applied

sciences
Article
Estimation of Work of Breathing from Respiratory
Muscle Activity In Spontaneous Ventilation:
A Pilot Study
Isabel Cristina Muñoz 1 , Alher Mauricio Hernández 1, * and Miguel Ángel Mañanas 2
1 Bioinstrumentation and Clinical Engineering Research Group—GIBIC, Bioengineering Department,
Engineering Faculty, Universidad de Antioquia UdeA, Calle 70 No. 52-21, Medellín 050010, Colombia;
[email protected]
2 Department of Automatic Control (ESAII), Biomedical Engineering Research Center (CREB) and Biomedical
Research Networking Center in Bioengineering, Biomaterials, and Nanomedicine, CIBER-BBN, Universitat
Politècnica de Catalunya, 08928 Barcelona, Spain; [email protected]
* Correspondence: [email protected]




Received: 22 March 2019; Accepted: 9 May 2019; Published: 16 May 2019


Abstract: Work of breathing (WOB) offers information that may be relevant to determine the patient’s
status under spontaneous mechanical ventilation in Intensive Care Unit (ICU). Nowadays, the most
reliable technique to measure WOB is based on the use of invasive catheters, but the use of qualitative
observations such as the level of dyspnea is preferred as a possible indicator of WOB level. In this
pilot study, the activity of three respiratory muscles were recorded on healthy subjects through surface
electromyography while they were under non-invasive mechanical ventilation, using restrictive and
obstructive maneuvers to obtain different WOB levels. The respiratory pattern between restrictive
and obstructive maneuvers was classified with the Nearest Neighbor Algorithm with a 91% accuracy
and a neural network model helped classify the samples into three WOB levels with a 89% accuracy,
Low: [0.3–0.8) J/L, Medium: [0.8–1.3] J/L and Elevated: (1.3–1.8] J/L, demonstrating the relationship
between the respiratory muscle activity and WOB. This technique is a promising tool for the healthcare
staff in the decision-making process when selecting the best ventilation settings to maintain a low
WOB. This study identified a model to estimate the WOB in different ventilatory patterns, being
an alternative to invasive conventional techniques.

Keywords: non-invasive ventilation; lung diseases; work of breathing; respiratory muscles; surface
electromyography; machine learning

1. Introduction
The work of breathing (WOB) has been defined as the energy required to accomplish the body´s
ventilatory demand, which in spontaneous breathing depends on the effort performed by respiratory
muscles [1]. Monitoring WOB levels provides several advantages for healthcare staff in the decision-making
process, making it possible to adjust the mechanical ventilator’s settings, and the pharmacological treatment;
furthermore, it helps understand the mechanism of the pathologies and assess the patient’s evolution
post-extubation [1–3].
Some authors have identified WOB values for different patients, both with and without mechanical
ventilation. WOB values in the range of 0.3 to 0.8 J/L are considered indicators of low respiratory demand
in patients under spontaneous mechanical ventilation or low WOB [4–6]. Sharp et al. [6] found that
96% of patients with a WOB lower than 0.8 J/L had a successful weaning process. On the other hand,
values in the range of 0.8–1.3 J/L are considered a population with high respiratory demand, such as
people with airway and pulmonary parenchymal abnormalities, intubated patients, people with dynamic

Appl. Sci. 2019, 9, 2007; doi:10.3390/app9102007 www.mdpi.com/journal/applsci

Appl. Sci. 2019, 9, 2007 2 of 18

hyperinflation, airflow obstruction (Chronic Obstructive Pulmonary Disease or COPD patients) or obese
people [4,7,8]. In addition, WOB values in the range of 1.3–1.8 J/L are considered elevated, making it
impossible to maintain an adequate spontaneous ventilation for a long period of time [5,8].
According to the state of the art, the use of an esophageal balloon catheter is required to accurately
measure the work of breathing. However, this technique is invasive and uncomfortable for the patient,
which restricts its use in clinical practice [9,10].
Several clinical strategies to detect an increase in work of breathing include distress or dyspnea, vital
signs, physical signs of increased breathing effort, breathing rate, tidal volume, and frequency/tidal volume
ratio [3,5,11]. However, these strategies are not reliable in all cases, which hampers its generalization [3].
Dyspnea is defined as “breathing discomfort”, which occurs when there is an imbalance between
the load imposed upon the respiratory muscles and the capacity of these to overcome it. For that reason,
many authors have successfully used surface electromyography (sEMG) signals of respiratory muscles
to quantify the level of dyspnea [12–14]. Therefore, a relationship between respiratory muscle activity
and work of breathing can be inferred.
Some authors have proposed indexes from sEMG signals of respiratory muscles, in the following
domains: time, frequency, and time-frequency [12,15–17]. However, the relationship between respiratory
muscle activity and work of breathing has not been proven yet, because those authors had no WOB
measurements. Furthermore, considering that multiple respiratory muscles are involved, patients
may have different ventilation patterns to achieve the same work, and to our knowledge this has not
been studied.
Some researchers have studied muscle activity patterns from sEMG signals and used different
machine learning algorithms to recognize activation patterns in several muscles, mainly of the limbs.
Chowdhury et al. [18] presented a review describing the use of sEMG, showing selected features,
used muscles, classification algorithms, objective, and the accuracy of the studies. A common objective
was the classification of patients with neuromuscular diseases and muscular fatigue using sEMG signals
of the Biceps brachii muscle. Consequently, classifiers like the Simple Fisher Linear Discriminant Analysis
(LDA), statistical K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Neural Networks (NN),
among others, were used.
In mechanical ventilation, in order to achieve an optimal configuration of the ventilator, some authors
have used Artificial Neural Networks to monitor parameters like total positive end-expiratory pressure
(PEEP) [19] and power of breathing or work of breathing per minute [11,20] from ventilatory signals.
In previous studies, our team used unsupervised learning techniques to find indexes that were related
to the patient’s condition or identify muscle respiratory patterns [15,17,21]. Based on the hierarchical cluster,
an unsupervised learning technique, Hernández et al. [21] characterized different muscular responses
attributed to changes in the mechanical ventilator settings. Using similar techniques, Salazar et al. [15]
identified muscle weakness in ventilated patients, and in a complementary study, patients who presented
failures in the weaning process were separated from patients with a successful weaning [17].
This paper presents a strategy to estimate work of breathing exerted by the subject during the
ventilation process using indexes from sEMG signals based on machine learning techniques to identify
and discriminate different respiratory patterns in spontaneous ventilation.

2. Materials and Methods

2.1. Experimental Design

The study was conducted according to the Helsinki Declaration with subsequent revisions.
The experimental design and written consent were approved by the Ethics Committee of the University
of Antioquia, approval report 15-59-664.
Two experiments were designed to achieve different levels of work of breathing. Experiment one
(E1) involved restrictive maneuvers to reduce the respiratory system compliance and experiment two
(E2) involved obstructive maneuvers to increase airway resistance.
 Appl. Sci. 2019, 9, 2007 3 of 18

Appl. Sci. 2019, 9, x 3 of 19

In both experiments, the participants were connected to a mechanical ventilator Hamilton G5
In both experiments,
(Hamilton the participants
Medical, Bonaduz, were connected
Switzerland) through atonon-invasive
a mechanical oronasal
ventilator mask.
Hamilton
TheG5ventilator was
(Hamilton Medical, Bonaduz, Switzerland) through a non-invasive oronasal mask. The
set in the spontaneous mode, with a PEEP of 0 cmH2O, pressure support of 0 cmH2O, and FiO2 of 21%. ventilator
was set in
In the
E1,spontaneous
a cuirass [22]mode,
was with
placeda PEEP
on theof chest
0 cmH2O,and pressure
abdomen support
of the of 0 cmH2O,changing
volunteers and FiO2 the pressure
of 21%.
in the cuirass with an external controlled pump, which allowed the reduction of the respiratory system
In E1, a cuirass [22] was placed on the chest and abdomen of the volunteers changing the
compliance. Four samples were recorded for each subject. The first sample was recorded without
pressure in the cuirass with an external controlled pump, which allowed the reduction of the
the cuirass,
respiratory followed
system by three
compliance. Four samples sequentially
samples were recorded with pressures
for each of 0,first
subject. The 10,sample
and 20wascmH2O inside
the cuirass.
recorded without the cuirass, followed by three samples sequentially with pressures of 0, 10, and 20
cmH2O In E2, the
inside twocuirass.
different external airflow resistors (Model 7100R; Hans Rudolph, Inc., Kansas City,
MO) were
In E2, twoused between
different the airflow
external oronasal mask and
resistors (Modelthe 7100R;
flow sensor. Three samples
Hans Rudolph, were City,
Inc., Kansas recorded in each
MO) were used between the oronasal mask and the flow sensor. Three samples
subject sequentially with resistive loads of 0, 5, and 20 cmH2O/L/s (see Figure 1). were recorded in each
subject sequentially with resistive loads of 0, 5, and 20 cmH2O/L/s (see Figure 1).

