6
Personal wearable monitor of the heart rate variability
OPENING P
APER
Therefore several energy savings-oriented measures were
taken into consideration at the design stage. Besides the use
of low-power circuitry, main contribution to the longevity of the
wearable recorder are: optimization of the ECG processing and
organization of the wireless transfer of tachogram series. Unfor
-
tunately, due to the required mobility of the subject implying the
omnidirectional radio wave propagation, the wireless connection
is the main contribution to the power consumption.
Materials and methods
Deriving the tachogram from the
ECGrecord
The heart rate reported or displayed for diag nostic purposes is
usually the average value of seven consecutive RR intervals.
Excluding two shorter and one longer intervals from the averag-
ing increases the HR report stability in the pre sence of missing
or extra detections. The instanta neous HR and its beat-to-beat
variations (HR vari ability, or HRV) have been studied for many
years, particularly in the framework of investigation of the ANS
balance. In periods of stable HR, there are small beat-to-beat
variations that result from the balance between the sympathetic
system and the parasympathetic system. For this reason any
pharmacologic agent that acts over ANS functions also inlu
-
ences the heart rate. Lower rhythms (si nus bradycardia, below
50 bpm), resulting from the prevalence of a parasympathetic
system, occur in states of deep relaxation or in athletes with en-
larged stroke volume. Higher rhythms (sinus ta chycardia, above
110 bpm) are caused by the dominance of a sympathetic system
and appear during physical effort or mental stress. The time-
domain parameters of the heart rate variability are thus crucial
indicators for the human surveillance in everyday life including
possible assessment of sleep disorders. Since the ANS inluences
mainly the sinoatrial node (SA) in the heart conductive system,
the HRV parameters aiming at assess ment of the ANS require
to be calculated from the NN (normal-to-normal) time intervals.
In case of occurrence of escape or complementary stimula tions,
these beats need to be excluded and re placed by a hypothetical
heart beat of sinoatrial origin.
The ECG record needs speciic processing to yield the
RR time series. This includes the tem poral localization of the
heart beats and basic clas siication of beat types. Single-lead
electrocardio gram is suficient to provide a reliable temporal
markers and usually also information of origin for each heart beat.
The use of three bipolar lead system (Holter type) is considered,
but not imple mented in the prototype for the reason of subject’s
comfort. The ECG is acquired with the sampling frequency of
500 sps and 12 bits resolution. The processing is performed
in real time with the use of mathematical signal transformation
revealing the common features of the QRS complex. Next, the
adaptive threshold is applied to discriminate the QRS section.
The localization of the R wave peak is further reined to 1 ms with
use of ive points-based parabola itting [5]. Beat origin labeling
needs further processing of the signal sections isolated in 100 ms
vicinity of the R wave peak. Certain features of the signal most
discriminative for atrial and ventricular beats [6] are calculated
and support the inal decision on the beat origin attribute. The
beats are next qualiied to the time-domain HRv analyzer. The
whole processing yields for each heartbeat a pair of numerical
de scriptors aggregated in a word:
–
the relative position coded in 12 bits,
–
the origin coded in remaining 4 bits.
–
For the reason of data integrity one beat type is reserved
for a synchronization time marker.
Since the ECG interpretation algorithm is em bedded and raw
ECG record is not available in normal operation mode, the per-
formance of the processing and pertinence of the sleep analysis
rely on the robustness of the beat detection and classiication
methods to interferences and noise. The algorithm applied as re
-
corder’s embedded software was developed as a part of real-time
Holter interpreter and complies with the essential performance
requirements (IEC60601-2-51). The wearable device is thus
considered reliable in standard recording conditions, therefore
Do'stlaringiz bilan baham: 
