diff --git a/research/index.md b/research/index.md index 01f0feb..39fdaef 100644 --- a/research/index.md +++ b/research/index.md @@ -5,7 +5,6 @@ group: research --- ## [Learning Lifespan Brain Anatomical Correspondence via Cortical Developmental Continuity Transfer](https://www.sciencedirect.com/science/article/pii/S1361841524002536?casa_token=PyunmY4ukk8AAAAA:3ljJmw3chie2GBAD2iq56kV_IsrocRM-XaqdBSHaZVQOhEny114H2kk-sBwpinfdqoTscxjO) -Due to the variability in cortical folding, neurodevelopmental stages, and limited neuroimaging data, inferring reliable lifespan anatomical correspondences is challenging. To address this, we leverage cortical developmental continuity and propose a transfer learning strategy: training the model on the largest age group and adapting it to other groups along the cortical trajectory. Evaluated on 1,000+ brains across four age groups (34 gestational weeks to young adults), results show that this strategy significantly improves performance in populations with limited samples and robustly infers complex anatomical correspondences across stages. Eigenmode +Due to the variability in cortical folding, neurodevelopmental stages, and limited neuroimaging data, inferring reliable lifespan anatomical correspondences is challenging. To address this, we leverage cortical developmental continuity and propose a transfer learning strategy: training the model on the largest age group and adapting it to other groups along the cortical trajectory. Evaluated on 1,000+ brains across four age groups (34 gestational weeks to young adults), results show that this strategy significantly improves performance in populations with limited samples and robustly infers complex anatomical correspondences across stages. + ## [BI-AVAN: A Brain-Inspired Adversarial Visual Attention Network for Characterizing Human Visual Attention from Neural Activity](https://ieeexplore.ieee.org/abstract/document/10636811)