Seminal
The magazine and atlas of EU-funded researchA new publication from the web and AI team at Scott Charnley

Atlas / Netherlands

European Research Council grant · Running

MINDSEYE Reading the mind’s eye: AI-inspired personalised brain models of mental imagery

Coordinator
Universiteit Maastricht, Netherlands
Runs
1 September 2024 to 31 August 2029
Status
Running
EU funding
€2.5m
Total cost
€2.5m
Partners
1 organisations in 1 countries
Type
European Research Council grant
Grant agreement
101141800

What the project is doing

In the project's own words, from its summary on CORDIS.

How do we create mental images? Despite extensive research, we still miss an overarching mechanistic understanding of mental imagery: How do the various brain regions involved in mental imagery contribute to the unified percept in our mind's eye? Why do we experience imagery so differently ranging from no to extremely vivid mental images? In this project, we propose a novel perspective on mental imagery, viewing it as a personalised computational process that takes into account individual brain characteristics.

The task of this computational process is to progressively transform abstract object descriptions (input) into sensory-like visual representations (output) through feedback connections in the brain’s processing hierarchy. To unravel the stages of this conversion process, we will employ advanced fMRI techniques to measure neural activity across the entire brain at an unparalleled level of detail. Using sophisticated analysis methods, we will decode topographic and semantic information about the content of mental images from activated brain areas.

To investigate the causal involvement of specific areas, we will transiently disrupt their activity using transcranial magnetic stimulation (TMS). Informed by the fMRI and TMS data, we will develop the first personalized AI-inspired neural network model of mental imagery. This model will simulate the emergence of sensory representations backwards through the processing hierarchy and predict the perceived vividness of generated images for each individual.

By implementing the model into a Brain Computer Interface, we will enable participants to see imagined objects on a screen during fMRI scanning. This opens up new applications, such as training the strength of mental imagery by providing neurofeedback based on predicted vividness. This interdisciplinary project, at the intersection of psychology, neuroscience, and AI, will provide an integrative framework of the generation of subjective experiences in the mind's eye.

Topics

Who is involved

OrganisationCityCountryRole
Universiteit MaastrichtMaastrichtNetherlandsCoordinator

Source: CORDIS, European Commission. Data of 5 Nov 2026, reused under the Commission's reuse policy. View on CORDIS · Grant DOI