this scoping review. The following six main uses of natural language processing were identified: diagnostic and predictive modeling, followed by specific linguistic phenomena, speech and communication analysis, social media and online content analysis, clinical and cognitive assessment, and linguistic feature analysis. (4) Conclusions: This review highlights the main uses of natural language processing in the field of schizophrenia and the need for more studies to validate the effectiveness of natural language processing in diagnosing and treating schizophrenia.

Clinical efficacy of pre-trained large language models through the lens of aphasia

The rapid development of large language models (LLMs) motivates us to explore how such state-of-the-art natural language processing systems can inform aphasia research. What kind of language indices can we derive from a pre-trained LLM? How do they differ from or relate to the existing language features in aphasia? To what extent can LLMs serve as an interpretable and effective diagnostic and measurement tool in a clinical context? To investigate these questions, we constructed predictive and correlational models, which utilize mean surprisals from LLMs as predictor variables. Using AphasiaBank archived data, we validated our models' efficacy in aphasia diagnosis, measurement, and prediction. Our finding is that LLMs-surprisals can effectively detect the presence of aphasia and different natures of the disorder, LLMs in conjunction with the existing language indices improve models' efficacy in subtyping aphasia, and LLMs-surprisals can capture common agrammatic deficits at both word and sentence level. Overall, LLMs have potential to advance automatic and precise aphasia prediction. A natural language processing pipeline can be greatly benefitted from integrating LLMs, enabling us to refine models of existing language disorders, such as aphasia.

Patient Perspectives on AI-Driven Predictions of Schizophrenia Relapses: Understanding Concerns and Opportunities for Self-Care and Treatment

Early detection and intervention for relapse is important in the treatment of schizophrenia spectrum disorders. Researchers have developed AI models to predict relapse from patient-contributed data like social media. However, these models face challenges, including misalignment with practice and ethical issues related to transparency, accountability, and potential harm. Furthermore, how patients who have recovered from schizophrenia view these AI models has been underexplored. To address this gap, we first conducted semi-structured interviews with 28 patients and reflexive thematic analysis, which revealed a disconnect between AI predictions and patient experience, and the importance of the social aspect of relapse detection. In response, we developed a prototype that used patients' Facebook data to predict relapse. Feedback from seven patients highlighted the potential for AI to foster collaboration between patients and their support systems, and to encourage self-reflection. Our work provides insights into human-AI interaction and suggests ways to empower people with schizophrenia.

Automated linguistic analysis in speech samples of Turkish-speaking patients with schizophrenia-spectrum disorders

Modern natural language processing (NLP) methods provide ways to objectively quantify language disturbances for potential use in diagnostic classification. We performed computerized language analysis in speech samples of 82 Turkish-speaking subjects, including 44 patients with schizophrenia spectrum disorders (SSD) and 38 healthy controls (HC). Exploratory analysis of speech samples involved 16 sentence-level semantic similarity features using SBERT (Sentence Bidirectional Encoder Representation from Text) as well as 8 generic and 8 part-of-speech (POS) features. The random forest classifier using SBERT-derived semantic similarity features achieved a mean accuracy of 85.6 % for the classification of SSD and HC. When semantic similarity features were combined with generic and POS features, the classifier's mean accuracy reached to 86.8 %. Our analysis reflected increased sentence-level semantic similarity scores in SSD. Generic and POS analyses revealed an increase in the use of verbs, proper nouns and pronouns in SSD while our results showed a decrease in the utilization of conjunctions, determiners, and both average and maximum sentence length in SSD compared to HC. Quantitative language features were correlated with the expressive deficit domain of BNSS (Brief Negative Symptom Scale) as well as with the duration of illness. These findings from Turkish-speaking interviews contribute to the growing evidence-based NLP-derived assessments in non-English-speaking patients.

Screening for psychotic experiences and psychotic disorders in general psychiatric settings: a systematic review and meta-analysis

Background

The absence of systematic screening for psychosis within general psychiatric services contribute to substantial treatment delays and poor long-term outcomes. We conducted a meta-analysis to estimate rates of psychotic experiences, clinical high-risk for psychosis syndrome (CHR-P), and psychotic disorders identified by screening treatment-seeking individuals to inform implementation recommendations for routine psychosis screening in general psychiatric settings.

Methods

PubMed and Web of Science databases were searched to identify empirical studies that contained information on the point prevalence of psychotic experiences, CHR-P, or psychotic disorders identified by screening inpatient and outpatient samples aged 12-64 receiving general psychiatric care. Psychotic experiences were identified by meeting threshold scores on validated self-reported questionnaires, and psychotic disorders and CHR-P by gold-standard structured interview assessments. A meta-analysis of each outcome was conducted using the Restricted Maximum Likelihood Estimator method of estimating effect sizes in a random effects model.

Results

41 independent samples (k=36 outpatient) involving n=25,751 patients (58% female, mean age: 24.1 years) were included. Among a general psychiatric population, prevalence of psychotic experiences was 44.3% (95% CI: 35.8-52.8%; 28 samples, n=21,957); CHR-P was 26.4% (95% CI: 20.0-32.7%; 28 samples, n=14,395); and psychotic disorders was 6.6% (95% CI: 3.3-9.8%; 32 samples, n=20,371).

Conclusions

High rates of psychotic spectrum illness in general psychiatric settings underscore need for secondary prevention with psychosis screening. These base rates can be used to plan training and resources required to conduct assessments for early detection, as well as build capacity in interventions for CHR-P and early psychosis in non-specialty mental health settings.

Understanding psychosis complexity through a syndemic framework: A systematic review

Psychotic conditions pose significant challenges due to their complex aetiology and impact on individuals and communities. Syndemic theory offers a promising framework to understand the interconnectedness of various health and social problems in the context of psychosis. This systematic review aims to examine existing literature on testing whether psychosis is better understood as a component of a syndemic. We conducted a systematic search of 7 databases, resulting in the inclusion of five original articles. Findings from these studies indicate a syndemic characterized by the coexistence of various health and social conditions, are associated with a greater risk of psychosis, adverse health outcomes, and disparities, especially among ethnic minorities and deprived populations. This review underscores the compelling need for a new paradigm and datasets that can investigate how psychosis emerges in the context of a syndemic, ultimately guiding more effective preventive and care interventions as well as policies to improve the health of marginalised communities living in precarity.

Speech markers to predict and prevent recurrent episodes of psychosis: A narrative overview and emerging opportunities

Preventing relapse in schizophrenia improves long-term health outcomes. Repeated episodes of psychotic symptoms shape the trajectory of this illness and can be a detriment to functional recovery. Despite early intervention programs, high relapse rates persist, calling for alternative approaches in relapse prevention. Predicting imminent relapse at an individual level is critical for effective intervention. While clinical profiles are often used to foresee relapse, they lack the specificity and sensitivity needed for timely prediction. Here, we review the use of speech through Natural Language Processing (NLP) to predict a recurrent psychotic episode. Recent advancements in NLP of speech have shown the ability to detect linguistic markers related to thought disorder and other language disruptions within 2-4 weeks preceding a relapse. This approach has shown to be able to capture individual speech patterns, showing promise in its use as a prediction tool. We outline current developments in remote monitoring for psychotic relapses, discuss the challenges and limitations and present the speech-NLP based approach as an alternative to detect relapses with sufficient accuracy, construct validity and lead time to generate clinical actions towards prevention.