Figure
Figure 1. Experimental
1. Experimental set-up.set-up. Two independent
Two independent experiment
experiment were performed:
were performed: in experiment
in experiment 1 (E1), 1 (E1), 22
volunteers
22 volunteers were
were enrolled,
enrolled, and and 4 samples
4 samples per volunteer
per volunteer were obtained;
were obtained; and in experiment
and in experiment 2 (E2), 32 2 (E2), 32
volunteers
volunteers werewere enrolled,
enrolled, and 3 samples
and 3 samples per volunteer
per volunteer were obtained.
were obtained.

Each sample consisted of a 3-minute recording, where the first two minutes were considered
an adaption period after a change of stimulus because there is high variability among subjects during
the transient. Therefore, only the last minute associated with stable state was analyzed in this study.
Appl. Sci. 2019, 9, 2007 4 of 18

As shown in Figure 1, the samples obtained per volunteer are recorded sequentially, e.g., in E2,
the volunteer began with an external resistor of 0 cmH2O for 3 minutes, and subsequently, the external
resistor was changed to 5 cmH2O; the 2 minutes following the change in resistance is considered the
adaption period or time needed to stabilize the ventilatory process.

2.2. Subjects
The inclusion criterion was: Adult male subjects. The exclusion criteria were: subjects with a body
mass index higher than 30, those undergoing medical treatment, with implanted electronic devices or
thoracic trauma, those who had ingested alcohol 48 h prior, those who frequently use hallucinogens
and who practiced yoga or pilates due to atypical breathing patterns [23].
A public and independent call for volunteers was made to participate in each experiment, looking
for at least twenty-two participants per experiment in accordance with Equation (1) [24].
p 2
 θ(1 − θ)zβ +0.5zα 
N =  (1)
0.5 − θ 

where α is the significance value, β the statistical test power and θ the probability that the
alternative hypothesis is met. With a reported variability of respiratory pattern of 0.15 (θ = 0.15) [25],
α = 0.05 (Zα = 1.645), and β = 0.99 (Zβ = 2.326), the sample size N should be at least 22.3. According to
this result, twenty-two volunteers (age of 25.95 ± 6.54 years, height of 173.18 ± 5.11 cm and weight of
72.90 ± 6.54 Kg) were enrolled to the restrictive maneuver experiment (E1) and thirty-two different
volunteers (age of 26.62 ± 5.79 years, height of 172.45 ± 4.48 cm and weight of 72.82 ± 9.43 Kg) were
enrolled to the obstructive maneuver experiment (E2) (see Figure 1).

2.3. Measurements and Signal Acquisition

2.3.1. Respiratory and Surface Electromyography Signals

Airflow, airway pressure and tidal volume signals were recorded using a datalogger attached to
the mechanical ventilator (Hamilton Medical, Bonaduz, Switzerland) and sEMG of respiratory muscles
was acquired with the biopotential amplifier Bagnoli-8 (Bagnoli EMG system, Delsys, Boston, US,
2003). A sampling frequency of 1024 Hz was used in both equipment for each single channel.
Ag-AgCl electrodes in bipolar configuration were placed on all subjects to record three respiratory
muscles, diaphragm (Dia), intercostal (Int) and sternocleidomastoid (Strn). A pair of electrodes was
used to record Dia, placed between the seventh and eighth intercostal spaces, located in the middle of
the mid axillary line and the external clavicular line [26]. Electrodes placed in the second and third
intercostal spaces, at the midclavicular line, were used to record Int [27], and electrodes between the
mastoid process and mid-point of the muscle were used to record Strn [26].

2.3.2. Respiratory Mechanics

To estimate the respiratory system compliance, an occlusion for 3 seconds was set in the ventilator
at the end of inspiration, the occlusion maneuver was performed by the mechanical ventilator during
the last minute in the first respiratory cycle. Once the respiratory system compliance was estimated,
an optimization algorithm was applied to calculate the airway resistance (see technical details in [28,29]).
After estimating the airway resistance and respiratory system compliance, the muscular pressure
(Pmus) was calculated with equation of motion (2) and WOB with Equation (3) for each inspiratory cycle.

1
Pmus = ∗V + R ∗ Q + PEEPT − Paw (2)
C T
Z t
1
WOBr = Pmus × Q dt (3)
VT 0
 Appl. Sci. 2019, 9, 2007 5 of 18

where Paw , Q, and VT represent pressure measured at the mouth, airflow, and tidal volume respectively.
PEEPT is total positive pressure at the end of expiration, C is respiratory system compliance, R is
airway resistance, and t is the inspiratory time [30].
According to the literature, depending of the subject’s effort and respiratory load, the WOB
takes values in different ranges that can be divides into three groups [4–8], G1: [0.3–0.8) J/L, G2:
[0.8–1.3] J/L, and G3: (1.3–1.8] J/L. The determined group was considered as the true value in the
classification process.

2.4. Data Analysis

2.4.1. Indexes from Surface Electromyography of Respiratory Muscles

From the last minute, variable epochs of at least thirty seconds of sEMG signals were selected,
avoiding motion artifacts per epoch. Each signal was filtered with a 4th-order Butterworth 60 Hz
bandstop filter and a FIR bandpass filter between 10 Hz and 500 Hz [31,32]. Furthermore, a 5th-order
RLS adaptive filter was used to reduce cardiac artifacts [33].
The ventilatory and sEMG signals were synchronized through an auxiliary signal recorded by
both the ventilator and EMG amplifier [34] in order to identify the respiratory cycles. All indexes were
calculated for each respiratory cycle during the epoch of thirty seconds, and the average was obtained
for each sample.
• Time Domain Indexes
The amplitude of sEMG signals was obtained through the Hilbert transformation and its maximum
was calculated as the difference between maximum and minimum values during the inspiratory phase
(see Figure 2) [16]. Information concerning muscle synchronization was obtained by calculating the
Appl. Sci. 2019, 9, x 6 of 19
Pearson’s correlation coefficient between sEMG amplitude signals [15] (see Table 1).

Figure
Figure 2. 2. sEMG
sEMG indexesinintime
indexes timeand
andfrequency
frequency domain
domain during
duringthe
theinspiratory
inspiratoryphase.
phase.MaxAmp,
MaxAmp, the
the maximum amplitude of the sEMG signal; HF and LF, the high and low frequency ranges for
maximum amplitude of the sEMG signal; HF and LF, the high and low frequency ranges for
calculating RHL.
calculating RHL.

Table 1. sEMG indexes in time and frequency domains for each one of the recorded muscles,
diaphragm (Dia), intercostal (Int) and sternocleidomastoid (Strn).

Index Time Domain Index Frequency Domain

Appl. Sci. 2019, 9, 2007 6 of 18

Table 1. sEMG indexes in time and frequency domains for each one of the recorded muscles, diaphragm
(Dia), intercostal (Int) and sternocleidomastoid (Strn).

Index Time Domain Index Frequency Domain

Maximum amplitude of the sEMG The central frequency of power spectral
MaxAmpj Fcj
signal of muscle j: Dia, Int or Strn density of sEMG signal of muscle j
Pearson correlation between pairs The ratio between high and low
Rj-k of amplitudes of sEMG signals of RHLj frequencies of power spectral density of
the muscles j and k sEMG signal of muscle j
Spectral coherence between pairs of
Cj-k
sEMG signals of the muscles j and k

• Frequency Domain Indexes

The power spectral density (PSD) of sEMG signals was estimated using the Burg method with
an 8th-order (see Figure 2). Using the PSD, the central frequency and ratio between high and low
frequencies, RHL (LF: 20–40 Hz and HF: 138–240 Hz) were calculated to quantify activation level
of different types of muscle fibers [17,21]. Finally, the coherence between sEMG signals of different
muscles was obtained to identify the muscle engagement in the frequency domain (see Table 1).

