DOI: 10.1192/j.eurpsy.2026.10515 ISSN: 0924-9338

Neural Engagement During a Video-Based Theory of Mind Task: A Pilot fMRI Study in Schizophrenia

I. París-Gómez, B. Yildirim, A. Serra-Mayoral, N. Ramiro-Sousa, P. Del Olmo, A. Aquino-Servin, S. García-Rojas, S. Leonangeli, S. Vargas, P. J. Mckenna, P. Fuentes-Claramonte, E. Pomarol-Clotet

Introduction

Theory of Mind (ToM) - the capacity to infer others’ mental states – is consistently impaired in patients with schizophrenia (SZ), and is a well-established predictor of social dysfunction and poor functional outcomes. While numerous functional imaging studies have examined ToM in SZ, few have employed videos, which may be important for ecological validity and a more nuanced understanding of social cognitive deficits in the disorder.

Objectives

This pilot study aimed to: 1) validate a novel video-based task for engaging the core ToM neural network, and 2) analyse differences in brain activation between SZ patients and healthy controls (HC).

Methods

15 HC and 6 SZ patients matched for sex, age and estimated premorbid IQ underwent fMRI while performing a video-based ToM task adapted from the Battery for Theory of Mind Assessment (BAT). The task included 16 videos, 8 showing social situations with a ToM component, and 8 control videos featuring people performing various non-social activities, each followed by a question (either ToM-related or related to a concrete aspect of the scene) (Figure 1). Data were analysed using FSL’s FEAT with a GLM approach (cluster-based threshold: z > 2.6, p < .05) to compare brain activity while responding to ToM relative to control questions.

Results

Behavioral performance was significantly worse in the SZ group compared to HCs on the ToM questions (p < .05). Both groups activated the canonical ToM network during the task, including the medial prefrontal cortex (mPFC) and the precuneus (both nodes of the default mode network), and the temporoparietal junction (TPJ). The comparison between patients and controls revealed four clusters of significant hyperactivation in the dorsomedial prefrontal cortex (dmPFC), precuneus/posterior cingulate cortex, the right inferior parietal lobule and the cerebellum (Figure 2).

Image 1:

Image 1: Long description.

Image 2:

Image 2: Long description.

Conclusions

This video-based task robustly engaged the canonical ToM network in both healthy subjects and SZ patients, supporting its utility for studying the neural correlates of ToM. Our results for the comparison between patients and controls are preliminary, but point to potential hyperactivation in parts of the default mode network in SZ patients.

Disclosure of Interest

None Declared

More from our Archive