A PROPOSAL OF A NOVEL CARDIORESPIRATORY

LONG-TERM MONITORING DEVICE

S. Lapi, E. Biagi, G. Borgioli, M. Calzolai, L. Masotti

Department of Electronics and Telecommunications, University of Florence, Florence, Italy

G. Fontana

Department of Internal Medicine, University of Florence, Florence, Italy

Keywords: Wearable accelerometer, Cardiorespiratory holter, Sleep monitoring.

Abstract: Monitoring of respiratory movements is an important feature in planning of medical care. We present here a

simple, portable, accelerometer-based device suitable for long term-monitoring of the breathing and heart

rates, along with postural changes, during sleep and wakefulness. Recordings of respiratory frequency, heart

rate, posture and voluntary cough were obtained from a group of volunteers who also participated in sleep

studies (6-8 hrs). A pair of capacitive MEMS tri-axial accelerometers was positioned at the level of the 10

rib along the mid-axillary line bilaterally; simultaneous recordings of respiratory movements, heart rate and

body position could be easily performed. The signal were digitized and used to detect body position and

relative movement between accelerometers. Conventional spirometry was performed in parallel when

appropriate. During resting breathing, qualitative analysis revealed that the accelerometric assessment of

respiratory pattern correlated well with that obtained by spirometry. Values of respiratory rates were

identical with the two techniques. Recordings of respiratory and cardiac activity during sleep were

satisfactorily obtained except for short lasting episodes corresponding to changes in body position. These

devices seem to be also suitable for detecting the motor pattern of cough.

1 INTRODUCTION

It has repeatedly been shown that detailed analysis

of breathing pattern can provide valuable

information regarding the respiratory system (Tobin,

1983). In patients with respiratory and sleep

disturbances, including the chronic obstructive

pulmonary disease (Pauwels, 2001) and the sleep

apnoea syndrome (Britton, 2003), continuous

monitoring of simple vital functions may provide

useful therapeutic options. COPD is a prevalent and

disabling condition that results in significant

personal impact to patients and their carers, and

financial cost to health services. (Pauwels, 2001)

More than half of these costs relates to hospital

admissions for acute exacerbation (Britton, 2003).

Reducing exacerbations and hospitalizations are

therefore key goals in COPD management (Pauwels,

2001).

Obstructive sleep apnoea syndrome (OSAS) is a

highly prevalent disorder (Panossian, 2009)

characterized by instability of the upper airway

during sleep, which results in markedly reduced

absent airflow at the nose/mouth. Episodes are

typically

accompanied by oxyhemoglobin

desaturation and terminated by brief micro-arousals

that result in sleep fragmentation (Panossian, 2009).

Despite having significant breathing problems

during sleep, most patients have no readily

detectable

respiratory abnormality while awake

(Redline, 1993).

Conceivably, continuous monitoring of

respiratory and heart rates, cough frequency and

motor pattern, and time spent daily in physical

activities may provide early information on

deterioration of clinical conditions in patients with

respiratory diseases, thus allowing prompt medical

intervention and possibly reducing both the severity

of disease exacerbations and the need of sudden

hospitalization of the patient (Deegan, 1995). In

addition, domestic sleep monitoring by means of a

simple device that allows simultaneous assessments

Lapi S., Biagi E., Borgioli G., Calzolai M., Masotti L. and Fontana G..

A PROPOSAL OF A NOVEL CARDIORESPIRATORY LONG-TERM MONITORING DEVICE.

DOI: 10.5220/0003136100380042

In Proceedings of the International Conference on Bio-inspired Systems and Signal Processing (BIOSIGNALS-2011), pages 38-42

ISBN: 978-989-8425-35-5

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

of cardiorespiratory rates and posture may represent

a useful tool in the management of OSAS patients.

Several devices have been proposed to measure

ventilation indirectly. Respiratory monitoring is

usually achieved by having a subject breathing

through a mouthpiece or face mask attached

to a

pneumotachograph or spirometer. Although these

devices permit

the accurate measurement of

ventilation and its variables, they

also alter the

pattern of breathing and minute ventilation

(McNicholas, 2008). They are not useful for

monitoring ventilation

in any circumstance in which

keeping a mouthpiece and nose clip

in place is too

difficult or

impossible, as it may be the case of

continuous monitoring at home and/or during sleep

(Hurst, 2009). Devices that measure changes in

thoracic volume (respiratory inductance

plethysmography), and airflow at the airway opening

(oro-nasal thermistors) have been fully validated and

largely employed in the clinical assessment of sleep

disturbances; however, it is widely recognised that

their use for recording respiratory activity during

daily life is impractical, mainly due to their limited

portability.

