Co-reference and the topology of meaning in schizophrenia and depression

Studies using semantic measures from language models in psychopathology have predominantly focused on lexical-conceptual meaning, while the referential aspect of meaning has been targeted less. Here, we compare the performance of an automatic coreference resolution model across healthy controls (HC, n = 36), schizophrenia spectrum disorder (SSD, n = 41), and major depressive disorder (MDD, n = 42). We tested the hypothesis that co-reference in these clinical populations would be harder to resolve for such models. We then relate model performance to semantic and coreferential structure, employing a topological approach with speech graphs of both referential and lexical-semantic meaning. Results confirmed lower model performance in both SSD and MDD compared to controls, and a reduced structural divergence between semantic and referential graphs in SSD. Within SSD, model performance negatively correlated with positive symptoms, while in MDD, it was associated with verbal memory measures, suggesting partially distinct underlying mechanisms. These results newly profile lexical-conceptual and referential-semantic alterations against each other, linking two dimensions of meaning in a more integrated foundational model of language in major psychopathology.

Authors

Institutions

Publication Details

Journal
Schizophrenia
Published
2026-09-30
DOI
https://doi.org/10.1038/s41537-026-00802-6
Primary Topic
Schizophrenia research and treatment
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
article

Co-reference and the topology of meaning in schizophrenia and depression

Philipp Homan, Claudio Palominos, Wolfram Hinzen, Iris E. Sommer et al.
Schizophrenia
Schizophrenia research and treatment
article

Co-reference and the topology of meaning in schizophrenia and depression

Philipp Homan, Claudio Palominos, Wolfram Hinzen, Iris E. Sommer, Frederike Stein, Karla Fröhlich, Thomas Brochhagen, Rieke Roxanne Mülfarth, Svenja Seuffert, Tilo Kircher
article en

Abstract

Studies using semantic measures from language models in psychopathology have predominantly focused on lexical-conceptual meaning, while the referential aspect of meaning has been targeted less. Here, we compare the performance of an automatic coreference resolution model across healthy controls (HC, n = 36), schizophrenia spectrum disorder (SSD, n = 41), and major depressive disorder (MDD, n = 42). We tested the hypothesis that co-reference in these clinical populations would be harder to resolve for such models. We then relate model performance to semantic and coreferential structure, employing a topological approach with speech graphs of both referential and lexical-semantic meaning. Results confirmed lower model performance in both SSD and MDD compared to controls, and a reduced structural divergence between semantic and referential graphs in SSD. Within SSD, model performance negatively correlated with positive symptoms, while in MDD, it was associated with verbal memory measures, suggesting partially distinct underlying mechanisms. These results newly profile lexical-conceptual and referential-semantic alterations against each other, linking two dimensions of meaning in a more integrated foundational model of language in major psychopathology.

Schizophrenia
Institució Catalana de Recerca i Estudis Avançats (ES), University Medical Center Groningen (NL), Philipps University of Marburg (DE), Universitat Pompeu Fabra (ES), University of Zurich (CH), ETH Zurich (CH)
Openalex Percentile: Top 10%
Schizophrenia research and treatment
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.