2.4.2. Work of Breathing Estimation

Figure 3 summarizes the methodology to estimate the level of WOB from sEMG signals. The first
part consists of determining which indexes from the sEMG signals are relevant in describing the
breathing pattern and the work of breathing. The second part includes two classification steps. The first
one identifies whether the breathing pattern corresponds to a restrictive or obstructive one. Once the
breathing pattern is identified, the subjects are classified into one of the three WOB levels, as established
in the literature [4–8]. Lastly, the work of breathing is estimated using information from respiratory
muscle sEMG signals.
Appl. Sci. 2019, 9, x 7 of 19

Figure 3.
Figure 3. Methodology
Methodology to to
estimate the the
estimate levellevel
of WOB from sEMG
of WOB signals.signals.
from sEMG 1—determining which of 15
1—determining which
of 15 indexes from the sEMG signals are relevant to describe the breathing pattern and work
indexes from the sEMG signals are relevant to describe the breathing pattern and the of of
the work
breathing. 2—classification steps: 2.1—identifying if the breathing pattern corresponds to a restrictive
breathing. 2—classification steps: 2.1—identifying if the breathing pattern corresponds to a restrictive
orobstructive
or obstructive one;
one; 2.2—classifying
2.2—classifyingthe thesubject
subjectinin
one
oneof of
three groups
three groupsof WOB. 3—estimating
of WOB. workwork
3—estimating of of
breathing. The results of each step of the methodology are presented in Figure 5.
breathing. The results of each step of the methodology are presented in Figure 5.

• Relevant Indexes Selection

• Relevant Indexes Selection
Before the first classification, a dimension reduction was performed to choose the most relevant
Before
indexes (seethe first classification,
Section a dimension
2.3.1). The following reduction
two algorithms of was performed
relevant indexesto choosewere
selection the most
used:relevant
the
indexes
stepwise(see Section 2.3.1).
regression and theThemutual
following two algorithms
information of relevant
criterion. indexes
Statistically selection
significant were used:
indexes were the
stepwise
examinedregression and the mutual
using Kruskal-Wallis information
test, followed criterion.
by the post hoc Statistically
Tukey-Kramersignificant
test forindexes
multiplewere
comparisons.
examined Additionally,
using a Mann–Whitney
Kruskal-Wallis test, followedtest by
on the
eachpost
pair hoc
of groups was done test
Tukey-Kramer to determine
for multiple
which groups differed from each other.
1. Stepwise Regression
The stepwise regression is a systematic method for adding and removing indexes based on their
statistical significance in a regression [35]. In this study, the p-value for determining the use of an
index was 0.05. The summarized sequential algorithm steps are [35]:
• Calculus of correlation coefficients of n-indexes with the response variable Y.
Appl. Sci. 2019, 9, 2007 7 of 18

comparisons. Additionally, a Mann–Whitney test on each pair of groups was done to determine which
groups differed from each other.
1. Stepwise Regression
The stepwise regression is a systematic method for adding and removing indexes based on their
statistical significance in a regression [35]. In this study, the p-value for determining the use of an index
was 0.05. The summarized sequential algorithm steps are [35]:

• Calculus of correlation coefficients of n-indexes with the response variable Y.

• Sort the n-indexes from highest to lowest correlation coefficient X = [x1, x2 . . . ..xn].
• Add index xi to the P vector.
• Calculus of regression and partial Fj -values between response variable Y and P, where j is the
index in the P vector.
• Selection of the lowest Fj -value and comparison between this and the F-value at α = 0.05 in
the F-distribution.
• If the Fj -value is higher than the F-value at α = 0.05, i is incremented by one, the index Pj is
removed from the P vector.
• Steps 3–6 are repeated until the last index enters the P vector.

The final P vector contains the relevant indexes.

2. Mutual Information
The mutual information gives an idea of the general statistical dependence between two
variables [36]. In this study, the mutual information was calculated between each one of the sEMG
indexes and the target group through Equation (4).

I (X; Y) = H (X) − H (X, Y) (4)

where X was each one of the indexes, Y the target group, and H the entropy of a random variable
defined by Equation (5).
X
H (X ) = − p(X)log2 p(x) (5)

where p(X) is the probability density function of index X.

The indexes were sorted from higher to lower mutual information, and each one was examined
with a multiple comparisons test.
3. Kruskal-Wallis and Post Hoc Analysis
The indexes selected with each one of the algorithms were examined with the Kruskal-Wallis test
followed by the post hoc Tukey-Kramer multiple comparisons correction and Mann–Whitney tests on
each pair of groups. Only the indexes that separate statistically (p-value < 0.05) in any of the groups
were considered in the machine learning process.
• Machine Learning
Steps 2.1 and 2.2 in Figure 3 identify the respiratory muscular pattern and the WOB
level, respectively.
Due to the fact that in clinical practice, knowing the ventilatory pattern is not common,
its identification is necessary before the WOB level classification. In this study, the ventilatory
pattern was defined according to the respiratory muscular activity, in order to know what muscles,
participate in the ventilatory process.
Several supervised learning techniques were selected according to their wide and successful
use in classifying biomedical data [18]. The techniques tested were: LDA, SVM, KNN, an ensemble
algorithm, and NN. The parameters of the three first classifiers (LDA, SVM, and KNN) were optimized
Appl. Sci. 2019, 9, 2007 8 of 18

using the Bayesian optimization through the option of the MATLAB algorithms. In the case of the
Neural Network, three hidden layer sizes, with default values of 10, 5 and 20, were tested. Table 2
shows the parameters obtained for greater accuracy (see Table 2).

Table 2. Information about configuration and validation of each of the classification algorithms.
The algorithms used in each of the cases, classification of the respiratory muscular pattern and WOB
level, are specified. LDA: Linear Discriminant Analysis, SVM: Support Vector Machine, KNN: K-Nearest
Neighbor, NN: Neural Network.

Respiratory Muscular Pattern

Classifier Algorithm Configuration Validation
Gamma = 0
Delta = 0
LDA K-fold 20%
Discriminant type: linear
Indexes: logarithmic transformed
Box Constraint = 962.47
Kernel Scale = 8.4
SVM K-fold 20%
Kernel Function: Gaussian
Indexes: Standardized
Distance: Euclidean
KNN Number of Neighbors: 2 K-fold 20%
Indexes: Standardized
WOB LEVEL
Experiment 1 Experiment 2
Classifier Algorithm Validation
Configuration Configuration
Gamma = 0.49
Delta = 0.06
LDA K-fold 20%
Discriminant type: linear
Indexes: logarithmic transformed
Box Constraint = 1 Box Constraint = 0.7416
Kernel Scale = 1 Kernel Scale = 110.77
SVM K-fold 20%
Kernel Function: Gaussian Kernel Function: Gaussian
Indexes: Standardized Indexes: Standardized
Distance: Euclidean
KNN Number of Neighbors: 2 K-fold 20%
Indexes: Standardized
Distance: Euclidean
META-CLASSIFIER KNN Number of Neighbors: 2 K-fold 20%
Indexes: Standardized
Training function: Levenberg-Marquardt
hidden layer: 1
hidden layer sizes: 10
NN K-fold 20%
input layer sizes: 4
output layer sizes: 3
Indexes: Standardized

