Ryota Kanai

I'm very excited to share our new mathematical framework for consciousness! co-authored with @oizumim and Chanseok Lim. We use principal bundle geometry to characterize the structure of qualia. I hope to find likeminded people to explore this new frontier.

I tried our own BMI + Mixed Reality robot control at Araya for the first time today. This system was built by @kaixhin , @RousslanDossa and the team. It was an amazing experience. I could intuitively pick up a mug just by imagining the action. Still some rough edges, but seeing EEG + gaze seamlessly drive a robot end-to-end was a powerful experience. Even with non-invasive EEG, it truly felt like experiencing a BMI.

I know they don't understand human language. But what if we could directly inject the meaning of a question into the brain of a mouse (or other animals) so they just understand the question and answer?

A random question. Could we ever ask a mouse a simple Yes/No question? I don't mean training them to press a "Yes" or "No" button for a reward. I mean genuinely asking something like "Are you hungry right now?" and getting a real answer. Is that possible?

It's awesome to just sit there with an MR headset, look at objects in front of you, pick an action by thinking about it, and have a robot perform the action for you. And it's totally possible with a non-invasive brain-robot interface 🧠🤖🥽

🚀 Exciting News! 🚀 The official Trusted Neurotechnology Initiative (TNI) website is now live! Promoting the responsible development of neurotechnology. Our foundational guidebooks and evidence book are now easily accessible in a seamless web format ✨ Link in the replies! 👇

@dimitrispp I just found out that the non-triviality I wrote in the first post may not work. But I think I can fix it. It's interesting to keep looking for good definitions of what I want to capture.

@dimitrispp I also noticed that it's straight forward to formulate something like partial information decomposition with this framework. So I want to define "Integrated Koopman NTIC" and find it in the brain.

Thinking about defining Koopman NTIC.

We can proceed scientifically by defining what can be operationalized and studied, without settling the full metaphysical question of what consciousness “really is.” This is similar to the transition from Newtonian physics to analytical mechanics, which got rid of the vague notion of force. Once we have a workable definition of consciousness, which not everyone would agree, we can build a powerful formal framework without requiring ontological closure. In the same spirit, the goal is to develop an analytical framework of consciousness based on clear, usable definitions, rather than waiting for a perfectly resolved concept.

@kanair Consciousness isn't a well-defined term. It therefore remains outside science. The challenge is to find useful definitions, as Newton did with the esoteric term "force" by defining it as mass times acceleration. Only then did it become scientific. x.com/modelsarereal/…

I still love studying consciousness, but I'm now more interested in abstract and mathematical ideas. It may be becoming a bit different from most research, and not many people like it. But sometimes I meet someone thinking similar things, and that makes me happy.

@MancerAI_ @Plinz Since your bio says "Physics → ML", this might be of your interest. Basically, we import ML maths (mechinterp) to neuroscience and consciousness research. osf.io/preprints/psya…

@kanair @Plinz This sounds very interesting, I would love to read (and discuss!)

@DamiDina I'm also interested in linking the koopman operator framework now to IIT-like consciousness research. arxiv.org/abs/1611.06664

@DamiDina For example, this is something we have been working on. This is useful for AI, neuroscience, and consciousness. osf.io/preprints/psya…

AAAI 2026 Spring Symposium - Machine Consciousness: Integrating Theory, Technology, and Philosophy open.substack.com/pub/cimcai/p/a…

@burnt_jester The shift is from empirical neuroscience to mathematical ideas of consciousness. It became more difficult to be excited about cognitive neuroscience findings compared to mathematical concepts that captures my intuition about consciousness.

@kanair What made you shift to abstract maths? The quantitative nature of maths vs the qualitative nature of consciousness?

@dimitrispp This might work as a method to identify an NTIC core in the brain.

@kanair That's intriguing

Exciting news! 🚀 Araya and Fluent BCI are teaming up to revolutionize neural speech decoding 🤝 We're advancing "Speech Decoding BCI" to restore or replace lost communication abilities. Check out the thread below for the full story 👇