Growth trajectory analysis
Investigating how individualized growth curves, rather than population-based percentiles, could improve early identification of growth restriction.
Investigating how individualized growth curves, rather than population-based percentiles, could improve early identification of growth restriction.
Researching machine learning approaches to interpreting complex fetal data, enabling detection of subtle patterns that indicate developing complications.
Exploring how combining multiple sensor modalities could provide complementary signals for fetal wellbeing assessment in non-clinical settings.
Studying how continuous monitoring data could integrate with existing clinical pathways, supporting clinical care.
Note. Our research is ongoing. We are currently in early-stage feasibility studies and are not making claims about clinical efficacy. We are committed to rigorous validation through proper clinical trials.
Current prenatal care relies on periodic snapshots; ultrasounds every few weeks, occasional non-stress tests. Yet conditions like fetal growth restriction (FGR) develop gradually, and warning signs often emerge between clinical visits.
Research shows that when FGR is detected early, outcomes improve dramatically; detection roughly halves stillbirth risk. But with limited sonographer availability, most women receive only 2–3 ultrasounds throughout pregnancy, leaving dangerous conditions unidentified.
Today, most mothers’ primary tool for monitoring their baby is “kick counting” – tracking fetal movements manually. But this method is time-intensive, subjective, and by the time movement patterns change noticeably, it’s often too late for intervention.
“There is urgent need to improve detection and monitoring of FGR to improve outcomes. Short-term recordings limit their utility.”Frontiers in Physiology, 2022
Stillbirth rate with and without antenatal fetal growth restriction (FGR) detection.
Methodological note. Improved outcomes in detected cases are largely due to earlier, well-timed delivery (median 10 days earlier than undetected cases). This is an observational association, not a causal claim about any single intervention. Presented for research context; Project Stardust does not make clinical efficacy claims about its own device.
50% of mothers worry about fetal movement during pregnancy and ~10% go to hospital at some point just for reassurance. Here’s what expectant parents told us they want:
Source: Berkeley Human-Centered Design parent interviews, 2025.
Our work builds on decades of research in maternal-fetal medicine, ultrasound imaging, and AI-assisted diagnostics.
We are enrolling people in their second or third trimester of pregnancy in a voluntary, IRB-reviewed feasibility study of an investigational wearable fetal monitoring device. Open to participants in the United States.