For linear discriminant analysis (LDA), the indexes of sEMG were logarithmically transformed
based on the fundamental assumption of normally distributed variables for this method. Indexes were
standardized for the algorithms where the assumption is not applicable [37] (see configuration details
in Table 2).
An ensemble technique was implemented to enhance the power of single classification techniques,
looking to combine single classifiers through different strategies. In this study, the stacking algorithm
was selected to combine three single classifiers or supervised methods (LDA, SVM, and KNN).
based on the fundamental assumption of normally distributed variables for this method. Indexes
were standardized for the algorithms where the assumption is not applicable [37] (see configuration
details in Table 2).
An ensemble technique was implemented to enhance the power of single classification
techniques, looking to combine single classifiers through different strategies. In this study, the
Appl. Sci. 2019, 9, 2007 9 of 18
stacking algorithm was selected to combine three single classifiers or supervised methods (LDA,
SVM, and KNN). The stacking algorithm has two levels (see Figure 4), the single classifiers (Ck) on
the first
The level where
stacking the indexes
algorithm has twoare the inputs
levels (Fn) and
(see Figure thesingle
4), the correct classes are
classifiers (Ckthe
) onoutputs
the first(Glevel
n); atwhere
this
level the single
the indexes areclassifiers
the inputs predict thethe
(Fn ) and classes (φk),
correct whichare
classes will
thebeoutputs
the inputs
(Gn );for
at the
thismeta-classifier
level the single
(supervised method)
classifiers predict theonclasses
the second
(ϕk ), level,
whichwhere
will bethe
theoutputs
inputs are
for also the correct classes
the meta-classifier [38]. Themethod)
(supervised meta-
classifier implemented
on the second in this
level, where thestudy
outputs wasare
a KNN model
also the and
correct the predicted
classes [38]. Theclasses (Φ) wereimplemented
meta-classifier compared
with thestudy
in this correct wasclasses.
a KNN model and the predicted classes (Φ) were compared with the correct classes.

Figure 4. Stacking algorithm representation. In the first level, the features of the sEMG signals enter
into the single classifiers (LDA, SVM or KNN); the results of these classifiers are the entry features of
the metaclassifier, in this case, a KNN classifier.

Once the hidden layer size was determined for the fourth implemented classification algorithm,
NN, one thousand neural networks were tested using a different randomized segmentation for training
and validation dataset (80% and 20% respectively). Also, the initial weights were randomized for
each one. The best NN was selected based on its relation to training and validation loss function.
The accuracy, specificity and sensitivity were calculated for the best algorithms according to
Subasi et al. [39]. In the case of the WOB level classification, the calculus was as follows:

Accuracy: number of samples correctly classified/number of total samples

Specificity: number correctly classified as normal WOB/number of total normal WOB.
Sensitivity (Group 2 only): number correctly classified as medium WOB/number of total medium WOB.
Sensitivity (Group 3 only): number correctly classified as elevated WOB/number of total elevated WOB.
Sensitivity (Group 2 and Group 3): number correctly classified as high WOB/number of total high WOB.

• Modeling to Estimate the WOB

From previous patient observations [15,17], our research group state the following hypothesis:
“high work of breathing implies high respiratory muscle activity, differentiating the diaphragm activity
from accessory muscles”. Equation (6) is proposed in this article as a model to quantify work of
breathing from the muscular activity of two accessory muscles and diaphragm.

α1 × MaxAmpDia + α2 × (MaxAmpInt + MaxAmpStrn )

WOBe = (6)
MaxAmpDia + MaxAmpInt + MaxAmpStrn

where MaxAmpDia, MaxAmpInt, and MaxAmpStrn are the maximum amplitudes of the surface
electromyographic signals of the diaphragm, intercostal and sternocleidomastoid muscles respectively.
α1 and α2 are coefficients that express the contribution of the main and accessory muscles respectively
to achieve the necessary work of an inspiratory cycle.
Appl. Sci. 2019, 9, 2007 10 of 18

Surface electromyography is a technique sensitive to size and type of tissue over the muscle,
which influences the level of attenuation of the signal [18]. The standardization of Equation (6) by the
sum of the maximum amplitudes was done to know the contribution of each muscle to the ventilatory
process, reducing the effect of different anatomy structures in the surface electromyography recording.
• Optimization of Model Parameters
An optimization algorithm based on sequential quadratic programming (SQP) was applied to
find α1 and α2 values in each WOB group (G). Equation (7) was the function to be minimized.

N
1 X
J= |WOBr − WOBe | (7)
N
i=1

where N is the number of samples in each group, WOBr is the real work of breathing represented by
Equation (3) and WOBe is the estimated work of breathing from muscular activity, using the proposed
Equation (6).
In the optimization algorithm, two nonlinear constraints were proposed to find parameters with
a physical meaning. The first constraint, Equation (8), looks for values of α1 and α2 that express the
contribution of the diaphragm and accessory muscles, the sum of α1 and α2 must be equal to the
proportion between the average work of breathing (WOBrG ) and the average maximum amplitude of
sEMG signals of all muscles (MaxAmpG ) for each group (G).

α1G + α2G = βG ; where G = 1, 2, 3

WOBrG (8)
βG =
MaxAmpG

The second constraint, Equation (9), ensures that the highest coefficient is equivalent to the highest
muscular activity. For example, if the maximum diaphragm amplitude is higher than the other two
maximum amplitudes (intercostal and sternocleidomastoid), α1 will be higher than α2 . In addition,
if any of the maximum amplitudes of accessory muscles is higher than the maximum diaphragm
amplitude, α2 will be higher than α1 .

MaxAmpDiaG > MaxAmpIntG & MaxAmpDiaG > MaxAmpStrnG =⇒ α1G > α2G
(9)
MaxAmpIntG > MaxAmpDiaG || MaxAmpStrnG > MaxAmpDiaG =⇒ α2G > α1G

The model implementation and the parameter optimization were performed for each data set
obtained from the Experiments 1 and 2. The validation technique was mainly K-fold 20%.

2.5. Statistical Analysis

For the normal distribution data, the mean and standard deviation are presented, and all other
data are characterized by the median and interquartile range. Normal distribution was tested with the
Lilliefors test. To compare two datasets, ANOVA and Kruskal-Wallis analyses were applied to the
normal and non-normal distribution respectively. To compare more than two data sets, a Tukey-Kramer
test was implemented. For all tests, statistical significance was assumed at p-value < 0.05.

3. Results
Seventy-six and eighty-three samples with work of breathing lower than 1.8 J/L (superior limit
according to literature) were obtained from E1 and E2, respectively. Therefore, one hundred and
fifty-nine samples were used in the training and testing of the algorithms. Figure 5 shows a scheme of
the best model proposed.
3. Results
Seventy-six and eighty-three samples with work of breathing lower than 1.8 J/L (superior limit
according to literature) were obtained from E1 and E2, respectively. Therefore, one hundred and fifty-
nineSci.
Appl. samples were used in the training and testing of the algorithms. Figure 5 shows a scheme11ofofthe
2019, 9, 2007 18
best model proposed.

Figure5.5.Scheme
Figure Schemeofof thethe
bestbest model
model foundfound to estimate
to estimate WOB. WOB. In the
In the first first
part, thepart,
samplesthe were
samples were
classified
classified
by a KNN by a KNN according
algorithm, algorithm,toaccording to the
the muscular muscular
pattern, withpattern, with of
an accuracy an91%.
accuracy
In theofsecond
91%. In the
part,
second
the part, were
samples the samples were
classified byclassified by a NN algorithm,
a NN algorithm, according according to WOB
to WOB level, level,
with with an accuracy
an accuracy of 89%.
In
ofthe last
89%. Inpart, the part,
the last WOBthe value
WOB was predicted
value with an accuracy
was predicted of 80.42%.
with an accuracy of 80.42%.

3.1.
3.1. Ventilatory
Ventilatory Pattern
Pattern Classification
Classification
The
Thefirst
firststep was
step classifying
was the the
classifying samples into ainto
samples respiratory pattern.
a respiratory Four out
pattern. of the
Four outfifteen
of theindexes
fifteen
from dimension reduction (stepwise regression and mutual information) were relevant
indexes from dimension reduction (stepwise regression and mutual information) were relevant in in differentiating
the samples between
differentiating the restrictive
the samples and obstructive
between the restrictivepatterns. The indexes patterns.
and obstructive were coherence betweenwere
The indexes the
diaphragm and the accessory muscles (C Dia-Int and C Dia-Strn ), the central frequency,
coherence between the diaphragm and the accessory muscles (CDia-Int and CDia-Strn), the central and ratio between
high and low
frequency, andfrequencies of thehigh
ratio between sternocleidomastoid
and low frequencies muscle (FCsternocleidomastoid
of the Strn and RHLStrn , respectively).
muscle (FCTable 3
Strn and
shows
RHLStrnthe median values
, respectively). and3interquartile
Table rangesvalues
shows the median of theseandfour indexes. Furthermore,
interquartile the four
ranges of these p-value that
indexes.
indicates significant differences between E1 and E2 is presented.
Furthermore, the p-value that indicates significant differences between E1 and E2 is presented.

