Nothing Special   »   [go: up one dir, main page]
Next Article in Journal
Serum Profiling of Proinflammatory Mediators in Inflammatory Bowel Disease: Indication for Use in Differential Diagnosis
Previous Article in Journal
Research on Range of Appropriate Spatial Scale of Underground Commercial Street Based on Psychological Perception Evaluation
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Applied Sciences
Volume 14
Issue 13
10.3390/app14135436
applsci-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

How to Optimize the Experimental Protocol for Surface EMG Signal Measurements Using the InterCriteria Decision-Making Approach

by
Maria Angelova
*,†,
Silvija Angelova
† and
Rositsa Raikova
Institute of Biophysics and Biomedical Engineering, Bulgarian Academy of Sciences, 105 Acad. G. Bonchev Str., 1113 Sofia, Bulgaria
*
Author to whom correspondence should be addressed.
†
These authors contributed equally to this work.
Appl. Sci. 2024, 14(13), 5436; https://doi.org/10.3390/app14135436
Submission received: 30 May 2024 / Revised: 19 June 2024 / Accepted: 20 June 2024 / Published: 22 June 2024
Browse Figures
Figure 1
<p>(<b>a</b>) Starting position; (<b>b</b>) position of maximal upper flexion with a wristband of 0.5 kg.</p> "> Figure 2
<p>A subject and raw EMG signals during TASK 5 performance of the FEP.</p> ">
Versions Notes

Abstract

:
The InterCriteria decision-making approach, known as InterCriteria analysis (ICrA), was applied here to optimize the experimental protocol when the surface electromyography (EMG) signals of upper arm human muscles are recorded. Ten healthy subjects performed cycling movements in the sagittal plane with and without added weight for ten, six, two, and one second, respectively, for each active phase. The EMG signals from six muscles or parts of muscles, namely m. deltoideus pars clavicularis and pars spinata, m. brachialis, m. anconeus, m. biceps brachii, and m. triceps brachii caput longum, were recorded. ICrA was used on the obtained data to find correlations between the sixteen different phases, eight for elbow flexion and eight for elbow extension. Based on the obtained results, we proposed an optimized experimental protocol (OEP) that omits slower and more difficult tasks while saving crucial data. The optimized protocol consists of seven, instead of ten, tasks and takes three minutes less than the time taken for the full experimental protocol (FEP). The lower number of movements in the OEP could prevent physical and psychical fatigue, discomfort, or even pain in the investigated subjects. In addition, the time to train subjects, as well as the time to process the surface EMG data, can be significantly reduced.

1. Introduction

InterCriteria analysis (ICrA) was elaborated by Atanassov et al. [1] in 2014. Intended as a solution to a specific production problem, it provoked unabated interest. ICrA establishes the correlation dependencies between chosen criteria and thus allows more time-consuming or more expensive measuring criteria to be removed from the evaluation process. As a result, the production costs are reduced. ICrA has numerous applications in different fields. Here, we will mention those related to biomedicine and quality of life due to the specific area of the current investigation.
Several scientific collectives have conducted studies on health indicators for healthy people and patients with specific diseases, such as Behterev’s disease, cardiac conditions, diabetes, ear inflammation, respiratory diseases, gastritis, and asthma. In Vankova et al. [2], ICrA was applied to establish the correlations between health-related indicators in a questionnaire survey among 1050 respondents. This research aimed to boost people’s health control and thus improve it. Zaharieva et al. [3,4,5] also use health indicators to discover ICrA dependencies in an experimental dataset of patients with Behterev’s disease. In addition, the decision-making approach helps to determine the appropriate treatment courses such as medicine, physiotherapeutic, and kinesitherapeutic programs. In the other four articles, ICrA was used as follows: to estimate the correlations between body composition indices associated with overweight in healthy students [6]; to support the process for criteria pre-selection for a large set of 88 ECG criteria [7] concerning patients with arrhythmia; to compare physiological rhythm curves [8]; and to search for relationships between the level of air pollution and the number of people with a variety of health problems seeking medical help [9].
ICrA was proven to be a promising tool for data analysis in several investigations on cancer diseases. Todinova et al. [10] apply this approach, along with correlation analysis, to define the interrelations between calorimetric parameters and statistical ones, acquired from blood plasma proteome thermograms of patients suffering from colorectal cancer. The authors show that both approaches complement each other and can help with deep analysis of calorimetric datasets. In Krumova et al. [11], the authors test the same decision-making method along, with two kinds of correlation analysis (Pearson’s and Spearman’s), for a large dataset. They use calorimetric and biochemical parameters, obtained from the serum proteome of multiple myeloma patients. Vassilev et al. [12] use real data from prostate carcinoma patients to verify two new InterCriteria analysis modifications.
The classical decision-making method supports four papers [13,14,15,16] investigating blood collection at the Centers for Transfusion Haematology in Bulgaria. Valuable results were obtained for correlations between center activities, the age and sex dependencies of the donors, transfusion hematology processes over the years, and also different blood type distributions across the population in the country.
The ICrA tool was applied in a kinesiological study for the first time in [17], focusing on a novel strategy for finding muscle interactions. The authors examined the repeatability and similarity of muscle synergies in healthy subjects using variable elbow motion velocities and applied or non-applied additional weight on the wrist.
Biomedical research such as that mentioned above is associated with certain obstacles. The study of human biomechanics represents a serious challenge in understanding and reproducing movement principles. Motor task protocol construction for functional assessment or movement pattern learning is not an easy initiative. Its implementation is sometimes complicated. An in-depth knowledge of normal human kinetics can aid in understanding the mechanisms of abnormal conditions as a result of disease and injury. The study of kinetic parameters requires conditions that are maximally identical for all participants. The correct performance of the motor tasks is of particular importance. For this purpose, the movements are required to be precisely explained.
The protocol used here is verbally explained, then demonstrated, and then performed independently by the subject. The person carries out as many reruns as is necessary to master it. Since the movements have a certain duration, the subject must make several repetitions, acquire a sense of time, and become accurate in their performance. Only then is the electromyography (EMG) activity recorded. This act of learning sometimes takes quite a lot of time. An excessively long experimental set can lead to slight mental fatigue and a loss of concentration. On the other hand, prolonged muscle work leads to the accumulation of muscle fatigue. To avoid fatigue, more frequent breaks are taken. All of this prolongs the experimental protocol and affects not only the protocol duration but also the quality of the received signals. A longer contact duration for the electrodes and skin stretching due to frequent movements create conditions for skin-electrode adhesion loosening. “Soft tissue artifacts” appear and require more serious filtration. Movement analysis is critical for the study of biomechanical components.
Both healthy people and patients with specific diseases participate in biomedical studies. A certain movement or a long-held position may cause discomfort or provoke pain and even numbness sometimes. Research can lead to changes in thought processes and emotions (confusion, feelings of stress, guilt, embarrassment, etc.) [18]. Therefore, it is appropriate to perform only the really necessary and informative movements. The investigators themselves may also prefer a shorter research protocol if the level of scientific output is guaranteed and the needed database is preserved. Just like volunteers, investigators actively participate in the whole process. While participants are involved in only one protocol, the researchers observe the process and gather data for at least ten to several hundred subjects. Sometimes, data analysis is a quite time-consuming process. Understanding the key position of time, Hanney et al. [19] have developed methods to assess time lags in biomedical research and propose potential ways to reduce them. Biomedical research is a long-term process with significant consumption of money and human resources. Only in the year 2023, the European Union spent a total of 8.4 billion on health research according to [20].
The topic of cost reduction is relevant in many branches. ICrA was also created to reduce production costs in the oil industry. Based on existing correlations between important criteria, some of the slower, more time-consuming, and expensive measurements can be eliminated without compromising the precision of the solution. ICrA can help with criteria number reduction after pairwise comparisons and determining the degrees of consonance and dissonance. A sufficiently high consonance degree is assumed as a possible indicator for elimination [21,22]. The opportunity that ICrA provides has been used to find reliable solutions for different tasks [23,24,25] and in various scientific fields [26,27,28,29].
In this paper, the InterCriteria decision-making approach is applied for the first time to optimize an experimental protocol (EP) of ten preselected tasks, aimed at measuring the surface EMG signals of human upper arm muscles. The purpose of this investigation is to obtain an optimized experimental protocol (OEP) that saves physical and mental fatigue in the investigated subjects, as well as time, money, and utilities, while the informative value of the EMG signals remains.

