Microvascular Research 89 (2013) 34–39

Contents lists available at ScienceDirect

Microvascular Research
journal homepage: www.elsevier.com/locate/ymvre

Complexity analysis of the microcirculatory-blood-flow response following

acupuncture stimulation
Hsin Hsiu a,⁎, Wei-Chen Hsu b, Chia-Liang Hsu c, Jian-Guo Bau d, Chao-Tsung Chen b, Yi-Syuan Liu e
a
Graduate Institute of Biomedical Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan
b
Department of Traditional Chinese Medicine, Taipei City Hospital RenAi Branch, Taipei, Taiwan
c
Graduate Institute of Applied Science and Technology, National Taiwan University of Science and Technology, Taipei, Taiwan
d
Department of Biomedical Engineering, Hungkuang University, Taichung, Taiwan
e
Department of Rehabilitation, Taipei City Hospital RenAi Branch, Taipei, Taiwan

a r t i c l e i n f o a b s t r a c t

Article history: Beat-to-beat cardiovascular variability analysis provides important information on the circulatory regulatory ac-
Accepted 13 June 2013 tivities. Changes in the arterial pulse transmission or the opening condition of arteriolar openings might change
Available online 24 June 2013 the fluctuation pattern of the MBF supply, and thus change the complexity property therein. We performed com-
plexity analysis of beat-to-beat laser Doppler flowmetry (LDF) signals to study the microcirculatory-blood-flow
(MBF) response at the needled site (Hegu acupoint) following acupuncture stimulation (AS). LDF signals were
measured in male healthy volunteers (n = 29). Each experiment involved recording a 20-minute baseline-
data sequence and two sets of effects data recorded 0–20 and 50–70 min after stopping AS. Approximate-
entropy (ApEn) analysis, which quantifies the unpredictability of fluctuations in a time series, was performed
on each 20-minute beat-to-beat LDF data sequence. The present findings indicate that AS can not only improve
the local blood supply but may also increase ApEn values and decrease MBF variability parameters. This was the
first attempt to apply complexity analysis to LDF signals in order to elucidate microcirculatory responses follow-
ing AS. The observed results are probably attributable to the contradictory effects on the MBF supply induced by
AS, which might interfere with the microcirculatory regulatory activities so as to increase the complexity of LDF
signals. The present findings could help to identify the mechanism underlying the effects of AS, might aid the de-
velopment of an index for monitoring the induced microcirculatory regulatory responses, and thus provide an
evidence-based connection between AS and modern physiology.
© 2013 Elsevier Inc. All rights reserved.

Introduction
The cardiovascular system comprises complex biophysical pro-
cesses with multiple levels of regulation (Humeau et al., 2009).
Beat-to-beat cardiovascular variability analysis can provide important
information on the circulatory autonomic control. For example, phys-
iological rhythms embedded in a beat-to-beat heart-rate (HR) data
sequence have been used to monitor autonomic regulation in cardio-
vascular control (Task Force, 1996). At the microcirculatory level, it
has been suggested that the microcirculatory blood flow (MBF) can
be driven into the capillaries through precapillary arteriolar openings
(AO) by the propelling force of the pulse pressure generated by the
heartbeat and transmitted along the artery (Chen et al., 2011; Hsiu
et al., 2008a, 2009a). The variability of the blood flow in arterioles
has been attributed to their ability to regulate their diameter in re-
sponse to external and internal changes, thereby protecting the
⁎ Corresponding author at: No. 43, Section 4, Keelung Road, Graduate Institute of
Biomedical Engineering, National Taiwan University of Science and Technology, Taipei
10607, Taiwan. Fax: +886 2 2730 3733.
E-mail address: hhsiu@mail.ntust.edu.tw (H. Hsiu).
brain against ischemia, capillary damage, and edema (Heistad and
Kontos, 1983).