Natural language processing for defining linguistic features in schizophrenia: A sample from Turkish speakers

Natural language processing (NLP) provides fast and accurate extraction of features related to the language of schizophrenia. We utilized NLP methods to test the hypothesis that schizophrenia is associated with altered linguistic features in Turkish, a non-Indo-European language, compared to controls. We also explored whether these possible altered linguistic features were language-dependent or -independent. We extracted and compared speech in schizophrenia (SZ, N = 38) and healthy well-matched control (HC, N = 38) participants using NLP. The analysis was conducted in two parts. In the first one, mean sentence length, total completed words, moving average type-token ratio to measure the lexical diversity, and first-person singular pronoun usage were calculated. In the second one, we used parts-of-speech tagging (POS) and Word2Vec in schizophrenia and control. We found that SZ had lower mean sentence length and moving average type-token ratio but higher use of first-person singular pronoun. All these significant results were correlated with the Thought and Language Disorder Scale score. The POS approach demonstrated that SZ used fewer coordinating conjunctions. Our methodology using Word2Vec detected that SZ had higher semantic similarity than HC and K-Means could differentiate between SZ and HC into two distinct groups with high accuracy, 86.84 %. Our findings showed that altered linguistic features in SZ are mostly language-independent. They are promising to describe language patterns in schizophrenia which proposes that NLP measurements may allow for rapid and objective measurements of linguistic features.

Towards a scalable approach to assess speech organization across the psychosis-spectrum -online assessment in conjunction with automated transcription and extraction of speech measures

Automatically extracted measures of speech constitute a promising marker of psychosis as disorganized speech is associated with psychotic symptoms and predictive of psychosis-onset. The potential of speech markers is, however, hampered by (i) lengthy assessments in laboratory settings and (ii) manual transcriptions. We investigated whether a short, scalable data collection (online) and processing (automated transcription) procedure would provide data of sufficient quality to extract previously validated speech measures. To evaluate the fit of our approach for purpose, we assessed speech in relation to psychotic-like experiences in the general population. Participants completed an 8-minute-long speech task online. Sample 1 included measures of psychometric schizotypy and delusional ideation (N = 446). Sample 2 included a low and high psychometric schizotypy group (N = 144). Recordings were transcribed both automatically and manually, and connectivity, semantic, and syntactic speech measures were extracted for both types of transcripts. 73%/86% participants in sample 1/2 completed the experiment. Nineteen out of 25 speech measures were strongly (r > 0.7) and significantly correlated between automated and manual transcripts in both samples. Amongst the 14 connectivity measures, 11 showed a significant relationship with delusional ideation. For the semantic and syntactic measures, On Topic score and the Frequency of personal pronouns were negatively correlated with both schizotypy and delusional ideation. Combined with demographic information, the speech markers could explain 11-14% of the variation of delusional ideation and schizotypy in Sample 1 and could discriminate between high-low schizotypy with high accuracy (0.72-0.70, AUC = 0.78-0.79) in Sample 2. The moderate to high retention rate, strong correlation of speech measures across manual and automated transcripts and sensitivity to psychotic-like experiences provides initial evidence that online collected speech in combination with automatic transcription is a feasible approach to increase accessibility and scalability of speech-based assessment of psychosis.

Developmental perspectives on the origins of psychotic disorders: The need for a transdiagnostic approach

Research on serious mental disorders, particularly psychosis, has revealed highly variable symptom profiles and developmental trajectories prior to illness-onset. As Dante Cicchetti pointed out decades before the term "transdiagnostic" was widely used, the pathways to psychopathology emerge in a system involving equifinality and multifinality. Like most other psychological disorders, psychosis is associated with multiple domains of risk factors, both genetic and environmental, and there are many transdiagnostic developmental pathways that can lead to psychotic syndromes. In this article, we discuss our current understanding of heterogeneity in the etiology of psychosis and its implications for approaches to conceptualizing etiology and research. We highlight the need for examining risk factors at multiple levels and to increase the emphasis on transdiagnostic developmental trajectories as a key variable associated with etiologic subtypes. This will be increasingly feasible now that large, longitudinal datasets are becoming available and researchers have access to more sophisticated analytic tools, such as machine learning, which can identify more homogenous subtypes with the ultimate goal of enhancing options for treatment and preventive intervention.

Automated analysis of speech as a marker of sub-clinical psychotic experiences

Automated speech analysis techniques, when combined with artificial intelligence and machine learning, show potential in capturing and predicting a wide range of psychosis symptoms, garnering attention from researchers. These techniques hold promise in predicting the transition to clinical psychosis from at-risk states, as well as relapse or treatment response in individuals with clinical-level psychosis. However, challenges in scientific validation hinder the translation of these techniques into practical applications. Although sub-clinical research could aid to tackle most of these challenges, there have been only few studies conducted in speech and psychosis research in non-clinical populations. This work aims to facilitate this work by summarizing automated speech analytical concepts and the intersection of this field with psychosis research. We review psychosis continuum and sub-clinical psychotic experiences, and the benefits of researching them. Then, we discuss the connection between speech and psychotic symptoms. Thirdly, we overview current and state-of-the art approaches to the automated analysis of speech both in terms of language use (text-based analysis) and vocal features (audio-based analysis). Then, we review techniques applied in subclinical population and findings in these samples. Finally, we discuss research challenges in the field, recommend future research endeavors and outline how research in subclinical populations can tackle the listed challenges.

Language and communication rehabilitation in patients with schizophrenia: A narrative review

Language impairments often appear in patients with schizophrenia and are potential targets for rehabilitation. Clinical practice and research should be intimately connected. The aim was to perform a narrative review of the assessment and intervention tools that have been used for the rehabilitation of schizophrenia patients with language and communication impairments. Two types of tools, general and specific, were developed for both purposes. General tools include the Positive and Negative Syndrome Scale for assessment, and the Integrated Psychological Therapy for intervention. The specific tools used to evaluate language and communication impairments include the Scale for the Assessment of Thought, Language and Communication, the Formal Thought Disorder scales (for caregivers and patients), and the Thought and Language Disorder scale. The most recent language-specific intervention tools include the Cognitive Pragmatic Treatment, Conecta-2, Let's talk! Multimodal Speech-Gesture training, Speech Therapy Intervention Group, and PragmaCom. These tools primarily involve psychopathology/psychiatry, psychology, linguistics, speech and language therapy, and nursing. In conclusion, a wide range of assessment and intervention tools are available for the rehabilitation of language and communication impairments associated with schizophrenia. An integrative and interdisciplinary approach should always be considered for rehabilitation of language and communication in patients with schizophrenia throughout their lifetime.

The Influence of Using Novel Predictive Technologies on Judgments of Stigma, Empathy, and Compassion among Healthcare Professionals

BACKGROUND

Our objective was to evaluate whether the description of a machine learning (ML) app or brain imaging technology to predict the onset of schizophrenia or alcohol use disorder (AUD) influences healthcare professionals' judgments of stigma, empathy, and compassion.

METHODS

We randomized healthcare professionals (N = 310) to one vignette about a person whose clinician seeks to predict schizophrenia or an AUD, using a ML app, brain imaging, or a psychosocial assessment. Participants used scales to measure their judgments of stigma, empathy, and compassion.

