Skip to main content

Table 2 The list of features extracted from HRV

From: Sleep stage classification using extreme learning machine and particle swarm optimization for healthcare big data

No. Name of feature Method of extraction Explanation
1 AVNN Time domain AVNN is mean of all normal sinus to normal sinus of NN intervals.
\(AVNN = \frac{1}{N}\mathop \sum \limits_{j = 1}^{N} RR_{j}\) where j is interval index and N is number of total intervals
2 SDNN Time domain SDNN is the standard deviation of NN intervals
\(SDNN\;\text{ = }\;\sqrt {\frac{1}{{N\;\text{ - }\;1}}\;\sum\limits_{{j\;\text{ = }\;1}}^{N} {\left( {RR_{j} \text{ - }\;\overline{{\text{RR}}} } \right)^{2} } }\)
3 RMSSD Time domain Root Mean Square of the Successive Differences (RMSSD)
\(RMSSD = \sqrt {\frac{1}{N - 1}\mathop \sum \limits_{j = 1}^{N - 1} \left( {RR_{j + 1} - RR_{j} } \right)^{2} }\)
4 SDSD Time domain The standard deviation of the successive difference between adjacent R-R intervals
5 NN50 Time domain The number of pairs of adjacent NN intervals differing by more than 50 ms
\(Number of \left( {RR_{j + 1} - RR_{j} } \right) > 50\)
6 pNN50 Time domain Percentage of adjacent NN intervals differing by more than 50 ms
\(pNN50 = \frac{NN50}{N - 1} \times 100\)
7 HRV Triangular Index Geometrical (The total number of RR intervals)/(the height of histogram of all RR intervals with 7.8125 ms bin)
8 SD1 Poincare \(SD1^{2} = \frac{1}{2}SDSD^{2}\) where SD1 = the standard deviation of perpendicular points to the line-of-identity
9 SD2 Poincare \(SD2^{2} = 2SDNN^{2} - \frac{1}{2}SDSD^{2}\) where SD2 = the standard deviation of parallel points along the line-of-identity
10 SD1SD2 Ratio Poincare \(SD1 SD2 Ratio = \frac{SD1}{SD2}\)
11 S Poincare \(S = \pi \times {\text{SD}}1 \times {\text{SD}}2\)
12 TP Frequency domain The total power (TP) of the HRV
13 VLF Frequency domain Very low frequency (VLF) is defined as the total power (TP) ranging from 0‒0.04 Hz
14 LF Frequency domain Low frequency is defined as the total power (TP) ranging from 0.04‒0.15 Hz
15 HF Frequency domain High frequency is defined as the total power (TP) ranging from 0.15‒0.4 Hz
16 LFHF Ratio Frequency domain The ratio of LF to HF indicates the sympathovagal balance.
17 LFnorm Frequency domain \(LFnorm = \frac{LF}{TP - VLF}\)
18 HFnorm Frequency domain \(HFnorm = \frac{HF}{TP - VLF}\)