Acupuncture stimulation (AS), an important treatment strategy in
oriental medicine, has been suggested to prevent or treat illness by
adjusting autonomic functions (Jansen et al., 1989). It has also been
found that AS can induce changes in microcirculatory perfusion of
the skin surface around the needled site (Hsiu et al., 2011a, 2011b)
or the internal organs (Tsuru and Kawakita, 2009). Laser Doppler
flowmetry (LDF) is a widely-used technique to monitor the microcir-
culation due to its advantages of a good frequency response, ease of
application, and tissue specificity, and is therefore well suited for
noninvasive investigations of the microvascular responses to AS. We
have previously noted that AS can increase the MBF and decrease
the MBF resistance at the needled site (Chen et al., 2011; Hsiu et al.,
2011a, 2011b). Differences in the induced sympathetic neural activi-
ties (SNA) between needling the Hegu acupoint and its nearby
nonacupoints have also be studied by analyzing the frequency con-
tent of skin LDF signals (Hsiu et al., 2011a).
Nonlinear analysis has been suggested to provide information
about the variability structure of skin blood-flow oscillations (Liao
et al., 2010). Several nonlinear analysis strategies, such as fractal

0026-2862/$ – see front matter © 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.mvr.2013.06.009
 H. Hsiu et al. / Microvascular Research 89 (2013) 34–39 35

analysis (Carolan-Rees et al., 2002; Esen and Esen, 2006) and

detrended fluctuation analysis (Esen et al., 2009), have been applied
to MBF signals with the aim of advancing the understanding of various
microvascular physiological and pathological conditions. Among vari-
ous types of nonlinear analysis, the signal complexity can be evaluated
by analyses of approximate entropy (ApEn) (Pincus, 1991). When ap-
plied to physiological signals, ‘complexity’ implies the absence of clear
temporal patterns. For example, calculations of entropy values can pro-
vide information about the mechanism underlying dynamic cerebral
microvascular autoregulation (Panerai, 2009). It has also been sug-
gested that changes in microcirculatory complexity with aging and dis-
ease depend upon the type of change required to maintain function
during the demands of a particular task (Vaillancourt and Newell,
2002).
The aim of the present study was to apply complexity analysis to
LDF signals in order to explain the MBF response at the needled site
(Hegu acupoint) following AS. One important characteristic of the
present analysis was that ApEn was calculated on a beat-to-beat
basis rather than the calculation being applied to the raw data.
Changes in many factors that affect the pulse transmission or the
AO condition, such as those in the heartbeat, arterial elastic proper-
ties, or the opening property of AO, may change the beat-to-beat
MBF supply, might change the fluctuation pattern of the MBF supply,
and thus change the complexity property therein. For example, many
of the capillary pathways in local vascular beds are closed at any
given instant, while others nearby continue to carry an active flow.
The pattern of AO opening and closing over time appears random
but could be dictated by induced changes in local regulatory activities
Fig. 1. Representative waveforms of ECG and LDF signals at Site 1. LDF flux and ECG are
(Lee and Fung, 1971), which can play an important role in in arbitrary units. The ECG signal was filtered by a low-pass filter (with a 50 Hz cut-off
maintaining a sufficient blood supply and thus the normal physiolog- to filter out the high-frequency noise) and a high-pass filter (with a 0.3 Hz cut-off to
ical condition of local tissues. Therefore, the present method of com- eliminate the baseline drift) to make the location of the R peak more prominent.
plexity analysis of the LDF signal may be useful for monitoring the
regulatory activities of local vascular beds, and thus help in under- (PCI-9111DG, Adlink Technology, Taiwan) operating at a sampling
standing the microcirculatory response following AS. rate of 1024 Hz (Hsiu et al., 2011a, 2011b).
The measuring sites of the LDF probes were at the left Hegu (LI4;
Materials and methods Site 1; an acupoint located between the thumb and the index finger
on the back of the hand, between the 1st and 2nd metacarpal
Experimental setup and data acquisition bones, in the middle of the 2nd metacarpal bone on the radial side)
and a nearby nonacupoint (Site 2; halfway between Hegu and Sanjian
Experiments were performed on male healthy volunteers aged [LI3]) for comparison. The LDF probes were held vertically onto the
20–27 years (all were students of the Department of Electrical Engi- skin surface by a holder with a radius of around 6 mm. The AS was
neering, Yuan Ze University) and without signs or symptoms of car- performed by a qualified Chinese-Medicine doctor working in the De-
diovascular or neurological disease (n = 29). The reason we used partment of Traditional Chinese Medicine, Taipei City Hospital RenAi
only male subjects and also subjects with ages within this range is Branch, who has a clinical experience of about 12 years. During the
to minimize the possible interference effects induced by menstrual measurement, the subject was supine on a measurement couch.