RESULTS

Participants randomized to the ML vignette endorsed less anger and more fear relative to the psychosocial vignette, and the brain imaging vignette elicited higher pity ratings. The brain imaging and ML vignettes evoked lower personal responsibility judgments compared to the psychosocial vignette. Physicians and nurses reported less empathy than clinical psychologists.

CONCLUSIONS

The use of predictive technologies may reinforce essentialist views about mental health and substance use that may increase specific aspects of stigma and reduce others.

Theory-Driven Analysis of Natural Language Processing Measures of Thought Disorder Using Generative Language Modeling

BACKGROUND

Natural language processing (NLP) holds promise to transform psychiatric research and practice. A pertinent example is the success of NLP in the automatic detection of speech disorganization in formal thought disorder (FTD). However, we lack an understanding of precisely what common NLP metrics measure and how they relate to theoretical accounts of FTD. We propose tackling these questions by using deep generative language models to simulate FTD-like narratives by perturbing computational parameters instantiating theory-based mechanisms of FTD.

METHODS

We simulated FTD-like narratives using Generative-Pretrained-Transformer-2 by either increasing word selection stochasticity or limiting the model's memory span. We then examined the sensitivity of common NLP measures of derailment (semantic distance between consecutive words or sentences) and tangentiality (how quickly meaning drifts away from the topic) in detecting and dissociating the 2 underlying impairments.

RESULTS

Both parameters led to narratives characterized by greater semantic distance between consecutive sentences. Conversely, semantic distance between words was increased by increasing stochasticity, but decreased by limiting memory span. An NLP measure of tangentiality was uniquely predicted by limited memory span. The effects of limited memory span were nonmonotonic in that forgetting the global context resulted in sentences that were semantically closer to their local, intermediate context. Finally, different methods for encoding the meaning of sentences varied dramatically in performance.

CONCLUSIONS

This work validates a simulation-based approach as a valuable tool for hypothesis generation and mechanistic analysis of NLP markers in psychiatry. To facilitate dissemination of this approach, we accompany the paper with a hands-on Python tutorial.

Evaluating the clinical utility of speech analysis and machine learning in schizophrenia: A pilot study

BACKGROUND

Schizophrenia is a serious mental disorder that significantly impacts social functioning and quality of life. However, current diagnostic methods lack objective biomarker support. While some studies have indicated differences in audio features between patients with schizophrenia and healthy controls, these findings are influenced by demographic information and variations in experimental paradigms. Therefore, it is crucial to explore stable and reliable audio biomarkers for an auxiliary diagnosis and disease severity prediction of schizophrenia.

METHOD

A total of 130 individuals (65 patients with schizophrenia and 65 healthy controls) read three fixed texts containing positive, neutral, and negative emotions, and recorded them. All audio signals were preprocessed and acoustic features were extracted by a librosa-0.9.2 toolkit. Independent sample t-tests were performed on two sets of acoustic features, and Pearson correlation on the acoustic features and Positive and Negative Syndrome Scale (PANSS) scores of the schizophrenia group. Classification algorithms in scikit-learn were used to diagnose schizophrenia and predict the level of negative symptoms.

RESULTS

Significant differences were observed between the two groups in the mfcc_8, mfcc_11, and mfcc_33 of mel-frequency cepstral coefficient (MFCC). Furthermore, a significant correlation was found between mfcc_7 and the negative PANSS scores. Through acoustic features, we could not only differentiate patients with schizophrenia from healthy controls with an accuracy of 0.815 but also predict the grade of the negative symptoms in schizophrenia with an average accuracy of 0.691.

CONCLUSIONS

The results demonstrated the considerable potential of acoustic characteristics as reliable biomarkers for diagnosing schizophrenia and predicting clinical symptoms.

Assessing coherence through linguistic connectives: Analysis of speech in patients with schizophrenia-spectrum disorders

BACKGROUND

Incoherent speech is a core diagnostic symptom of schizophrenia-spectrum disorders (SSD) that can be studied using semantic space models. Since linguistic connectives signal relations between words, they and their surrounding words might represent linguistic loci to detect unusual coherence in speech. Therefore, we investigated whether connectives' measures are useful to assess incoherent speech in SSD.

METHODS

Connectives and their surrounding words were extracted from transcripts of spontaneous speech of 50 SSD-patients and 50 control participants. Using word2vec, two different cosine similarities were calculated: those of connectives and their surrounding words (connectives-related similarity), and those of free-of-connectives words-chunks (non-connectives similarity). Differences between groups in proportion of five types of connectives were assessed using generalized logistic models, and connectives-related similarity was analyzed through non-parametric multivariate analysis of variance. These features were evaluated in classification tasks to differentiate between groups.

RESULTS

SSD-patients used less contingency (e.g., because) (p = .008) and multiclass connectives (e.g., as) (p < .001) than control participants. SSD-patients had higher minimum similarity of multiclass (adj-p = .04) and temporality connectives (e.g., after) (adj-p < .001), narrower similarity-range of expansion (e.g., and) (adj-p = .002) and multiclass connectives (adj-p = .04), and lower maximum similarity of expansion connectives (adj-p = .005). Using connectives' features alone, SSD-patients and controls could be distinguished with 85 % accuracy.

DISCUSSION

Our results show that SSD-speech can be distinguished from speech of control participants with high accuracy, based solely on connectives' features. We conclude that including connectives could strengthen computational models to categorize SSD.

Language network self-inhibition and semantic similarity in first-episode schizophrenia: A computational-linguistic and effective connectivity approach

INTRODUCTION

A central feature of schizophrenia is the disorganization and impoverishment of language. Recently, we observed higher semantic similarity in first-episode-schizophrenia (FES) patients. In this study, we investigate if this aberrant similarity relates to the 'causal' connectivity between two key nodes of the word production system: inferior frontal gyrus (IFG) and the semantic-hub at the ventral anterior temporal lobe (vATL).

METHODS

Resting-state fMRI scans were collected from 60 participants (30 untreated FES and 30 healthy controls). The semantic distance was measured with the CoVec semantic tool based on GloVe. A spectral dynamic causal model with Parametrical Empirical Bayes was constructed modelling the intrinsic self-inhibitory and extrinsic-excitatory connections within the brain regions. We estimated the parameters of a fully connected model with the semantic distance as a covariate.

RESULTS

FES patients chose words with higher semantic similarity when describing the pictures compared to the HC group. Among patients, an increased semantic similarity was related with an increase in intrinsic connections within both the vATL and IFG, suggesting that reduced 'synaptic gain' in these regions likely contribute to aberrant sampling of the semantic space during discourse in schizophrenia.

CONCLUSIONS

Lexical impoverishment relates to increased self-inhibition in both the IFG and vATL. The associated reduction in synaptic gain may relate to reduced precision of locally generated neural activity, forcing the choice of words that are already 'activated' in a lexical network. One approach to improve word sampling may be via promoting synaptic gain via supra-physiological stimulation within the Broca's-vATL network; this proposal needs verification.

