Michele Svanera

Diagnostic Labels and Measurement Timing Drive Systematic Inconsistency in ADNI Neuroimaging Data

This paper identifies systematic biases in ADNI neuroimaging data that affect machine learning predictive modeling and proposes a framework to address these issues.

Delineating In-Vivo T1-Weighted Intensity Profiles Within the Human Insula Cortex Using 7-Tesla MRI

This paper uses 7-Tesla MRI to identify distinct T1-weighted intensity profiles within the human insula cortex, revealing compartment-specific signal variations and highlighting atlas-dependent parcellation.

CEREBRUM-7T: Fast and Fully-volumetric Brain Segmentation of 7 Tesla MR Volumes

CEREBRUM-7T is a fast, fully-volumetric brain segmentation method for 7T MRI that outperforms its training labels and is suitable for neuroimaging studies.

A Self-Supervised Deep Neural Network for Image Completion Resembles Early Visual Cortex fMRI Activity Patterns for Occluded Scenes

This paper compares a self-supervised CNN for image completion with fMRI data from early visual cortex, finding the CNN's decoder pathway more similar to brain activity than a supervised model.

Transfer learning of deep neural network representations for fMRI decoding

This paper transfers deep neural network features to fMRI data to improve decoding performance by reconstructing CNN features from imaging data.