Subsense—an artificial intelligence (AI)-native neurotechnology company developing non-surgical, nanoparticle-based brain-computer interfaces (BCIs)—has announced a major expansion of its AI-driven NanoBCI strategy, bringing together AI, nanotechnology and neural interfaces to advance a new category of BCI technology. Unlike traditional approaches that either rely on implanted hardware or capture neural activity from outside the skull, Subsense is developing a nanoparticle-based interface intended to eventually enable neural activity to be read or stimulated without permanently implanted electrodes.
The company is using AI across the development process to design, score and evaluate nanoparticle candidates, and model how they could be navigated to specific areas of the brain. Rather than applying AI only after neural data are collected, Subsense is using these data upstream to help determine what the underlying interface itself should look like. This extends the role of AI beyond software and neural-data analysis into the materials and delivery layers of the BCI itself, the company states.
Subsense also says it is applying AI at the materials-development stage, where nanoparticle composition, structure and field-driven behaviour can be evaluated computationally before synthesis. Subsense further claims to be “the first in the world” to develop an integrated computational workflow that evaluates and ranks magnetoelectric nanoparticle candidates across five coupled design dimensions—colloidal stability, core magnetostriction, shell piezoelectric response, core-shell interface strain transfer, and biocompatibility—and has built two- and three-dimensional physics simulations of nanoparticle navigation so that candidates can be prioritised before experimental testing.
Additionally, Subsense has been awarded US$250,000 in Google Cloud credits to scale these computational workflows, explore a broader nanoparticle design space, and strengthen the AI infrastructure supporting its research and development (R&D). This support is helping the company expand the computational resources behind its AI-enabled design and simulation work.
“We are combining AI with physics-based modelling to help identify promising nanoparticle architectures, understand how they may respond to externally applied magnetic fields, and guide which candidates move forward into experimental validation,” said Parminder Mankoo, head of AI at Subsense. “The goal is to narrow a highly complex design space computationally so that our laboratory work can focus on the candidates with the strongest potential.”
“Most of the BCI field has been forced into a trade-off,” added Tetiana Aleksandrova, co-founder and chief executive officer (CEO) of Subsense. “You can get closer to the neurons and obtain better signals, but doing that generally requires surgery—or you can stay outside the skull, but accept limitations in what you can access. We started Subsense because we believe that trade-off should be challenged.
“Our nanoparticle-based approach is designed to challenge that trade-off by creating a path to neural access that does not require surgery. AI helps us move faster and with greater precision in R&D by allowing us to explore nanoparticle design, navigation and performance computationally before committing to a physical candidate. It accelerates how we develop the technology.”
Subsense is currently developing separate nanoparticle systems for stimulation and neural reading. For stimulation, the company is working with magnetoelectric nanoparticles designed to respond to externally generated magnetic fields and produce localised electrical effects. For neural reading, Subsense is studying plasmonic nanoparticle approaches intended to generate optical signals that could ultimately be detected by external hardware.
The two programmes are being advanced in parallel as Subsense works toward a longer-term architecture that could eventually combine neural reading and stimulation within the same platform. The company remains in the preclinical stage, and is engaging with the US Food and Drug Administration (FDA) as it develops its research and regulatory strategy, as noted in a press release.
As it advances its platform, Subsense has assembled a multidisciplinary advisory network spanning deep brain stimulation (DBS) and neuromodulation, functional neurosurgery, movement disorders, amyotrophic lateral sclerosis (ALS), neural-interface engineering, regulatory strategy, and medtech commercialisation. This group is intended to help Subsense pressure-test its technology against the same scientific, clinical and practical standards applied to established implantable neurotechnology, the release adds.
“AI is becoming increasingly capable of helping us understand and engineer systems of extraordinary complexity, and the brain is perhaps the most complex system we can work with,” commented Ray Kurzweil, who recently joined Subsense as a product and vision advisor. “Subsense’s approach brings AI into the development of the interface itself, while nanotechnology creates a potential path beyond the limitations of surgically implanted hardware.”