We present here a simple, mini-invasive,

accelerometer-based device that can be used for long

term-monitoring of respiratory movements and

cardiac activity, thus allowing the detection of the

breathing frequency and heart rate, along with the

ongoing postural changes, during sleep and

wakefulness. The device can also provide useful

information on the motor pattern of cough and other

expulsive efforts. In this connection, it seems worth

to recall that the cough reflex is an important airway

defensive mechanism and the assessment of its

functionality is becoming increasingly important,

especially in elderly patients who are at risk of

aspiration pneumonia (Perez, 1985). The intensity of

a cough effort, generally indexed in terms of

expiratory flow rate and/or electromyographic

abdominal muscle activity (Que, 2002) is of pivotal

importance in the assessment of cough effectiveness.

Specifically, the purposes of this paper are i) to

determine whether the accelerometer-based device is

suitable for short- and long-term recordings of

respiratory and heart rates in conditions such as

sleeping and daily activities; ii) to obtain preliminary

information on the feasibility of non-invasive

recording of more complex respiratory motor acts

such as the cough.

2 MATERIALS AND METHODS

The proposed device consists of two MEMS

capacitive tri-axial accelerometers (MMA7260Q,

Freescale Semiconductor) which were chosen for

their high sensitivity (800 mV/g @1.5 g).

Due to manufacturing process and physical

structure of each sensor, measures could be affected

by a systematic error; to offset the output of the

accelerometers used in the experimentation, a

calibration procedure was devised.

Accelerometric sensors were mounted on two

small circuit boards, as illustrated in figure 1,

containing a first conditioning module and a voltage

regulator.

Figure 1: Accelerometer mounted on the circuit board.

Note the small size of the device compared with a 2 Euro

cent coin.

Accelerometric sensors were positioned

bilaterally on the skin of the anterior thoracic wall at

the level of the 10

rib along the mid-axillary line,

by using paper adhesive bandage. In preliminary

experimental sessions, several different attempts at

identifying the most suitable thoracic area for sensor

positioning were carried out. The margin of the 10

rib on the mid-axillary line was selected as it turned

out to provide a respiratory signal consistent with

that obtained with conventional spirometry.

Both sensors were connected with a flexible flat

cable to a data acquisition board, containing

analogue conditioning circuits, A/D converter, a real

time clock for acquisition management and power

supply. The acquisition board is embedded in a

12 cm x 7 cm x 3 cm lightweight plastic case

equipped with a clasp on the back side for hanging

the device to the subject’s belt in order to improve

monitoring device’s portability.

Acceleration signals were sampled at 340 Hz,

stored in a removable mass memory device and later

transferred to the computer for off-line processing,

which included digital filtering, offset correction and

study of frequency components.

During offline processing two digital FIR filters

were used to extract the respiratory and cardiac

components from the raw signal, respectively in the

A PROPOSAL OF A NOVEL CARDIORESPIRATORY LONG-TERM MONITORING DEVICE

[0.1, 0.6] Hz and [1, 15] Hz frequency range.

However, the use of a digital filter “per se” is not

effective in removing signal components originating

from body movements unrelated to respiratory

activity; in facts, the frequency components of non-

respiratory body movements are generally lower

than 15 Hz and therefore comprised in the range of

respiratory and cardiac signals. In consequence, the

influence of non-respiratory thoracic movements

was minimised by considering the difference ‘d’

between the sensor’s resultant vectors ‘R1’ and ‘R2’

(figure 2). By subtracting the common components

acquired from the two accelerometers, a signal of

purely respiratory origin was obtained. Once

positioned on the chest wall, however, the reference

co-ordinate systems of the two sensors were not

aligned, leading to a systematic error due to an

imbalance between common components. To avoid

imbalance, we developed a procedure that rotates the

reference system of one sensor in order to align it to

the reference system of the other sensor. With this

procedure, the accelerometric signals deriving from

non-respiratory thoracic movements were

minimised.

(a) (b)

Figure 2: Sensors location on the chest wall (a),

representation of resultant vectors ‘R1’, ‘R2’ and their

difference ‘d’ (b).

We recruited six subjects of both sexes (age

range 19 – 64) among faculty personnel and

students; none of them complained of any relevant

respiratory or cardiac disturbance and agreed to

participate in accelerometric recordings of

respiratory frequency and heart rate. Three other

subjects were selected to perform a series of

maximum voluntary cough efforts starting either

from near total lung capacity or from functional

residual capacity. Accelerometric measurements in

awaken patients were always performed in parallel

with conventional spirometry by using a mouthpiece

connected to a pneumotachograph which provided

standard outputs of respiratory volume and flow.

Nine additional subjects participated in sleep studies

lasting 6-8 hrs during which simultaneous

recordings of posture, respiratory and heart rate by

means of the accelerometers were obtained.

