|
A Machine Learning View of Cognitive Neuroscience| title | A Machine Learning View of Cognitive Neuroscience |
|---|
| start_date | 2026/04/24 |
|---|
| schedule | 11h |
|---|
| online | no |
|---|
| location_info | visioconférence Big Blue Button |
|---|
| summary | Neuroimaging has changed our understanding of the human brain. Providing us with in-vivo measurements of the human brain's structure and function has opened a window into the relationships across anatomy, physiology, cognition and disease. Nonetheless, it remains largely a science of consent: these relationships exist if most of the neuroimaging community exists. The reproducibility crisis, prompted by the analysis of small samples and the misuse of statistical tools, has been instrumental in keeping neuroimaging a consent-based science.
In the first section of this talk, I will show how current machine learning approaches, such as natural language processing and large language models, can be harnessed to analyze the neuroscience literature and quantify the confidence the community has in different neuroimaging hypotheses.
Secondly, I will show how we can develop interpretable and reproducible results using novel neuroimaging analysis paradigms based on machine-learning paradigms. Through the development of new approaches for large-scale Bayesian Modelling based on deep learning architectures such as normalizing flows, I will show a novel analysis of the relationship between brain anatomy, function, and cognition at the population and individual levels. |
|---|
| responsibles | Bawden |
|---|
Workflow history| from state (1) | to state | comment | date |
|---|
| submitted | published | | 2026/04/23 07:10 UTC |
| |
|