A living-neuron computer places cultured neurons inside a computational loop. Electronics stimulate the network, electrodes record its activity and software decides how those signals affect a task. The biological network contributes dynamics and plasticity; conventional digital processing remains in the surrounding electronics and software.
Two-dimensional cultures
DishBrain and CL1 grow neurons close to planar electrodes. The geometry gives relatively direct electrical access and is easier to instrument than a dense three-dimensional organoid.
Three-dimensional networks
3D-MIND uses a flexible electrode structure embedded through a three-dimensional neuronal culture. Organoid computers take a different 3D route, allowing neural tissue to self-organise into organoid structures and then interfacing from outside or around the tissue.
What “programming” means here
Code normally controls stimulation, recording, decoding and feedback. The neural tissue changes through its own dynamics and plasticity. There is no instruction set comparable with a CPU, and identical stimulation does not guarantee an identical internal state.
Current limitation
The decisive problem is system-level usefulness. A neuronal culture may consume little energy locally while the complete apparatus requires amplifiers, pumps, temperature control, signal processing and conventional compute. Useful comparisons therefore need whole-system measurements.