Probing Language Brain Processes in Psychosis

We are thrilled to share that our lab is joining forces with Dr. Lena Palaniyappan and the DIALOG research team as collaborators on a transformative project funded by the Wellcome Trust. This ambitious initiative—Disorganized Language as a Target for Intervention in Psychosis (DIALOG)—seeks to uncover the brain mechanisms behind disrupted language in psychotic disorders, a long-overlooked but profoundly disabling symptom.

At the Montreal Neurological Institute, our lab will contribute by leveraging time-resolved neuroimaging methods, particularly magnetoencephalography (MEG), to study how the brain processes natural speech in real time. This work builds directly on the framework we developed in our 2020 Neuron paper (Donhauser & Baillet), where we showed how predictive processes in the brain during speech comprehension can be systematically related to those used in artificial language models.

Now, by aligning these computational tools with MEG recordings from individuals with psychosis, we aim to decode where and when language predictions go awry in the brain. This will help us probe not only the what of disorganized speech, but also the how—in terms of the neural computations that may be disrupted.

We are especially excited about the opportunity to scale this work across large, diverse clinical datasets and patient populations. The DIALOG initiative will draw from neuroimaging and language data in over 3,000 individuals, creating a unique opportunity to map predictive neural computations across the full spectrum of psychotic disorders.

Looking forward, we see this as a powerful step toward personalized approaches to mental health. By bringing together neuroscience, computational linguistics, and psychiatry, we hope to make lasting contributions to how we understand—and eventually treat—psychosis.

Read the full press release from the project’s coordinating site at the Douglas Research Centre.

Next
Next

A Landmark Collaboration to Probe the Foundations of Consciousness.