2. Materials and Methods

2.1. Experimental Procedure

Fifteen volunteers with no evidence of current musculoskeletal disorders or those related to changes in muscle strength, pain, or a reduced range of motion were invited to participate in the study. After a narrow check-up of the recorded video and EMG files, only ten of them passed for further consideration—four men and six women. The main reasons for excluding the participants were serious EMG artifacts, spikes, incorrect starting and ending positions, and non-compliance with the set rhythm of the motor task execution. All of them were right-handed and performed a sequence of ten tasks with their right hand. The experimental protocol was approved by the Scientific Council of the Institute of Biophysics and Biomedical Engineering. On the day of the study, the whole procedure was explained to the subjects in detail, and they signed informed consent. Before each motor task, performance instructions were given, the movement was shown, and the participant performed several trials on their own before the recording of the EMG data and video began. The examination was conducted in a sitting position on a chair with a backrest and without armrests. The EMG transmitter was placed on the waist of the examined person, and electrodes were stuck to the muscle fibers of six muscles or parts of muscles. The system used was Telemyo 2400 G2 from Noraxon Inc. (Scottsdale, Arizona, USA), the surface electrodes were “Skintact Premier” F-301 (Archenweg, Austria), and the diameter of the circles was 30 mm. For better electrode adhesion according to the standards of EMG signal gathering, the skin was prepared in the following order—hair removal (if needed), alcohol whipping, and conductive gel application. These muscles were targeted—pars acromialis and pars spinata of m. deltoideus, m. biceps brachii, caput longum of m. triceps brachii, the surface lateral part of m. brachialis, and m. anconeus. Two elbow angles were recorded using a 2D goniometer. The sampling frequency during the experiments was 1500 Hz.
A complete motor task protocol was designed to study the kinesiology of the upper limb in detail. For the present study, only the first half of the protocol was studied, where movement performance is in the sagittal plane. Each task was completed in 1 min.
Motor tasks in the full experimental protocol (FEP):
A person’s starting position is always the same. The volunteer is seated on the chair with their shoulders leveled and their arms resting calmly beside their body and their palms facing their hips. Their knee joints are bent at 90 degrees, and their feet are placed firmly on the floor. Their gaze is directed forward.
TASK 1. Hold the starting position for 1 min. The goal here is to see the power and stability of the EMG signal.
TASK 2. Within 1 min, the examiner takes the subject’s arm and passively moves it into several positions. From these positions, the volunteer performs the maximal isometric contractions against the maximal manual resistance to elicit the maximal muscle response. Resistance is applied to the distal part of the bone lever. This task is further used for the normalization of the EMG signals.
The next six motor tasks are cyclic movements in the sagittal plane, divided into four phases. The first phase is flexion from the starting position to the possible maximum upper position. The thumb points in the direction of movement. The second phase is the pose of holding the reached upper position for 5 s. The third phase is extension from the maximal upper flexion to the starting position. The fourth phase is the pose of holding the starting position for 5 s. The number of repetitions for cyclic movements is limited to 1 min, i.e., the subject carries out as many repetitions as possible in this time. The duration of the holding positions is always 5 s, and the active phases have different durations.
Further on in the paper, we denote the active 1st and 3rd phases of each task (from TASK 3 to TASK LOAD 10) as fsp (coming from flexion in the sagittal plane) and esp (coming from extension in the sagittal plane). The number before the abbreviations shows the seconds taken for the execution of each active phase. “W” after the abbreviations indicates added weight.
TASK 3. Cyclic movement in the sagittal plane with a duration of the active 1st (10fsp) and 3rd (10esp) phases of 10 s. In other words, TASK 3 consists of the following four phases: 10fsp, holding position, 10esp, and holding position. The phases are performed in the same sequence for the rest of the tasks (from TASK 4 to TASK LOAD 10). They will not be given here in detail to avoid repetitions.
TASK 4. Cyclic movement in the sagittal plane with a duration of the active 1st (6fsp) and 3rd (6esp) phases of 6 s.
TASK 5. Cyclic movement in the sagittal plane with a duration of the active 1st (2fsp) and 3rd (2esp) phases of 2 s.
TASK 6. Cyclic movement in the sagittal plane with a duration of the active 1st (1fsp) and 3rd (1esp) phases of 1 s.
For the last four tasks, a load of 0.5 kg is attached to the wrist. This is an anatomical and comfortable wristband.
TASK LOAD 7. Cyclic movement in the sagittal plane with a duration of the active 1st (10fspW) and 3rd (10espW) phases of 10 s.
TASK LOAD 8. Cyclic movement in the sagittal plane with a duration of the active 1st (6fspW) and 3rd (6espW) phases of 6 s.
TASK LOAD 9. Cyclic movement in the sagittal plane with a duration of the active 1st (2fspW) and 3rd (2espW) phases of 2 s.
TASK LOAD 10. Cyclic movement in the sagittal plane with a duration of the active 1st (1fspW) and 3rd (1espW) phases of 1 s.
For clarity, Figure 1 presents one subject performing TASK LOAD 7.
Data processing
Each recorded trial was stored as a separate file for convenient post-processing. After initial observation of the records, only one cyclic movement was selected from each task. Filtration was applied—a Butterworth high-pass filter, 4th-order (cut-off frequency of 20 Hz), and a Butterworth low-pass filter, 4th-order (cut-off frequency of 250 Hz). The electromyographic signal for TASK 2 was rectified, and six maximal values for each muscle were calculated. These are the normalization coefficients for the EMG signals received from the other eight movements.
The same filtration and rectification were used for the active tasks (from TASK 3 to TASK LOAD 10). Normalization was applied using the abovementioned coefficients. After that, just one experimental cycle of flexion–pose–extension–pose was chosen visually. This is a maximally good cycle with an exact duration and trajectory and no signal interference present. The time points (the exact second) where the movement freezes in a pose and pose transitions into movement were determined. This processing was carried out manually, and the indications from the recorded angles were used as a reference. They denote four periods (phases). The EMG was rectified and smoothed with twenty samples for each. The area under the obtained curve was calculated. These values were taken into further consideration.
Figure 2 shows the screen with the EMG and angle measurements during TASK 5.