period and different ages. The subjects were all Taiwan natives, Each assessment involved making the following recordings: the
were lightly clothed, supine, and were allowed to stabilize for at 20-minute baseline (M0) data was recorded prior to AS, after which
least 20 min before commencing recording. The environmental tem- the acupuncture needle (stainless steel needle, gauge #30, 7.5 cm in
perature was within 23–25 °C during the entire measuring period. length; Chianhuei, Taiwan) was inserted into the left Hegu acupoint
The institutional ethics committee at Taipei City Hospital approved to a depth of around 4 mm for 20 min without applying any form of
the study protocol, and each volunteer provided written informed stimulation. The De-Qi feeling following AS (a special feeling, includ-
consent before entering the study. Tea, coffee, alcohol, and smoking ing distension of soreness or numbness, as described mutually by the
were forbidden on the day before experiments. All subjects did not subjects and the acupuncturist) was experienced in all the subjects.
exercise or consume food for at least 1 h before each experiment The needle was then withdrawn to stop AS, with the data being
(Hsiu et al., 2009b, 2011a, 2011b). recorded from 0 to 20 min (M1) and from 50 to 70 min (M2) after
The ECG and LDF signals were measured simultaneously and non- stopping AS. In several of our previous works, 20-minute AS has
invasively (representative waveforms are shown in Fig. 1). ECG sig- been noted to be able to induce prominent MBF responses, either in
nals were measured by surface electrodes, and acquired by a healthy (Hsiu et al., 2011a) or stroke subjects (Hsiu et al., 2011b).
preamplifier (lead II, RA-LL; 6600-series, Gould, USA). LDF (VP1 Using AS with a shorter time might not be sufficient to induce prom-
probe; MBF3, Moor Instruments, UK) was used to measure the micro- inent MBF response, whereas it might be difficult to remain the phys-
circulatory flux with a time constant of 0.001 s, a cut-off frequency of iological stability of the subjects during the whole experimental
14.9 kHz and a sampling frequency of 40 Hz. The laser operating periods when using AS with a longer time.
wavelength and output power were 780 nm and less than 1.6 mW, Before the baseline and after the M2 period, fundamental physiolog-
respectively. Subjects were asked to relax and breathe naturally ical parameters of the subject, including the heart rate (HR), systolic
throughout the measurement period so as to avoid motion artifacts. blood pressure (SBP), and diastolic blood pressure (DBP) were mea-
The signals were connected to an analog-to-digital converter card sured by a sphygmomanometer (MediGuard 150i, Rossmax, Taiwan).
36 H. Hsiu et al. / Microvascular Research 89 (2013) 34–39
A thermistor was attached to the skin surface to monitor the skin surface
temperature near the LDF measurement site (see Fig. 2). The resistance
of the thermistor was transformed into voltage using a custom-made cir-
cuit, which was also sampled by the analog-to-digital converter card. The
temperature stability during the baseline period was considered accept-
able when the variation was less than 1.0 °C.
Signal analysis
Data files containing spikes that were too large relative to the
mean flux value were discarded, since this implied the presence of
motion artifacts. Several parameters of the LDF signal were calculated
to elucidate the microcirculatory effects induced by AS. The MMBF
(mean component of the flux) was defined as the mean value of the
laser Doppler flux signals during each 20-minute measurement peri-
od. In determining the parameter of the pulsatile LDF waveform, the
LDF flux signal was filtered by an 11th-order Chebyshev digital
high-pass filter with a cut-off frequency of 0.01 Hz to eliminate the
baseline drift. The two neighboring R peaks of the ECG signal were
used to help in identifying the cut points of each pulse for LDF signals
to define each “pulse” of the microcirculatory flux. The cut point was
defined as the minimum LDF flux signal within a 300-ms search win-
dow located after the corresponding ECG R-peak. For each pulse, the
mean component was defined as the average value, the pulse compo-
nent was defined as the difference between the maximum and the
minimum of the LDF flux signal, and PMBF was defined as the mean
value of LDF pulse component during each 20-minute measurement
period. These parameters were used to calculate the MBF variability
(MBFV) index [coefficient of variation (CV; mean/SD) of the mean
MBF (MMBFCV; defined as CV of the mean component of all the
pulses in the 20-minute period)] and the CV of the pulsatile MBF
(PMBFCV; defined as the CV of the pulse component of all the pulses
in the 20-minute period) (Hsiu et al., 2010a, 2010b).
