Synaptrix Labs has released the first public demo of Atlas, its non-invasive brain-computer interface. The behind-the-scenes video, posted on 12 March, shows a user controlling a wheelchair by thinking about the direction they want to move. No implant. No surgery. A headset on the scalp and a deep learning stack doing the rest.

The demo is early-stage by the company’s own admission, but the implications are significant. If a non-invasive device can achieve reliable, real-time wheelchair navigation from scalp-measured EEG alone, it would represent a meaningful step toward making BCI-driven mobility accessible at consumer scale, not just within clinical research settings.

Deep Learning First, Better Sensors Later

Synaptrix Labs was founded in 2023 by CEO Aryan Govil, who spent five years in a clinical research lab at NYU Langone, and CTO Eric Yao, whose background is in time-series forecasting and quantitative modelling. The company is headquartered in Manhattan with seven full-time employees and has secured seed funding led by Mark Cuban, estimated between $8 million and $10 million.

The company’s thesis is deliberately contrarian within the BCI industry. Most competitors invest heavily in improving sensor hardware to get cleaner neural signals. Synaptrix is doing the opposite: treating signal quality as a software problem and prioritising AI-driven noise reduction. Govil’s position is that existing EEG hardware already captures enough useful information. The bottleneck is decoding, not recording.

Atlas builds on the company’s proprietary AI platform, NeuroDiffusion, which filters noise artefacts from raw EEG data. The system uses advanced signal processing, new classes of deep learning models, and large-scale neural datasets to isolate true cortical intent from the surrounding neural noise. The result is stable decoding of motor signals from the scalp surface, without requiring the electrode-to-neuron proximity that invasive implants provide.

From Neuralis to Atlas

The company’s first product, Neuralis, is a wearable EEG headset designed for wheelchair users. It uses dry sensors and on-device AI processing to translate brain signals from the motor cortex into navigation commands. Neuralis is the subject of an upcoming clinical trial in partnership with Columbia University to assess real-world effectiveness for individuals with motor impairments.

Atlas appears to represent the next evolution of this technology. Where Neuralis focused on a specific wearable form factor, Atlas positions the underlying decoding platform as extensible: the same system that drives wheelchair navigation can be applied to cursor control, assistive communication, prosthetics, and what Synaptrix describes as other programmable brain-driven applications.

The distinction matters commercially. A single decoding platform that generalises across multiple output modalities is a fundamentally different proposition from a single-purpose medical device. It suggests Synaptrix is building toward a platform business, not just a product.

The Non-Invasive Race Intensifies

Atlas arrives at a moment when the non-invasive BCI space is rapidly gaining credibility. Recent research from UC Berkeley, the Max Planck Institute, and Graz University of Technology demonstrated that scalp-based BCIs can approach the decoding accuracy of implanted devices when combined with sophisticated machine learning. The performance gap that once justified the risk of brain surgery is narrowing with every generation of algorithms.

Synaptrix is not alone in this race. Companies like Neurable, Cognixion, and NextMind (acquired by Snap) have all pursued non-invasive approaches with varying degrees of commercial traction. But Synaptrix’s deep learning-first philosophy, combined with Cuban’s backing and the Columbia clinical trial, positions it as one of the more credible entrants in a field that has historically struggled to move beyond laboratory demos.

The company has also secured a licensing agreement with IDUN Technologies, an EEG hardware manufacturer, to implement NeuroDiffusion’s artefact removal AI directly in their sensors. That kind of third-party integration suggests the technology is maturing beyond internal prototypes.

What to Watch

The Atlas demo is compelling but early. Synaptrix acknowledges this is still developmental work, and the jump from controlled demo to reliable daily-use device is significant. The Columbia clinical trial for Neuralis will be an important validation point. If the underlying decoding technology proves robust with diverse users in uncontrolled environments, the path to broader applications becomes considerably more credible.

The company has FDA approval for its clinical trial, which it obtained faster than many invasive BCI studies, precisely because a non-invasive approach sidesteps the regulatory complexity of surgical implants. That speed advantage could prove just as important as the technology itself.

For the broader BCI industry, Atlas reinforces a trend that has been building for the past two years: the most commercially viable path to mass-market brain-computer interfaces may not run through the operating theatre. It may run through a headset you charge overnight.