3 RESULTS

As expected, accelerometric recordings of

respiratory activity provided a precise measure of

respiratory rate in all subjects tested (figure 3, upper

panel); similar consideration were also valid for

heart rate recordings (figure 4). Furthermore, with

the subject seated in the upright position, qualitative

analysis of the recordings revealed that the pattern of

breathing obtained by considering the difference ‘d’

between the two accelerometers was fairly coherent

with that obtained with the spirometer (figure 3,

lower panel). In more detail, the different patterns of

rapid shallow breathing and deep breathing were

correctly detected by the accelerometers (figure 3).

Figure 3: Upper panel: respiratory signals from the

accelerometer pair: X, Y and Z are the signals originating

from each of the two tri-axial accelerometers. For the

purpose of morphological comparisons, all traces are in

arbitrary units and displayed at different levels. Lower

panel: morphological comparison of the spirometric

volume (in blue) and of the accelerometer signal

difference (in red) during normal breathing, rapid shallow

breathing and deep breathing. Signals were differently

amplified to allow superimposition.

As shown in figure 5, the accelerometric tracings

recorded during voluntary cough efforts of variable

intensity turned out to be clearly distinguishable

from those of resting breathing. The intensity of

BIOSIGNALS 2011 - International Conference on Bio-inspired Systems and Signal Processing

Figure 4: 15 s recording of cardiorespiratory activity by

means of an accelerometer in an awake subject at rest.

Traces derive from the Y axis of an accelerometer

positioned on the chest wall projection of the cardiac area

and were digitally filtered (see methods) to obtain the

signals of the ongoing respiratory (in blue) and cardiac

activities (in red). Traces are in arbitrary units and were

similarly amplified.

Figure 5: Upper panel: accelerometric recordings of three

single cough efforts obtained by means of the

accelerometer pair: X, Y and Z are the signals originating

from each of the two tri-axial accelerometers. Traces are

in arbitrary units and displayed at different levels to

facilitate comparisons of their morphology. Lower panel:

morphological comparison of spirometric flow (in blue,

cough expulsions marked by downward swings) and

accelerometer signal difference (in red, expulsions marked

by upward swings) during three single cough efforts.

Signals were differently amplified to facilitate

morphological comparison.

each cough event can be appreciated in terms of the

magnitude of the corresponding spirometric flow

signal. At the present stage of development it seems

likely that the accelerometric signals are suitable to

detect cough appearance but not to quantify its

intensity.

Long term accelerometric recordings of

cardiorespiratory activity and postural behaviour

during sleep were successfully accomplished in all

of the nine normal subjects. Respiratory and cardiac

activity

could be recorded only during stable body

Figure 6: Upper panel: 7.5 h recording of cardiorespiratory

and postural activities during sleep in one representative

subject using an accelerometer. The signal of respiratory

activity was derived from a single axis of one of the two

accelerometers (see methods). Sudden variations of signal

mark the changes in body position that occurred

throughout the recording period. Middle and lower panels:

60 s - time window of respiratory and cardiac activities

obtained from the upper panel tracing.

posture; in fact, during changes in body position, the

signal related to general movement masked the

cardiorespiratory signal (figure 6, upper panel). As a

consequence, the device allowed us to accurately

detect the number of postural adjustments that

occurred during the recording period.

4 DISCUSSION

This study shows that the use of an accelerometer

pair positioned on the chest wall seems to be suitable

for non-invasive detection of cardiorespiratory and

postural behaviours, both during wakefulness and

sleep; furthermore, the device may also be useful for

recording the number and the motor pattern of cough

events.

Any signal derived from each of the sensors can

provide information about the frequency of cardiac

activity. The latter is displayed as a sequence of

peaks corresponding to the systole. Further analyses

revealed that the Z axis is the most sensitive one in

picking vibrations of the chest wall caused by heart

activity, especially when the sensor is positioned on

the chest wall area corresponding to heart projection.

Qualitative analysis also revealed that the signals

obtained by the two accelerometric sensors,

particularly the difference signal, strictly matched

with the respiratory thoracic movements. The

preliminary results seem also to suggest that not

only respiratory frequency, but also the amplitude of

each breath, is correctly detected by the

accelerometer pair. Indeed, values of respiratory rate

obtained by spirometry and accelerometers were

A PROPOSAL OF A NOVEL CARDIORESPIRATORY LONG-TERM MONITORING DEVICE

identical, in keeping with previous findings

(Reinvuo, 2006) and (Hung, 2008). However, this

study is the first to show that voluntary changes in

the depth of breathing can also be accurately

detected by the device.

This study is also the first to demonstrate that the

motor pattern of voluntary cough can successfully be

distinguished from that of normal breathing by

means of an accelerometer pair. In addition, the

accelerometric pattern of cough was fairly coherent

with that recorded spirometrically, and revealed high

frequency components due to speed of the expulsive

activity and related chest wall vibrations. However,

inaccuracies still exist as far as the accelerometric

evaluation of cough intensity is concerned.

