This technology uses a smartphone app and microphone placed on a pregnant woman’s abdomen to record and analyze fetal movement sounds with AI, providing at-home, non-invasive monitoring and alerts for both patients and clinicians to improve prenatal care.
Background
Fetal movement monitoring is a critical aspect of prenatal care, as changes in fetal activity can be an early indicator of fetal compromise or distress, including the risk of stillbirth. Late stillbirth remains a significant global health issue, with millions of cases annually, particularly in low- and middle-income countries where access to specialized care is limited. Current clinical guidelines recommend surveillance methods such as ultrasound biophysical profiles and non-stress tests, especially for high-risk pregnancies.
However, these methods require specialized equipment, trained personnel, and in-person clinic visits, making them inaccessible for many populations. Furthermore, most stillbirths occur in pregnancies previously considered low risk, highlighting the need for more widespread, accessible, and continuous monitoring solutions.
Despite the recognized importance of fetal movement as a marker of fetal well-being, current approaches to monitoring are fraught with limitations. Maternal perception of fetal movement, though commonly used, is highly unreliable studies show that pregnant individuals detect only about a third of movements observed on ultrasound. Definitions of decreased fetal movement (DFM) are inconsistent, leading to confusion and variability in clinical response. Clinic-based monitoring is episodic and resource-intensive, lacking evidence from randomized controlled trials that it effectively reduces stillbirth rates.
These limitations underscore a critical gap: there is not widely available, objective, and user-friendly method for continuous fetal movement assessment outside of clinical settings, leaving many pregnancies inadequately monitored and at risk for adverse outcomes.
Technology description
This technology is a smartphone-based system for non-invasive, at-home monitoring of fetal movements using advanced machine learning analysis of audio recordings. By securing a smartphone microphone on the maternal abdomen, the system captures acoustic signals generated by fetal movements. These recordings are preprocessed for noise reduction and signal normalization, after which a suite of engineered features—spanning time, frequency, and time-frequency domains (such as amplitude envelope, spectral centroid, and Mel-spectrograms)—are extracted.
Deep learning models, including convolutional and recurrent neural networks, are trained on these features using synchronized ultrasound data as ground truth, enabling accurate detection, counting, and classification of fetal movements, including protective actions like breathing or hiccups. The results are displayed to the user via a mobile application interface and can be transmitted to clinicians, with processing performed either locally on the device or remotely in the cloud. The system also integrates patient metadata (e.g., BMI, placental location) to enhance accuracy and provide real-time feedback on recording quality.
What differentiates this technology is its unique combination of accessible consumer hardware, sophisticated acoustic feature engineering tailored to the transient, low-frequency nature of fetal sounds, and robust AI validation against gold-standard ultrasound data. Unlike traditional fetal monitoring, which relies on costly, clinic-based equipment and subjective maternal perception, this solution democratizes fetal health surveillance by enabling objective, continuous monitoring anywhere a smartphone is available. Its scalable architecture supports both edge and cloud processing, making it suitable for global deployment, including low-resource settings.
The system’s integration of clinical dashboards, automated alerts, and longitudinal tracking provides actionable insights for both patients and healthcare providers, potentially enabling earlier intervention and reducing stillbirth rates. Comprehensive clinical validation and a modular, portable software design further distinguish this technology as a practical, globally impactful advancement in prenatal care.
Markets
- Diagnostic instruments
- Diagnostic measurements
Benefits
- Enables non-invasive, at-home fetal movement monitoring using widely available smartphones, increasing accessibility and convenience for pregnant individuals.
- Provides objective, continuous, and accurate detection and classification of fetal movements through advanced machine learning and deep learning analysis of audio signals.
- Supports early identification of decreased fetal movements, a critical indicator of fetal distress, potentially reducing late stillbirth rates globally.
- Delivers real-time feedback and visualizations to patients via a user-friendly mobile app, enhancing maternal engagement and awareness.
- Facilitate clinical decision-making by transmitting quantitative fetal movement data and alerts to healthcare providers through an integrated clinical dashboard.
- Employs a scalable system architecture supporting both local (edge) and cloud processing, ensuring flexibility and broad deployment in diverse healthcare settings.
- Incorporates sophisticated noise cancellation and signal quality assessment to ensure reliable data capture even in non-clinical environments.
- Offers longitudinal monitoring capabilities to track fetal movement patterns over time, enabling detection of subtle changes indicative of fetal compromise.
Commercial applications
- At-home fetal movement monitoring
- Remote prenatal care support
- Clinical decision support for obstetricians
- Early stillbirth risk detection
- Low-resource prenatal screening
Opportunity
The University of Texas at Austin is seeking a commercial partner to license this patented technology. This system utilizes a smartphone microphone on the maternal abdomen to capture fetal movement acoustics. Audio signals undergo preprocessing, feature extraction, and analysis by deep neural networks, trained with synchronized ultrasound data. It objectively detects, counts, and assesses fetal movements, including protective ones. Results are output via a graphical interface and can be transmitted to clinicians, with processing occurring locally or remotely.
Patent
