Neuromorphic Cameras Achieve Kilohertz Brain Imaging via New Self Supervised Reconstruction Algorithms

New research from South Korea uses event based neuromorphic cameras and self supervised algorithms to reconstruct high speed brain activity at 1000Hz.

By: AXL Media

Published: Apr 17, 2026, 7:16 AM EDT

Source: Information for this report was sourced from EurekAlert!

Repurposing Neuromorphic Sensors for Biological Discovery

A collaborative research team has successfully transitioned event based cameras from their traditional roles in robotics and autonomous driving to the field of functional neuroimaging. Unlike standard cameras that capture full images at fixed intervals, these neuromorphic sensors independently record only sharp changes in brightness at the pixel level. This fundamental shift in acquisition allows for the monitoring of subtle fluorescence changes tied to blood flow and neuronal activity without the data bottlenecks associated with conventional frame based systems. Published in the journal PhotoniX, the study demonstrates that these sensors can move beyond simple motion tracking to provide a scientifically rigorous view of the living brain.

Overcoming the Challenges of Small Signal Detection

Applying event cameras to biological systems presented a significant technical hurdle, as the sensors were originally designed to detect large contrast changes and fast motion. To address this, researchers from Seoul National University and the Korea Advanced Institute of Science and Technology conducted a quantitative optical characterization of the sensors. They focused on how reliably the camera could report slow, small amplitude variations in brightness that are typical of calcium signals and vascular dynamics. This calibration process established a critical link between binary, asynchronous events and the continuous intensity changes that neuroscientists require for accurate data interpretation.

Achieving Kilohertz Resolution in Cortical Vascular Imaging

The integration of the event camera framework into mouse models allowed the team to monitor cortical vascular dynamics with unprecedented speed. By comparing event streams to traditional widefield measurements, the researchers proved the sensor could track rapid fluctuations in blood flow at an effective rate of 1 kilohertz. This kilohertz temporal resolution is achieved while maintaining a large field of view, a feat that usually overwhelms the bandwidth of standard cameras. The ability to record at this speed without the overhead of full frame readouts suggests a new path forward for studying fast physiological processes in vivo.