Neural Noise

University of California BCI and neuroprosthetics research

2 min read

  • UC labs reported notable progress in brain-computer interfaces and neuroprosthetics (September 2025).
  • The work signals institutional pipeline and partnerships for BCI execution (University of California, tier-1). 1

Gardner updates

  • UC labs reported significant progress in BCI and neuroprosthetics research, with relevance to pipeline and partnership execution. 1

Weekly enrichment (2026-07-20)

  • The underlying item is a University of California news roundup (September 17, 2025) surveying BCI and neuroprosthetics progress across UC campuses, with AI as a recurring enabling theme.2
  • At UC San Francisco, neurologist Karunesh Ganguly trained an AI to account for day-to-day drift in a participant’s brainwaves, letting a paralyzed user control a robotic arm accurately for months without daily recalibration, with plans to test the system at home.2
  • A UCLA team (Jonathan Kao’s lab), publishing in Nature Machine Intelligence, built a noninvasive “AI co-pilot” BCI: a 64-channel EEG cap plus custom decoders and a camera-equipped AI that infers user intent under shared autonomy.3
  • The UCLA study tested four participants (three without motor impairment and one paralyzed by a T5-level spinal cord injury); all completed tasks significantly faster with AI assistance.3
  • For the paralyzed participant, the AI co-pilot improved performance about 3.9-fold, and he completed the robotic-arm task in roughly six and a half minutes with the AI whereas he could not finish it at all without it.3
  • At UC Davis, the Neuroprosthetics Lab (co-directed by Sergey Stavisky and neurosurgeon David Brandman) showed a speech BCI could also drive a computer cursor; the BrainGate2 participant adapted to cursor control in under 40 seconds to click apps and open links.4
  • The UC Davis lab had previously reported the most accurate speech BCI of its kind, translating brain signals to speech with up to about 97% accuracy (2024), and the roundup also notes UC Santa Barbara bionic-vision work and an FDA-approved adaptive deep brain stimulation therapy for Parkinson’s from UCSF’s Simon Little.2

Footnotes

  1. https://news.google.com/rss/articles/CBMinAFBVV95cUxPNnB5RkRSaTNfNk5LVXJ0QmxsczlFQU9XQ2FoMDdsWXlEWE83bk9kQy1aUTNGMHdqMUx6RDVyTTBpemlCTmtEZG9yemdIYXdrc0poVFV1d2EwdFVuejdkaEk1NGdQeTNvVkJrTGFVSmQ5QlNQTjJraVZ5eGlQWlpRWUFDbzI0MzlGSXNGWmF3TXkwQ0I5cm9aQ1J2aXo?oc=5 ↩ ↩2

  2. https://www.universityofcalifornia.edu/news/thrilling-progress-brain-computer-interfaces-uc-labs ↩ ↩2 ↩3

  3. https://newsroom.ucla.edu/releases/ai-brain-computer-interface-interprets-user-intent-ucla ↩ ↩2 ↩3

  4. https://www.ucdavis.edu/news/engineers-neurosurgeons-help-restore-autonomy-people-paralysis ↩

Graph View