2.2. The ICrA Decision-Making Approach

The classical ICrA combines two fundamental mathematical approaches—index matrices (IMs) and intuitionistic fuzzy sets (IFSs) [1]. The IMs approach represents an algebraic apparatus for data array processing with various dimensions, while the IFSs approach is a mathematical tool for uncertainty treatment. ICrA was developed to serve industry and scientific investigations to solve common problems, where it is necessary to work with datasets for multiple objects measured against different criteria.
Let us consider the following initial IMA:
Obj1…Objk…Objl…Objn
Cr1eCr1,Obj1…eCr1,Objk…eCr1,Objl…eCr1,Objn
……………………
IMA =CrieCri,Obj1…eCri,Objk…eCri,Objl…eCri,Objn
……………………
CrjeCrj,Obj1…eCrj,Objk…eCrj,Objl…eCrj,Objn
……………………
CrmeCrm,Obj1…eCrm,Objk…eCrm,Objl…eCrm,Objn
where Cr1, …, Crm denote the IM criteria; Obj1, …, Objn represent the IM objects; and eCr1,Obj1 … eCrm,Objn are the IM elements. IM elements can be real numbers or other objects that are commensurate with relation R to the other e-objects, so that for each i, j, k, R(eCrk,Obji, eCrk,Objj) is defined. Let R ¯ be the dual relation of R in the sense that if R is satisfied, then R ¯ is not satisfied, and vice versa is always true.
Let N k , l µ and N k , l υ be the number of cases in which R(eCrk,Obji, eCrk,Objj) and R(eCrl,Obji, eCrl,Objj) and, respectively, R(eCrk,Obji, eCrk,Objj) and R ¯ (eCrl,Obji, eCrl,Objj) are simultaneously satisfied.
It is obvious that:
N k , l µ + N k , l υ   ≤   n ( n − 1 ) 2
For each k, l such that 1 ≤ k < l ≤ m and for n ≥ 2, we define the following:
µ C r k , C r l = 2 N k , l µ n ( n − 1 ) ,   υ C r k , C r l = 2 N k , l υ n ( n − 1 )
The resulting ordered pair µ C r k , C r l ,   υ C r k , C r l is an intuitionistic fuzzy pair (IFP). An IFP is an intuitionistic fuzzy evaluation of the relations established between two criteria Crk and Crl. Thus, the index matrix A, which gives the relationships between objects and the evaluating criteria, can be transformed into another index matrix A* that contains the relations only between the criteria.
Cr1…Crm
IMA* =Cr1 µ C r 1 , C r 1 , υ C r 1 , C r 1 … µ C r 1 , C r m , υ C r 1 , C r m
…………
Crm µ C r m , C r 1 , υ C r m , C r 1 … µ C r m , C r m , υ C r m , C r m
IMA* sets the degrees of correspondence and non-correspondence between the criteria Cr1, …, Crm. Also, for each pair of criteria, the degree of uncertainty is taken into account. Correlation dependencies in ICrA are in the IFP form, with their values in the interval [0;1].
The last step of the algorithm is to determine the degree of correlation between the criteria depending on the threshold values for μ and ν. These correlations are known as positive or negative consonance and dissonance.
Let 0 ≤ α ≤ 1 and 0 ≤ β ≤ 1 be numbers such that α + β ≤ 1. The two criteria Crk and Crl are as follows:
  • In positive consonance when µ C r k , C r l > α and υ C r k , C r l < β;
  • In negative consonance when µ C r k , C r l < β and υ C r k , C r l > α;
  • In dissonance otherwise.
Table 1 shows the scale for positive and negative consonance, as well as dissonance, according to [21].
As can be seen from Table 1, two criteria are in positive consonance when the calculated µ-value is larger than 0.75 and hits the interval (0.75, 1]. Negative consonance is observed if 0 ≤ µ ≤ 0.25. In the other cases, when 0.25 < µ ≤ 0.75, dissonance appears.
Obviously, the larger the α-value and the smaller the β-value, the fewer the criteria that can be simultaneously associated with a positive consonance correlation. From a practical point of view, the most informative cases for the relations between criteria are those in which the positive or negative consonance is as great as possible, while cases of dissonance give less information and can be omitted.
The ICrA algorithm can be run in different ways, including using the MATLAB and Excel environments, but here, ICrAData software version 2.5 is used [30]. For the purposes of the current investigation, the most common µ-biased algorithm for criteria comparison is selected. ICrAData software is freely available for users and can be reached at http://intercriteria.net/software/, accessed on 15 April 2024.

