Inducing Dyslexia in Vision Language Models

ICLR 2026 Conference Submission9475 Authors

Keywords: VLMs, Dyslexia, Reading, Cognition, Causal hypothesis testing, NeuroAI

TL;DR: We model dyslexia in vision-language models by selectively perturbing word processing units, reproducing core reading deficits while preserving general language and visual abilities.

Abstract:

Dyslexia, a neurodevelopmental disorder characterized by persistent reading difficulties, is often linked to reduced activity of the visual word form area in the ventral occipito-temporal cortex. Traditional approaches to studying dyslexia, such as behavioral and neuroimaging methods, have provided valuable insights but remain limited in their ability to test causal hypotheses about the underlying mechanisms of reading impairments. In this study, we use large-scale vision-language models (VLMs) to simulate dyslexia by functionally identifying and perturbing artificial analogues of word processing. Using stimuli from cognitive neuroscience, we identify visual-word-form-selective (VWF-selective) units within VLMs and demonstrate that targeted ablation of these units, unlike ablation of random units, leads to selective impairments in reading tasks while general visual and language comprehension abilities remain intact.
In particular, the resulting model matches dyslexic humans' phonological deficits without a significant change in orthographic processing. Additionally, the model’s VWF-selective units predict human-VWFA neural responses better than random units and the ablated model mirrors dyslexic behavior in font sensitivity.
Taken together, our modeling results replicate key characteristics of dyslexia and establish a computational framework for investigating reading disorders.

Primary Area: applications to neuroscience & cognitive science

Submission Number: 9475