Tutorial Spotlight

AI in Neuroimaging: From Research to the Clinic

A 3-hour tutorial translating AI-driven brain imaging advances into clinical workflows.

Read the Abstract Expected Impact

Abstract

Recent developments in AI for neuroimaging have dramatically advanced both structural and functional modalities, offering new possibilities for diagnosis, prognosis, and understanding of brain disorders. In structural neuroimaging, AI models, particularly deep learning architectures, have enabled automated detection and segmentation of lesions related to various diseases with high accuracy. These methods have accelerated large-scale analyses in population studies enabling robust quantification of disease-related biomarkers. In functional neuroimaging, AI facilitates extraction of meaningful connectivity patterns from fMRI and EEG data, supporting disease classification, cognitive state decoding, and prediction of clinical outcomes. Multimodal fusion approaches that combine structural MRI, fMRI, PET, and other imaging modalities provide comprehensive insights into brain organization and pathology.AI-driven preprocessing methods, including motion correction, denoising, and harmonization across scanners, further enhance reproducibility and clinical applicability.

Despite these advances, several challenges remain. Model interpretability and explainability are critical for clinical adoption, as "black-box" predictions hinder trust. Generalization across populations, imaging centers, and scanner types remains difficult.Limited availability of annotated datasets, particularly for rare conditions, constrains model training, while ethical, regulatory, and data privacy considerations continue to impact deployment.

This 3-hour tutorial, "AI in Neuroimaging: From Research to the Clinic", provides participants with a comprehensive overview of AI applications in brain imaging, from automated biomarker discovery to predictive modeling for neurological disorders.This tutorial is part of the ICVGIP 2025 scientific program, contributing an in-depth focus on AI-driven neuroimaging translation alongside the conference's broader agenda on computer vision, graphics, and image processing.

The tutorial will cover both theoretical foundations and practical workflows, highlighting how AI tools can accelerate diagnosis, improve clinical decision-making, and enable personalized patient care.

Learning Outcomes

By the end of this tutorial, participants will be able to:

  1. Understand key AI techniques used in neuroimaging, including deep learning, machine learning, and multimodal data integration.
  2. Identify clinically relevant biomarkers and assess AI-driven methods for early detection of neurological disorders.
  3. Recognize challenges and best practices in translating AI research into clinical neuroimaging workflows.

Tutorial Outline

Duration: 3 hours
Speakers: TBD

Expected Impact

Enhanced Knowledge: Participants will gain a clear understanding of AI applications in neuroimaging research and clinical practice.

Bridging Research and Clinic: The tutorial promotes translation of computational innovations into clinical workflows.

Community Building: Facilitates networking among researchers, clinicians, and industry professionals, fostering collaboration and future research opportunities.

Registration

Attendance at the tutorial requires registration through the ICVGIP 2025 conference system. Conference registration grants access to this session and the wider technical program; updated timelines and fee details will be announced shortly.

Please keep an eye on the official conference communication channels for deadlines related to early-bird and on-site registration.

ICVGIP 2025 Registration

A direct link to the registration portal will be shared here as soon as it is available.

Organising Committee