Roya Melanie Hüppi

This study investigates the effectiveness of automatic speech analysis in differentiating between patients with psychotic disorders and healthy control individuals with low or high levels of schizotypy within a German-speaking cohort. A range of acoustic and linguistic features was extracted from audio recordings and transcriptions obtained through structured interviews, open discourse, and picture description tasks, in which each participant described at least six pictures. These features served as input to train Random Forest classifiers, with linguistic features achieving consistent area under the receiver operating characteristic curve values of around 0.8–0.9 across all interview formats. In contrast, acoustic features performed significantly worse, particularly in the picture description task. Using linguistic features, we showed that increasing the number of described pictures significantly improved model accuracy, with performance plateauing after the combination of three pictures. Subsequent feature importance analysis revealed that patients produced simpler language, shorter sentences, less speech overall, and exhibited lower semantic coherence compared to healthy controls. These findings highlight the potential of automatic linguistic analysis as a scalable and objective tool for the detection of psychotic disorders in German-speaking populations.

Introduction: The course of psychotic disorders typically involves relapses. Early warning signs vary between individuals and are difficult to detect in clinical practice, especially in outpatient settings. Speech provides a quantitative clinical marker for detecting such early warning signs. The EU Horizon project TRUSTING (A TRUSTworthy speech-based AI monitoring system for the prediction of relapse in individuals with schizophrenia) aims to develop and evaluate a speech-based monitoring system for predicting imminent psychotic relapses. The study will examine the potential for prospective relapse prediction, and feasibility and usability of the monitoring system. Methods and analysis: In this multicentre observational study, n=240 remitted and at-risk-of-relapse adults with psychotic disorders and a comparison group with n=120 healthy participants (matched by age and sex) will be examined at six sites and in six different languages (German, French, Dutch, English, Czech and Turkish). The follow-up period is 6 months. The TRUSTING smartphone app will be used to collect weekly voice recordings through speech tasks; information on medication adherence, substance use, mood, anxiety and sleep quality; and motor data from a tapping task. Primary endpoints encompass model performance for relapse prediction, user adherence, transcription quality, usability of recordings and overall system usability. The primary analysis of user adherence, transcription quality, usability of recordings and overall system usability will be an unadjusted description of the respective proportions using 95% Wilson confidence intervals. Regarding relapse prediction, the predictive value of the risk estimates for relapse occurrence will be assessed using the area under the receiver operating characteristic curve. Exploratory analysis will be performed on potential speech-based markers associated with relapse risk. Ethics and dissemination: This study has been approved by swissethics (Business Administration System for Ethics Committees number: 2025–01177). Findings from this project will be disseminated through peer-reviewed journal publications and presentations at relevant scientific conferences, as well as at public events related to mental health. Trial Registration: ClinicalTrials.gov ID: NCT07397975 .

Formal thought disorder (FTD) is a core psychosis feature. Disentangling its dimensions requires tasks simple enough for formal modeling yet sensitive enough to capture individual variation across the psychosis spectrum. The semantic verbal fluency task offers precisely this: a structured behavioral trace of semantic memory sampling, amenable to computational analysis using distributed word embeddings. We hypothesized that this sampling process is governed by two dissociable mechanisms mapping onto FTD dimensions: initial retrieval drive (d ₀ ), quantifying the motivational resource sustaining production, and semantic search precision (α), quantifying how strongly similarity to the preceding word constrains each retrieval step from near-random to highly structured. We hypothesized that reduced d ₀ would track negative psychosis symptoms and alogia, while degraded α would track language disorganization and left inferior longitudinal fasciculus (ILF) fractional anisotropy. We tested these predictions in a primary (N = 120) and an independent replication sample (N = 249) of German-speaking individuals across the psychosis spectrum. Both parameters decreased with greater psychosis severity and, in the primary sample, they dissociated regarding their clinical correlates. d ₀ correlated negatively with negative symptoms, general psychopathology, and poverty of speech, consistent with a computational signature of alogia. α correlated negatively with positive symptoms and cognitive flexibility, and, in individuals with psychosis, positively with left ILF fractional anisotropy. The association between d ₀ and negative symptoms was replicated in the independent sample. These findings pave the way for mechanistic, automatically derived FTD markers capturing subclinical variation across the psychosis spectrum and mapping onto underlying cognitive and neural processes.