ApEn analysis, which quantifies the unpredictability of fluctua-
tions in a time series, was performed on each 20-minute sequence
of the mean (ApEn_MMBF) and the pulse components (ApEn_PMBF)
of the beat-to-beat LDF data sequence. For a sequence u (n), the ApEn
(we used pattern length m = 2; criterion of similarity r = 0.2 × SD)
can be calculated as follows:
m mþ1
ApEn ¼ Φ ðr Þ−Φ ðr Þ
Fig. 2. Locations of measurement sites. Measurement and needling sites: ‘□’, Hegu
acupoint (Site 1); ‘×’, a nearby nonacupoint (Site 2); ‘○’, location of the thermistor.
where
m −1
X
N−mþ1
m
Φ ðr Þ≡ðN−m þ 1Þ lnC i ðr Þ
i¼1
Cm
i ðr Þ ¼ number of xðjÞ such that d½xðiÞ; xðjÞ ≤ r=ðN−m þ 1Þ
xðiÞ ¼ ½uðiÞ; …; uði þ m−1Þ; xðjÞ ¼ ½uðjÞ; …; uðj þ m−1Þ
where N = number of pulses in the beat-to-beat LDF data sequence;
d represents difference between elements (Panerai, 2009; Pincus,
1991).
Oscillations of the microvascular blood perfusion signal can be
separated into different components by spectral analysis (Bernardi
et al., 1989; Hsiu et al., 2011a, 2011b; Kvandal et al., 2006; Kvernmo
et al., 1999). Calculating the occupied proportions of spectral power
for each frequency bands may help to monitor the activities of differ-
ent physiological rhythms, and hence provides a promising tool to
study the underlying microcirculatory regulatory mechanism (Hsiu
et al., 2008b, 2011a, 2011b). Wavelet transform with Morlet mother
wavelet was applied to the measured LDF signals to improve the
low-frequency resolution. Periodic oscillations with five characteristic
frequency peaks were observed from 0.0095 to 1.6 Hz, with the posi-
tions of these peaks falling within the following frequency bands:
0.0095–0.02, 0.02–0.06, 0.06–0.15, 0.15–0.4, and 0.4–1.6 Hz (defined
as FR1–FR5, respectively); which are suggested to be influenced by
the endothelial activity of the vessel wall, the neurogenic activity of
the vessel wall, the intrinsic myogenic activity of vascular smooth
muscle, the respiration, and the heartbeat, respectively (Bernardi
et al., 1989; Kvandal et al., 2006). The energy density within each fre-
quency band was calculated, and the relative energy contribution
(REC) in each frequency band (from FR1 to FR5) was defined as the
ratio between the total energy density within that band and the
total energy density of the entire spectrum from 0.0095 to 1.6 Hz
(Hsiu et al., 2011b; Kvernmo et al., 1999).
All signal processing was performed with MATLAB (MathWorks,
Natick, MA, USA). Two-tailed paired-t test or simple regression anal-
ysis was used to verify the statistical significance. Differences were
considered significant when p b 0.05.
Results
The HR, SBP, DBP, and the baseline temperature were 76.1 ±
10.9 beats/min (mean ± SD), 109.0 ± 6.1 mm Hg, 71.2 ± 4.2 mm Hg,
and 31.5 ± 1.2 °C, respectively, before the LDF measurement, and
69.7 ± 10.4 beats/min, 111.8 ± 8.6 mm Hg, 72.4 ± 6.4 mm Hg, and
31.2 ± 1.1 °C, after the LDF measurement. Only the HR changed signifi-
cantly (a decrease) between before and after the LDF measurement
(p b 0.05 by two-tailed paired t-test).