Table 3. Restrictive and obstructive breathing pattern classification. The median values and interquartile
Table 3. Restrictive and obstructive breathing pattern classification. The median values and
ranges of each index are shown.
interquartile ranges of each index are shown.

IndexIndex Experiment
Experiment P-value
P-Value
Restrictive (E1) Obstructive (E2)
Restrictive (E1) Obstructive (E2)
CDia-Int 0.67 [0.64–0.73] 0.75 [0.72–0.80] <0.01
CDia-Int
CDia-Strn 0.67[0.64–0.73]
0.72 [0.70–0.76] 0.75[0.72–0.80]
0.65 [0.62–0.70] <0.01 <0.01
FC (Hz) 77.84 [68.24–86.71] 62.37 [58.46–67.88] <0.01
CDia-StrnStrn 0.72 [0.70–0.76] 0.65[0.62–0.70] <0.01
RHLStrn 169.76 [107.71–229.17] 92.07 [55.00–131.52] <0.01
FCStrn (Hz) 77.84[68.24–86.71] 62.37[58.46–67.88] <0.01
RHLStrnaccessory169.76[107.71–229.17]
The intercostal muscle supports the diaphragm 92.07[55.00–131.52] <0.01 more than
during obstructive maneuvers
the sternocleidomastoid, with the highest coherence between the diaphragm and intercostal, which
indicates synchronization of these muscles in the ventilatory process. In contrast, the sternocleidomastoid
supports the diaphragm during restrictive maneuvers, with the highest coherence between diaphragm
and sternocleidomastoid, and highest central frequency and ratio between high and low frequencies for
the sternocleidomastoid, which indicates that this muscle has a higher engagement in the ventilatory
process of restrictive maneuvers.
The classification capabilities of the machine learning algorithms to distinguish restrictive and
obstructive patterns were 86.16%, 90.57% and 90.85% for LDA, SVM, and KNN, respectively, with
KNN being the best algorithm; therefore, the latter was selected for the final model (see Figure 5).

3.2. WOB Level Classification

The second step was classifying the samples in one level of work of breathing. After applying the
stepwise regression and mutual information tests, the MaxAmpDia and MaxAmpInt were relevant
in identifying the level of WOB for E1, the MaxAmpDia and MaxAmpStrn for E2, and CDia-Int was
Appl. Sci. 2019, 9, 2007 12 of 18

relevant for both experiments. Table 4 shows the median values of each of the indexes, with the p-value
for each pair of groups.

Table 4. Median and standard deviation of real work of breathing values for each group. G1: Low
WOB; G2: medium-high WOB; G3: elevated WOB. Furthermore, the parameters of each model and the
selected index values for classification model are shown.

Restrictive (E1 ) P-Value

Parameter/Index
G1 G2 G3 G1 - G2 G1 - G3 G2 - G3
WOBr (J/L) 0.61 ± 0.13 1.02 ± 0.14 1.52 ± 0.13 <0.05 <0.05 <0.05
0.024 0.032 0.054
MaxAmpDia (mV) * <0.05 <0.05
[0.022–0.030] [0.022–0.049] [0.040–0.068]
0.020 0.032 0.034
MaxAmpInt (mV) <0.05 <0.05 *
[0.016–0.027] [0.020–0.043] [0.020–0.049]
0.018 0.021 0.021
MaxAmpStrn (mV) * * *
[0.010–0.024] [0.014–0.036] [0.017–0.041]
CDia-Int 0.66 [0.62–0.68] 0.67 [0.65–0.75] 0.71 [0.67–0.83] * <0.05 *
α1 0.55 ± 10-5 1.07 ± 0.04 1.55 ± 10–5 - - -
α2 0.45 ± 10-5 0.93 ± 0.04 1.45 ± 10–5 - - -
Obstructive (E2 )
WOBr (J/L) 0.63 ± 0.12 1.03 ± 0.14 1.51 ± 0.12 <0.05 <0.05 <0.05
0.028 0.031 0.050
MaxAmpDia (mV) * <0.05 <0.05
[0.021–0.038] [0.023–0.047] [0.038–0.083]
0.016 0.019 0.023
MaxAmpInt (mV) * * *
[0.010–0.029] [0.013–0.036] [0.015–0.045]
0.016 0.017 0.030
MaxAmpStrn (mV) * <0.05 <0.05
[0.010–0.023] [0.013–0.022] [0.019–0.048]
CDia-Int 0.79 [0.75–0.84] 0.75 [0.72–0.78] 0.73 [0.69–0.75] <0.05 <0.05 *
α1 0.55 ± 10–5 1.05 ± 10–5 1.55 ± 10–5 - - -
α2 0.45 ± 10–5 0.95 ± 10–5 1.45 ± 10–5 - - -
* There are no statistically significant differences between groups.

It should be noted that MaxAmpDia, MaxAmpInt and MaxAmpStrn median values increase
when the work of breathing increases. MaxAmpDia separates G1 and G2 from G3 in both experiments.
MaxAmpInt separates G1 from G2 and G3 in E1. MaxAmpStrn separates G1 and G2 from G3 in E2.
In addition, CDia-Int separates G1 from G3 in E1 and G1 form G2 and G3 in E2.
It is noted that the features are not related in a linear manner, since the accuracy, specificity, and
sensitivity were lower for the LDA model. For the case of the NN model, it is noted that the volunteers
with elevated WOB (group 3) were different from the other groups in the restrictive database (E1), with
a sensitivity of 100%, in contrast to 80% for the obstructive database (E2), such as group 1 or low WOB.
However, in medium WOB, a sensitivity of 100% was achieved in the obstructive database, in contrast
to 91.67% in the restrictive database. The above could indicate that during restrictive maneuvers, it was
possible to achieve a greater number of samples for level 3 of WOB.
The selection of the neural network was according to the training and validation loss function
relation, which is shown in Figure 6 for the restrictive and obstructive dataset. It is possible to observe
that depending on the dataset obtained for validation (20%) and training (80%), the model could be
overfitted, as is the case with the points located in the fourth quadrant in the figure, where it is possible
to observe a training loss function of zero but a validation loss function of almost one, while in the
Appl. Sci. 2019, 9, 2007 13 of 18

third
Appl.quadrant,
Sci. 2019, 9,there
x is the NN model, where the loss function of both datasets can be considered14low.
of 19
Therefore, this NN was selected.