Happy thoughts: What computational assessment of connectedness and emotional words can inform about early stages of psychosis

In recent years, different natural language processing tools measured aspects related to narratives' structural, semantic, and emotional content. However, there is a need to better understand the limitations and effectiveness of speech elicitation protocols. The graph-theoretical analysis applied to short narratives reveals lower connectedness associated with negative symptoms even in the early stages of psychosis, but emotional topics seem more informative than others. We investigate the interaction between connectedness and emotional words with negative symptoms and educational level in participants with and without psychosis. For that purpose, we used a speech elicitation protocol based on three positive affective pictures and calculated the proportion of emotional words and connectedness measures in the first-episode psychosis (FEP) group (N: 24) and a control group (N: 33). First, we replicated the association between connectedness and negative symptoms (R2: 0.53, p: 0.0049). Second, the more positive terms, the more connected the narrative was, exclusively under psychosis and in association with education, pointing to an interaction between symptoms and formal education. Negative symptoms were independently associated with connectedness, but not with emotional words, although the associations with education were mutually dependent. Together, education and symptoms explained almost 70 % of connectedness variance (R2: 0.67, p < 0.0001), but not emotional expression. At this initial stage of psychosis, education seems to play an important role, diminishing the impact of negative symptoms on the narrative connectedness. Negative symptoms in FEP impact narrative connectedness in association with emotional expression, revealing aspects of social cognition through a short and innocuous protocol.

The SPEAK study rationale and design: A linguistic corpus-based approach to understanding thought disorder

AIM

Psychotic symptoms are typically measured using clinical ratings, but more objective and sensitive metrics are needed. Hence, we will assess thought disorder using the Research Domain Criteria (RDoC) heuristic for language production, and its recommended paradigm of "linguistic corpus-based analyses of language output". Positive thought disorder (e.g., tangentiality and derailment) can be assessed using word-embedding approaches that assess semantic coherence, whereas negative thought disorder (e.g., concreteness, poverty of speech) can be assessed using part-of-speech (POS) tagging to assess syntactic complexity. We aim to establish convergent validity of automated linguistic metrics with clinical ratings, assess normative demographic variance, determine cognitive and functional correlates, and replicate their predictive power for psychosis transition among at-risk youths.

METHODS

This study will assess language production in 450 English-speaking individuals in Australia and Canada, who have recent onset psychosis, are at clinical high risk (CHR) for psychosis, or who are healthy volunteers, all well-characterized for cognition, function and symptoms. Speech will be elicited using open-ended interviews. Audio files will be transcribed and preprocessed for automated natural language processing (NLP) analyses of coherence and complexity. Data analyses include canonical correlation, multivariate linear regression with regularization, and machine-learning classification of group status and psychosis outcome.

CONCLUSIONS

This prospective study aims to characterize language disturbance across stages of psychosis using computational approaches, including psychometric properties, normative variance and clinical correlates, important for biomarker development. SPEAK will create a large archive of language data available to other investigators, a rich resource for the field.

Speech disturbances in schizophrenia: Assessing cross-linguistic generalizability of NLP automated measures of coherence

INTRODUCTION

Language disorders - disorganized and incoherent speech in particular - are distinctive features of schizophrenia. Natural language processing (NLP) offers automated measures of incoherent speech as promising markers for schizophrenia. However, the scientific and clinical impact of NLP markers depends on their generalizability across contexts, samples, and languages, which we systematically assessed in the present study relying on a large, novel, cross-linguistic corpus.

METHODS

We collected a Danish (DK), German (GE), and Chinese (CH) cross-linguistic dataset involving transcripts from 187 participants with schizophrenia (111DK, 25GE, 51CH) and 200 matched controls (129DK, 29GE, 42CH) performing the Animated Triangles Task. Fourteen previously published NLP coherence measures were calculated, and between-groups differences and association with symptoms were tested for cross-linguistic generalizability.

RESULTS

One coherence measure, i.e. second-order coherence, robustly generalized across samples and languages. We found several language-specific effects, some of which partially replicated previous findings (lower coherence in German and Chinese patients), while others did not (higher coherence in Danish patients). We found several associations between symptoms and measures of coherence, but the effects were generally inconsistent across languages and rating scales.

CONCLUSIONS

Using a cumulative approach, we have shown that NLP findings of reduced semantic coherence in schizophrenia have limited generalizability across different languages, samples, and measures. We argue that several factors such as sociodemographic and clinical heterogeneity, cross-linguistic variation, and the different NLP measures reflecting different clinical aspects may be responsible for this variability. Future studies should take this variability into account in order to develop effective clinical applications targeting different patient populations.

Lower cohesion and altered first-person pronoun usage in the spoken life narratives of individuals with schizophrenia

Usage of computational tools to quantify language disturbances among individuals with psychosis is increasing, improving measurement efficiency and access to fine-grained constructs. However, few studies apply automated linguistic analysis to life narratives in this population. Such research could facilitate the measurement of psychosis-relevant constructs such as sense of agency, capacity to organize one's personal history, narrative richness, and perceptions of the roles that others play in one's life. Furthermore, research is needed to understand how narrative linguistic features relate to cognitive and social functioning. In the present study, individuals with schizophrenia (n = 32) and individuals without a psychotic disorder (n = 15) produced personal life narratives within the Indiana Psychiatric Illness Interview. Narratives were analyzed using the Coh-Metrix computational tool. Linguistic variables analyzed were indices of connections within causal and goal-driven speech (deep cohesion), unique word usage (lexical diversity), and pronoun usage. Individuals with schizophrenia compared to control participants produced narratives that were lower in deep cohesion, contained more first-person singular pronouns, and contained fewer first-person plural pronouns. Narratives did not significantly differ between groups in lexical diversity, third-person pronoun usage, or total word count. Cognitive-linguistic relationships emerged in the full sample, including significant correlations between greater working memory capacity and greater deep cohesion and lexical diversity. In the schizophrenia group, social problem-solving abilities did not correlate with linguistic variables but were associated with cognition. Findings highlight the relevance of psychotherapies which aim to promote recovery among individuals with psychosis through the construction of coherent life narratives and increasing agency and social connectedness.

Linguistic profile automated characterisation in pluripotential clinical high-risk mental state (CHARMS) conditions: methodology of a multicentre observational study

INTRODUCTION

Language is usually considered the social vehicle of thought in intersubjective communications. However, the relationship between language and high-order cognition seems to evade this canonical and unidirectional description (ie, the notion of language as a simple means of thought communication). In recent years, clinical high at-risk mental state (CHARMS) criteria (evolved from the Ultra-High-Risk paradigm) and the introduction of the Clinical Staging system have been proposed to address the dynamicity of early psychopathology. At the same time, natural language processing (NLP) techniques have greatly evolved and have been successfully applied to investigate different neuropsychiatric conditions. The combination of at-risk mental state paradigm, clinical staging system and automated NLP methods, the latter applied on spoken language transcripts, could represent a useful and convenient approach to the problem of early psychopathological distress within a transdiagnostic risk paradigm.

METHODS AND ANALYSIS

Help-seeking young people presenting psychological distress (CHARMS+/- and Clinical Stage 1a or 1b; target sample size for both groups n=90) will be assessed through several psychometric tools and multiple speech analyses during an observational period of 1-year, in the context of an Italian multicentric study. Subjects will be enrolled in different contexts: Department of Neuroscience, Rehabilitation, Ophthalmology, Genetics, Maternal and Child Health (DINOGMI), Section of Psychiatry, University of Genoa-IRCCS Ospedale Policlinico San Martino, Genoa, Italy; Mental Health Department-territorial mental services (ASL 3-Genoa), Genoa, Italy; and Mental Health Department-territorial mental services (AUSL-Piacenza), Piacenza, Italy. The conversion rate to full-blown psychopathology (CS 2) will be evaluated over 2 years of clinical observation, to further confirm the predictive and discriminative value of CHARMS criteria and to verify the possibility of enriching them with several linguistic features, derived from a fine-grained automated linguistic analysis of speech.