Fig. 3 reveals that the MMBF and PMBF increased significantly com-
pared to the baseline values during M1. Fig. 4 shows that the ApEn
values of MMBF and PMBF increased significantly following AS at Site
1 (the AS site). MBFV parameters at Site 1, as illustrated in Fig. 5,
MMBFCV and PMBFCV decreased significantly during M1, whereas
there was only a borderline-significant decrease in PMBFCV during
M2. The results of ApEn and MBFV parameters at a nearby non-AS site
(Site 2) are also shown in Figs. 4 and 5. There were not any significant
changes in these parameters between baseline (M0) and effect periods
(M1 or M2).
Fig. 6 shows the results of linear regression analysis performed on
the correlation between changes in MMBFCV and ApEn-MMBF during
M1 at Site 1. The regression lines had coefficients of determination
(R2) larger than 0.5 (p b 0.05 by F-test), indicating that there was
an inverse correlation between them.
Changes in the spectra of blood-flow signals following AS at Site 1
are shown in Fig. 7. The REC of FR2 decreased significantly and that of
 H. Hsiu et al. / Microvascular Research 89 (2013) 34–39 37

Fig. 3. Changes in MMBF and PMBF following AS. Data are mean and SD values. ‘*’ in-
dicates p b 0.05 compared with the baseline value by two-tailed paired t-test. Com-
pared to the baseline values, the MMBF and PMBF significantly increased during M1,
and returned to their baseline values during M2.

FR5 increased significantly during M1. The powers in these two bands
returned to their baseline values during M2.

Discussion

The present study found that MMBF and PMBF increased at the
needled site, which was consistent with the findings of several previ-
ous studies (Hsiu et al., 2009a; Jansen et al., 1989). The improved Fig. 5. Comparison of MBFV parameters following AS. ‘*’ and ‘+’ indicate p b 0.05 and
blood supply could be relevant with the adjustment of physiological 0.05 b p b 0.1, respectively, compared with the baseline value by two-tailed paired
t-test. Following AS at Site 1, MMBFCV and PMBFCV significantly decreased during
function at local tissues by AS.
M1; moreover, there was a borderline-significant decrease in PMBFCV during M2.
The present findings illustrate that AS can not only improve MMBF
and PMBF responses but may also change ApEn values. It was revealed
that the ApEn parameters of LDF signals significantly increased, which
0.1 illustrated the increased complexity of these signals following AS. The
Site 1 * * underlying mechanism may be the microcirculatory response induced
0.08 by inserting the needle into the local tissue. It has been shown previous-
ly that AS can induce local vasodilation, perhaps through decreasing
SNA (Knardahl et al., 1998; Lossius et al., 1993; Thomas et al., 1992), im-
0.06 * *
proving endothelial dysfunction (Park et al., 2010), or the secretion of
M0 various kinds of vasoactive substances (Jansen et al., 1989; Sato et al.,
0.04 2000). However, the inserted needle could also apply a contact force
M1
to local tissue, which could slightly compress the vessel walls of nearby
M2 vascular beds and impede the blood supply therein. These two contra-
0.02
dictory effects on the MBF supply might interfere with the microcircu-
latory regulatory activities so as to increase the complexity of LDF
0 signals, and thereby lead to increases in ApEn parameters. Moreover, a
ApEn_MMBF ApEn_PMBF pressure-induced-vasodilation response might also be induced by the
needle pressing on local tissues, and thus further increases the MBF
complexity.
Spectral analysis of LDF signals is a noninvasive method for studying
the mechanism of the microcirculatory regulation underlying the
responses to various types of stimulation, including acupuncture
(Bernardi et al., 1989; Hsiu et al., 2008b, 2009b, 2011a; Kvandal et al.,
M0 2006). Among the five spectral bands, FR2 is believed to be the most im-
portant frequency range for the neural activities of local microvascular
M1 beds (Kvandal et al., 2006), and thus is particularly useful in noninva-
M2 sive assessments of SNA. Many previous studies have found decreased
SNA following AS using various techniques, including thermography
(Thomas et al., 1992), microneurography (Knardahl et al., 1998), and
heart rate variability (Lossius et al., 1993). As revealed in Fig. 7 and by
our previous studies (Hsiu et al., 2011a, 2011b), the RECs of FR2 signif-
icantly decreased following AS, which implies a decrease in SNA in local
Fig. 4. Comparison of ApEn values of MMBF and PMBF following AS. ‘*’ indicates
vascular beds. This may further lead to local vasodilation and hence
p b 0.05 compared with the baseline value by two-tailed paired t-test. Following AS, could be correlated with the improved MMBF and PMBF noted in the
the ApEn values of MMBF and PMBF significantly increased at Site 1. present study.