Figure 6. Validation and training loss function relation for the one thousand NN of experiment 1
Figure
(E1: 6. Validation
restrictive dataset) and
and training loss2function
experiment relation for
(E2: obstructive the one thousand NN of experiment 1 (E1:
dataset).
restrictive dataset) and experiment 2 (E2: obstructive dataset).
The neural network selected for the final model (see Figure 4 and Table 5) obtained an accuracy
of 89%,The neural network
a specificity selected
of 83.67%, for the final
a sensitivity model(group
of 96.36% (see Figure
2 and4 group
and Table
3), a5)sensitivity
obtained an accuracy
of 90.12%
of 89%,
(group 2), aand
specificity of 83.67%,
a sensitivity a sensitivity
of 89.66% (group 3)of(see
96.36%
Table(group
6). 2 and group 3), a sensitivity of 90.12%
(group 2), and a sensitivity of 89.66% (group 3) (see Table 6).
Table 5. Classification accuracies, specificities, and sensitivities obtained for each of the machine
learning algorithms for both
Table experiments.
6. Confusion matrixInfor
thethe
case of neural
final networks,
classification theof
model accuracy is determined
Figure 4.
using 20% of the dataset for validation and the between-brackets accuracy is determined using 100% of
the dataset. Real Values
Group 1 Group 2 Group 3 Total
Sensitivity
0 (%) Sensitivity (%)
AccuracyGroup
(%) 1 Specificity
41 (%) 4 45
Algorithm (Group 2) (Group 3)
Group 2 5 73 3 81
Predicted Values
E1 E2 E1 E2 E1 E E1 E2
Group 3 3 4 26 2 33
56.58 55.42
LDA
Total 57.6949 17.39 81 63.89 75.56
29 35.71
159 53.33
SVM 47.37 56.63 0 0 100 100 0 0
KNN 39.47 54.22 34.62 52.17 41.67 64.44 42.86 26.67
3.3. WOB estimation
STACKING
35.53 46.99 53.85 17.39 25.00 77.78 21.43 0
(LDA-SVM-KNN)
The last step was estimating the work of breathing value. Figure 7 shows the relation between
STACKING
40.79 48.19 42.31 17.39 44.44 75.56 28.57 13.33
the (LDA-KNN-KNN)
real work of breathing (WOBr) and the estimated work of breathing (WOBe) with coefficients α1
STACKING
and α2 (see Table 4) of the groups
47.37 found
37.35with42.31
the NN model.
65.22 The real WOB
66.67 24.44and the
7.14estimated
33.33 WOB
(SVM-KNN-KNN)
have aSTACKING
Pearson correlation coefficient higher than 80%.
50.00 43.37 34.62 34.78 66.67 62.22 35.71 0
Figure 6 shows the dispersion of the data, where it is possible to observe that this is smaller for
(LDA-SVM-KNN-KNN)
G3, offering 80.00 82.35
NNhigher precision in the prediction92.31
for cases82.61
that have a risky Work
91.67 100 of Breathing
100 (near
80 the
[93.42] [91.57]
superior level). It is also possible to observe the samples that were classified with an incorrect WOB
level, being points placed near the next or before the neighboring level, exactly in the overlapping
Table 6. Confusion matrix for the final classification model of Figure 4.
region.
Real Values
Group 1 Group 2 Group 3 Total
Group 1 41 4 0 45
Predicted Values Group 2 5 73 3 81
Group 3 3 4 26 33
Total 49 81 29 159
 Appl. Sci. 2019, 9, 2007 14 of 18

3.3. WOB Estimation

The last step was estimating the work of breathing value. Figure 7 shows the relation between the
real work of breathing (WOBr) and the estimated work of breathing (WOBe) with coefficients α1 and
α2 (see Table 4) of the groups found with the NN model. The real WOB and the estimated WOB have
Appl. Sci. 2019, correlation
a Pearson 9, x coefficient higher than 80%. 15 of 19

Figure 7. Real work of breathing (WOBr) compared with the estimated work of breathing (WOBe) in
Figure 7. Real work of breathing (WOBr) compared with the estimated work of breathing (WOBe) in
each one of the groups. The linear regression analysis is presented.
each one of the groups. The linear regression analysis is presented.
Figure 6 shows the dispersion of the data, where it is possible to observe that this is smaller for G3,
4.offering
Discussion higher precision in the prediction for cases that have a risky Work of Breathing (near the
superior level).known
It is widely It is alsothat
possible
the useto of
observe the samples
mechanical that were
ventilation classified with
is a treatment an incorrect
that causes WOB level,
complications
being points placed near the next or before the neighboring level, exactly
when used for prolonged periods, increasing patient mortality [40]. The work of breathing (WOB) in the overlapping region.
is
an accepted index to know the state of the ventilatory mechanics and respiratory muscles,
4. Discussion
nevertheless, the monitoring of WOB during spontaneous ventilation is not easy to perform [10].
This
It ispilot study
widely knownseeksthat
to estimate
the use ofthe WOB in both
mechanical restrictive
ventilation and obstructive
is a treatment situations,
that causes taking
complications
advantage
when used offor
its prolonged
relationship with the
periods, respiratory
increasing muscle
patient activity.
mortality Indexes
[40]. The workfromofrespiratory muscleis
breathing (WOB)
sEMG signalsindex
an accepted weretoused know to the
classify each
state of the one of the muscular
ventilatory mechanicspatterns (restrictive
and respiratory and obstructive
muscles, nevertheless,
maneuvers)
the monitoringand each
of WOB one during
of the levels of WOB.ventilation is not easy to perform [10].
spontaneous
Three
This indexes
pilot study from the to
seeks frequency
estimatedomain,
the WOBFC, RHL restrictive
in both of the sternocleidomastoid and coherence
and obstructive situations, taking
(C) between the
advantage diaphragm
of its relationship andwith
the accessory muscles
the respiratory discriminated
muscle the patterns
activity. Indexes frombetween restrictive
respiratory muscle
and
sEMGobstructive (see used
signals were Tableto3).classify
The FC andone
each RHL of the
of the sternocleidomastoid
muscular patterns (restrictivewere and
higher in the
obstructive
restrictive
maneuvers) pattern
and each thanonein thelevels
of the obstructive
of WOB.one, which indicates that the activity of the
sternocleidomastoid
Three indexes from is located in the high
the frequency frequency
domain, FC, RHL band, therefore,
of the it has a high and
sternocleidomastoid activity [17,21].
coherence (C)
The coherence
between between the
the diaphragm anddiaphragm and any
the accessory accessory
muscles muscles makes
discriminated it possible
the patterns to infer
between the level
restrictive and
ofobstructive
engagement of Table
(see the involved
3). The accessory
FC and RHL muscle,
of thethe sternocleidomastoid
sternocleidomastoid werein the restrictive
higher case. On
in the restrictive
the contrary,
pattern thanininthetheobstructive
obstructive case, thewhich
one, muscle that supports
indicates theactivity
that the ventilatory
of the process is the intercostalis
sternocleidomastoid
muscle,
locatedasinobserved
the high in the coherence
frequency band, value between
therefore, it hasthe diaphragm
a high activityand intercostal,
[17,21]. which is between
The coherence higher
inthe
thisdiaphragm
pattern than and inany
the restrictive one.
accessory muscles makes it possible to infer the level of engagement of the
Concerning WOB classification, two of the calculated indexes presented a statistically significant
contribution to discriminate the samples: the amplitude and the coherence (see Table 4). The median
of the maximum amplitude of sEMG signals of the diaphragm increased with increments in work of
breathing in both experiments, confirming that this is the principal muscle in the ventilation process.
A differentiated pattern was found related to the accessory muscle activity, the more significant
Appl. Sci. 2019, 9, 2007 15 of 18

involved accessory muscle, the sternocleidomastoid in the restrictive case. On the contrary, in the
obstructive case, the muscle that supports the ventilatory process is the intercostal muscle, as observed
in the coherence value between the diaphragm and intercostal, which is higher in this pattern than in
the restrictive one.
Concerning WOB classification, two of the calculated indexes presented a statistically significant
contribution to discriminate the samples: the amplitude and the coherence (see Table 4). The median
of the maximum amplitude of sEMG signals of the diaphragm increased with increments in work of
breathing in both experiments, confirming that this is the principal muscle in the ventilation process.
A differentiated pattern was found related to the accessory muscle activity, the more significant
accessory muscle for the restrictive maneuver (E1) was the intercostal muscle (maximum amplitude
of sEMG) and in the case of the obstructive maneuver (E2), the significant changes were due to the
sternocleidomastoid muscle. For the same reason, the coherence between the diaphragm and the
intercostal increased in E1 and decreased in E2.
The findings show that during the ventilatory process of a subject with a restrictive pattern,
the muscles with permanent activation are the diaphragm as the main muscle and the sternocleidomastoid
as the accessory muscle engaged. However, as the work of breathing increases, there is also an increase in
the intercostal muscle activity. Conversely, in a subject with an obstructive pattern, where the muscles are
permanently activated are the diaphragm and the intercostal, but when the work of breathing increases,
the muscle that showed increments was the sternocleidomastoid.
To identify the muscular pattern between restrictive or obstructive, the best technique was the
KNN, obtaining an accuracy of 91%. Meanwhile, to identify the level of work of breathing, LDA,
SVM and KNN showed accuracies lower than 60%. Combining the properties of some of these
classifiers, the accuracy reached is lower than 80%. However, the neural network was the algorithm
with the greatest accuracy, achieving a mean of 89% for the all dataset (see Table 5). It should be noted
that the accuracy of the classification of WOB level depends on the real WOB (WOBr), and this was
calculated from values of compliance, resistance, and muscular pressure, obtained with an occlusion
maneuver and optimization algorithms, thus WOBr may have an estimation error [28,29]. In future
work, the classifiers should be validated to increase the accuracy, recording the real WOB with invasive
methods in patients with invasive mechanical ventilation.
The proposed algorithm might be useful in critical care medicine, due to its ability to identify the
level of work of breathing in a non-invasive manner. With a specificity of 83.67% and a sensitivity
(Group 2 and Group 3) of 96.36% (see Figure 5 and Table 6), it is possible to predict who could have
a successful weaning, which in terms of low WOB is in line with the study of Kirton et al. [8], where
96% of patients with a work of breathing lower than 0.8 J/L had a successful weaning process.
The model proposed in this study relates work of breathing with the maximum amplitudes of
sEMG signals of the diaphragm and two accessory muscles, the intercostal and sternocleidomastoid.
However, due to the normalization process, it was necessary to find different coefficients according to
the level of WOB. The coefficients indicate the contribution to the ventilatory process of each of the
muscles. In all cases, the muscle with the most contribution was the diaphragm with 55% for the first
group and 53% for the other groups; additionally, the remaining contribution was distributed between
the other respiratory muscles.
The coefficients (α1 y α2) in both experiments were similar, because the minimum error for all the
dataset was in the median value of the group. In that case, it was possible to consider only one model
for both cases, restrictive and obstructive patterns.
The proposed approach estimates the value of work of breathing with an accuracy of 80.43%.
Other authors have shown non-invasive techniques to estimate the work of breathing. Vicario et al. [41]
presented a model that estimates the work of breathing of mechanically ventilated patients, using
ventilatory signals such as airway pressure, volume, and airflow. Their model offers favorable results
in cases in which the patients have normal and controlled respiratory efforts, but it has limitations
for uncontrolled efforts. Banner et al. [11,20] presented an Artificial Neuronal Network to estimate
Appl. Sci. 2019, 9, 2007 16 of 18