ETHICS AND DISSEMINATION

The methodology described in this study adheres to ethical principles as formulated in the Declaration of Helsinki and is compatible with International Conference on Harmonization (ICH)-good clinical practice. The research protocol was reviewed and approved by two different ethics committees (CER Liguria approval code: 591/2020-id.10993; Comitato Etico dell'Area Vasta Emilia Nord approval code: 2022/0071963). Participants will provide their written informed consent prior to study enrolment and parental consent will be needed in the case of participants aged less than 18 years old. Experimental results will be carefully shared through publication in peer-reviewed journals, to ensure proper data reproducibility.

TRIAL REGISTRATION NUMBER

DOI:10.17605/OSF.IO/BQZTN.

Language Analytics for Assessment of Mental Health Status and Functional Competency

BACKGROUND AND HYPOTHESIS

Automated language analysis is becoming an increasingly popular tool in clinical research involving individuals with mental health disorders. Previous work has largely focused on using high-dimensional language features to develop diagnostic and prognostic models, but less work has been done to use linguistic output to assess downstream functional outcomes, which is critically important for clinical care. In this work, we study the relationship between automated language composites and clinical variables that characterize mental health status and functional competency using predictive modeling.

STUDY DESIGN

Conversational transcripts were collected from a social skills assessment of individuals with schizophrenia (n = 141), bipolar disorder (n = 140), and healthy controls (n = 22). A set of composite language features based on a theoretical framework of speech production were extracted from each transcript and predictive models were trained. The prediction targets included clinical variables for assessment of mental health status and social and functional competency. All models were validated on a held-out test sample not accessible to the model designer.

STUDY RESULTS

Our models predicted the neurocognitive composite with Pearson correlation PCC = 0.674; PANSS-positive with PCC = 0.509; PANSS-negative with PCC = 0.767; social skills composite with PCC = 0.785; functional competency composite with PCC = 0.616. Language features related to volition, affect, semantic coherence, appropriateness of response, and lexical diversity were useful for prediction of clinical variables.

CONCLUSIONS

Language samples provide useful information for the prediction of a variety of clinical variables that characterize mental health status and functional competency.

Semantic and Acoustic Markers in Schizophrenia-Spectrum Disorders: A Combinatory Machine Learning Approach

BACKGROUND AND HYPOTHESIS

Speech is a promising marker to aid diagnosis of schizophrenia-spectrum disorders, as it reflects symptoms like thought disorder and negative symptoms. Previous approaches made use of different domains of speech for diagnostic classification, including features like coherence (semantic) and form (acoustic). However, an examination of the added value of each domain when combined is lacking as of yet. Here, we investigate the acoustic and semantic domains separately and combined.

STUDY DESIGN

Using semi-structured interviews, speech of 94 subjects with schizophrenia-spectrum disorders (SSD) and 73 healthy controls (HC) was recorded. Acoustic features were extracted using a standardized feature-set, and transcribed interviews were used to calculate semantic word similarity using word2vec. Random forest classifiers were trained for each domain. A third classifier was used to combine features from both domains; 10-fold cross-validation was used for each model.

RESULTS

The acoustic random forest classifier achieved 81% accuracy classifying SSD and HC, while the semantic domain classifier reached an accuracy of 80%. Joining features from the two domains, the combined classifier reached 85% accuracy, significantly improving on separate domain classifiers. For the combined classifier, top features were fragmented speech from the acoustic domain and variance of similarity from the semantic domain.

CONCLUSIONS

Both semantic and acoustic analyses of speech achieved ~80% accuracy in classifying SSD from HC. We replicate earlier findings per domain, additionally showing that combining these features significantly improves classification performance. Feature importance and accuracy in combined classification indicate that the domains measure different, complementing aspects of speech.

Latent Factors of Language Disturbance and Relationships to Quantitative Speech Features

BACKGROUND AND HYPOTHESIS

Quantitative acoustic and textual measures derived from speech ("speech features") may provide valuable biomarkers for psychiatric disorders, particularly schizophrenia spectrum disorders (SSD). We sought to identify cross-diagnostic latent factors for speech disturbance with relevance for SSD and computational modeling.

STUDY DESIGN

Clinical ratings for speech disturbance were generated across 14 items for a cross-diagnostic sample (N = 334), including SSD (n = 90). Speech features were quantified using an automated pipeline for brief recorded samples of free speech. Factor models for the clinical ratings were generated using exploratory factor analysis, then tested with confirmatory factor analysis in the cross-diagnostic and SSD groups. The relationships between factor scores and computational speech features were examined for 202 of the participants.

STUDY RESULTS

We found a 3-factor model with a good fit in the cross-diagnostic group and an acceptable fit for the SSD subsample. The model identifies an impaired expressivity factor and 2 interrelated disorganized factors for inefficient and incoherent speech. Incoherent speech was specific to psychosis groups, while inefficient speech and impaired expressivity showed intermediate effects in people with nonpsychotic disorders. Each of the 3 factors had significant and distinct relationships with speech features, which differed for the cross-diagnostic vs SSD groups.

CONCLUSIONS

We report a cross-diagnostic 3-factor model for speech disturbance which is supported by good statistical measures, intuitive, applicable to SSD, and relatable to linguistic theories. It provides a valuable framework for understanding speech disturbance and appropriate targets for modeling with quantitative speech features.

Emergence of Language Related to Self-experience and Agency in Autobiographical Narratives of Individuals With Schizophrenia

BACKGROUND AND HYPOTHESIS

Disturbances in self-experience are a central feature of schizophrenia and its study can enhance phenomenological understanding and inform mechanisms underlying clinical symptoms. Self-experience involves the sense of self-presence, of being the subject of one's own experiences and agent of one's own actions, and of being distinct from others. Self-experience is traditionally assessed by manual rating of interviews; however, natural language processing (NLP) offers automated approach that can augment manual ratings by rapid and reliable analysis of text.

STUDY DESIGN

We elicited autobiographical narratives from 167 patients with schizophrenia or schizoaffective disorder (SZ) and 90 healthy controls (HC), amounting to 490 000 words and 26 000 sentences. We used NLP techniques to examine transcripts for language related to self-experience, machine learning to validate group differences in language, and canonical correlation analysis to examine the relationship between language and symptoms.

STUDY RESULTS

Topics related to self-experience and agency emerged as significantly more expressed in SZ than HC (P < 10-13) and were decoupled from similarly emerging features such as emotional tone, semantic coherence, and concepts related to burden. Further validation on hold-out data showed that a classifier trained on these features achieved patient-control discrimination with AUC = 0.80 (P < 10-5). Canonical correlation analysis revealed significant relationships between self-experience and agency language features and clinical symptoms.

CONCLUSIONS

Notably, the self-experience and agency topics emerged without any explicit probing by the interviewer and can be algorithmically detected even though they involve higher-order metacognitive processes. These findings illustrate the utility of NLP methods to examine phenomenological aspects of schizophrenia.

