Sunday, Jun 23: 9:00 AM - 5:30 PM
1036
Educational Course - Full Day (8 hours)
COEX
Room: Grand Ballroom 101
The neuroimaging community represents a collection of scientists who use imaging to address questions concerning development, cognition and clinical neuropsychiatry. Recently, both reproducibility and individual variability have emerged as prominent themes common to all domains within neuroimaging. Advances in image acquisition, processing and analysis softwares have improved our ability to map the cerebral cortex, which varies markedly in its structure and function across individuals. At the same time, not accounting for this variation is likely a key factor that has contributed to issues of reproducibility that are facing the neuroimaging community and has also limited the discovery potential of neuroimaging.
Central to all of these issues are spatial localisation and cross-subject correspondence i.e. making sure we are accurately representing and comparing corresponding neuroanatomical organization across subjects. An extensive body of literature has demonstrated that surface-based approaches better account for the structural and functional variability of the cerebral cortex across individuals compared to volumetric representations. Although pre-processed surface data are made publicly available by large neuroimaging consortia, transitioning to surface-based analyses for individual investigators’ own datasets is non-trival, hence the need for the training provided by this course.
This course will introduce fundamental concepts and toolboxes necessary for cortical surface analyses. The learning outcomes for this course are:
1. Describe how image acquisition affects the accuracy of data represented on the cortical surface
2. Understand how to use surface representations to model cortical structure and function
3. Describe how surface-based analyses can be applied to one’s own research question
4. Critically appraise neuroimaging studies that utilise surface-based approaches and contrast them with volume-based approaches
By the end of the course, attendees should be able to:
1. Describe which neuroimaging data need to be acquired to accurately analyse cortical signals
2. Describe the processing steps that are required to get data onto surfaces, including how these steps need to be tailored for developmental and clinical cohorts
This course is designed for all neuroimaging scientists interested in applying surface-based methods to their own work, especially those who are working with their own in-house datasets. We anticipate that most attendees will be early career researchers (students/post-docs), but we welcome attendees from all career stages.
Presentations
This lecture will cover the following:
- Introduction to how to acquire and preprocess MRI data for a surface-based study
- Illustrate the effect these choices have on surface-based modelling of structure and function
This lecture will cover the following:
- Introduction to cross-subject registration using cortical surfaces, and its benefits over volume-based cortical registration
- What are cortical parcellations and why are they useful?
- Generalising HCP pipelines to other datasets with more standard acquisition protocols
Presenter
Logan Williams, King's College London London, London
United Kingdom
This lecture will cover the following:
- Highlight the challenges specific to imaging the infant brain
- How to use Infant FreeSurfer
This lecture will cover the following:
- Different representations of brain function (soft vs. hard parcellation, static vs. dynamic functional connectivity)
- Approaches to data harmonisation
This lecture will cover the following:
- Methods for statistical inference on the cortical surface
- Statistical benefits of different cortical representations, e.g. parcellated vs. dense
Presenter
Anderson Winkler, Dr., University of Texas Rio Grande Valley
Human Genetics
Brownsville, TX
United States
This lecture will cover the following:
- Methods of capturing individual variation in cortical organisation
- Methods for linking cortical variation to behaviour/cognition
Presenter
Ruby Kong, National University of Singapore Singapore, Singapore
Singapore
This lecture will cover the following:
- Approaches to biomechanical modelling
- Case study: Insights into cortical folding mechanisms and evolutionary expansion
Presenter
Katja Heuer, Institut Pasteur Paris, Paris
France
This lecture will cover the following:
- Combined mapping of in vivo and ex vivo cortical data
- Demonstration of the toolboxes available for these analyses
Presenter
Sofie Valk, Max Planck Institute for Human Cognitive and Brain Sciences Leipzig
Germany
This lecture will cover the following:
- Limitations of current tools for surface-based analyses
- The theory of translating deep learning from Euclidean domains to surface manifolds
- A tutorial on how to implement deep learning for cortical surfaces (with code examples)
This lecture will cover the following:
- Focal cortical dysplasias – subtle epilepsy-causing lesions that can be neurosurgically removed.
- Why use cortical surfaces for detecting focal cortical dysplasias?
- How are these tools being used for clinical decision making?