Most previous computational studies of psychiatric speech have concentrated on lexical-conceptual meaning, using word embeddings and semantic similarity measures to quantify how ideas relate to one another across a stretch of discourse. That work has produced striking results, including demonstrations that semantic coherence degrades in psychosis in ways that defy simple similarity accounts. Yet meaning has at least two dimensions. Lexical-conceptual meaning concerns what words and concepts denote and how they cluster semantically. Referential meaning, by contrast, concerns how a speaker anchors those concepts to particular individuals and situations, maintaining a stable web of entities across sentences. When someone says that a neighbor complained and then adds that she would not stop, the listener must resolve that she refers back to the neighbor. Failures of this binding process, long noted clinically in formal thought disorder, had never been systematically quantified with modern language models across diagnostic groups until now.

The research team tested a specific hypothesis: that coreference would be harder for the automatic model to resolve in the speech of clinical participants than in controls. They assembled three groups of participants who produced spoken narratives, comprising 36 healthy controls, 41 individuals with schizophrenia spectrum disorder and 42 with major depressive disorder. The coreference resolution model was applied to transcripts of this speech, and its performance, essentially how accurately it could link referring expressions to the correct antecedents, was compared across groups. The results confirmed the prediction. Model performance was significantly lower in both the schizophrenia and depression groups relative to healthy controls, indicating that the referential threads running through clinical discourse are genuinely harder to track, even for a machine system trained on vast amounts of typical language.

What elevates the study beyond a simple group difference is its topological approach. The researchers constructed speech graphs, network representations in which linguistic units become nodes and the relations between them become edges, separately for referential structure and for lexical-semantic structure. Graph theory then allowed them to characterize the topology of each kind of meaning and, crucially, to compare the two. In healthy speech, the referential graph and the semantic graph diverge in characteristic ways: the web of entities a speaker tracks and the web of concepts they deploy are related but structurally distinguishable systems. In the schizophrenia group, the researchers found a reduced structural divergence between these two graphs, suggesting that the normal differentiation between referential and lexical-semantic organization is blunted. Meaning, in effect, loses some of its articulated internal architecture in schizophrenia spectrum disorder.

The correlational analyses sharpened the clinical picture considerably. Within the schizophrenia group, poorer model performance on coreference was negatively associated with positive symptoms, meaning that participants with more severe positive symptoms, the hallucinations, delusions and disorganized thinking that historically define psychosis, showed the greatest referential disruption. This aligns with a long clinical tradition linking formal thought disorder to breakdowns in the cohesion of discourse, but it now provides a quantitative, model-based handle on the phenomenon. Within the depression group, by contrast, coreference performance was associated with verbal memory measures rather than with symptom severity in the positive-symptom domain. This dissociation is theoretically important. It implies that although both disorders show referential difficulty at the group level, the cognitive machinery underlying that difficulty may differ: in schizophrenia it tracks the psychotic process itself, while in depression it may reflect the well-documented memory and executive impairments that accompany the illness.

The technical machinery behind these conclusions deserves attention, because it illustrates how computational linguistics is becoming a precision instrument for psychopathology. Coreference resolution is one of the harder problems in natural language understanding. A model must recognize that a definite description, a pronoun, a name and even an elliptical phrase can all denote the same individual, and it must do so across sentence boundaries where ambiguity abounds. When such a model, calibrated on ordinary language, stumbles more often on clinical speech, the errors are not random noise; they signal that the discourse itself provides weaker or more inconsistent cues about entity identity. Combined with speech graph topology, which quantifies properties such as connectivity and structural divergence between representational layers, the method yields a multidimensional profile of how meaning is organized, or disorganized, in an individual speaker.

The study forms part of a broader research program in which language is treated not merely as a symptom carrier but as a measurable biomarker of brain function. The same Barcelona group and their collaborators have previously shown progressive changes in descriptive discourse in first-episode schizophrenia using computational semantics, and related work has connected reduced linguistic coherence in psychosis to altered large-scale cortical hierarchy. The new findings extend this agenda by profiling two dimensions of meaning against each other rather than in isolation. The authors describe this as linking lexical-conceptual and referential-semantic alterations in a more integrated foundational model of language in major psychopathology. In practical terms, it means future computational assays of speech could in principle distinguish referential breakdown from semantic drift, refining both diagnosis and the tracking of illness course.