Psychotic disorders as a framework for precision psychiatry

People with psychotic disorders can show marked interindividual variations in the onset of illness, responses to treatment and relapse, but they receive broadly similar clinical care. Precision psychiatry is an approach that aims to stratify people with a given disorder according to different clinical outcomes and tailor treatment to their individual needs. At present, interindividual differences in outcomes of psychotic disorders are difficult to predict on the basis of clinical assessment alone. Therefore, current research in psychosis seeks to build models that predict outcomes by integrating clinical information with a range of biological measures. Here, we review recent progress in the application of precision psychiatry to psychotic disorders and consider the challenges associated with implementing this approach in clinical practice.

Acoustic speech markers for schizophrenia-spectrum disorders: a diagnostic and symptom-recognition tool

BACKGROUND

Clinicians routinely use impressions of speech as an element of mental status examination. In schizophrenia-spectrum disorders, descriptions of speech are used to assess the severity of psychotic symptoms. In the current study, we assessed the diagnostic value of acoustic speech parameters in schizophrenia-spectrum disorders, as well as its value in recognizing positive and negative symptoms.

METHODS

Speech was obtained from 142 patients with a schizophrenia-spectrum disorder and 142 matched controls during a semi-structured interview on neutral topics. Patients were categorized as having predominantly positive or negative symptoms using the Positive and Negative Syndrome Scale (PANSS). Acoustic parameters were extracted with OpenSMILE, employing the extended Geneva Acoustic Minimalistic Parameter Set, which includes standardized analyses of pitch (F0), speech quality and pauses. Speech parameters were fed into a random forest algorithm with leave-ten-out cross-validation to assess their value for a schizophrenia-spectrum diagnosis, and PANSS subtype recognition.

RESULTS

The machine-learning speech classifier attained an accuracy of 86.2% in classifying patients with a schizophrenia-spectrum disorder and controls on speech parameters alone. Patients with predominantly positive v. negative symptoms could be classified with an accuracy of 74.2%.

CONCLUSIONS

Our results show that automatically extracted speech parameters can be used to accurately classify patients with a schizophrenia-spectrum disorder and healthy controls, as well as differentiate between patients with predominantly positive v. negatives symptoms. Thus, the field of speech technology has provided a standardized, powerful tool that has high potential for clinical applications in diagnosis and differentiation, given its ease of comparison and replication across samples.

Linguistic markers and basic self-disturbances among adolescents at risk of psychosis. A qualitative study

Background

Language impairments are key features of schizophrenia spectrum disorders, and have also been suggested to signal enhanced psychosis risk. Incoherence, derailment, and monotonous speaking are however closely related to psychosis onset, and thus not very early markers. Recent phenomenologic-psychiatric studies claim that basic self-disturbance (BSD) may represent more useful early markers.

Methods

We searched for distinctive irregular linguistics of 30 CHR outpatient adolescents, aged 12-18 years. Standard instruments established psychosis risk and BSD. Participants chose three personal and well manifested BSD phenomena. Ninety verbatim statements were analyzed and grouped into higher order clusters of linguistic irregularities.

Findings

We identified five clusters of irregular language features: distinctive words, describing an atmosphere of unreality; irregular use of prepositions, indicating experiential detachment; shifts of personal pronouns, indicating identity confusion; near-literal use of metaphors and conjunctions indicating existential insecurity, and idiosyncratic use of adjectives indicating perceptual transcendence.

Interpretation

The adolescents provided naturalistic descriptions of experiences that were markedly twisted and almost ineffable. This unique irregular "BSD -language" was highly meaningful in its proper context, expressing informative characteristics of first-personal experiential alterations, essential for early detection. The features may additionally represent precursors of psychosis transition, useful for clinical decision-making.

Funding

Foundation Dam, Oslo, Norway (Grant Number 2017/FO143368).

Understanding psychiatric illness through natural language processing (UNDERPIN): Rationale, design, and methodology

Introduction

Psychiatric disorders are diagnosed through observations of psychiatrists according to diagnostic criteria such as the DSM-5. Such observations, however, are mainly based on each psychiatrist's level of experience and often lack objectivity, potentially leading to disagreements among psychiatrists. In contrast, specific linguistic features can be observed in some psychiatric disorders, such as a loosening of associations in schizophrenia. Some studies explored biomarkers, but biomarkers have yet to be used in clinical practice.

Aim

The purposes of this study are to create a large dataset of Japanese speech data labeled with detailed information on psychiatric disorders and neurocognitive disorders to quantify the linguistic features of those disorders using natural language processing and, finally, to develop objective and easy-to-use biomarkers for diagnosing and assessing the severity of them.

Methods

This study will have a multi-center prospective design. The DSM-5 or ICD-11 criteria for major depressive disorder, bipolar disorder, schizophrenia, and anxiety disorder and for major and minor neurocognitive disorders will be regarded as the inclusion criteria for the psychiatric disorder samples. For the healthy subjects, the absence of a history of psychiatric disorders will be confirmed using the Mini-International Neuropsychiatric Interview (M.I.N.I.). The absence of current cognitive decline will be confirmed using the Mini-Mental State Examination (MMSE). A psychiatrist or psychologist will conduct 30-to-60-min interviews with each participant; these interviews will include free conversation, picture-description task, and story-telling task, all of which will be recorded using a microphone headset. In addition, the severity of disorders will be assessed using clinical rating scales. Data will be collected from each participant at least twice during the study period and up to a maximum of five times at an interval of at least one month.

Discussion

This study is unique in its large sample size and the novelty of its method, and has potential for applications in many fields. We have some challenges regarding inter-rater reliability and the linguistic peculiarities of Japanese. As of September 2022, we have collected a total of >1000 records from >400 participants. To the best of our knowledge, this data sample is one of the largest in this field.

Clinical Trial Registration

Identifier: UMIN000032141.

Machine Learning and Non-Affective Psychosis: Identification, Differential Diagnosis, and Treatment

PURPOSE OF REVIEW

This review will cover the most relevant findings on the use of machine learning (ML) techniques in the field of non-affective psychosis, by summarizing the studies published in the last three years focusing on illness detection and treatment.

RECENT FINDINGS

Multiple ML tools that include mostly supervised approaches such as support vector machine, gradient boosting, and random forest showed promising results by applying these algorithms to various sources of data: socio-demographic information, EEG, language, digital content, blood biomarkers, neuroimaging, and electronic health records. However, the overall performance, in the binary classification case, varied from 0.49, which is to be considered very low (i.e., noise), to over 0.90. These results are fully justified by different factors, some of which may be attributable to the preprocessing of the data, the wide variety of the data, and the a-priori setting of hyperparameters. One of the main limitations of the field is the lack of stratification of results based on biological sex, given that psychosis presents differently in men and women; hence, the necessity to tailor identification tools and data analytic strategies. Timely identification and appropriate treatment are key factors in reducing the consequences of psychotic disorders. In recent years, the emergence of new analytical tools based on artificial intelligence such as supervised ML approaches showed promises as a potential breakthrough in this field. However, ML applications in everyday practice are still in its infancy.

