> ## Content Index
> Fetch the complete content index at: https://www.neurabc.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Gestala Founder Phoenix Peng in Conversation with a Fields Medalist and a Nobel Laureate
- URL: https://www.neurabc.com/gestala-founder-phoenix-peng-in-conversation-with-a-fields-medalist-and-a-nobel-laureate/
- Published: 2026-11-10T04:03:40.000Z
- Updated: 2026-11-10T04:03:40.000Z
- Author: zhangguanghua

Phoenix Peng, founder and CEO of Gestala, recently met separately in the United States with Fields Medalist Terence Tao and Karl Deisseroth—recipient of the 2026 Nobel Prize in Physiology or Medicine and a pioneer of optogenetics. Two conversations spanning very different disciplines ultimately pointed to an underlying question the brain-computer interface industry has long under-addressed: we can now record massive amounts of neural signal, yet we are still far from truly “understanding the brain.”

The signal here is heavier than “another founder met a Nobel laureate.” Today’s BCI narrative is dominated by “capturing more signals and decoding them more accurately,” but Peng’s discussion with Deisseroth shifted the focus from observation to intervention, and from intervention back to mechanism. Functional ultrasound imaging (fUSI) indirectly reflects neural activity through hemodynamic changes, offering wide-field, high-resolution observation—but what it reads is still correlation. Optogenetics, by contrast, can precisely activate or inhibit specific cell types and neural circuits, providing clear causal variables. Using optogenetic intervention as a causal reference for fUSI signals—combined with electrophysiology and large-scale imaging—creates a path to build verifiable mappings between “local circuit changes, whole-brain network responses, and observed signals.”

This matters especially for ultrasound brain-computer interfaces. Unlike implanted electrodes that cover only local regions, ultrasound—through phased-array technology—can penetrate the skull non-invasively and form millimeter-level foci inside the brain, theoretically giving it whole-brain read-and-write spatial freedom. But “being able to read the whole brain” is not the same as “understanding the whole brain.” If we only see that blood-flow signals changed in a certain region, without knowing which neurons, which state, or which network pathway is involved, ultrasound modulation stays stuck at the level of empirical stimulation. In the conversation, Peng articulated a clear expectation for a Brain Foundation Model: it should not merely learn patterns from data, but must be testable through intervention experiments, answering “what happens if we change this?”

In research-paradigm terms, this pushes brain-computer interfaces from “signal engineering” toward “causal modeling of neuroscience.” The industry has long trained decoders on correlation. What will actually determine clinical efficacy going forward is the ability to predict intervention outcomes. Emotions, pain, and addiction are governed by cross-regional circuits, not single-region switches; the same location and the same parameters can produce entirely different results under different neural states. That means closed-loop ultrasound therapy must follow a logic of: first judge the brain’s current network state, then decide when, where, and how to stimulate, and finally continuously monitor whether the system evolves in the intended direction.

Gestala ties this technical thread directly to its own product roadmap. As China’s first ultrasound brain-computer interface company, it positions itself as an ultrasound-based platform with whole-brain read-and-write capability—doing both signal reading and neural modulation, while exploring the convergence of NeuroAI and brain science. The dialogue with top scientists is, in effect, laying the scientific foundation for that roadmap: without causal-level mechanistic support, ultrasound modulation cannot easily expand from single-point analgesia to more complex indications such as psychiatric disorders and stroke rehabilitation.

What deserve greater industry attention is how this exchange raises the competitive bar for ultrasound BCIs. Previously the race was about device miniaturization, focusing precision, and clinical progress. The next required capability is a cross-disciplinary “read–write–explain” closed loop—ultrasound engineering, optical intervention, electrophysiology, computational neuroscience, and foundation models must finally sit at the same table. Peng framed his notes around “from understanding the world to understanding the brain,” with an implicit message: the bottleneck for the next stage of brain-computer interfaces is not that models are not big enough, but that causal neuroscience knowledge is not deep enough.

For China’s ultrasound BCI track, this conversation is also a reminder: racing for registration certificates and first-in-human cases matters, but what will truly create separation is who builds a verifiable “brain causal model” first. Gestala is trying to redefine itself—from a hardware company into a composite of “frontier brain-science lab + ultrasound platform + NeuroAI.” The path is long and uncertain, but if it works, ultrasound BCI will no longer be “non-invasive version of electrical stimulation”—it will become new infrastructure for understanding the brain, treating brain disease, and connecting artificial intelligence.

This article is based on the public founder’s notes of Phoenix Peng, “In Conversation with the Nobel and Fields Medal Laureates: From Understanding the World to Understanding the Brain,” and has been independently interpreted and reconstructed as industry analysis.