Story perspectives

Brain Organoids Master Balance Task, Lack Long-Term Memory

2/20/2026

Summary

Story summary

Researchers at the University of California, Santa Cruz trained brain organoids to solve the inverted pendulum problem.
The organoids learned to balance a virtual pole with electrical feedback, achieving 46% success via adaptive reinforcement learning.
With random feedback, success fell to 4.5%, showing adaptive reinforcement raised performance.
Performance returned to baseline after a rest period, indicating a lack of long-term memory.
The findings were published in Cell Reports.

1 of 1

Story summary

Researchers at the University of California, Santa Cruz trained brain organoids to solve the inverted pendulum problem.
The organoids learned to balance a virtual pole with electrical feedback, achieving 46% success via adaptive reinforcement learning.
With random feedback, success fell to 4.5%, showing adaptive reinforcement raised performance.
Performance returned to baseline after a rest period, indicating a lack of long-term memory.
The findings were published in Cell Reports.