Deconstructing heterogeneity in schizophrenia through language: a semi-automated linguistic analysis and data-driven clustering approach

Previous works highlighted the relevance of automated language analysis for predicting diagnosis in schizophrenia, but a deeper language-based data-driven investigation of the clinical heterogeneity through the illness course has been generally neglected. Here we used a semiautomated multidimensional linguistic analysis innovatively combined with a machine-driven clustering technique to characterize the speech of 67 individuals with schizophrenia. Clusters were then compared for psychopathological, cognitive, and functional characteristics. We identified two subgroups with distinctive linguistic profiles: one with higher fluency, lower lexical variety but greater use of psychological lexicon; the other with reduced fluency, greater lexical variety but reduced psychological lexicon. The former cluster was associated with lower symptoms and better quality of life, pointing to the existence of specific language profiles, which also show clinically meaningful differences. These findings highlight the importance of considering language disturbances in schizophrenia as multifaceted and approaching them in automated and data-driven ways.

Identifying psychiatric manifestations in schizophrenia and depression from audio-visual behavioural indicators through a machine-learning approach

Schizophrenia (SCZ) and depression (MDD) are two chronic mental disorders that seriously affect the quality of life of millions of people worldwide. We aim to develop machine-learning methods with objective linguistic, speech, facial, and motor behavioral cues to reliably predict the severity of psychopathology or cognitive function, and distinguish diagnosis groups. We collected and analyzed the speech, facial expressions, and body movement recordings of 228 participants (103 SCZ, 50 MDD, and 75 healthy controls) from two separate studies. We created an ensemble machine-learning pipeline and achieved a balanced accuracy of 75.3% for classifying the total score of negative symptoms, 75.6% for the composite score of cognitive deficits, and 73.6% for the total score of general psychiatric symptoms in the mixed sample containing all three diagnostic groups. The proposed system is also able to differentiate between MDD and SCZ with a balanced accuracy of 84.7% and differentiate patients with SCZ or MDD from healthy controls with a balanced accuracy of 82.3%. These results suggest that machine-learning models leveraging audio-visual characteristics can help diagnose, assess, and monitor patients with schizophrenia and depression.

Interconnectedness and (in)coherence as a signature of conspiracy worldviews

Conspiracy theories may arise out of an overarching conspiracy worldview that identifies common elements of subterfuge across unrelated or even contradictory explanations, leading to networks of self-reinforcing beliefs. We test this conjecture by analyzing a large natural language database of conspiracy and nonconspiracy texts for the same events, thus linking theory-driven psychological research with data-driven computational approaches. We find that, relative to nonconspiracy texts, conspiracy texts are more interconnected, more topically heterogeneous, and more similar to one another, revealing lower cohesion within texts but higher cohesion between texts and providing strong empirical support for an overarching conspiracy worldview. Our results provide inroads for classification algorithms and further exploration into individual differences in belief structures.

Social media and schizophrenia: An update on clinical applications

Social media has redesigned the landscape of human interaction, and data obtained through these platforms are promising for schizophrenia diagnosis and management. Recent research shows mounting evidence that machine learning analysis of social media content is capable of not only differentiating schizophrenia patients from healthy controls, but also predicting conversion to psychosis and symptom exacerbations. Novel platforms such as Horyzons show promise for improving social functioning and providing timely access to therapeutic resources. Social media is also a considerable means to assess and lessen the stigma surrounding schizophrenia. Herein, the relevant literature pertaining to social media and its clinical applications in schizophrenia over the past five years are summarized, followed by a discussion centered on user feedback to highlight future directions. Social media provides valuable contributions to a multifaceted digital phenotype that may improve schizophrenia care in the near future.

Who does what to whom? graph representations of action-predication in speech relate to psychopathological dimensions of psychosis

Graphical representations of speech generate powerful computational measures related to psychosis. Previous studies have mostly relied on structural relations between words as the basis of graph formation, i.e., connecting each word to the next in a sequence of words. Here, we introduced a method of graph formation grounded in semantic relationships by identifying elements that act upon each other (action relation) and the contents of those actions (predication relation). Speech from picture descriptions and open-ended narrative tasks were collected from a cross-diagnostic group of healthy volunteers and people with psychotic or non-psychotic disorders. Recordings were transcribed and underwent automated language processing, including semantic role labeling to identify action and predication relations. Structural and semantic graph features were computed using static and dynamic (moving-window) techniques. Compared to structural graphs, semantic graphs were more strongly correlated with dimensional psychosis symptoms. Dynamic features also outperformed static features, and samples from picture descriptions yielded larger effect sizes than narrative responses for psychosis diagnoses and symptom dimensions. Overall, semantic graphs captured unique and clinically meaningful information about psychosis and related symptom dimensions. These features, particularly when derived from semi-structured tasks using dynamic measurement, are meaningful additions to the repertoire of computational linguistic methods in psychiatry.

Computational analysis of spoken language in acute psychosis and mania

OBJECTIVES

This study aimed to (1) determine the feasibility of collecting behavioral data from participants hospitalized with acute psychosis and (2) begin to evaluate the clinical information that can be computationally derived from such data.

METHODS

Behavioral data was collected across 99 sessions from 38 participants recruited from an inpatient psychiatric unit. Each session started with a semi-structured interview modeled on a typical "clinical rounds" encounter and included administration of the Positive and Negative Syndrome Scale (PANSS).

ANALYSIS

We quantified aspects of participants' verbal behavior during the interview using lexical, coherence, and disfluency features. We then used two complementary approaches to explore our second objective. The first approach used predictive models to estimate participants' PANSS scores from their language features. Our second approach used inferential models to quantify the relationships between individual language features and symptom measures.

RESULTS

Our predictive models showed promise but lacked sufficient data to achieve clinically useful accuracy. Our inferential models identified statistically significant relationships between numerous language features and symptom domains.

CONCLUSION

Our interview recording procedures were well-tolerated and produced adequate data for transcription and analysis. The results of our inferential modeling suggest that automatic measurements of expressive language contain signals highly relevant to the assessment of psychosis. These findings establish the potential of measuring language during a clinical interview in a naturalistic setting and generate specific hypotheses that can be tested in future studies. This, in turn, will lead to more accurate modeling and better understanding of the relationships between expressive language and psychosis.

Going deep into schizophrenia with artificial intelligence

Despite years of research, the mechanisms governing the onset, relapse, symptomatology, and treatment of schizophrenia (SZ) remain elusive. The lack of appropriate analytic tools to deal with the heterogeneity and complexity of SZ may be one of the reasons behind this situation. Deep learning, a subfield of artificial intelligence (AI) inspired by the nervous system, has recently provided an accessible way of modeling and analyzing complex, high-dimensional, nonlinear systems. The unprecedented accuracy of deep learning algorithms in classification and prediction tasks has revolutionized a wide range of scientific fields and is rapidly permeating SZ research. Deep learning has the potential of becoming a valuable aid for clinicians in the prediction, diagnosis, and treatment of SZ, especially in combination with principles from Bayesian statistics. Furthermore, deep learning could become a powerful tool for uncovering the mechanisms underlying SZ thanks to a growing number of techniques designed for improving model interpretability and causal reasoning. The purpose of this article is to introduce SZ researchers to the field of deep learning and review its latest applications in SZ research. In general, existing studies have yielded impressive results in classification and outcome prediction tasks. However, methodological concerns related to the assessment of model performance in several studies, the widespread use of small training datasets, and the little clinical value of some models suggest that some of these results should be taken with caution.

