Researchers Use AI and Connectome to Predict Brain Cell Activity

Researchers have made a groundbreaking breakthrough by combining the power of artificial intelligence with the connectome, a map of neurons and their connections, to predict brain cell activity without making a single measurement in a living brain. This achievement has the potential to transform how neuroscientists generate and test hypotheses about how the brain works.

A team of researchers from HHMI’s Janelia Research Campus and the University of Tübingen used the fruit fly optic lobe connectome to build a detailed deep mechanistic network simulation of the fly visual system. Led by Srini Turaga, Janne Lappalainen, and Jakob Macke, the team created an AI simulation that can predict the activity of every neuron in the circuit, accurately reproducing more than two dozen experimental studies performed over the past two decades. This new method has the potential to accelerate scientific discovery, and HHMI’s $500 million investment in AI@HHMI will support similar projects in the life sciences.

Predicting Brain Cell Activity with AI and the Connectome

The human brain is a complex and intricate system, comprising billions of neurons that communicate with each other through trillions of connections. Understanding how these neurons work together to enable behavior has been a longstanding challenge in neuroscience. Recently, researchers have made significant progress in this area by combining the power of artificial intelligence (AI) with the connectome, a detailed map of neurons and their connections. This innovative approach enables scientists to predict the activity of individual neurons without making a single measurement in a living brain.

The Connectome: A Map of Neurons and Their Connections

The connectome is a comprehensive map of the neural connections within a specific brain region or system. It provides a detailed understanding of how neurons are connected, which is essential for deciphering the neural code that underlies behavior. The fruit fly visual system connectome, used in this study, is an exemplary example of such a map. By analyzing the connectivity patterns within this connectome, researchers can infer the functional properties of individual neurons and their role in the visual processing pathway.

AI-Driven Simulation of Neural Activity

The researchers employed deep learning methods to build a detailed mechanistic network simulation of the fruit fly visual system. This simulation was constrained by the connectome data, ensuring that each neuron and synapse in the model corresponded to a real neuron and synapse in the brain. Although the dynamics of every neuron and synapse were unknown, the team used the connectome data to infer these parameters. By combining this information with knowledge about the circuit’s goal – motion detection – they created an AI simulation that can predict the activity of every neuron in the circuit.

Predicting Neural Activity without Measurements

The new model predicts the neural activity produced by 64 neuron types in the fruit fly visual system in response to visual input. Notably, it accurately reproduces more than two dozen experimental studies performed over the past two decades. This achievement demonstrates the potential of AI-driven simulations to transform how neuroscientists generate and test hypotheses about brain function. By enabling researchers to predict neural activity using only the connectome, this approach can accelerate scientific discovery and provide new insights into the brain’s workings.

Implications for Neuroscience Research

The new research provides a strategy for turning the wealth of connectome data generated by Janelia and other research institutions into an advanced understanding of the living brain. By bridging the gap between the static snapshot of the connectome and the dynamics of real-life computation in the living brain, this approach can facilitate the development of more accurate models of brain function. Furthermore, it showcases the potential of AI to accelerate scientific discovery, as demonstrated by HHMI’s $500 million investment in AI-driven projects in the life sciences.

Future Directions

The success of this study paves the way for future research directions. For instance, researchers can now use the model to simulate any experiment and generate detailed predictions that can be tested in the lab. The identification of cells not known to be involved in motion detection previously provides new avenues for investigation. Moreover, the approach can be extended to other brain regions and systems, enabling a more comprehensive understanding of brain function and behavior.