An online tool for showcasing the possibilities of Dynamical Detrended Fluctuation Analysis for studying heart rate variability, developed by the Quantum Control and Dynamics group at Tampere University. The method is particularly well-suited for highly non-stationary conditions, such as during physical exercise, or detecting transient changes in the heart rate variability.
Try it Read moreExplore some of our example data to get started. You may also upload your own data of RR intervals or synchronize from other services to collect data from wearable devices. Please review our privacy policy regarding your data.
Review the time series given as input.
Explore the dynamic RRI correlations as functions of time and scale, together with overlaid time series data. Raw data is available below for custom analysis.
The dynamic RRI correlations aggregated as a function of different variables, such as the heart rate.
DDFA results visualized on the travelled route.
The beating of the human heart is not regular, but instead follows fractal-like patterns. In fact, complex variations in the heart rate are characteristic of a healthy heart, and deviations in this heart rate variability (HRV) are induced for example by cardiac diseases or physical exercise. HRV is studied by analyzing the time between successive heart beats, also known as RR intervals. The complex behavior of the RR intervals (RRIs) may be characterized by a method called detrended fluctuation analysis (DFA), which measures statistical self-affinity by scaling exponents.
Conventionally the scaling exponents are determined for short- (4-16 beats) and long scale (16-64 beats) correlations. However, the extents of these ranges are arbitrary and in practice often the correlations are not satisfactorily described by just two exponents. The RRI correlations may also change in time, particularly as the heart rate responds to different external conditions, such as physical exercise or sleep. Therefore, we introduce dynamical DFA (DDFA), which determines the scaling exponents as the function of both scale and time.
Read more from M. Molkkari et. al., Dynamical heart beat correlations during running, Scientific Reports 10, 13627 (2020).
Esa is a professor of physics, who has published about 130 journal articles, including recent innovations in time-series analysis of physiological signals. He is the leader of the Quantum Control and Dynamics group at Tampere University.
Matti is a researcher in the doctoral program of engineering and natural sciences at Tampere University. He has background in physics with expertise in computational methods, mathematical modelling and data analysis. Matti is the lead developer of the DDFA methods.
Janne is an experienced software engineer and data analyst with a PhD in computational physics. He has developed and published several numerical methods and software. Janne performed the initial integration of our methods with Amazon Web Services.
Send an email to
Send an email to