Construct validity for computational linguistic metrics in individuals at clinical risk for psychosis: Associations with clinical ratings

Language deficits are prevalent in psychotic illness, including its risk states, and are related to marked impairment in functioning. It is therefore important to characterize language impairment in the psychosis spectrum in order to develop potential preventive interventions. Natural language processing (NLP) metrics of semantic coherence and syntactic complexity have been used to discriminate schizophrenia patients from healthy controls (HC) and predict psychosis onset in individuals at clinical high-risk (CHR) for psychosis. To date, no studies have yet examined the construct validity of key NLP features with respect to clinical ratings of thought disorder in a CHR cohort. Herein we test the association of key NLP metrics of coherence and complexity with ratings of positive and negative thought disorder, respectively, in 60 CHR individuals, using Andreasen's Scale of Assessment of Thought, Language and Communication (TLC) Scale to measure of positive and negative thought disorder. As hypothesized, in CHR individuals, the NLP metric of semantic coherence was significantly correlated with positive thought disorder severity and the NLP metrics of complexity (sentence length and determiner use) were correlated with negative thought disorder severity. The finding of construct validity supports the premise that NLP analytics, at least in respect to core features of reduction of coherence and complexity, are capturing clinically relevant language disturbances in risk states for psychosis. Further psychometric study is required, in respect to reliability and other forms of validity.

Automatic language analysis identifies and predicts schizophrenia in first-episode of psychosis

Automated language analysis of speech has been shown to distinguish healthy control (HC) vs chronic schizophrenia (SZ) groups, yet the predictive power on first-episode psychosis patients (FEP) and the generalization to non-English speakers remain unclear. We performed a cross-sectional and longitudinal (18 months) automated language analysis in 133 Spanish-speaking subjects from three groups: healthy control or HC (n = 49), FEP (n = 40), and chronic SZ (n = 44). Interviews were manually transcribed, and the analysis included 30 language features (4 verbal fluency; 20 verbal productivity; 6 semantic coherence). Our cross-sectional analysis showed that using the top ten ranked and decorrelated language features, an automated HC vs SZ classification achieved 85.9% accuracy. In our longitudinal analysis, 28 FEP patients were diagnosed with SZ at the end of the study. Here, combining demographics, PANSS, and language information, the prediction accuracy reached 77.5% mainly driven by semantic coherence information. Overall, we showed that language features from Spanish-speaking clinical interviews can distinguish HC vs chronic SZ, and predict SZ diagnosis in FEP patients.

Improving the Applicability of AI for Psychiatric Applications through Human-in-the-loop Methodologies

OBJECTIVES

Machine learning (ML) and natural language processing have great potential to improve efficiency and accuracy in diagnosis, treatment recommendations, predictive interventions, and scarce resource allocation within psychiatry. Researchers often conceptualize such an approach as operating in isolation without much need for human involvement, yet it remains crucial to harness human-in-the-loop practices when developing and implementing such techniques as their absence may be catastrophic. We advocate for building ML-based technologies that collaborate with experts within psychiatry in all stages of implementation and use to increase model performance while simultaneously increasing the practicality, robustness, and reliability of the process.

METHODS

We showcase pitfalls of the traditional ML framework and explain how it can be improved with human-in-the-loop techniques. Specifically, we applied active learning strategies to the automatic scoring of a story recall task and compared the results to a traditional approach.

RESULTS

Human-in-the-loop methodologies supplied a greater understanding of where the model was least confident or had knowledge gaps during training. As compared to the traditional framework, less than half of the training data were needed to reach a given accuracy.

CONCLUSIONS

Human-in-the-loop ML is an approach to data collection and model creation that harnesses active learning to select the most critical data needed to increase a model's accuracy and generalizability more efficiently than classic random sampling would otherwise allow. Such techniques may additionally operate as safeguards from spurious predictions and can aid in decreasing disparities that artificial intelligence systems otherwise propagate.

Progressive changes in descriptive discourse in First Episode Schizophrenia: a longitudinal computational semantics study

Computational semantics, a branch of computational linguistics, involves automated meaning analysis that relies on how words occur together in natural language. This offers a promising tool to study schizophrenia. At present, we do not know if these word-level choices in speech are sensitive to the illness stage (i.e., acute untreated vs. stable established state), track cognitive deficits in major domains (e.g., cognitive control, processing speed) or relate to established dimensions of formal thought disorder. In this study, we collected samples of descriptive discourse in patients experiencing an untreated first episode of schizophrenia and healthy control subjects (246 samples of 1-minute speech; n = 82, FES = 46, HC = 36) and used a co-occurrence based vector embedding of words to quantify semantic similarity in speech. We obtained six-month follow-up data in a subsample (99 speech samples, n = 33, FES = 20, HC = 13). At baseline, semantic similarity was evidently higher in patients compared to healthy individuals, especially when social functioning was impaired; but this was not related to the severity of clinically ascertained thought disorder in patients. Across the study sample, higher semantic similarity at baseline was related to poorer Stroop performance and processing speed. Over time, while semantic similarity was stable in healthy subjects, it increased in patients, especially when they had an increasing burden of negative symptoms. Disruptions in word-level choices made by patients with schizophrenia during short 1-min descriptions are sensitive to interindividual differences in cognitive and social functioning at first presentation and persist over the early course of the illness.

Use of social media data to diagnose and monitor psychotic disorders: systematic review and perspectives (Preprint)

Background

Schizophrenia is a disease associated with high burden, and improvement in care is necessary. Artificial intelligence (AI) has been used to diagnose several medical conditions as well as psychiatric disorders. However, this technology requires large amounts of data to be efficient. Social media data could be used to improve diagnostic capabilities.

Objective

The objective of our study is to analyze the current capabilities of AI to use social media data as a diagnostic tool for psychotic disorders.

Methods

A systematic review of the literature was conducted using several databases (PubMed, Embase, Cochrane, PsycInfo, and IEEE Xplore) using relevant keywords to search for articles published as of November 12, 2021. We used the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) criteria to identify, select, and critically assess the quality of the relevant studies while minimizing bias. We critically analyzed the methodology of the studies to detect any bias and presented the results.

Results

Among the 93 studies identified, 7 studies were included for analyses. The included studies presented encouraging results. Social media data could be used in several ways to care for patients with schizophrenia, including the monitoring of patients after the first episode of psychosis. We identified several limitations in the included studies, mainly lack of access to clinical diagnostic data, small sample size, and heterogeneity in study quality. We recommend using state-of-the-art natural language processing neural networks, called language models, to model social media activity. Combined with the synthetic minority oversampling technique, language models can tackle the imbalanced data set limitation, which is a necessary constraint to train unbiased classifiers. Furthermore, language models can be easily adapted to the classification task with a procedure called “fine-tuning.”

Conclusions

The use of social media data for the diagnosis of psychotic disorders is promising. However, most of the included studies had significant biases; we therefore could not draw conclusions about accuracy in clinical situations. Future studies need to use more accurate methodologies to obtain unbiased results.

Natural language processing in psychiatry: the promises and perils of a transformative approach

A person's everyday language can indicate patterns of thought and emotion predictive of mental illness. Here, we discuss how natural language processing methods can be used to extract indicators of mental health from language to help address long-standing problems in psychiatry, along with the potential hazards of this new technology.