3. Results

Real EMG data from six surface muscles, obtained from ten healthy volunteers, are used in this study. The investigated subjects followed a protocol of cyclic motor tasks for elbow movement consisting of four phases—full flexion, full extension, and two poses between them. The active phases have different time durations, and they are performed without and with added weight. To answer the question of whether some of these cyclic motor tasks can be omitted due to the similarity in the calculated area under the EMG curves (after filtration, rectification, and normalization), the correlation dependences between eight flexion and eight extension phases are detected using ICrA. Having the answer, we will be able to exclude from the EP slower and harder tasks. A slower task will be omitted if its flexion and extension phases simultaneously participate in consonance pairs with flexion and extension phases from a faster task. Thus, the number of studied cyclic movements will be reduced, and the FEP will be optimized.
For each subject (S1, S2, …, S10), an initial index matrix was created with the criteria (by rows) of the respective flexion phases from TASK 3 to TASK LOAD 10 and the six muscles as objects (by columns). Altogether, ten IMs are constructed. The results for twenty-eight flexion-phase pairs for the ten subjects are summarized in Table 2.
As can be seen from Table 2, a positive consonance (according to Table 1) is observed for ten pairs altogether, namely 10fsp-6fsp, 10fsp-2fsp, 6fsp-2fspW, 1fsp-6fspW, 1fsp-2fspW, 10fspW-6fspW, 10fspW-1fspW, 6fspW-2fspW, 6fspW-1fspW, and 2fspW-1fspW. For three criteria pairs (10fsp-6fsp, 6fsp-2fspW, and 10fspW-6fspW), strong positive and positive consonance was found. In the remaining seven pairs (10fsp-2fsp, 1fsp-6fspW, 1fsp-2fspW, 10fspW-1fspW, 6fspW-2fspW, 6fspW-1fspW, and 2fspW-1fspW), besides strong positive and positive consonance, weak positive consonance also appears for some of the investigated subjects. Due to the detected consonance relations, the tasks containing correlated flexion phases will be considered in the next step for possible exclusion from the FEP.
The results in Table 2 show relations in positive consonance, as well as relations in dissonance, for the remaining eighteen criteria pairs. The included phases in these pairs are considered independent, and thus the tasks containing them will remain in the EP.
It is interesting to note that five criteria pairs from the movements with loads are in positive consonance (10fspW-6fspW, 10fspW-1fspW, 6fspW-2fspW, 6fspW-1fspW, and 2fspW-1fspW). Two pairs in positive consonance are observed for movements without added weight (10fsp-6fsp, 10fsp-2fsp). Three pairs in positive consonance were found for movements without and with added load (6fsp-2fspW, 1fsp-6fspW, 1fsp-2fspW). One can conclude that the added weight influences the muscle interactions.
In the next step, in a similar way, the extension phases from TASK 3 to TASK LOAD10 were the criteria, and the six muscles were objects in the initial IMs. Again, ten IMs are constructed, one for each subject. The ICrA results for twenty-eight extension-phase pairs are presented in Table 3.
As shown in Table 3, for the extension phases from TASK 3 to TASK LOAD 10, analogous ICrA calculations were made. The following relations in positive consonance (according to Table 1) were observed: 10esp-6esp, 10espW-6espW, 10espW-2espW, and 2espW-1espW. Between 10esp and 6esp as well as 10espW and 6espW, strong positive and positive consonance is observed, while for the pairs 10espW-2espW and 2espW-1espW, weak positive consonance is also detected. The correlated extension phases determine tasks for consideration and possible removal from the FEP.
Going deeper into the results presented in Table 3, dissonance was found for some of the subjects in the remaining twenty-four pairs. The phases in these pairs can be assumed to be independent, and thus they will be excluded from future consideration.
Again, more criteria pairs in positive consonance for extension phases with added loads are found (10espW-6espW, 10espW-2espW, 2espW-1espW) compared to those without loads (10esp-6esp). Hence, more muscle interactions appear when the load is added.
For clarity, the observed positive consonance for the phases of flexion and extension is summarized in Table 4. Table 4 also presents the six tasks that will be compared and considered for omission from the FEP.
According to Table 4, there are more criteria pairs in positive consonance for flexion than for extension. Ten criteria pairs are in positive consonance for the flexion phases, while for extension, they are only four. As can be seen from Table 4, there are three flexion-phase pairs (10fsp-6fsp, 10fspW-6fspW, 6fspW-1fspW) corresponding to three extension-phase pairs from the same tasks (10esp-6esp, 10espW-6espW, 6espW-1espW) simultaneously in positive consonance. For the remaining criteria pairs, such a coincidence is not observed. When the two criteria in a pair are in positive consonance, according to ICrA, they can be considered dependent. Therefore, the slower or more difficult-to-perform flexion and extension phases can be omitted from the experimental sequence. In other words, the tasks containing easier and faster phases to perform will remain in the experimental protocol instead of slower and more difficult ones.

