BEIJING, Feb. 5, 2026 /PRNewswire/ -- A research team led by Prof. Hesheng Liu, founder and Chief Scientific Officer of Galaxy Brain Scientific Inc., has published a study in Nature that pinpoints the core functional circuit underlying Parkinson's disease (PD). The findings redefine the biological mechanism of PD and provide a scientific basis for developing precision neural circuit stimulation strategies for treating the condition.
PD currently affects over 13 million people worldwide and continues to pose a major clinical challenge, as existing treatments ranging from long-term medication to invasive deep brain stimulation (DBS) are often limited by diminishing efficacy or surgical risks. For decades, the disease was conceptualized as a movement disorder arising from dysfunction of the basal ganglia. We are tackling this big challenge head-on: clinical ready solutions that solve brain disorder. This study— an international collaboration between Galaxy Brain Scientific Inc., Washington University in St. Louis, Tsinghua University, Peking University, and Harvard University—analyzed precision functional neuroimaging data from over 800 participants to reveal a different story.
The researchers identified severe dysfunction in the somato-cognitive action network (SCAN)—a brain network essential for planning, coordinating, and executing actions—as the core feature of PD. In patients, this network shows abnormally high functional connectivity with deep brain regions, a signature not observed in other movement disorders such as essential tremor. "Our work shows that the disease is rooted in a much broader network dysfunction," noted Prof. Hesheng Liu. "The SCAN is hyperconnected to key subcortical regions in PD, and this abnormal hyperconnectivity disrupts not only movement but also related cognitive and automatic functions."
A critical finding of the research is that all existing effective therapies for PD share a common mechanism: they reduce this abnormally high connectivity between the SCAN and deep brain regions, essentially normalizing the circuit. "Our findings establish Parkinson's as a SCAN disorder. By targeting this network with personalized precision, we can now treat the disease more effectively than ever—potentially slowing or reversing its progression, rather than just suppressing symptoms," said co-author Nico U. Dosenbach, MD, PhD, the David M. & Tracy S. Holtzman Professor of Neurology at WashU Medicine.
This pioneering study was enabled by Galaxy Brain Scientific's proprietary personalized Brain Functional Sectors (pBFS) technology and its precision circuit stimulation system. The company's China NMPA -approved software and hardware system offers individualized precision targeting and non-invasive TMS stimulation with millimeter accuracy.
Galaxy Brain Scientific is already moving to bring these insights to the clinical frontline, having begun a pivotal registration trial for Class III devices dedicated to treating PD. Beyond Parkinson's, the company is also pioneering the application of this technology to treat other complex brain disorders, including Autism and Alzheimer's Disease.
** The press release content is from PR Newswire. Bastille Post is not involved in its creation. **
Galaxy Brain Scientific's Technology Enables Landmark Parkinson's Study Published in Nature, Redefining Disease Mechanism
SHANGHAI, April 3, 2026 /PRNewswire/ -- As AI becomes essential in K–12 education, many teachers face a practical question: How to teach AI through hands-on, project-based learning without advanced coding skills?
A recent workshop in Kathmandu University offers a clear answer. Conducted by DFRobot, an innovator in STEAM education, the three-hour workshop at the School of Engineering brought together teachers and department heads from across disciplines. Participants completed two hands-on AI projects while exploring how to translate these experiences into their own classroom teaching.
From Awareness to Classroom Practice
According to the Computer Science Teachers Association, while most educators support AI in the curriculum, many lack confidence in teaching it effectively. This workshop directly addressed that gap—moving from awareness to practical implementation.
Learning AI Through Hands-On Building
The training adopted a face-to-face, project-based approach, combining technical learning with pedagogy. It focused on deploying AI capabilities—such as speech and vision—on edge devices, connecting abstract concepts to interactive classroom applications.
Using UNIHIKER K10 hardware and Mind+ graphical programming software, participants completed two progressive projects:
Voice Interaction
Using the UNIHIKER K10 and the Mind+ graphical programming platform software, the workshop introduced voice-based human–machine interaction as an accessible entry point into AI learning. Teachers began by creating a simple voice-controlled system with commands such as "turn on the light," experiencing a fundamental AI interaction loop: wake → recognize → execute.
Building on this foundation, they extended the system to control on-screen movement through voice directions, transforming a basic function into a more engaging and interactive experience. Rather than focusing on complex programming, the training emphasized how simple voice commands can be translated into practical classroom applications.
Through this process, teachers gained a clear understanding of real-time voice interaction and its classroom potential.
Vision Interaction
The workshop introduced the HUSKYLENS AI vision sensor and its face recognition capabilities, helping teachers understand how machines perceive the world through visual data. Trainers explained the core workflow of face recognition—including detection, alignment, encoding, and matching—providing a clear technical framework for classroom application. Building on this foundation, teachers connected HUSKYLENS with the Mind+ programming environment and implemented real-time recognition tasks. Extending this approach, they developed a "Smart Pet" system by integrating HUSKYLENS with the UNIHIKER K10. The system could recognize different types of cats—such as Orange Tabby, Striped Tabby, and Siamese—and respond with corresponding interactive states.
Through this hands-on process, teachers gained a clear understanding of computer vision concepts and how visual input can drive interactive systems, enabling more engaging and project-based learning in the classroom.
A Model That Can Be Applied in AI Classrooms
The workshop highlights a practical approach to AI education that can be replicated across schools:
Accessible: Entry-level coding is all that's needed — suitable for K–12 learners.
Adaptable: Projects can be transformed into games, smart systems, or classroom tools
Structured: Aligned with project-based learning and real-world problem-solving
As part of the assessment, each teacher was asked to outline how the projects could be adapted for their own classrooms—highlighting a core objective of the training: not just using tools, but enabling curriculum design and effective knowledge transfer.
More importantly, it emphasized a critical shift—from understanding AI to applying it. By experiencing the full creation process, teachers gained the confidence to bring AI into their own classrooms.
** This press release is distributed by PR Newswire through automated distribution system, for which the client assumes full responsibility. **
How to Teach AI in the Classroom: A Hands-On Teacher Training Workshop at Kathmandu University
