What We Brought to Bordeaux and What OHBM Sent Home with Us: The TNI Lab at OHBM 2026

The Annual Meeting of the Organization for Human Brain Mapping took place from 14 to 18 June 2026 at the Palais 2 l’Atlantique in Bordeaux. Bringing together researchers from neuroimaging, neuroscience, medicine, statistics and artificial intelligence, OHBM offered an exciting mix of scientific perspectives that closely reflects the interdisciplinary work of our lab.

Our eleven contributions shared a common goal: extracting clinically meaningful information from complex brain images while preserving the perspective of the individual patient. The projects ranged from imaging biomarkers and network-based atrophy patterns in autoimmune limbic encephalitis to functional reorganization after hemispherotomy. We also presented transformer-based methods for faster and more robust white-matter analysis, including approaches designed for brains with lesions, as well as a prospective evaluation of deep-learning-based focal cortical dysplasia

detection across 3- and 7-tesla MRI. Further studies addressed brain asymmetry in Rasmussen’s encephalitis, possible glymphatic alterations in focal epilepsy, ultra-low- field postictal diffusion imaging, seizure-related changes in connectome identifiability and a multimodal, multi-field-strength epilepsy dataset.

Across the program, we encountered many topics that connected directly with our research. Longitudinal and multi-dataset studies emphasized the importance of understanding how brains and diseases change over time. Presentations on deep learning addressed the challenge of creating models that remain reliable across scanners, populations and neurological conditions. Sessions on multimodal imaging explored how different measurements can be combined, while work on virtual brain twins pointed towards models that may eventually simulate disease processes and interventions rather than merely describe scans.

What made OHBM especially memorable, however, was the extraordinary scientific exchange. Poster sessions turned into detailed discussions about methodological decisions, unexpected results and the practical limitations of working with heterogeneous clinical data. Researchers from clinical, computational and engineering backgrounds often approached the same result from completely different directions, producing new questions and ideas that would not have emerged within a single discipline.

Conversations with our collaborator Martin Reuter and members of the

FastSurfer/Deep-MI and MELD London communities were particularly valuable for our work on anatomical analysis and lesion detection. More broadly, the conference gave us the opportunity to reconnect with colleagues, meet researchers whose work we had previously known only from publications and explore new possibilities for collaboration.

Spending time together as a lab was just as valuable as attending the formal program. Between presentations and scientific conversations, we celebrated each other’s work, shared new ideas and experienced the energy of a large international research community. OHBM 2026 was not only an opportunity to present our work. It was an inspiring week of ambitious science, open discussion and genuine community. The conference reinforced the importance of developing methods that remain robust across lesions, scanners and field strengths, while also showing how much progress depends on

collaboration across disciplines.