Accelerometric long-term recordings of

cardiorespiratory activity during sleep in normal

subjects provide some advantages compared with

standard techniques. Although not formally tested

here, it seems obvious that our device offers better

portability than the conventional ones and that the

possibility of simultaneous recording of body

posture may represent an important tool in the

evaluation of sleep disturbances.

It could be argued that, in order to monitor

cardiorespiratory activity, the use of an

accelerometer-based device is somewhat limited by

the interfering signals of body motion, especially

during daily activities. Whilst we acknowledge that

this limitation may exist at least to some extent, we

also feel that the recording a cardiorespiratory signal

“disturbed” by that originating from body

movements is “per se” of clinical usefulness. It may

be inferred that patients with respiratory

disturbances must present a ratio of resting-to-

activity time that is inversely related to the general

clinical condition. Therefore, an increase in the

above time ratio could be interpreted as an index of

a deteriorating clinical condition. In addition, in

patients with respiratory diseases, the detection of an

increase in cardiac and respiratory activity at rest

could also point to an increased metabolic demand

such as in the event of a respiratory exacerbation.

5 CONCLUSIONS

Simultaneous recordings of respiratory movements,

heart rate and body position can easily be

accomplished by using pairs of tri-axial

accelerometers; these devices seem to be also

suitable for the detection of the motor pattern of

cough.

The device can be employed for daytime and

nocturnal long-term monitoring thanks to its small

dimensions, small weight and easy positioning of

sensors on the chest wall, that warrant non-invasive

measurements. It could be employed in the diagnosis

of sleep disturbances such as the sleep apnoea

syndrome, or in the monitoring of the elderly. Even

at the present stage of development, the device

presented here appears to be ready for accurate and

reliable long-term sleep studies. Being easily

portable and not bulky, the device seems to be

particularly suitable for sleep studies in the domestic

environment.

REFERENCES

Tobin, M. J., Chadha, T. S., Jenouri, G., Birch, S. J.,

Gazeroglu, H. B., Sackner, M. A., 1983. Breathing

patterns. 1. Normal subjects. In Chest; 84 (2): 202-5.

Pauwels, R. A., Buist, A. S., Calverley, P. M., Jenkins, C.

R., Hurd, S. S., 2001. Global strategy for the

diagnosis, management, and prevention of chronic

obstructive pulmonary disease. NHLBI/WHO Global

Initiative for Chronic Obstructive Lung Disease

(GOLD) Workshop summary. In Am J Respir Crit

Care Med; 163(5):1256-76.

Britton, M., 2003. The burden of COPD in the U.K.:

results from the Confronting COPD survey. In Respir

Med; 97 Suppl C:S71-9.

Panossian, L. A., Avidan, A. Y., 2009. Review of sleep

disorders. In Med Clin North Am.; 93(2):407-25.

Redline, S., Young, T., 1993. Epidemiology and natural

history of obstructive sleep apnea. In Ear Nose Throat

J; 72(1):20-1, 24-6.

Deegan, P. C., McNicholas, W. T., 1995. Pathophysiology

of obstructive sleep apnoea. In Eur Respir

J;8(7):1161-78.

McNicholas, W. T., 2008. Diagnosis of obstructive sleep

apnea in adults. In Proc Am Thorac Soc;5(2):154-60.

Hurst, J. R., Wedzicha, J. A., 2009. Management and pre-

vention of chronic obstructive pulmonary disease

exacerbations: a state of the art review. In BMC Med;7:40.

Perez, W., Tobin, M. J., 1985. Separation of factors

responsible for change in breathing pattern induced by

instrumentation. In J Appl Physiol; 59 (5):1515-20.

Que, C. L., Kolmaga, C., Durand, L. G., Kelly, S. M.,

Macklem, P. T., 2002. Phonospirometry for non-

invasive measurement of ventilation: methodology and

preliminary results. In J Appl Physiol;93(4):1515-26.

Reinvuo, T., Hannula, M., Sorvoja, H., Alasaarela, E.,

Myllyla, R., 2006. Measurement of respiratory rate

with high-resolution accelerometer and emfit pressure

sensor. In Sensors Applications Symposium, 2006.

Proceedings of the 2006 IEEE , vol., no., pp. 192- 195,

Hung, P. D., Bonnet, S., Guillemaud, R., Castelli, E., Yen,

P. T. N., 2008. Estimation of respiratory waveform

using an accelerometer. In Biomedical Imaging: From

Nano to Macro, 2008. ISBI 2008. 5th IEEE

International Symposium on, vol., no., pp.1493-1496,

14-17 May

BIOSIGNALS 2011 - International Conference on Bio-inspired Systems and Signal Processing