the power of breathing or work of breathing per minute from parameters like spontaneous minute
ventilation, inspiratory trigger pressure, inspiratory flow rise time, intrinsic PEEP and respiratory
muscle pressure, where the last two parameters depend on the respiratory system compliance and
airway resistance calculated from controlled ventilations. The model presented in this study can
predict the WOB at different levels during total spontaneous ventilation and has the advantage that
it does not need ventilatory signals, i.e., it is possible to know the work of breathing in situations
that do not require mechanical ventilation or devices to record ventilatory signals that can affect the
respiratory mechanics.
This pilot study suggests that the work of breathing can be estimated by measuring the
activity of respiratory muscles using surface electromyography. In accordance with other authors
who inferred it in previous works, like Salazar M, et al. [15], who found a promising index to
assess the participation of respiratory muscle on spontaneous breathing test in poisoned patients,
Schmidt M, et al. [14], who showed that measurements of accessory inspiratory muscles by sEMG can
contribute to a reliable non-invasive indicator of respiratory effort in mechanically ventilated patients,
and Hernandez et al. [21], who showed that changes in indexes from sEMG of the diaphragm and
sternocleidomastoid are related to changes in the level of PEEP.
Furthermore, the work of breathing was estimated in different ventilatory patterns in spontaneous
ventilation, which corroborates the close relationship, not only of the diaphragm, but of the accessory
muscles with the ventilatory process. This offers a promising technique in different areas, mainly
in critical care medicine, because the information obtained with this model will not only indicate
the state of the patient during mechanical ventilation, but it will also help medical staff decide
whether or not a patient needs treatment with mechanical ventilation, determine the right extubation
moment and optimize the mechanical ventilator settings. However, for future work, a database of
patients in ICU would be recorded to validate the model and its possible use in the daily routine of
healthcare professionals.

Author Contributions: Design of the work, I.C.M. and A.M.H.; the data acquisition, I.C.M.; analysis, I.C.M.,
A.M.H., M.Á.M.; writing—original draft preparation, I.C.M.; writing—review and editing, A.M.H., M.Á.M., and
I.C.M.; final approval of the version to be published, A.M.H. and M.Á.M.
Funding: Research was partially supported by PGR-CODI-2015-7851 “Análisis de la actividad muscular
respiratoria en ventilación mecánica no invasiva y su relación con la configuración del ventilador”. Grant
by University of Antioquia (Medellín, Antioquia, Colombia). By FP44842-450-2018 “Desarrollo de aplicaciones
móviles para el aprendizaje del sistema respiratorio: Conceptos básicos, diagnóstico de enfermedades, terapia
y rehabilitación”. Grant by Sistema General de Regalías fondos de CTeI de la Gobernación de Antioquia,
administrados a través del patrimonio autónomo fondo nacional de financiamiento para la ciencia, la tecnología y
la innovación francisco José de Caldas. And by the Spanish Ministry of Economy and Competitiveness- Spain
project DPI2017-83989-R.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Grinnan, D.C.; Truwit, J.D. Clinical review: Respiratory mechanics in spontaneous and assisted ventilation.
Crit. Care 2005, 9, 472–484. [CrossRef]
2. Brochard, L.; Martin, G.S.; Blanch, L.; Pelosi, P.; Belda, F.J.; Jubran, A.; Gattinoni, L.; Mancebo, J.; Ranieri, V.M.;
Richard, J.C.M.; et al. Clinical review: Respiratory monitoring in the ICU—A consensus of 16. Crit. Care
2012, 16, 219. [CrossRef] [PubMed]
3. Tulaimat, A.; Patel, A.; Wisniewski, M.; Gueret, R. The validity and reliability of the clinical assessment of
increased work of breathing in acutely ill patients. J. Crit. Care 2016, 34, 111–115. [CrossRef] [PubMed]
4. Aboussouan, L. Respiratory Failure and the Need for Ventilatory Support. In Egan’s Fundamentals of
Respiratory Care, 11th ed.; Kacmarek, R., Stoller, J., Heuer, A., Eds.; Elsevier: St. Louis, MO, USA, 2017;
pp. 971–986.
5. Piraino, T. Monitoring the Patient in the Intensive Care Unit. In Egan’s Fundamentals of Respiratory Care,
11th ed.; Kacmarek, R., Stoller, J., Heuer, A., Eds.; Elsevier: St. Louis, MO, USA, 2017; pp. 1154–1189.
Appl. Sci. 2019, 9, 2007 17 of 18

