Login / Signup

Linear and Non-Linear Heart Rate Variability Indexes from Heart-Induced Mechanical Signals Recorded with a Skin-Interfaced IMU.

Milena ČukićChiara RomanoFrancesca De TommasiMassimiliano CarassitiDomenico FormicaEmiliano SchenaCarlo Massaroni
Published in: Sensors (Basel, Switzerland) (2023)
Heart rate variability (HRV) indexes are becoming useful in various applications, from better diagnosis and prevention of diseases to predicting stress levels. Typically, HRV indexes are retrieved from the heart's electrical activity collected with an electrocardiographic signal (ECG). Heart-induced mechanical signals recorded from the body's surface can be utilized to record the mechanical activity of the heart and, in turn, extract HRV indexes from interbeat intervals (IBIs). Among others, accelerometers and gyroscopes can be used to register IBIs from precordial accelerations and chest wall angular velocities. However, unlike electrical signals, the morphology of mechanical ones is strongly affected by body posture. In this paper, we investigated the feasibility of estimating the most common linear and non-linear HRV indexes from accelerometer and gyroscope data collected with a wearable skin-interfaced Inertial Measurement Unit (IMU) positioned at the xiphoid level. Data were collected from 21 healthy volunteers assuming two common postures (i.e., seated and lying). Results show that using the gyroscope signal in the lying posture allows accurate results in estimating IBIs, thus allowing extracting of linear and non-linear HRV parameters that are not statistically significantly different from those extracted from reference ECG.
Keyphrases
  • heart rate variability
  • heart rate
  • heart failure
  • atrial fibrillation
  • blood pressure
  • high glucose
  • diabetic rats
  • electronic health record
  • high resolution
  • soft tissue
  • mass spectrometry
  • deep learning