NeuroWave: Biological Signal AI Without a Computer

NeuroWave turns the human body into the interface, translating measurable biological signals into AI-driven insight, no computer required.

Comment

Here’s a Markdown + LaTeX formatted draft for your Project Story – About the project that fits the Devpost style and frames NeuroWave as visionary yet scientifically grounded:

About the Project

Inspiration

The starting point for NeuroWave came from a simple but radical question:

What if AI didn’t need a keyboard, a screen, or even a computer — but could interface directly with the biological signals of the human body?

Modern computing devices imitate — but rarely match — the elegance of natural information processing. The human body already operates as a complex network of oscillators, resonators, and adaptive systems. Our aim was to reconnect technology to this original “hardware,” using measurable, reproducible science rather than speculation.

What We Learned

Through research and experimentation, we confirmed that:

  • The human brain and vestibular system operate in well-defined frequency bands (delta, theta, alpha, beta, gamma).
  • Water-based fluids in the brain and inner ear can be modeled as resonant oscillators.
  • Periodic external stimuli can entrain neural activity — a phenomenon documented in auditory steady-state response (ASSR) and photic driving studies.
  • The Kuramoto model of coupled oscillators provides a robust mathematical framework for simulating and analyzing phase synchronization in such systems.

We also learned how to integrate gpt-oss-120b as a transparent, reproducible reasoning engine that analyzes signal patterns without “black box” opacity.

How We Built It

  1. Biophysical Modeling
  • We represent neural and vestibular oscillators as phases $\theta_i$ with natural frequencies $\omega_i$.

  • Synchronization dynamics follow the Kuramoto equation:

    $$ \frac{d\theta_i}{dt} = \omega_i + \frac{K}{N} \sum_{j=1}^N \sin(\theta_j - \theta_i) $$

  • Coherence is quantified as:

    $$ R(t) = \left| \frac{1}{N} \sum_{j=1}^N e^{i\theta_j} \right| $$

    where $R \approx 1$ indicates high synchronization.

  1. Signal Acquisition
  • Non-invasive sensors capture biological resonance patterns and environmental reference signals.
  • Data is pre-processed into structured JSON for the AI.
  1. AI Analysis (gpt-oss)
  • The gpt-oss-120b model interprets evidence (max coherence, time windows, driving frequency).
  • Outputs are deterministic and reproducible for identical inputs.
  1. Visualization
  • While the focus is analytical, we provide a simple 3D interface for exploring coherence over time.

Challenges We Faced

  • Data integrity: Biological signals are noisy; filtering without distorting the resonance signature was critical.
  • Hardware constraints: Running large reasoning models locally required optimization and pruning.
  • Scientific clarity: We avoided unverifiable claims, framing the work in the language of neuroscience and physics to ensure credibility.
  • Integration: Bridging low-level physics simulations with high-level AI reasoning required careful interface design.

What Makes It Unique

NeuroWave is not just a model or a dataset — it is a framework for human–AI interaction that:

  • Works without a traditional computer interface.
  • Is rooted in measurable science and open-source reproducibility.
  • Has potential applications in accessibility, remote communication, creative arts, and beyond.

Built With

Share this project:

Updates