Congrats to our 2025 vEM Awesome image competition winners!

We had two categories this year:
       1. The Power of vEM – Stunning images that highlight the capabilities of volume EM
       2. Bringing Meaning to the Data – Images that show computational techniques of processing data

---

Category 1: The Power of vEM

1st place: Mário Silva (Universidade Federal de Minas Gerais, Brazil)

Image Description:
"Just Keep Swimming" in the Neuropil Sea
​
3D ultrastructural reconstruction of a motor neuron (green) from the mouse spinal cord using serial section array tomography by grid tape EM (ATUMtome).
Technique: Serial section array tomography with scanning electron microscopy (SEM)

---

2nd Place: Vanessa Chiappini - University of Turin 

Image Description:
Astrocytes: the Coral of the Brain. Astrocytic processes reconstructed from mouse hippocampus
​Technique: SBF-SEM

---

3rd Place: Georg Kislinger, Fereniki Moschogiannaki - DZNE

Image Description:
From sections to volume: a quiet neural tangle:
3D rendering of a translucent amber axon entwined with vivid-purple oligodendrocyte endoplasmic reticulum, mounted on a pedestal composed of stacked FIB-SEM images.
Technique: FIB-SEM

---

Category 2: Bringing Meaning to the Data

1st place: Hélène Roberge - VIB-BioImaging Core Leuven Belgium

Image Description:
CLEM in action: 3D registration and segmentation of contacts sites. Correlation of 3D structured illumination microscopy (SIM) with 3D FIB-SEM using Mitotracker as a fiducial marker enabled the precise alignment of fluorescence signals with ultrastructural features. This 3D registration allowed subsequent segmentation of a region of interest with contact sites (ER in green, lysosomes in red, and mitochondria in magenta), highlighting their dysregulation in the context of Alzheimer’s disease.


Computational technique: 
We developed a computational pipeline to bring meaning to our CLEM data by combining 3D registration, segmentation and visualisation. We used Mitotracker fiducials to precisely align structured illumination microscopy (SIM) with volume EM (FIB-SEM), mapping fluorescence signals onto ultrastructural features. We then segmented organelles (ER, lysosomes, mitochondria) to extract contact sites and quantify their spatial relationships. This workflow transforms raw multimodal images into interpretable maps of subcellular interactions, revealing dysregulation in Alzheimer’s disease.

---

2nd Place: Marcel Sayre - Macquarie University, Australia

Image Description:
A biological compass: modelling the head direction circuit of a tropical sweat bee. Synaptic connectivity of neurons segmented from serial block-face electron microscopy image data (top left) was used to create an anatomically constrained model (top right node graph) of a network that represents head direction in the bee brain (black horizontal plots).
​
Computational technique: 
The population activity of head direction neurons reconstructed from vEM was simulated using a firing-rate model

---

3rd Place: Veronica Falconieri Hays - Falconieri Visuals

Image Description:
Towards a universal AI model in volume EM:
Massive numbers of volume EM datasets can be used to train a generalist or universal AI model, which can then output high-quality segmentations of cellular features like mitochondria from any EM image.
​
Computational technique: Training and use of AI model for segmenting cellular features