.png)
Get an exclusive early-customer discount with code DECODED10. Peri is FSA & HSA eligible, with no subscriptions or monthly costs..
Get an exclusive early-customer discount with code DECODED10. Peri is FSA & HSA eligible, with no subscriptions or monthly costs..
Peri algorithms are built from our own real-world datasets of women in perimenopause. Because research has historically overlooked women, especially those in midlife, we had to build the dataset that didn’t exist. Our algorithms are trained on real physiological data collected directly from women experiencing perimenopausal symptoms, ensuring accuracy grounded in lived experience rather than assumptions.
Peri’s algorithms were developed using data from 120 women transitioning through perimenopause, who were experiencing symptoms including hot flashes and night sweats, sleep disruption, and anxiety. Each participant wore the Peri device continuously while self-reporting symptoms, completing validated anxiety questionnaires, and wearing polysomnography (the gold standard for sleep tracking) at night to provide clinical context for the sensor data.
200+ physiological features. One clear picture of your body.
Peri extracts unique features from motion, optical, electrodermal, and temperature sensors to build our proprietary digital biomarkers. These biomarkers capture measurable physiological patterns that allow us to detect perimenopause symptoms with high precision.
Think about the heart rate and HRV insights you get from today’s wearables. Not long ago, those seemed impossible. A single optical sensor produced thousands of noisy data points that looked unusable until someone discovered the hidden physiological features inside that signal and applied the right algorithms to decode it and give users a heart rate or HRV result in their app.
We’ve done exactly the same for perimenopause.
Using 200+ physiological features drawn from all sensors, our algorithms decode the invisible patterns behind hot flashes, night sweats, sleep disruption, anxiety, and cycle changes, giving you a clear, objective picture of what your body is experiencing, every day and every night.
Hot Flash Detection
F1 score: 74%
(F1 tells you how reliable the algorithm’s decisions are.)
The Peri hot flash algorithm is derived from rapid biosignal changes captured from multiple sensors and validated against real-time symptom logging. Many people assume hot flashes are detected just by tracking skin temperature, but Peri doesn’t work that way, it is tuned to complex biosignal patterns from multiple sensors, not a single temperature reading.
Night Sweat Detection
F1 score: 67%
Peri also identifies night-time vasomotor episodes by detecting rapid biosignal changes from multiple sensors while you sleep. This algorithm is specifically developed for night sweats and uses different features than the hot flash algorithm. In a similar manner as the hot flashes, this algorithm has been validated using real-time logging and is not based in skin temperature.
Both hot flash and night sweat algorithms are symptom specific and can be differentiated from other sweat related events such as exercise.
Sleep Monitoring
F1 score: 84% for sleep vs. wake
Peri classifies sleep and wake, tracks disturbances, and captures tossing and turning throughout the night. Our sleep algorithm was validated against home polysomnography (PSG), the gold standard in sleep monitoring, showing 96% agreement in detecting sleep and wake.
Cycle Tracking
Peri detects the biphasic temperature pattern (follicular → luteal) that defines the menstrual cycle, giving you passive, objective insight into your cycle without logging or tracking symptoms. Our cycle algorithm is validated against real-world menstrual cycle data.
Anxiety Tracking
Using only multisensory data, with no need for self-reporting, Peri predicts anxiety levels with very low error, closely matching established anxeity questionnaire scores (mean absolute error of 0.13). This reveals how symptoms and cycle phases may influence your emotional wellbeing.
.png)