6. Sharp, J.T.; Henry, J.P.; Sweany, S.K.; Meadows, W.R.; Pietras, R.J. The Total Work of Breathing in Normal and
Obese Men. J. Clin. Investig. 1964, 43, 728–739. [CrossRef]
7. Diehl, J.L.; Mercat, A.; Guérot, E.; Aïssa, F.; Teboul, J.L.; Richard, C.; Labrousse, J. Helium/oxygen mixture
reduces the work of breathing at the end of the weaning process in patients with severe chronic obstructive
pulmonary disease. Crit. Care Med. 2003, 31, 1415–1420. [CrossRef]
8. Kirton, O.C.; DeHaven, C.B.; Morgan, J.P.; Windsor, J.; Civetta, J.M. Elevated imposed work of breathing
masquerading as ventilator weaning intolerance. Chest 1995, 108, 1021–1025. [CrossRef] [PubMed]
9. Umbrello, M.; Formenti, P.; Longhi, D.; Galimberti, A.; Piva, I.; Pezzi, A.; Mistraletti, G.; Marini, J.J.;
Iapichino, G. Diaphragm ultrasound as indicator of respiratory effort in critically ill patients undergoing
assisted mechanical ventilation: A pilot clinical study. Crit. Care 2015, 19, 161. [CrossRef]
10. Khirani, S.; Polese, G.; Aliverti, A.; Appendini, L.; Nucci, G.; Pedotti, A.; Colledan, M.; Lucianetti, A.;
Baconnier, P.; Rossi, A. On-line monitoring of lung mechanics during spontaneous breathing: A physiological
study. Respir. Med. 2010, 104, 463–471. [CrossRef] [PubMed]
11. Banner, M.J.; Tams, C.G.; Euliano, N.R.; Stephan, P.J.; Leavitt, T.J.; Martin, A.D.; Al-Rawas, N.; Gabrielli, A.
Real time noninvasive estimation of work of breathing using facemask leak-corrected tidal volume during
noninvasive pressure support: Validation study. J. Clin. Monit. Comput. 2016, 30, 285–294. [CrossRef]
[PubMed]
12. Schmidt, M.; Chiti, L.; Hug, F.; Demoule, A.; Similowski, T. Surface electromyogram of inspiratory muscles:
A possible routine monitoring tool in the intensive care unit. Br. J. Anaesth. 2011, 106, 913–914. [CrossRef]
13. Schmidt, M.; Banzett, R.B.; Raux, M.; Morélot-Panzini, C.; Dangers, L.; Similowski, T.; Demoule, A. Unrecognized
suffering in the ICU: Addressing dyspnea in mechanically ventilated patients. Intensive Care Med. 2014, 40,
1–10. [CrossRef]
14. Schmidt, M.; Kindler, F.; Gottfried, S.B.; Raux, M.; Hug, F.; Similowski, T.; Demoule, A. Dyspnea and surface
inspiratory electromyograms in mechanically ventilated patients. Intensive Care Med. 2013, 39, 1368–1376.
[CrossRef]
15. Salazar, M.B.; Hernandez, A.M.; Mananas, M.A. Assessment of mechanically ventilated patients intoxicated
with organophosphates by a novel surface electromyographic index. J. Crit. Care 2017, 41, 260–267. [CrossRef]
16. Salazar, M.B.; Hernandez, A.M.; Mananas, M.A.; Zuluaga, A.F. Potential clinical application of surface
electromyography as indicator of neuromuscular recovery during weaning tests after organophosphate
poisoning. Rev. Bras. Ter. Intensiv. 2017, 29, 253–258.
17. Munoz, I.C.; Hernandez, A.M.; Alonso, J.F.; Mananas, M.Á.; Atehortúa, L.H. Assessment of weaning indexes
based on diaphragm activity in mechanically ventilated subjects after cardiovascular surgery. A pilot study.
Rev. Bras. Ter. Intensiv. 2017, 29, 213–221.
18. Chowdhury, R.; Reaz, M.; Ali, M.; Bakar, A.; Chellappan, K.; Chang, T. Surface Electromyography Signal
Processing and Classification Techniques. Sensors 2013, 13, 12431–12466. [CrossRef]
19. Perchiazzi, G.; Rylander, C.; Pellegrini, M.; Larsson, A. Monitoring of total positive end-expiratory pressure
during mechanical ventilation by artificial neural networks. J. Clin. Monit. Comput. 2017, 31, 551–559.
[CrossRef]
20. Banner, M.J.; Euliano, N.R.; Brennan, V.; Peters, C.; Layon, A.J.; Gabrielli, A. Power of breathing determined
noninvasively with use of an artificial neural network in patients with respiratory failure. Crit. Care Med.
2006, 34, 1052–1059. [CrossRef]
21. Hernández, A.M.; Salazar, M.B.; Muñoz, I.C. Efecto del incremento del PEEP en la actividad muscular
respiratoria en sujetos sanos bajo ventilación espontánea. IATREIA 2016, 29, 280–291.
22. Muñoz, I.C.; Urrego, D.A.; Vallejo, A.F.; Hernández, A.M. Device for simulation of restrictive pathologies
in healthy subjects with non-invasive mechanical ventilation. Rev. Fac. Ing. Univ. Antioq. 2018, 1, 19–26.
[CrossRef]
23. Stănescu, D.C.; Nemery, B.; Veriter, C.; Maréchal, C. Pattern of breathing and ventilatory response to CO2 in
subjects practicing hatha-yoga. J. Appl. Physiol. 1981, 51, 1625–1629. [CrossRef]
24. Chakraborti, S.; Gibbons, J.D. Nonparametric Statistical Inference, 4th ed.; Marcel Dekker, Inc.: New York, NY,
USA, 2003.
25. Hernández, A.M. Sistema de Control. Respiratorio Ante Estímulos y Patologías. Análisis, Modelado y Simulación,
1st ed.; Publicia: Saarbrücken, Germany, 2007.
Appl. Sci. 2019, 9, 2007 18 of 18

26. Alonso, J.F.; Mañanas, M.A.; Rojas, M.; Bruce, E.N. Coordination of respiratory muscles assessed by means
of nonlinear forecasting of demodulated myographic signals. J. Electromyogr. Kinesiol. 2011, 21, 1064–1073.
[CrossRef]
27. Chien, M.Y.; Wu, Y.T.; Chang, Y.J. Assessment of diaphragm and external intercostals fatigue from surface
EMG using cervical magnetic stimulation. Sensors 2008, 8, 2174–2187. [CrossRef] [PubMed]
28. Muñoz, I.C.; Hernández, A.M. Noninvasive approach to estimate ventilatory mechanics in spontaneous
breathing with different PEEP and pressure support values: Validation with mechanical simulation.
In Proceedings of the VII Latin American Congress on Biomedical Engineering CLAIB 2016, Bucaramanga,
Santander, Colombia, 26–28 October 2016; pp. 241–244.
29. Munoz, I.C.; Hernández, A.M. Cambios en la mecánica ventilatoria debidos a variaciones de la PEEP y la
presión soporte: Estudio en sujetos sanos bajo ventilación mecánica no invasiva. Rev. Fac. Med. 2017, 65,
321–328. [CrossRef]
30. Becher, T.; Schädler, D.; Rostalski, P.; Zick, G. Determination of respiratory system compliance during
pressure support ventilation by small variations of pressure support. J. Clin. Monit. Comput. 2018, 32,
741–751. [CrossRef]
31. De Luca, C.J.; Donald, G.L.; Kuznetsov, M.; Roy, S.H. Filtering the surface EMG signal: Movement artifact
and baseline noise contamination. J. Biomech. 2010, 43, 1573–1579. [CrossRef]
32. Merletti, R.; Parker, P.A. Electromyography: Physiology, Engineering, and Non-Invasive Applications; John Wiley
& Sons, Inc.: Hoboken, NJ, USA, 2004.
33. Haykin, S.S. Adaptive Filter Theory, 5th ed.; Pearson Education: Upper Saddle River, NJ, USA, 2014.
34. Camacho, A.; Hernandez, A.M.; Londono, Z.; Serna, L.Y.; Mananas, M.A. A synchronization system for
the analysis of biomedical signals recorded with different devices from mechanically ventilated patients.
In Proceedings of the 2012 Annual International Conference of the IEEE Engineering in Medicine and Biology
Society (EMBC), San Diego, CA, USA, 28 August–1 September 2012; pp. 1944–1947.
35. Draper, N.R.; Smith, H. Selecting the “Best” Regression Equation. In Applied Regression Analysis, 3rd ed.;
Draper, N.R., Smith, H., Eds.; John Wiley & Sons, Inc.: New York, NY, USA, 1998; pp. 335–336.
36. Tourassi, G.D.; Frederick, E.D.; Markey, M.K.; Floyd, C.E. Application of the mutual information criterion for
feature selection in computer-aided diagnosis. Med. Phys. 2001, 28, 2394–2402. [CrossRef]
37. Hwang, W. Data Mining in Ergonomics. In International Encyclopedia of Ergonomics and Human Factors, 2nd ed.;
Karwowski, W., Ed.; Taylor & Francis Group: Louisville, KY, USA, 2006; pp. 3077–3081.
38. Polikar, R. Ensemble based systems in decision making. IEEE Circuits Syst. Mag. 2006, 6, 21–45. [CrossRef]
39. Subasi, A.; Yilmaz, M.; Ozcalik, H.R. Classification of EMG signals using wavelet neural network. J. Neurosci.
Methods 2006, 156, 360–367. [CrossRef] [PubMed]
40. Guler, H.; Kilic, U. The development of a novel knowledge-based weaning algorithm using pulmonary
parameters: A simulation study. Med. Biol. Eng. Comput. 2017, 56, 373–384. [CrossRef] [PubMed]
41. Vicario, F.; Albanese, A.; Wang, D.; Karamolegkos, N.; Chbat, N.W. Constrained optimization for noninvasive
estimation of work of breathing. In Proceedings of the 37th Annual International Conference of the IEEE
Engineering in Medicine and Biology Society (EMBC), Milano, Italy, 25–29 August 2015; pp. 5327–5330.

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).