Neuralink brain implant trials reach 50,000 hours: What to know

In a major update, Neuralink says it has begun putting more than 50,000 hours of natural recordings from clinical participants to work optimizing interface precision while streamlining setup and setting a new brain-computer interface performance record.

As mentioned in an October 1 blog post, Neuralink said participants have streamed more than 50,000 hours of neural data during the two years since its clinical trial began. It was noted that the first participant alone recorded more than 9,000 hours.

P9 said the new mode, “ Better than the prior model. Smoother, more directionally accurate, felt effortless. Didn’t have to push or anything, it just glided where I wanted it to go.”

The company has now developed tailored neural feature extractors leveraging extensive participant recordings collected from each person using those models for cursor control.

This approach now lets the system learn neural oscillations from past recordings instead of starting each task from scratch. Neuralink says these encoders turn noisy neural signals into more stable vector representations that can make decoding more accurate and durable.

Nonetheless, Neuralink's longer-term goal is to deploy a framework that merges federated learning with continual domain adaptation. In one experiment the company transferred a model from P9 to P2 while keeping 99% of its weights frozen despite static parameters, the model still surpasses P2’s current decoding baseline.