King’s College London Researchers Develop Hybrid 2D-3D Neural Organoid System to Overcome Scaling and Reproducibility Barriers

King’s College London researchers reveal a new 2D-3D organoid method that improves reproducibility and electrical recording for drug testing. Learn more.

By: AXL Media

Published: Mar 28, 2026, 4:49 AM EDT

Source: Information for this report was sourced from King's College London

Bridging the Gap Between Dimensionality and Complexity

Neural organoids represent a frontier in neuroscience, offering a window into how brain tissue responds to drugs and genetic variations. However, the field has long struggled with a trade-off between the structural diversity of 3D models and the precision of 2D systems. While 3D organoids capture the cellular variety of a developing brain, they are notoriously difficult to replicate and even harder to monitor internally. According to Professor Deepak Srivastava, the inability to consistently record electrical activity from deep within 3D tissue has limited the scalability of functional genomic studies until now.

The Limitations of Traditional 3D and 2D Models

The primary challenge in organoid research stems from the inherent variability of lab-grown 3D tissues. Each organoid develops its own unique mixture of cell types, which, while healthy, creates "noise" that complicates drug testing. Furthermore, researchers are typically restricted to recording electrical signals from the surface of these spheres. Conversely, traditional 2D networks allow for easy monitoring of many neurons simultaneously but lack the essential support cells and diverse neuron types that characterize a real human brain. This lack of complexity often renders 2D results less applicable to actual clinical scenarios.

A Hybrid Approach to Reducing Biological Variance

To solve these issues, the King’s College London team developed a method to extract the "best of both worlds" by first growing 3D organoids and then breaking them down into individual cells. This process, known as dissociation, allows researchers to harvest a diverse mixture of developing neurons that can then be plated onto a 2D surface. By pooling cells from multiple different organoids, the team successfully averaged out the variation between individual samples. This resulted in multiple, highly similar neural networks placed side-by-side, effectively transforming a variable biological process into a standardized, high-throughput assay.