Aging is often accompanied by cognitive decline, but the extent, timing, and severity of this process is subject to large inter-individual variability. Predicting cognitive decline along a continuum, encompassing healthy age-related decline, mild cognitive impairment, and dementia, allows for more precise individual-level predictions. Previous work has demonstrated that machine learning (ML) models combining risk factors, clinical, neuropsychological, and structural magnetic resonance imaging (MRI) data can predict continuous cognitive decline. However, generalizability across independent datasets has rarely been evaluated. This study aimed to replicate previous findings using an independent dataset and to assess the models’ generalizability across acquisition sites. Multi-target random forest regression models were employed to predict annual rates of decline of the Clinical Dementia Rating Scale Sum of Boxes (CDR-SOB) and the Mini-Mental State Examination (MMSE). Results showed that adding structural MRI data to non-brain data enhanced model performance in the large-scale novel ADNI ( N = 1237) cohort, as previously attested for OASIS-3 ( N = 662). Further, detectable reductions in performance were measured when models were tested across datasets compared to within datasets. This performance degradation was not explained by distributional shifts of the target variables. Additionally, models trained on the most important features of the respective training set achieved nearly identical performance as those trained on all features when tested externally, suggesting high redundancy among predictors. In summary, multimodal ML models predicting continuous cognitive decline generalize partially to unseen cohorts, with statistically significant performance degradation, whereas unimodal models trained on structural MRI features do not.

Introduction: The course of psychotic disorders typically involves relapses. Early warning signs vary between individuals and are difficult to detect in clinical practice, especially in outpatient settings. Speech provides a quantitative clinical marker for detecting such early warning signs. The EU Horizon project TRUSTING (A TRUSTworthy speech-based AI monitoring system for the prediction of relapse in individuals with schizophrenia) aims to develop and evaluate a speech-based monitoring system for predicting imminent psychotic relapses. The study will examine the potential for prospective relapse prediction, and feasibility and usability of the monitoring system. Methods and analysis: In this multicenter observational study, n = 240 remitted and at-risk-of-relapse adults with psychotic disorders and a comparison group with n = 120 healthy participants (matched by age and sex) will be examined at six sites and in six different languages (German, French, Dutch, English, Czech, and Turkish). The follow-up period is 6 months. The TRUSTING smartphone app will be used to collect weekly voice recordings through speech tasks; information on medication adherence, substance use, mood, anxiety, and sleep quality; and motor data from a tapping task. Primary endpoints encompass model performance for relapse prediction, user adherence, transcription quality, usability of recordings, and overall system usability. The primary analysis of user adherence, transcription quality, usability of recordings, and overall system usability will be an unadjusted description of the respective proportions using 95% Wilson confidence intervals. Regarding relapse prediction, the predictive value of the risk estimates for relapse occurrence will be assessed using the area under the receiver operating characteristic curve. Exploratory analysis will be performed on potential speech-based markers associated with relapse risk. Ethics and dissemination: This study has been approved by swissethics (BASEC number: 2025-01177). Findings from this project will be disseminated through peer-reviewed journal publications and presentations at relevant scientific conferences, as well as public events related to mental health. ARTICLE SUMMARY: Strengths and limitations of this study: – International multicenter study spanning six sites, six languages, and five countries, enabling evaluation of the cross-linguistic generalizability of speech-based relapse prediction models in psychosis. – Human oversight enabling head-to-head comparison between human judgment and machine-generated predictions of relapse risk and ensuring the study’s safety and trustworthiness. – Involvement of people with lived experience of psychosis in both study and system design. – Inclusion of a matched control group to study intra- and interindividual variations in speech features over multiple measurements. – Insights into the feasibility of implementing artificial intelligence (AI)-based transcription and speech analysis in routine mental healthcare, and exploration of novel speech-based markers associated with relapse risk to enhance prediction, understanding, and prevention of relapse in the future.

Semantic language dysfunction is a hallmark of early psychosis, yet the underlying brain structural correlates are largely unexplored. In particular, it is unclear whether core deficits arise from disruptions to semantic representation, which refers to the stored knowledge of word meanings, or to semantic control, which entails top-down mechanisms that guide the retrieval and selection of context-appropriate semantic information. By dissociating semantic representation-related from semantic control-related performance, we aimed to identify the preferential impairment in early psychosis and its structural correlates in the ventral and dorsal language streams. We investigated N = 120 individuals across the psychosis spectrum: N = 40 individuals with early psychosis, N = 40 individuals with high schizotypy, and N = 40 individuals with low schizotypy. Participants with high and low schizotypy constituted the non-clinical comparison group. All participants completed tasks designed to isolate semantic representation-related and semantic control-related processes. Given the importance of accurate delineation, this study employed meticulous manual fiber tractography of diffusion tensor imaging (DTI) data to ensure reliable evaluation of ventral and dorsal pathway microstructure. Compared to individuals with high and low schizotypy, individuals with early psychosis showed pronounced deficits in semantic control-related performance, while the semantic representation-related measure remained largely intact. Mean diffusivity in the left inferior fronto-occipital fasciculus and left uncinate fasciculus was lower in the early psychosis group than in individuals with schizotypy. In the early psychosis group, fractional anisotropy in the left arcuate fasciculus was negatively correlated with semantic control-related performance, but no DTI measure was associated with the semantic representation-related measure. These results underscore semantic control-related performance as a core deficit in early psychosis and extend the conventional view that semantic processing is subserved primarily by ventral pathways. The arcuate fasciculus appears implicated in semantic control-related processes, indicating a more integrated interplay of dorsal and ventral streams in semantic language processing.