4. Discussion

Experimental protocols for EMG signal measurements are widely applied for an in-depth understanding of biomechanical problems. They are also implemented in rehabilitation, sports, medical, and scientific research, as well as in robotics and devices for controlling human limbs, etc. From the experiment to the real implementation of the results, there is a long way to go, using many studies and large databases. Furthermore, when we discuss human research, human treatment and absolute safety insurance are mandatory, according to the Declaration of Helsinki [31]. Hence, painless and less exhausting studies are demanded for biomedical EPs.
Decision-making in biomedicine is a complicated problem that requires compromises between different conflicting goals. For instance, the choice between a longer and more exhausting but informative EP and a shorter and easier EP that could be performed by a larger group of volunteers and patients with motor deficits must be made. The recently developed ICrA approach, as a decision making-method, can give the answer. Therefore, ICrA is applied to optimize the number of tasks included in the FEP for surface EMG signal measurements.
The optimization of the full experimental protocol follows seven steps, given below:
  • Preliminary preparation for EMG data measurements: Determination of the purpose of the research; selection of surface muscles for EMG signals recording; determination of the type and sequence of movements in the FEP; determination of inclusion and exclusion criteria for the participants.
  • Selecting suitable subjects for the study.
  • FEP execution, including training on and performing the cyclic movements.
  • EMG data processing, including the choice of the best cyclic movement (BCM). BCM separation into four phases; application of filters; signal rectification; normalization and calculation of area under the obtained curve.
  • ICrA application for optimization of the FEP, based on the calculated values for the flexion and extension phases.
  • Objective (based on ICrA) and subjective (based on the researcher’s opinion) assessment of the results.
  • Obtaining the OEP.
First, the ICrA correlations between eight flexion phases (presented in Table 2) from the considered cyclic movements (TASK 3 to TASK LOAD 10) were calculated. Next, ICrA was applied to estimate the relations between eight extension phases (shown in Table 3) from the same movements (TASK 3 to TASK LOAD 10). The purpose was to determine the tasks that could be excluded from the FEP.
We seek the correlations in positive consonance simultaneously appearing for both the flexion and extension phases from the same task. The following correlations are found: 10fsp-6fsp and 10esp-6esp (from TASK 3 and TASK 4), 10fspW-6fspW and 10espW-6espW (from TASK LOAD 7 and TASK LOAD 8), and 2fspW-1fspW and 2espW-1espW (from TASK LOAD 9 and TASK LOAD 10). In the next stage, we decide which phase included in the correlated pair should remain and which to drop. We can easily see that the 10fsp/10esp phases are more time-consuming (10 s) than 6fsp/6esp (only 6 s). Here, the longer execution time is an exclusion criterion. The same logic is used for the phases from movements with added weight: 10fspW/10espW and 6fspW/6espW, as well as 2fspW/2espW and 1fspW/1espW. Therefore, the phase 6fsp/6esp from TASK 4, 6fspW/6espW as part of TASK LOAD 8, and 1fspW/1espW from TASK LOAD 10 will remain in the OEP. On the other hand, 10fsp/10esp, 10fspW/10espW, and 2fspW/2espW are phases from TASK 3, TASK LOAD 7, and TASK LOAD 9 that will be removed from the FEP. Thus, the execution of the slower tasks is omitted from the OEP. The decision to exclude one task from the FEP is made only if there is a positive consonance simultaneously appearing between the flexion and extension phases with equal load and duration. Even more, we decided to perform the faster tasks since Park et al. [32] showed that for the investigated subjects, it is difficult to perform smooth rhythmic movements very slowly. The reason is related to motor control organization and is not a consequence of muscle or neural pathway weaknesses.
In Table 5, full and optimized EPs are presented, as well as the runtime durations for each task and protocol.
As can be seen from Table 5, the runtime duration of the FEP was reduced 1.43 times, i.e., the OEP can save 30% of the FEP execution time. The pre-task training takes a different time for each participant, and it is not included in the calculations here. This means the time that would be saved is much more. The results presented here prove that ICrA is a suitable tool for the optimization of the experimental protocol for surface EMG signal measurements.
In our future work, we plan to validate the outcomes obtained here for the optimization of the horizontal plane protocol. In addition, different decision-making methods will be compared. Also, the OEP will be applied in a series of studies to a larger group of subjects and patients with motor deficits.

5. Conclusions

In this paper, ICrA was applied to assess the correlations between eight flexion and eight extension phases from cyclic movement tasks, differing in time duration and added weight. Active-phase pairs that showed positive consonance were taken for future consideration. Objective analysis, based on ICrA, was followed by subjective analysis at the discretion of the researchers. As a result, some excessively slow and, for that reason, harder-to-execute movement tasks are excluded from the full experimental protocol. The OEP for surface EMG signal measurements consists of seven, instead of ten, movement tasks and saves 30% of the runtime duration of the FEP. In addition, the OEP can prevent human fatigue while averting the loss of crucial data.
The proposed optimization for the EP for surface EMG signal measurements using the ICrA decision-making approach might be useful in preliminary preparation for the study of target groups with many volunteers or patients with motor disabilities. Also, the procedure may be helpful for other biomedical experimental protocols or even for EPs in different scientific areas.