The clinical implications are potentially far-reaching. Speech samples are cheap, noninvasive and increasingly easy to collect, and language-based digital biomarkers are already being developed for early detection and monitoring of psychosis risk, including within European projects such as TRUSTING, which supported this work alongside the German Research Foundation. If coreference resolution performance and speech graph topology can be computed reliably from a few minutes of natural speech, they could complement clinical interviews with objective, quantifiable measures. The differential associations reported here, positive symptoms in schizophrenia and verbal memory in depression, hint that such measures might eventually help discriminate between conditions that overlap superficially, or track distinct therapeutic targets within a single patient over time. Much validation remains to be done, including longitudinal studies and replication across languages and clinical settings, but the direction of travel is clear.

The study also carries conceptual weight for theories of language in the mind and brain. Referential capacity, the ability to maintain a coherent cast of entities across discourse, has been proposed by some theorists, including Hinzen, to be a distinctive vulnerability point in schizophrenia, connected to the very architecture of human grammar and thought. The observation that referential and semantic graphs normally diverge structurally but converge in schizophrenia gives that proposal an empirical, topological form. Meanwhile, the depression findings caution against a single unified account of linguistic disruption in psychiatry: referential difficulty is not a psychosis-specific signature when it also appears, with different cognitive correlates, in mood disorder. Disentangling these threads is precisely what the new integrated framework is designed to do.

For now, the study stands as a demonstration that the fine grain of meaning, down to the humble pronoun and its antecedent, can be measured, modeled and related to symptoms and cognition in major mental illness. As language models grow more capable and clinical speech datasets grow larger, the boundary between computational linguistics and psychiatry is dissolving, and with it may come a new generation of tools that read the structure of thought directly from the structure of speech. The participants in the FOR2107 consortium whose narratives made this analysis possible contributed to a result that is as conceptually elegant as it is clinically promising: meaning, in its referential dimension, has a measurable topology, and in psychosis that topology changes.

Subject of Research: Computational analysis of coreference and speech graph topology in schizophrenia and major depressive disorder

Article Title: Co-reference and the topology of meaning in schizophrenia and depression

Article References: Co-reference and the topology of meaning in schizophrenia and depression. (n.d.). https://doi.org/10.1038/s41537-026-00802-6

Image Credits: AI Generated

DOI: 10.1038/s41537-026-00802-6

Keywords: schizophrenia, major depressive disorder, coreference resolution, natural language processing, speech graphs, computational psychopathology, formal thought disorder, positive symptoms, verbal memory, semantic topology, language biomarkers, digital psychiatry

For decades, clinicians and researchers have grouped the heterogeneous experiences of schizophrenia into broad domains. Positive symptoms such as hallucinations and delusions were separated from negative symptoms like blunted affect and social withdrawal, while disorganized speech and behavior formed a third cluster. These divisions, formalized in instruments such as the Positive and Negative Syndrome Scale, shaped drug development trials, genetic association studies, and the diagnostic criteria in the DSM and ICD. Yet the assumption that these categories reflect distinct underlying biological mechanisms has never been firmly established, and critics have argued that the dimensions are artifacts of the measurement instruments rather than discoveries about the illness itself.

Graph analysis, the approach at the heart of the new study, treats symptoms as nodes in a network and statistical associations between them as edges. Instead of asking whether a set of preassigned items load onto a particular factor, exploratory graph analysis applies algorithms derived from network science to detect communities of tightly interconnected symptoms. Methods such as the walktrap algorithm combined with regularization techniques can identify clusters that emerge purely from the pattern of relationships in the data. Researchers can then compare these data-driven communities against the conventional dimensions to see whether the traditional structure holds up under scrutiny.

