We develop multimodal machine learning methods and computational analysis techniques that bring together single cell omics, spatial omics, histology, and neuroimaging data, with the goal of better understanding the brain in health and disease and moving towards personalised diagnosis and treatments. Towards this aim, we bring new advancements in data analysis and machine learning from computational biology and computer vision to neuroscientific data to advance our understanding of neurovascular and neurodegenerative phenotypes.

We are currently focussing on the following key topics:

Spatial omics analysis pipelines for brain tissue

We aim to gain a deeper understanding of the cellular organisation of the brain by analysing gene and protein expression in a spatial context. Using spatial omics datasets such as 10x Visium and MERFISH we extract spatial gene and protein expression patterns and relate them to morphology, single-cell RNA sequencing data, and disease variables. For this, we have recently developed the spatial omics analysis toolbox Squidpy and the subcellular analysis method Campa in collaboration with Fabian Theis at Helmholtz Munich. In collaboration with the Dichgans lab, we are currently developing a best-practice pipeline for cellsegmentation from MERFISH data.

Key publications

• Palla, Spitzer et al, Nature Methods 2022

• Spitzer, Berry et al, Nature Methods 2023

Multimodal data analysis and representation learning

We aim to combine datasets across diseases and modalities into a multimodal and multiscale brain atlas. By integrating data from different sources and diseases, we hope to identify features that are conserved across diseases and study correlations between modalities. Towards this aim, we are currenly focussing on learning general representations from spatial omics data, and combining these with histological representations that our collaboration partner Christian Schiffer at Forschungszentrum Jülich is learning. This work is done in the context of the Helmholtz Imaging Project Xbrain.

In an orthogonal line of research we are investigating how we can use the rich information from spatial omics data to inform histological analysis by for example building models to predict cell types from DAPI stains. In addition, we recently build a multi-modal cross-species single-cell pancreatic atlas, and aim to apply similar techniques to brain data.

Key publications

•  Schiffer et al, arXiv 2025

•  Yang, Spitzer et al, Nature Communications 2025

•  Spitzer et al, MICCAI 2018

•  Spitzer et al, ISBI 2027

Neuroimaging representation learning

We establish pipelines for extracting interpretable features from neuroimaging data using advanced deep learning algorithms. In collaboration with Anna Dewenter, we are developing an algorithm to detect and segment cortical superficial siderosis (cSS) on brain MRI scans to facilitate automated cSS assessment in large datasets. Our long-term goal is to assess the prevalence of superficial siderosis as an estimate of cerebral amyloid angiopathy on the population level. In addition, as part of the MELD project, we are using graph neural networks to identify focal cortical dysplasia lesions in a large cohort of epilepsy patients. Our goal with these projects is to develop and apply advanced neuroimage analysis techniques that will help us to better understand structural brain changes associated with neurodegenerative diseases. 

Key publications

•  Ripart et al, Jama Neurology 2025

•  Spitzer et al, MICCAI 2023

•  Spitzer, Ripart et al, Brain 2022

Our group is also associated with the Computational Health Center at Helmholtz Munich and part of the Munich cluster for Systems Neurology (SyNergy), where we are coordinating the Omics Analysis&Integration Hub.

We are currently hiring a postdoctoral researcher for multimodal omics and image analysis. 
Please look at the job posting or contact hannah.spitzer@med.uni-muenchen.de to learn more.