Author Contributions

Conceptualization, M.A. and S.A.; methodology, M.A., S.A. and R.R.; software, M.A.; validation, S.A. and M.A.; formal analysis, M.A. and S.A.; investigation, M.A., S.A. and R.R.; resources, S.A., M.A. and R.R.; data curation, M.A. and S.A.; writing—original draft preparation, M.A. and S.A.; writing—review and editing, M.A., S.A. and R.R.; visualization, S.A. and M.A.; supervision, R.R.; project administration, S.A. and M.A.; funding acquisition, M.A. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board (or Ethics Committee) of the Institute of Biophysics and Biomedical Engineering, Bulgarian Academy of Sciences (202ND/28.02.2022).

Informed Consent Statement

Informed consent was obtained from all the subjects involved in the study.

Data Availability Statement

The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Atanassov, K.T.; Mavrov, D.; Atanassova, V.K. Intercriteria decision making: A new approach for multicriteria decision making, based on index matrices and intuitionistic fuzzy sets. Issues IFSs GNs 2014, 11, 1–8. [Google Scholar]
  2. Vankova, D.; Sotirova, E.; Bureva, V. An application of the Intercriteria method approach to health-related quality of life. Notes Int. Fuzzy Sets 2015, 21, 40–48. [Google Scholar]
  3. Zaharieva, B.; Doukovska, L.; Ribagin, S.; Michalikova, A.; Radeva, I. Intercriteria analysis of Behterev’s kinesitherapy program. Notes Int. Fuzzy Sets 2017, 23, 69–80. [Google Scholar]
  4. Zaharieva, B.; Doukovska, L.; Ribagin, S.; Radeva, I. InterCriteria approach to Behterev’s disease analysis. Notes Int. Fuzzy Sets 2017, 23, 119–127. [Google Scholar]
  5. Zaharieva, B.; Doukovska, L.; Ribagin, S.; Radeva, I. InterCriteria analysis of data obtained from patients with Behterev’s disease. Int. J. Bioaut. 2020, 24, 5–14. [Google Scholar] [CrossRef]
  6. Ribagin, S.; Grozeva, A.; Popova, G.; Stoyanova, Z. InterCriteria analysis of body composition measurements data, associated with obesity among college students. Notes Int. Fuzzy Sets 2019, 25, 78–82. [Google Scholar] [CrossRef]
  7. Jekova, I.; Vassilev, P.; Stoyanov, T.; Pencheva, T. InterCriteria analysis: Application for ECG data analysis. Mathematics 2021, 9, 854. [Google Scholar] [CrossRef]
  8. Atanassov, K.; Pencheva, T. InterCriteria analysis approach as a tool for promising decision making in physiological rhythms. In 2019–20 MATRIX Annals; MATRIX Book Series; de Gier, J., Praeger, C.E., Tao, T., Eds.; Springer: Cham, Switzerland, 2021; pp. 279–285. [Google Scholar]
  9. Fidanova, S.; Zhivkov, P.; Roeva, O. InterCriteria analysis applied on air pollution influence on morbidity. Mathematics 2022, 10, 1195. [Google Scholar] [CrossRef]
  10. Todinova, S.; Mavrov, D.; Krumova, S.; Marinov, P.; Atanassova, V.; Atanassov, K.; Taneva, S.G. Blood plasma thermograms dataset analysis by means of InterCriteria and correlation analyses for the case of colorectal cancer. Int. J. Bioaut. 2016, 20, 115–124. [Google Scholar]
  11. Krumova, S.; Todinova, S.; Mavrov, D.; Marinov, P.; Atanassova, V.; Atanassov, K.; Taneva, S.G. InterCriteria analysis of calorimetric data of blood serum proteome. Biochim. Biophys. Acta (BBA)-Gen. Subj. 2017, 1861, 409–417. [Google Scholar] [CrossRef]
  12. Vassilev, P.; Todorova, L.; Popov, E.; Georgieva, R.; Slavov, C.; Atanassov, K. Two new modifications of the InterCriteria analysis. Proc. Bulg. Acad. Sci. 2023, 76, 23–34. [Google Scholar] [CrossRef]
  13. Andreev, N.; Sotirova, E.; Ribagin, S. Intercriteria analysis of data from the centers for transfusion hematology in Bulgaria. Comptes Rendus L’acade’mie Bulg. Sci 2018, 72, 982–990. [Google Scholar]
  14. Andreev, N.; Vassilev, P.; Ribagin, S.; Sotirov, S. InterCriteria analysis of data for blood collection in the Transfusion Hematology Department. University Hospital “St. Anna”, Sofia. Notes Int. Fuzzy Sets 2019, 25, 88–95. [Google Scholar] [CrossRef]
  15. Andreev, N.; Atanassova, V. InterCriteria analysis of the blood group distribution of patients of Saint Anna hospital in 2015–2019. In Advances and New Developments in Fuzzy Logic and Technology, IWIFSGN’2019—The Eighteenth International Workshop on Intuitionistic Fuzzy Sets and Generalized Nets, Warsaw, Poland, 24–25 October 2019; Advances in Intelligent Systems and Computing; Atanassov, K.T., Atanassova, V., Kacprzyk, J., Kałuszko, A., Krawczak, M., Owsiński, J.W., Sotirov, S.S., Sotirova, E., Szmidt, E., Zadrożny, S., Eds.; Springer: Cham, Switzerland, 2021; Volume 1308, pp. 158–165. [Google Scholar]
  16. Andreev, N.; Atanasov, K.; Bureva, V. InterCriteria analysis of blood collection data. In Yearbook of the Informatic Section, Union of Scientists in Bulgaria; 2019/2020; Volume 10, pp. 30–53. (In Bulgarian)
