Maternal-Centered AI

An Agentic AI System for
Perinatal Women

Nyx Health, integrating intelligent wearables with AI Agents, enables continuous monitoring,
early detection, and timely intervention, closing the maternal healthcare loop from
passive biosensing to personalized behavioral interventions, making
proactive maternal mental health care possible at scale.

140+
Maternal Biomarkers
Monitored
1.5M+
Real Pregnancies in
Training Data
40%
Reduction in Maternal Morbidity
Explore Novel Solutions
01 / Passive Monitoring

Living Laboratory
within AI Rings.

Maternal-Centered Biosensors

Nyx Health integrates PPG, IMU, ECG, and EDA biosensors with ultra-low-power circuit design to enable continuous monitoring of 140+ maternal biomarkers, closing the maternal psychophysiology feedback loop. In parallel, Nyx Health has designed a novel dual-ring structural architecture. The inner ring incorporates elastic component with pressure and humidity biosensors, allowing self-adjustment to dynamic changes in finger circumference caused by edema and weight fluctuations throughout pregnancy. The outer ring utilizes ceramic materials to avoid allergic reactions and injury to infants during breastfeeding.

Low-Profile Design Passive Monitoring 80% Maternal Adherence
02 / Signal Processing

Data Stream within
Digital Twins.

Maternal Mental & Physical Health States

By fusing heterogeneous data streams from embedded biosensors and smartphones, Nyx Health constructs adaptive digital twins of each mother's psychophysiological state. Our multi-stage filtering pipeline mitigates noise from motion and artifacts, while cross-modal modeling accurately imputes missing data across PPG, ECG, and EDA streams to ensure a highly robust and personalized foundation for continuous maternal care and proactive mental and physical health monitoring.

Robust Modeling Digital Twins Personalized Representations
Digital Twin Visualization
03 / Multimodal Machine Learning

Maternal Pathology within
Multimodal Models.

Multimodal Risk Prediction Models

Nyx Health integrates risk prediction models with LLMs under a Mix-of-Experts (MoE) architecture for early detection and interpretation of APOs & PMADs. Our models are fine-tuned on 1.5 million real women's health datasets—purpose-built to address the pervasive, long-standing male-physiology bias in wearable AI through maternal-centered and gender-responsive algorithms. By incorporating female-specific physiological baseline variations, our AI avoids misinterpretation of normal changes like HRV and STS during pregnancy, ensuring high diagnostic accuracy for every mother across diverse backgrounds and physiological profiles.

Early Detection Mix-of-Experts LLMs Women-Based Dataset
Multimodal Risk Prediction
04 / Medical Reasoning

Health Coaches for
Maternal Healthcare.

Individualized Behavioral Interventions

Nyx Health collaborates with the Seleni Institute to model validated behavioral intervention pathways, collected from 1.5 million pregnancies over 15 years, translated into AI Agents. Integrated with Chain-of-Thought (CoT) reasoning and MCP, Nyx Health securely coordinates across smartphones, wearables, and home environments — delivering real-time, personalized lifestyle interventions that reduce maternal morbidity from PMADs by at least 40%.

Timely Intervention AI Agents CoT + MCP Seleni Institute
Medical Reasoning Flow
05 / Human-AI Interaction

Agentic AI for Maternal
Adherence & Effectiveness.

Bidirectional Human-AI Alignment

In partnership with the Seleni Institute's network of 7,500 trained health coaches, Nyx Health builds a bidirectional Human-AI alignment framework connecting pregnant women, clinicians, coaches, and family members. AI is positioned not as a replacement for human care, but as an amplifier — overcoming temporal and spatial constraints to deliver continuous, personalized support and reduce maternal morbidity from PMADs.

Individual Effectiveness Bidirectional Human-AI Alignment Family-In-The-Loop
Human-AI Alignment Framework

Modeling the Psychophysiology
Feedback Loop
between APOs and PMADs.

A closed-loop maternal AI system — from biosensor wearables to clinical-grade intervention — designed to monitor, model, detect, and intervene at every stage of the perinatal journey.

📡
Continuous Monitoring
Biosensors capture continuous psychophysiological signals via intelligent rings and smartphones
🧬
Robust Modeling
Digital health twin fuses multimodal data streams into an adaptive living health profile
🔍
Early Detection
Multi-LLM MoE models identify early risk signals associated with PMADs and APOs
🤖
Timely Intervention
The embedded AI Agents deliver personalized behavioral interventions in real-time

Nyx Health Edge & Moat.

Four structural advantages that compound into an insurmountable lead in maternal AI.

🔬
Non-Invasive & Passive Biosensing
Sensor Stack + Maternal HCI
  • EDA tracks emotional and stress responses, capturing maternal psychological states
  • UWB enables contactless monitoring of HRV, movement, and proximity during sleep
  • Adaptive ring sizing adjusts for late-pregnancy swelling and weight fluctuations
🧠
Psychophysiology Feedback Loop
Pathology LLM
  • Captures interactions between physiological and psychological signals of APOs and PMADs
  • Fine-tuned on women-centered AI to address male-biased training datasets
  • Closed-loop system: Monitoring → Detection → Diagnosis → Intervention
🏥
Exclusive Clinical Data
Duke Clinical Research Institute
  • 1M+ female wearable records across physiological and psychological states
  • 36,000 adverse pregnancy outcomes across White, Black, Asian, and Latin populations
  • 80% of clinical studies conducted at Prince of Wales Hospital, 20% at Duke
🌍
Cross-Country Division of Labor
Clinical Flywheel
  • Xiaomi-grade wearable supply chain — only chain integrating UWB and EDA into a ring
  • Top-tier engineers prototype, validate, and deploy at rapid speed
  • Achieves half the FDA trial cost while delivering twice the end-to-end process speed

International Academia-Industry Collaborations.

Ready to change maternal health?

We're building the future of perinatal care. If you're a clinician, researcher, or investor — let's talk.

Get in Touch