forked from vitaldb/pyvital
-
Notifications
You must be signed in to change notification settings - Fork 0
/
ecg_hrv.py
152 lines (127 loc) · 5.28 KB
/
ecg_hrv.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
import arr
import numpy as np
import math
cfg = {
'name': 'ECG - Heart Rate Variability',
'group': 'Medical algorithms',
'desc': 'Calculate Heart Rate Variability. Approximately 60-second data is required for calculating HF component and 120-second for LF. To calculate VLF, a longer signal is needed.',
'reference': 'Heart rate variability. Standards of measurement, physiological interpretation, and clinical use. European Heart Journal (1996)17,354-381',
'overlap': 2, # 2 sec overlap for HR=30
'interval': 300, # 5 min
'inputs': [{'name': 'ecg', 'type': 'wav'}],
'outputs': [
{'name': 'SDNN', 'type': 'num', 'unit': 'ms', 'min': 0, 'max': 100},
{'name': 'RMSSD', 'type': 'num', 'unit': 'ms', 'min': 0, 'max': 10},
{'name': 'pNN50', 'type': 'num', 'unit': '%', 'min': 0, 'max': 5},
{'name': 'NNI', 'type': 'num', 'unit': 'ms', 'min': 500, 'max': 2500},
{'name': 'TP', 'type': 'num', 'unit': 'ms2', 'min': 0, 'max': 200000},
{'name': 'VLF', 'type': 'num', 'unit': 'ms2', 'min': 0, 'max': 200000},
{'name': 'LF', 'type': 'num', 'unit': 'ms2', 'min': 0, 'max': 10000},
{'name': 'HF', 'type': 'num', 'unit': 'ms2', 'min': 0, 'max': 10000},
{'name': 'LF_HF', 'type': 'num', 'unit': '', 'min': 0, 'max': 100}
]
}
def run(inp, opt, cfg):
data = arr.interp_undefined(inp['ecg']['vals'])
srate = inp['ecg']['srate']
rlist = arr.detect_qrs(data, srate) # detect r-peaks
# remove qrs before and after overlap
new_rlist = []
for ridx in rlist:
if cfg['overlap'] <= ridx / srate:
new_rlist.append(ridx)
rlist = new_rlist
ret_rpeak = [{'dt': ridx / srate, 'val': 1} for ridx in rlist]
# average qrs
qrs_width = int(0.1 * srate)
qrslist = []
for ridx in rlist:
qrslist.append(data[ridx - qrs_width: ridx + qrs_width])
avg_qrs = np.mean(np.array(qrslist), axis=0)
# correlation coefficient
celist = []
for qrs in qrslist:
ce = arr.corr(qrs, avg_qrs)
celist.append(ce)
# rr interval (ms)
rri_list = np.diff(rlist) / srate * 1000
nni_list = [] # nn interval (ms)
ret_nni = []
for i in range(len(rlist) - 1):
if celist[i] < 0.9 or celist[i+1] < 0.9:
continue
# median RR interval nearest 10 beats
med_rri = np.median(rri_list[max(0, i-5): min(len(rri_list), i+5)])
rri = rri_list[i]
if med_rri * 0.5 <= rri <= med_rri * 1.5:
nni_list.append(rri)
ret_nni.append({'dt': rlist[i+1] / srate, 'val': rri})
# make time domain nni_data function by linear interpolation (200 hz)
nni_srate = 200
nni_data = [None] * int(math.ceil(len(data) / srate * nni_srate))
for nni in ret_nni:
nni_data[int(nni['dt'] * nni_srate)] = nni['val']
nni_data = arr.interp_undefined(nni_data)
# hamming window
nni_data *= np.hamming(len(nni_data))
vlf = 0 # <= 0.04 Hz
lf = 0 # 0.04-0.15 Hz
hf = 0 # 0.15-0.4 Hz
# A power spectral density (PSD) takes the amplitude of the FFT, multiplies it by its complex conjugate and normalizes it to the frequency bin width.
# This allows for accurate comparison of random vibration signals that have different signal lengths.
psd = abs(np.fft.fft(nni_data)) ** 2 / (len(nni_data) * nni_srate) # power density per bin (ms2/hz) from fft
psd *= 2 # In order to conserve the total power,
# multiply all frequencies that occur in both sets -- the positive and negative frequencies -- by a factor of 2.
# Zero frequency (DC) and the Nyquist frequency do not occur twice
for k in range(len(nni_data)):
f = k * nni_srate / len(nni_data)
if f < 0.0033:
pass
elif f < 0.04:
vlf += psd[k]
elif f < 0.15:
lf += psd[k]
elif f < 0.4:
hf += psd[k]
else:
break
# multiply the width (hz) to get the area under curve
vlf *= nni_srate / len(nni_data)
lf *= nni_srate / len(nni_data)
hf *= nni_srate / len(nni_data)
tp = vlf + lf + hf
# lf_arrnorm = 0
# hf_arrnorm = 0
# if tp - vlf:
# lf_arrnorm = lf / (tp-vlf)
# hf_arrnorm = hf / (tp-vlf)
lf_hf = 0
if hf:
lf_hf = lf / hf
sdnn = np.std(nni_list)
dnni_list = abs(np.diff(nni_list)) # Difference between adjacent nn intervals
nn50 = 0
ret_dnni = []
for i in range(len(dnni_list)):
dnni = dnni_list[i]
if dnni > 50:
nn50 += 1
ret_dnni.append({'dt': ret_nni[i+1]['dt'], 'val': dnni})
pnn50 = nn50 * 100 / len(dnni_list)
rmssdnni = 0
if len(dnni_list) > 0:
for dnni in dnni_list:
rmssdnni += dnni * dnni
rmssdnni = (rmssdnni / len(dnni_list)) ** 0.5
dt_last = cfg['interval']
return [
[{'dt': dt_last, 'val': sdnn}],
[{'dt': dt_last, 'val': rmssdnni}],
[{'dt': dt_last, 'val': pnn50}],
ret_nni,
[{'dt': dt_last, 'val': tp}],
[{'dt': dt_last, 'val': vlf}],
[{'dt': dt_last, 'val': lf}],
[{'dt': dt_last, 'val': hf}],
[{'dt': dt_last, 'val': lf_hf}]
]