bioRxiv Science⌕ Search

Biology subjects

Zöllei, L.

Publications and source records attributed to Zöllei, L..

2 recordsLinked to original sources

The Infant Brainstem - A Multimodal Multiscale Postmortem Imaging Pipeline

Brainstem disorders in human infants - including sudden infant death syndrome (SIDS), the leading cause of postnatal infant mortality in the United States - are characterized by cellular and molecular abnormalities that conventional neuroimaging cannot detect. A substantial challenge arises from the fact that the immature myelination of the infant brain severely degrades MRI contrast, leaving the discrete nuclei and white matter tracts of the brainstem poorly resolved at the scale where pathology occurs. We introduce a postmortem imaging pipeline that bridges this gap by integrating four spatially registered modalities: whole-brain magnetic resonance imaging (MRI) (550 m), brainstem-specific MRI (150 m), polarization-sensitive optical coherence tomography (PSOCT, 10 m), and histology with immunohistochemistry (1.88 m). Our central finding is that PSOCT provides excellent tissue contrast independent of myelination state - directly overcoming the principal limitation of MRI in the infant brain -while enabling three-dimensional visualization of nuclei and tracts at resolutions 15 to 55 times finer than MRI alone. Histology provides cellular-level ground truth and validates the optical contrasts. Applied here to a normative 34-day-old infant brainstem, this pipeline establishes a generalizable framework for studying infant brainstem neuroanatomy in three dimensions, with particular relevance to disorders such as SIDS where gross anatomy is intact but cellular abnormalities remain the target of investigation.

neuroscience↗

Fast segmentation with the NextBrain histological atlas

Structural brain analysis at the subregion level offers critical insights into healthy aging and neurodegenerative diseases. The NextBrain histological atlas was recently introduced to support such fine-grained investigations, but its existing Bayesian segmentation framework remains computationally prohibitive, particularly for large-scale studies. We present a new, open-source tool that dramatically accelerates segmentation using a hybrid approach combining: machine learning, contrast-adaptive segmentation; target-specific image synthesis; and fast diffeomorphic registration (all three with GPU support). Our method enables highly granular segmentation of brain MRI scans of any resolution and contrast (in vivo or ex vivo) at a fraction of the computational cost of the original method (<5 minutes on a GPU). We validate our tool on four different modalities (in vivo MRI, ex vivo MRI, HiP-CT, and photography) across a total of approximately 4,000 brain scans. Our results demonstrate that the accelerated approach achieves comparable accuracy to the original method in terms of Dice scores, while reducing runtime by over an order of magnitude. This work enables high-resolution anatomical analysis at unprecedented scale and flexibility, providing a practical solution for large neuroimaging studies. Our tool is publicly available in FreeSurfer (https://surfer.nmr.mgh.harvard.edu/fswiki/HistoAtlasSegmentation).

neuroscience↗