ASCENT: Annotation-Free Self-Supervised Contrastive Embeddings for 3D Neuron Tracking in Fluorescence Microscopy

Abstract Tracking individual neurons in dense 3D microscopy recordings of fast-moving, deforming organisms is a critical challenge in neuroscience. This task is often hindered by complex tissue motion and the need for laborious manual annotation. Here, we introduce ASCENT, an annotation-free computational framework that learns robust representations of neuronal identity that are invariant to tissue deformation. ASCENT trains a Neuron Embedding Transformer (NETr) through a self-supervised contrastive scheme on augmented data that mimics tissue deformations. NETr generates highly discriminative embeddings for each neuron by combining visual appearance with positional information, refined by the context of all other neurons in the frame. On a challenging benchmark C. elegans datasets, ASCENT achieves state-of-the-art tracking accuracy, surpassing supervised methods. The method is robust to image noise, highly data-efficient, and generalizes across different imaging conditions. By eliminating the annotation bottleneck, ASCENT provides a practical and scalable solution for the robust analysis of whole-brain neural dynamics in behaving animals.