Biocomputers

DishBrain

DishBrain connected cultured mouse and human cortical neurons to a Pong-like closed loop. See how the experiment worked and what changed during gameplay.

Schematic of cultured neurons connected to a multi-electrode array
DishBrain used a closed loop between a planar neural culture, an electrode array and a simulated environment.Biocomputers original schematicasset record

DishBrain connected cultured mouse or human cortical neurons to a high-density multielectrode array and embedded the culture in a real-time virtual environment modelled on Pong. Electrical stimulation encoded information about the game state; recorded neuronal activity controlled the paddle.

DishBrain closed-loop Pong experiment showing neurons on a high-density multielectrode array connected to a simulated environment
Published overview of the DishBrain closed loop: stimulation represented the game state, recorded neural activity controlled the paddle and feedback returned through the electrode array.Kagan et al. 2022 · CC BY 4.0asset record

The 2022 Neuron paper reported changes in gameplay performance under structured feedback. The work is frequently compressed into the phrase “brain cells learned Pong”, but the scientific claim is narrower: adaptive behaviour emerged in a constrained closed-loop preparation.

Evidence status

Peer-reviewed research demonstration. The electrophysiological and behavioural results are established by the paper; broader language around intelligence, sentience or agency remains more interpretive.

Biological substrate
Mouse and human cortical neuronal cultures
Task
Closed-loop Pong environment
Interface
High-density multielectrode array
Published
Neuron, 2022