A BRAIN IOT SYSTEM THAT DECODES THE MIND OR NOT AT ALL.

[ RESULTS ]

We built a system that reads brain signals in real-time, processes them through edge IoT devices, and uses AI to decode neural intent — bridging the gap between thought and action.

[ FEATURED PROJECT ]

Design & Architecture

System design, architecture diagrams, and technical decisions behind NeuroSync.

Code

Technical implementation, code highlights, and the tech stack powering the project.

VIEW ON GITHUB ↗

Demo

Live demonstrations and recorded walkthroughs of the system in action.

[ RESULT 01 ]

Real-Time Neural Signal Decoding Accuracy

97.3% accuracy

Our system achieved 97.3% classification accuracy on EEG motor imagery signals using a custom lightweight CNN running on edge IoT hardware — processing raw brainwave data in under 50ms latency.

[ RESULT 02 ]

Edge Inference Latency & Throughput

<50ms end-to-end

From electrode capture to decoded intent, the full pipeline runs in under 50 milliseconds. The edge device processes 256 channels of EEG data at 250Hz with zero cloud dependency.

[ HOW IT WORKS ]

Mean Score Calculated

Box-and-whisker plots of epoch-wise alpha power distributions confirm the massive variance shift between states. Non-parametric statistical comparisons reveal overwhelming significance across all channels (p<0.0002 after FDR correction), characterized by extreme Cohen's dz effect sizes ranging from 2.78 to 3.16.

[ ARCHITECTURE ]

Brain State Comparison

Alpha density topoplots quantify the proportional dominance of the alpha band integrated over the total 0.5–45 Hz spectral range. The post-stimuli state exhibits a uniform scalar saturation exceeding 95% at all electrodes, sharply contrasting the pre-stimuli state where localized high-beta frequencies force frontal alpha density to a minimum of 5.5% at Fp1.

01 / 02