  17. Angelova, S.; Angelova, M.; Raikova, R. Estimating surface EMG activity of human upper arm muscles using InterCriteria Analysis. Math. Comput. 2024, 29, 8. [Google Scholar] [CrossRef]
  18. Guidance on Assessing and Minimizing Risk in Human Research, and When They Are Prolonged the Risk Rises. Available online: https://irb.unm.edu/library/documents/guidance/assessing-and-minimizing-risk-in-human-research.pdf (accessed on 18 March 2024).
  19. Hanney, S.R.; Castle-Clarke, S.; Grant, J.; Guthrie, S.; Henshall, C.; Mestre-Ferrandiz, J.; Pistollato, M.; Pollitt, A.; Sussex, J.; Wooding, S. How long does biomedical research take? Studying the time taken between biomedical and health research and its translation into products, policy, and practice. Health Res. Policy Syst. 2015, 13, 1. [Google Scholar] [CrossRef]
  20. Attracting Life Science Investments in Europe. Fourth Edition. 2023. Available online: https://www.europabio.org/wp-content/uploads/2023/10/Life-Science-Attractiveness-2023-October-22-Final.pdf (accessed on 21 March 2024).
  21. Atanassov, K.T.; Atanassova, V.K.; Gluhchev, G. InterCriteria analysis: Ideas and problems. Notes Int. Fuzzy Sets 2015, 21, 81–88. [Google Scholar]
  22. Atanassov, K.T. Index Matrices: Towards an Augmented Matrix Calculus; Springer International Publishing: Cham, Switzerland, 2014. [Google Scholar]
  23. Roeva, O.; Fidanova, S.; Ganzha, M. InterCriteria analysis of the evaporation parameter influence on ant colony optimization algorithm: A workforce planning problem. In Results of the Workshop on Computational Optimization; Fidanova, S., Ed.; Springer International Publishing: Cham, Switzerland, 2021; Volume 920, pp. 89–109. [Google Scholar]
  24. Mucherino, A.; Fidanova, S.; Ganzha, M. Ant Colony Optimization With Environment Changes: An Application to GPS Surveying. In Proceedings of the 2015 Federated Conference on Computer Science and Information Systems (FedCSIS), Lodz, Poland, 13–16 September 2015. [Google Scholar]
  25. Castillo, O.; Ramirez, E.; Roeva, O. Water cycle algorithm augmentation with fuzzy and intuitionistic fuzzy dynamic adaptation of parameters. Notes Int. Fuzzy Sets 2017, 23, 79–94. [Google Scholar]
  26. Cuvalcioglu, G.; Bureva, V.; Michalikova, A. Intercriteria analysis applied to university ranking system of Turkey. Notes Int. Fuzzy Sets 2019, 25, 90–97. [Google Scholar] [CrossRef]
  27. Kacprzyk, A.; Sotirov, S.; Sotirova, E.; Shopova, D.; Georgiev, P. Application of intercriteria analysis in the finance and accountancy positions. Notes Int. Fuzzy Sets 2017, 23, 84–90. [Google Scholar]
  28. Parvathi, R.; Atanassova, V.; Doukovska, L.; Yuvapriya, C.; Indhurekha, K. InterCriteria analysis of rankings of Indian universities. Notes Int. Fuzzy Sets 2018, 24, 99–109. [Google Scholar] [CrossRef]
  29. Krawczak, M.; Bureva, V.; Sotirova, E.; Szmidt, E. Application of the intercriteria decision making method to universities ranking. Nov. Develop. Uncert. Repres. Proc. 2016, 401, 365–372. [Google Scholar]
  30. Ikonomov, N.; Vassilev, P.; Roeva, O. ICrAData—Software for intercriteria analysis. Int. J. Bioautom. 2018, 22, 1–10. [Google Scholar] [CrossRef]
  31. Word Medical Association. Declaration of Helsinki: Ethical Principles for Medical Research Involving Human Subjects; WMA: Helsinki, Finland, 1964. [Google Scholar]
  32. Park, S.W.; Marino, H.; Charles, S.K.; Sternad, D.; Hogan, N. Moving slowly is hard for humans: Limitations of dynamic primitives. J. Neurophysiol. 2017, 118, 69–83. [Google Scholar] [CrossRef] [PubMed]
Applsci 14 05436 g001
Figure 1. (a) Starting position; (b) position of maximal upper flexion with a wristband of 0.5 kg.
Figure 1. (a) Starting position; (b) position of maximal upper flexion with a wristband of 0.5 kg.
Applsci 14 05436 g001
Applsci 14 05436 g002
Figure 2. A subject and raw EMG signals during TASK 5 performance of the FEP.
Figure 2. A subject and raw EMG signals during TASK 5 performance of the FEP.
Applsci 14 05436 g002
Table 1. ICrA µ-scale for positive consonance, dissonance, and negative consonance.
Table 1. ICrA µ-scale for positive consonance, dissonance, and negative consonance.
Meansµ-Values
positive consonance strong positive(0.95, 1.00]
positive(0.85, 0.95]
weak positive(0.75, 0.85]
dissonance (0.25; 0.75]
negative consonance weak negative(0.15, 0.25]
negative(0.05, 0.15]
strong negative[0.00, 0.05]
Table 2. Results for different flexion phases after ICrA application.
Table 2. Results for different flexion phases after ICrA application.
Flexion Phases
in Sagittal Plane		S1S2S3S4S5S6S7S8S9S10
10fsp-6fsp10.87110.9310.930.930.930.87