38 H. Hsiu et al. / Microvascular Research 89 (2013) 34–39
Fig. 6. Results of linear regression analysis of relative changes between MMBFCV and ApEn_MMBF during 5-minute segments in M1 at Site 1. The presented results are those in the
first and second 5-minute segments of M1. Relative changes are defined as [effect value] / [baseline (M0) value]. There were significant correlations (p b 0.05 by F-test) with neg-
ative slopes. For the remaining 5-minute segments of M1, the correlation line was y = −0.88x + 2.11 with R2 = 0.45 for the third segment, and y = −0.63x + 1.70 with R2 =
0.25 for the fourth segment. It illustrates that during M1, there was an inverse changing trend between MMBFCV and ApEn_MMBF not only for the average of the whole group (as
revealed in Figs. 4 and 5), but also among all the individual data points.
Changes in MBFV parameters (MMBFCV and PMBFCV) were sug-
gested to be attributable to various microcirculatory regulatory
mechanisms acting on the on–off tuning activities of AO, and hence
could reflect different regulatory vasomotions of local vascular beds.
Significant correlations between changes in the MBF supply and the
beat-to-beat variability within the MBF signal have been revealed
by applying local-heating stimulation to humans21 and administering
angiotensin II to rats (Hsiu et al., 2010a); MBFV parameters might
therefore be useful for evaluating cardiovascular regulatory activity.
Among the five frequency bands of the vasomotion measured by the
LDF signals recorded in the present study, the only one to exhibit a
significant increase was FR5 (which might be attributable to the im-
proved PMBF), and the only one with a significant decrease was
FR2. In the present study, PMBFCV and MMBFCV both significantly
decreased at the needled site. This implies that the decreased SNA fol-
lowing AS could suppress the local vasomotion and thus decrease the
MBFV parameters as noted in this study.
Fig. 6 reveals significant correlations between relative changes in
MMBFCV and ApEn parameters. The trend in the changes in these pa-
rameters (i.e., the negative correlation slopes between individual data
points) was similar to the comparison revealed in Figs. 4 and 5, and
may thus represent further support for the present conjecture.
Fig. 7. Changes in the REC (relate energy contribution) of each band in the LDF spec-
trum at Site 1. Periodic oscillations with five characteristic frequency peaks were ob-
served from 0.0095 to 1.6 Hz, with the positions of these peaks falling within the
following frequency bands: 0.0095–0.02, 0.02–0.06, 0.06–0.15, 0.15–0.4, and
0.4–1.6 Hz (defined as FR1–FR5, respectively). The REC in each frequency band was de-
fined as [(total energy density in that frequency band) / (total energy density in the
entire 0.0095-to-1.6-Hz spectrum)]. ‘*’ indicates p b 0.05 compared with the baseline
value by Dunnett's repeated-measures ANOVA. The REC significantly decreased in
FR2 and significantly increased in FR5.
The effects on the ApEn parameters (Fig. 4) were sustained longer
than those on the MBFV parameters (Fig. 5). Since neural control can
rapidly adjust the properties of vessel walls (e.g., their stiffness or ra-
dius) so as to respond quickly when facing various stimuli that may
influence the MBF supply, it is also possible that the change could di-
minish more rapidly than other regulatory mechanisms when the
stimulation is removed. Since SNA returned toward the baseline
value, other mechanisms might participate in the microcirculatory
regulation so as to sustain the increase in the MBF complexity longer
(until M2). Moreover, based on this conjecture, it might take a while
for the microcirculatory regulatory activities to reach the steady state,
and therefore the length of the settling time might be correlated with
the regulation ability of the recovery condition of local tissues, espe-
cially when facing the threat of harmful stimulation or diseases. Fu-
ture studies could focus on verifying whether understanding this
mechanism could aid the development of a new index to improve
the monitoring of disease progression and evaluate treatment efficacy
in various diseases related to an abnormal MBF supply.
In conclusion, to the best of our knowledge this study was the first
to apply complexity analysis to the beat-to-beat MBF signals in order
to examine the microcirculatory responses following AS. It was re-
vealed that AS can not only improve the MBF supply at the needled
site but may also increase the complexity and decrease the variability
parameters of the MBF signals. These responses may be attributable
to the microcirculatory regulatory activities induced by AS. The pres-
ent findings could help to elucidate the mechanism underlying re-
sponses to AS, and might aid the development of an index for
monitoring the induced microcirculatory regulatory responses, and
thus provide an evidence-based connection between AS and modern
physiology.
Acknowledgments
The authors would like to thank the National Science Council and
Department of Health, Taipei City Government for partial support of
this work.
References
Bernardi, L., Rossi, M., Fratino, P., Finardi, G., Mevio, E., Orlandi, C., 1989. Relationship
between phasic changes in human skin blood flow and autonomic tone. Microvasc.
Res. 37 (1), 16–27.
Carolan-Rees, G., Tweddel, A.C., Naka, K.K., Griffith, T.M., 2002. Fractal dimensions of
laser Doppler flowmetry time series. Med. Eng. Phys. 24 (1), 71–76.
Chen, C.T., Huang, S.M., Hsiu, H., Hsu, W.C., Lin, F.C., Hu, H.F., 2011. Using a blood pres-
sure harmonic variability index to monitor the cerebral blood flow condition in
stroke patients. Biorheology 48 (3), 219–228.
 H. Hsiu et al. / Microvascular Research 89 (2013) 34–39
Esen, F., Esen, H., 2006. Detrended fluctuation analysis of laser Doppler flowmetry time
series: the effect of extrinsic and intrinsic factors on the fractal scaling of microvas-
cular blood flow. Physiol. Meas. 27 (11), 1241–1253.
Esen, F., Aydin, G.S., Esen, H., 2009. Detrended fluctuation analysis of laser Doppler
flowmetry time series. Microvasc. Res. 78 (3), 314–318.
Heistad, D.D., Kontos, H.A., 1983. Cerebral circulation. In: Shepherd, J.T. (Ed.), Hand-
book of Physiology. American Physiological Society, Bethesda, MD, pp. 137–182.
Hsiu, H., Chao, P.T., Hsu, W.C., Jan, M.Y., Wang Lin, Y.Y., Wang, W.K., 2008a. The possible
role of arterial radial dilatation on the heart rate and blood pressure matching. P. I.
Mech. Eng. H 222 (H5), 773–779.
Hsiu, H., Hsu, W.C., Chang, S.L., Hsu, C.L., Huang, S.M., Wang Lin, Y.Y., 2008b. Microcir-
culatory effect of different skin contacting pressures around the blood pressure.
Physiol. Meas. 29, 1421–1434.
Hsiu, H., Hsu, W.C., Hsu, C.L., Jan, M.Y., Wang Lin, Y.Y., 2009a. Effects of acupuncture at
the Hoku acupoint on the pulsatile laser Doppler signal at the heartbeat frequency.
Lasers Med. Sci. 24 (4), 553–560.
Hsiu, H., Hsu, W.C., Hsu, C.L., Huang, S.M., Wang Lin, Y.Y., 2009b. Microcirculatory changes
by laser Doppler after infrared heating over acupuncture points — relevance to mox-
ibustion. Photomed. Laser Surg. 27 (6), 855–861.
Hsiu, H., Hsu, C.L., Jan, M.Y., 2010a. Relations between beat-to-beat microcirculatory
blood flow and variations therein. Photomed. Laser Surg. 28 (6), 785–792.
Hsiu, H., Huang, S.M., Hsu, T.L., 2010b. Evaluation of the function of arteriolar opening
by variability in microcirculatory blood flow following angiotensin II administra-
tion in rats. Biorheology 47 (3), 239–253.
Hsiu, H., Huang, S.M., Chen, C.T., Hsu, C.L., Hsu, W.C., 2011a. Acupuncture stimulation
causes bilaterally different microcirculatory effects in stroke patients. Microvasc.