The appeal of this methodology lies in its ability to sidestep some of the assumptions baked into classical factor analysis. Exploratory graph analysis has been shown in simulation studies to recover the correct number of dimensions more reliably than older techniques, particularly when samples are moderate in size or when the underlying factors are correlated. In psychometrics more broadly, the technique has spread rapidly, finding applications in depression, personality research, and quality-of-life measurement. Its arrival in schizophrenia research signals a broader shift toward network approaches that view mental disorders as systems of interacting elements rather than reflections of single latent causes.

Applying these tools to symptom ratings from people with schizophrenia, the researchers examined whether positive, negative, and disorganized symptoms genuinely form separable communities, or whether alternative configurations better describe the clinical reality. The stakes of this question are considerable. If symptom dimensions overlap more than assumed, clinical trials that measure only one domain may miss treatment effects that ripple across the network. If, on the other hand, the dimensions are robust, they remain valid targets for precision psychiatry approaches that aim to match patients to treatments based on symptom profiles rather than categorical diagnoses.

The study also speaks to a persistent puzzle in schizophrenia genetics. Genome-wide association studies have identified hundreds of genetic variants that contribute to risk, but connecting those variants to specific symptom dimensions has proven difficult, with studies of symptom genetics often yielding inconsistent results. Part of the inconsistency may stem from measurement: if the dimensions themselves are unstable across cohorts, instruments, and statistical methods, then genetic analyses built on top of them inherit that instability. Establishing whether the classical structure is reproducible under modern, assumption-light methods is therefore a prerequisite for meaningful biological discovery.

Network perspectives bring additional conceptual benefits. They make it possible to identify bridge symptoms, items that connect otherwise separate communities and may act as pathways through which dysfunction spreads. In depression research, for example, bridge symptoms such as sleep disturbance have been proposed as links between anxiety and mood clusters. In schizophrenia, identifying which symptoms bridge positive and negative domains could point to intervention targets with the broadest downstream impact, and could help explain why some patients deteriorate along multiple dimensions simultaneously while others remain relatively circumscribed in their difficulties.

The methodological rigor demanded by graph analysis is also part of the story. Exploratory graph analysis relies on estimating a regularized partial correlation network, typically through the graphical least absolute shrinkage and selection operator, which sets small spurious associations to zero and leaves a sparse network whose community structure can be extracted. Stability checks, such as bootstrapped estimates of edge weights and centrality indices, are essential to ensure that the detected communities are not statistical mirages. Cross-validation across independent samples provides a further safeguard, and the field has increasingly insisted on such replication before accepting any new dimensional structure as credible.

What makes this study timely is the convergence of several trends in psychiatric science. The National Institute of Mental Health’s Research Domain Criteria initiative has pressed researchers to move beyond diagnostic categories toward dimensions grounded in behavior and biology. Meanwhile, large-scale datasets and improved computational tools have made it feasible to test the architecture of psychopathology with unprecedented statistical power. Revisiting the symptom dimensions of schizophrenia with contemporary network methods is a natural next step in this program, and the findings carry implications that extend from the clinic to the genetics laboratory.

For clinicians, the message is that the familiar dimensional map of schizophrenia remains a useful, but not infallible, guide. For researchers, the study demonstrates that the tools of network science can interrogate long-standing psychiatric constructs in ways that classical psychometrics could not, potentially revealing where the traditional categories deserve preservation and where they require revision. As the field moves toward biologically informed classification, studies of this kind serve as a bridge, testing whether the clinical vocabulary accumulated over a century of observation can withstand the scrutiny of modern data science, and helping to ensure that future research into the causes and treatments of schizophrenia rests on foundations that can bear the weight.

Subject of Research: Symptom dimensions in schizophrenia analyzed with exploratory graph analysis

Article Title: Revisiting symptom dimensions in schizophrenia with exploratory graph analysis

Article References: Illing, S., & Leucht, S. (2026). Revisiting symptom dimensions in schizophrenia with exploratory graph analysis. Schizophrenia, 12(1), Article 72. https://doi.org/10.1038/s41537-026-00799-y

Image Credits: AI Generated

DOI: 10.1038/s41537-026-00799-y

Keywords: schizophrenia, symptom dimensions, exploratory graph analysis, network analysis, psychometrics, positive symptoms, negative symptoms, disorganization, precision psychiatry, psychiatric classification, statistical methods, mental health