Fig 1.
Schematic diagram of the difference between the traditional method and the proposed method with a 1 min time increment.
(a) Traditional method of monitoring temporal dynamics of HRV: HRV analysis was performed in each neighbor analysis window, which means the time interval between two analyzed results is 5min, i.e. we can only observe the status changes taking place in the second window at the 10min time point. (b) The proposed overlapping window method of monitoring temporal dynamics of HRV, HRV analysis is performed in each overlapped analysis window.
Fig 2.
Time course of RMSSD calculated from signal segments created by sliding window with six time increments.
With the time increment increasing, the time courses of the measure become smoother, resulting in losing the dynamical information. When increasing the time increment to 5min, the measure could not provide any information about the temporal dynamics in the representative synthetic signal.
Fig 3.
Time course of SPLF components derived from signal segments created by sliding window with six time increments.
With the time increment increasing, the time courses of the measure become smoother, resulting in losing the dynamical information. When increasing the time increment to 5 min, the measure could not provide any useful information about the temporal dynamics in the representative synthetic signal.
Fig 4.
Time course of SPHF components derived from signal segments created by sliding window with six time increments.
With the time increment increasing, the time courses of the measure become smoother, resulting in losing the dynamical information. When increasing the time increment to 5 min, the measure could not provide any useful information about the temporal dynamics in the representative synthetic signal.
Fig 5.
Time courses of HRV measures on human RR-interval series with1-min increment (red solid line) and 5-min increment (green dotted line).
The five rows from top to bottom are the time course of RR interval series and four measures, SDNN, RMSSD, SPLF and SPHF, respectively. It can be seen that when the RR-interval data changed from meditation state to non-meditation state, all the four time courses corresponding to 1-min increment began to be obviously changed after one minute, whereas those corresponding to a 5-min increment were constant until the 5 minutes later.
Fig 6.
Time course of the SPLF on human RR-interval series with1-min increment (red solid line) and 5-min increment (green dotted line).
M and N labelled on the signal represent Meditation and Non-meditation state, respectively. It can be seen that 1-min-increment-based HRV analysis measure gives an earlier responses to the changes of meditation states than 5-min-increment-based measure.
Fig 7.
Time course of the SPLF on human RR-interval series with1-min increment (red solid line) and 5-min increment (green dotted line).
M and N labelled on the signal represent Meditation and Non-meditation state, respectively. It can be seen that a 5-min-increment-based measure was hardly to capture the happenings of 2-min non-meditation state among the meditation states, 1-min-increment-based measure could clearly catch the short-time changes of meditation state.
Fig 8.
Time courses of measures of HRV analysis on ECG data from a dog with1-min increment (red solid line) and 5-min increment (green dotted line).
Four rows from top to bottom correspond to the SDNN (standard deviation of the normal-to-normal intervals), RMSSD (square root of the mean squared differences of successive intervals), SPLF (power spectrum in low frequency), and SPHF (power spectrum in high frequency) components. Following the application of atropine, all the four measures of HRV analysis with 1-min increment instantly decreased and reached their minimum values within about seven minutes, and then gradually increased. With a 5-min increment, the four measures began to decrease about three minutes later after giving atropine.