Res. 81 (3), 289–294.
Hsiu, H., Hsu, W.C., Hsu, C.L., Huang, S.M., 2011b. Assessing the effects of acupuncture
by comparing needling the Hegu acupoint and needling nearby nonacupoints by
spectral analysis of microcirculatory laser Doppler signals. Evid. Based Comple-
ment. Alternat. Med. 435928 (2011).
Humeau, A., Buard, B., Chapeau-Blondeau, F., Rousseau, D., Mahe, G., Abraham, P., 2009.
Multifractal analysis of central (electrocardiography) and peripheral (laser Dopp-
ler flowmetry) cardiovascular time series from healthy human subjects. Physiol.
Meas. 30 (7), 617–629.
Jansen, G., Lundeberg, T., Kjartansson, J., Samuelson, U.E., 1989. Acupuncture and sen-
sory neuropeptides increase cutaneous blood flow in rats. Neurosci. Lett. 97,
305–309.
39
Knardahl, S., Elam, M., Olausson, B., Wallin, G., 1998. Sympathetic nerve activity after
acupuncture in humans. Pain 75, 19–25.
Kvandal, P., Landsverk, S.A., Bernjak, A., Stefanovska, A., Kvernmo, H.D., Kirkeboen, K.A.,
2006. Low-frequency oscillations of the laser Doppler perfusion signal in human
skin. Microvasc. Res. 72 (3), 120–127.
Kvernmo, H.D., Stefanovska, A., Kirkeboen, K.A., Kvernebo, K., 1999. Oscillations in the
human cutaneous blood perfusion signal modified by endothelium-dependent and
endothelium-independent vasodilators. Microvasc. Res. 57 (3), 298–309.
Lee, J.L., Fung, Y.C., 1971. Flow in non-uniform small blood vessels. Microvasc. Res. 3,
272–287.
Liao, F., Garrison, D.W., Jan, Y.K., 2010. Relationship between nonlinear properties of
sacral skin blood flow oscillations and vasodilatory function in people at risk for
pressure ulcers. Microvasc. Res. 80 (1), 44–53.
Lossius, K., Eriksen, M., Walløe, L., 1993. Fluctuations in blood flow to acral skin in
humans: connection with heart rate and blood pressure variability. J. Physiol.
460, 641–655.
Panerai, R.B., 2009. Complexity of the human cerebral circulation. Philos. Transact. A
Math Phys. Eng. Sci. 367 (1892), 1319–1336.
Park, J.M., Shin, A.S., Park, S.U., Sohn, I.S., Jung, W.S., Moon, S.K., 2010. The acute effect of
acupuncture on endothelial dysfunction in patients with hypertension: a pilot,
randomized, double-blind, placebo-controlled crossover trial. J. Altern. Comple-
ment. Med. 16 (8), 883–888.
Pincus, S.M., 1991. Approximate entropy as a measure of system complexity. Proc. Natl.
Acad. Sci. U. S. A. 88, 2297–2301.
Sato, A., Sato, Y., Shimura, M., Uchida, S., 2000. Calcitonin gene-related peptide pro-
duces skeletal muscle vasodilation following antidromic stimulation of unmyelin-
ated afferents in the dorsal root in rats. Neurosci. Lett. 283, 137–140.
Task Force of the European Society of Cardiology and the North American Society of
Pacing and Electrophysiology, 1996. Heart rate variability. Standards of measure-
ment, physiological interpretation, and clinical use. Circulation 93, 1043–1065.
Thomas, D., Collins, S., Strauss, S., 1992. Somatic sympathetic vasomotor changes doc-
umented by medical thermographic imaging during acupuncture analgesia. Clin.
Rheumatol. 11, 55–59.
Tsuru, H., Kawakita, K., 2009. Acupuncture on the blood flow of various organs mea-
sured simultaneously by colored microspheres in rats. Evid. Based Complement.
Alternat. Med. 6 (1), 77–83.
Vaillancourt, D.E., Newell, K.M., 2002. Changing complexity in human behavior and
physiology through aging and disease. Neurobiol. Aging 23 (1), 1–11.