10fsp-2fsp10.870.9310.930.870.80.80.930.8
10fsp-1fsp10.60.870.870.80.870.670.930.930.8
10fsp-10fspW0.80.530.8710.870.870.8710.870.8
10fsp-6fspW0.870.60.870.9310.870.8710.870.8
10fsp-2fspW930.60.870.930.870.870.870.930.930.8
10fsp-1fspW10.60.870.870.80.870.670.930.870.8
6fsp-2fspW110.93110.870.870.870.870.93
6fsp-1fspW10.730.870.870.870.870.7310.870.93
6fsp-10fspW0.80.670.8710.80.870.930.930.80.93
6fsp-6fspW0.870.730.870.930.930.870.930.930.80.93
6fsp-2fspW0.930.730.870.930.80.870.93110.93
6fsp-1fspW10.730.870.870.730.870.7310.930.93
2fsp-1fspW10.730.930.870.8710.870.8711
2fsp-10fspW0.80.670.9310.810.930.80.930.87
2fsp-6fspW0.870.730.930.930.9310.930.80.931
2fsp-2fspW0.930.730.930.930.810.930.870.871
2fsp-1flspW10.730.930.870.7310.870.870.81
1fsp-10fspW0.80.80.870.870.6710.80.930.930.87
1fsp-6fspW0.870.8710.930.810.80.930.931
1fsp-2fspW0.930.8710.930.810.810.871
1fsp-1fspW11110.731110.81
10fspW-6fspW0.930.930.870.930.8711110.87
10fspW-2fspW0.730.930.870.930.87110.930.80.87
10fspW-1fspW0.80.80.870.870.9310.80.930.870.87
6fspW-2fspW0.81110.87110.930.81
6fspW-1fspW0.870.8710.930.810.80.930.871
2fspW-1fspW0.930.8710.930.9310.810.931
Table 3. Results for different extension phases after ICrA application.
Table 3. Results for different extension phases after ICrA application.
Extension Phases
in Sagittal Plane		S1S2S3S4S5S6S7S8S9S10
10esp-6esp0.87111110.930.870.930.93
10esp-2esp0.870.930.810.870.870.930.80.670.87
10esp-1esp0.730.930.80.930.80.870.730.730.80.8
10esp-10espW0.730.670.9310.930.87110.80.8
10esp-6espW0.80.67110.870.870.930.930.730.8
10esp-2espW0.80.870.930.930.730.870.930.930.870.87
10esp-1espW0.670.930.8710.870.870.80.80.870.67
6esp-2espW0.870.930.810.870.8710.80.730.93
6esp-1espW0.730.930.80.930.80.870.80.730.870.87
6esp-10espW0.870.670.9310.930.870.930.870.870.73
6esp-6espW0.80.67110.870.8710.930.80.73
6esp-2espW0.930.870.930.930.730.8710.930.930.8
6esp-1espW0.80.930.8710.870.870.870.80.930.73
2esp-1espW0.870.8710.930.9310.80.670.870.93
2esp-10espW0.730.730.7310.9310.930.80.730.8
2esp-6espW0.80.730.811110.870.670.8
2esp-2espW0.930.80.730.930.87110.870.80.87
2esp-1elspW0.80.870.81110.870.730.80.8
1esp-10espW0.60.60.730.930.8710.730.730.870.73
1esp-6espW0.670.60.80.930.9310.80.80.80.73
1esp-2espW0.80.80.730.870.9310.80.80.930.8
1esp-1espW0.80.860.80.930.9310.930.930.930.73
10espW-6espW0.9310.9310.9310.930.930.931
10espW-2espW0.80.80.870.930.810.930.930.930.93
10espW-1espW0.670.730.9310.9310.80.80.930.87
6espW-2espW0.730.80.930.930.871110.870.93
6espW-1espW0.60.730.871110.870.870.870.87
2espW-1espW0.870.930.80.930.8710.870.8710.8
Table 4. Comparison between flexion- and extension-phase pairs in positive consonance (PC).
Table 4. Comparison between flexion- and extension-phase pairs in positive consonance (PC).
ICrA Detected PC
between
Flexion Phases		ICrA Detected PC
between
Extension Phases		Detected ConsonanceTasks for Further
Consideration
10fsp-6fsp 10esp-6espstrong PC, PCTASK 3–TASK 4
10fsp-2fsp
6fsp-2fspW
1fsp-6fspW
1fsp-2fspW
10fspW-6fspW10espW-6espWstrong PC, PCTASK LOAD 7–TASK LOAD 8
10fspW-1fspW
10espW-2espW
6fspW-2fspW
6fspW-1fspW
2fspW-1fspW2espW-1espWstrong PC, PC, weak PCTASK LOAD 9–TASK LOAD 10
Table 5. Full and optimized experimental protocols.
Table 5. Full and optimized experimental protocols.
Full EP Runtime [min]Optimized EP Runtime [min]
TASK 11TASK 11
TASK 21TASK 21
TASK 31--
TASK 41TASK 41
TASK 51TASK 51
TASK 61TASK 61
TASK LOAD 71--
TASK LOAD 81TASK LOAD 81
TASK LOAD 91--
TASK LOAD 101TASK LOAD 101
Total runtime, [min]10Total runtime, [min]7
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Angelova, M.; Angelova, S.; Raikova, R. How to Optimize the Experimental Protocol for Surface EMG Signal Measurements Using the InterCriteria Decision-Making Approach. Appl. Sci. 2024, 14, 5436. https://doi.org/10.3390/app14135436

AMA Style

Angelova M, Angelova S, Raikova R. How to Optimize the Experimental Protocol for Surface EMG Signal Measurements Using the InterCriteria Decision-Making Approach. Applied Sciences. 2024; 14(13):5436. https://doi.org/10.3390/app14135436

Chicago/Turabian Style

Angelova, Maria, Silvija Angelova, and Rositsa Raikova. 2024. "How to Optimize the Experimental Protocol for Surface EMG Signal Measurements Using the InterCriteria Decision-Making Approach" Applied Sciences 14, no. 13: 5436. https://doi.org/10.3390/app14135436

APA Style

Angelova, M., Angelova, S., & Raikova, R. (2024). How to Optimize the Experimental Protocol for Surface EMG Signal Measurements Using the InterCriteria Decision-Making Approach. Applied Sciences, 14(13), 5436. https://doi.org/10.3390/app14135436

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop