Introduction
Asthma is characterized by episodic and reversible lower airway obstruction.1 The post-bronchodilator airway reversibility could offer useful objective information for a diagnosis of asthma and evaluating response to therapy.1,2 An increase of ≥ 12% in forced expiratory volume in the first second (FEV1) after inhalation of bronchodilators indicates airway reversibility.3 However, measuring FEV1 can be challenging due to difficulties in meeting quality control standards, and it has limitations, such as requiring forced maximal inhalation and exhalation maneuvers, which can affect airway calibre.4–6
Compared to adults, the success rate of conducting spirometry is lower in children, particularly in preschool-aged children, where the success rate is less than half.7 Moreover, the baseline lung function of children can be also affected by multiple factors such as previous lower respiratory tract infections, the timing of previous wheeze,8 which in turn impacts the evaluation of the results of bronchodilator response (BDR). In many healthcare settings, asthma in children is a clinical diagnosis based on parental reported symptoms,9 but this approach carries a significant risk of misdiagnosis,10,11 resulting in over- and undertreatment of asthma, poor symptom control, and in delayed commencement of asthma treatment.12
Since respiratory muscle EMG reflects the work of breathing, the overall stiffness or compliance of the lungs induced by changes in the airway smooth muscle activity, EMG could be used to reflect change in load or increase in compliance after bronchdilation.13–16 Our previous studies showed that respiratory muscles electric activities recorded from chest wall surface electrodes are a noninvasive technique, requires minimal active cooperation. Moreover, it could offer objective information that indirectly reflect the changes in the lower airway resistance induced by histamine provocation test17 or exercise18 in adults. However, no data are available on surface respiratory muscle EMG and BDR in children with acute asthma. The aims of the present study were to determine whether surface respiratory muscle EMG could be feasible and used to assess the BDR and differentiate children with acute asthma from healthy subjects.
Methods
Subjects
A case–control study was performed on uncontrolled asthmatic children matched by age to healthy controls recruited from the community. Uncontrolled asthmatic children aged 3–14 years were recruited in the department of pediatrics of the first affiliated hospital of Guangzhou Medical University. These asthmatic children were diagnosed by a pediatric physician based on history of recurrent attacks of dyspnea with perceptible wheezing and one of the objective tests (eg, IgE, FeNO, or methacholine provocation test), and meeting the 2022 Global Initiative for Asthma,1 and remained in an uncontrolled state at this visit. The uncontrolled state was defined as having poorly controlled symptoms, along with an increase in FEV1 of at least 12%, or presenting with a wheezing symptom that improved after inhaling a bronchodilator. The history of respiratory conditions for asthma children were also collected during this visiting. This study protocol was reviewed and approved by ethic committee of the first affiliated Hospital of Guangzhou Medical University, approval number [2020 K-123], and written informed consent was obtained from participants’ parent to participate in the study.
Measurement
All subjects were asked to lay on the bed to measure diaphragm electromyogram (EMGdi), parasternal intercostal EMG (EMGpara) and airflow. EMGdi was simultaneously recorded from two pairs of electrodes (3M Health Care, D-41453 Neuss, Germany) on the right side as described by previous studies.17–20 EMGpara was recorded using bipolar electrodes placed bilaterally 3 cm from the midpoint of the sternum in the second intercostal space.21 Airflow was recorded by a digital flow meter (Siargo, Inc. Santa Clara, CA 95054, USA) via mask further connected to an additional tube which was 2 cm in diameter and 7 cm in length to increase inspiratory load to improve EMGdi quality. The EMGdi and airflow were recorded by the respiratory signal processor (RA-16, Respiratory Medical Science Co. Ltd, Guangzhou, China) simultaneously. The subjects were lied down throughout the test to minimize influences caused by posture changes.
Spirometry measurements (Cosmed Micro Quark, Cosmed Italy) were performed based on the recommendation of ATS/ERS.22 Sufficient time was offered to the subjects for practicing maneuvers of the forced spirometry to ensure quality of the pulmonary function test. For FEV1 repeatability, the difference between the largest and the next largest FEV1 is <0.150 L for those older than 6 years of age and <0.100 L or 10% of the largest FEV1, whichever is greater, for those aged 6 years or younger.22 A minimum of two technically acceptable measurements were required for those who are able to perform spirometry in both healthy group (Healthy-A) and asthma group (Asthma-A) but not for those who are unable to perform spirometry after more than ten manoeuvres or twenty minutes practice in both healthy group (Healthy-U) and asthma group (Asthma-U).
Protocol
Surface respiratory muscle EMG, FEV1 and wheezing were successively recorded before and after the inhalation of salbutamol. EMG signals and airflow were systematically recorded before spirometry to avoid the effect produced by forced spirometry maneuvers on bronchial tone. It is difficult for children to maintain stable breathing for 3 minutes; therefore, brief interruptions were allowed during EMG signal recording. Salbutamol was delivered via a mouthpiece chamber: 2 puffs (200 µg) for children aged 3–5 years and 4 puffs (400 µg) for those older than 5 years. Timing between salbutamol and EMG measurements was at least 15min. The recording time for surface respiratory muscle EMG and FEV1 was collected, respectively, for the former including preparation and recording sufficient analyzed EMG signals, for the latter from the beginning of training to obtaining three valid measurements.
Data and Statistical Analysis
Respiratory muscle EMG signal was automatically converted to Root Mean Square (RMS) by the respiratory signal processor (RA-16, Respiratory Medical Science Co. Ltd, Guangzhou, China).23–25 Peak value of RMS of EMG free from ECG was selected. Tidal volume (VT) was calculated from integrated flow and was further standardized with body weight (mL/kg). Data from fifteen breathing cycles during tidal breaths were averaged for each time points (before and after the inhalation of salbutamol). EMGdi and EMGpara combined with VT were expressed as EMGdi/VT and EMGpara/VT. The changes of FEV1 (ΔFEV1%), EMGdi (ΔEMGdi%), EMGpara (ΔEMGpara%), EMGdi/VT (ΔEMGdi/VT%) and EMGpara/VT (ΔEMGpara/VT) before and after inhalation of bronchodilator were selected for further analysis. EMGdi recorded from one pair representing better signal-to-noise ratio and less baseline variation was regarded as the primary signal for analysis. For each subject, the same pair of electrodes were used consistently before and after inhalation of salbutamol.
All parameters were assessed for normality by the Shapiro–Wilk test. Normally distributed data was presented as mean (standards or range), while data distributed non-normally was represented with mean (minimum ~ maximum). Wilcoxon Signed rank test and Mann–Whitney U-test were used for statistical analysis of two independent or paired groups. For categorical variables, contingency tables were used (2 with Fisher exact test). ROC (Receiver Operating Characteristic) curves were used to evaluate the diagnosis value of ΔEMGdi%, ΔEMGpara%, ΔEMGdi/VT% and ΔEMGpara/VT% for asthma, and the difference between each two area under ROC curves (AUC) among those were compared with the method of DeLong et al.26 P<0.05 was considered statistically significant.
Results
Thirty-two healthy children from community who agreed to participate in the study were classified as healthy; four were excluded (two presented an acute respiratory tract infection and two refused to perform spirometry). Therefore, data from 28 healthy children and 14 asthmatics were available for analysis. Most demographic characteristics showed no significant differences, but there was a statistically significant difference in gender ratio between the two groups (p=0.047; Table 1). As anticipated, the prevalence of rhinitis and eczema was significantly higher among children with asthma compared to healthy controls.
|
Table 1 Demographic Characteristics of Healthy and Asthmatic Children
|
In this study, eighteen out of forty-two children were unable to perform the spirometry procedure correctly. All healthy participants underwent spirometry for the first time, whereas 6 asthmatic children had prior experience with the test. The proportion of healthy subjects unable to obtain valid lung function results (15 out of 28) was significantly higher than that for asthmatic children (3 out of 14), but these participants were all under six years of age. Among those who performed FEV1 maneuver and obtained valid results, the duration of EMG recording was significantly shorter compared to that of the pulmonary function tests (8 ± 1 min vs 15 ± 4 min, p < 0.01). Furthermore, the timing for subjects who were able to perform the pulmonary function tests was comparable to that of those who could not (8 ± 1 min vs 9 ± 1 min).
EMGdi, EMGpara and EMGdi/VT all exhibited a statistically significant reduction following the administration of salbutamol in patients with uncontrolled asthma, except for EMGpara/VT that showed non-significant reduction (Table 2). Conversely, no significant alterations were detected in EMGdi, EMGpara and EMGdi/VT, and EMGpara/VT in healthy subjects before and after salbutamol inhalation. Importantly, the changes in ΔEMGdi%, ΔEMGpara% and ΔEMGdi/VT%, with the exception of ΔEMGpara/VT%, post-salbutamol inhalation were significantly greater in asthmatic patients compared to healthy controls (Table 2). Figure 1 presents ROC curves for ΔEMGdi%, ΔEMGpara%, ΔEMGdi/VT% and ΔEMGpara/VT%. The optimal cut-off values for these indices, along with their corresponding sensitivities and specificities for differentiating uncontrolled asthmatics from healthy children, are detailed in Table 3. Notably, the ROC analysis indicated that a BDR of 12.4% in EMGdi% provided the highest discriminative capacity, achieving a sensitivity of 100% and a specificity of 96% (Table 3).
 |
Table 2 Changes in Diaphragm EMG (EMGdi), the Ratio of EMGdi to Tidal Volume (EMGdi/VT), Parasternal Intercostal EMG (EMGpara) and the Ratio of EMGpara to Tidal Volume (EMGdi/VT) Before and After Inhaled Salbutamol
|
 |
Table 3 Area Under Receiver Operating Characteristic (ROC) Curves (AUC), Sensitivity and Specificity of Bronchodilator Response (Calculate as % of Change From Basal) Using ROC
|
 |
Figure 1 ROC curves for ΔEMGdi%, ΔEMGpara%, ΔEMGdi/VT%, and ΔEMGpara/VT%. ΔEMGdi%, ΔEMGpara% and ΔEMGdi/VT% could distinguish asthmatics from the healthy using ROC curves, but the highest value was the bronchodilator response of 12.4% in ΔEMGdi with 100% of sensitivity, 96% of specificity.
|
In sub-group analysis, there were no significant differences in EMGdi, EMGpara, EMGdi/VT and EMGpara/VT before and after inhalation of salbutamol in both the Healthy-A and Healthy-U groups in healthy children. There were also no significant differences in ΔEMGdi%, ΔEMGpara%, ΔEMGdi/VT% and ΔEMGpara/VT% between the Healthy-A and Healthy-U groups (Figure 2). Regarding asthmatic patients, obvious difference in EMGdi, EMGpara, EMGdi/VT and EMGpara/VT before and after inhalation of salbutamol were observed in both the Asthma-A and Asthma-U groups (Figure 2), and the trend was consistent for both sub-groups. However, ΔEMGdi, ΔEMGpara, ΔEMGdi/VT and ΔEMGpara/VT in the Asthma-U group were larger than those in the Asthma-A group (Figure 2).
 |
Figure 2 Comparison between participants able and unable to perform spirometry. In healthy group, two sub-groups were divided based on whether spirometry are able (Healthy-A, n=13) or unable (Healthy-U, n=15) to be performed; similarly, two sub-groups were divided based on whether spirometry are able (Asthma-A, n=11) or unable (Asthma-U, n=3) to be performed in the asthma group. In subgroup analysis, the trend and degree of change in EMGdi, EMGpara before and after inhalation of bronchodilator between Healthy-A and Healthy-U subgroups were both consistent. In relation to asthmatic subjects, the trend of change in EMGdi, EMGpara before and after inhalation of bronchodilator between Asthma-A and Asthma-U groups was consistent, but the changes in the Asthma-U group were larger than those in the Asthma-A group.
|
Four asthmatics in Asthma-A group showed mild-to-moderate airway obstruction, while all healthy children in Healthy-A group had normal lung function. The results of FEV1, FVC, FEV1/FVC and FEF25-75% before and after BDR were presented in the Table 4. ΔFEV1% in the Asthma-A group was larger than that in the Healthy-A group (17.6±5.4% vs 5.5±3.7%, p<0.001). There was a significant relationship between ΔFEV1% and ΔEMGdi% (r=0.762, p<0.001) and between ΔFEV1% and ΔEMGdi/VT% (r=0.603, p<0.01), but no correlation was found with ΔEMGpara and ΔEMGpara/VT.
 |
Table 4 Changes in Forced Vital Capacity (FVC), Forced Expiratory Volume in 1 second (FEV1), the FEV1/FVC Ratio, and Forced Expiratory Flow During the Middle Half of the Forced Vital Capacity (FEF25-75%) Before and After Inhaled Salbutamol
|
Discussion
In this study, all participants were able to record high-quality EMGdi and EMGpara signals; but 18 of them were unable to complete the pulmonary function tests. Among those who successfully underwent pulmonary function assessment, the duration of surface EMG recording was significantly shorter than the time required for the pulmonary function procedures. Both EMGdi and EMGpara could differentiate the healthy from uncontrolled asthmatics through assessing the BDR, but EMGdi is more useful.
Consistent with previous studies,27,28 our data indicate that the age at which pulmonary function test could not be performed is under 6 years. The proportion of healthy subjects who were unable to perform pulmonary function tests was higher than that of asthma patients. This may be attributed to the fact that nearly half of the asthma patients in this study had prior experience with pulmonary function procedures. Early exposure to pulmonary function testing might enable children to be familiar with the correct techniques of deep breathing and rapid exhalation, thus increasing the odds of a successful test.
This study is the first to compare the duration of EMG signals recording with that of pulmonary function testing. Our data indicate that the time required to record EMG signals is significantly shorter than the time needed to collect FEV1 results. Although the analysis of EMGdi still requires some time, it is possible to achieve automatic removal of ECG and then automatic analysis for EMGdi with the improvement of technology.29 The advantage in operation time could provide more possibilities for the future clinical promotion of this technology.
To improve the signal-to-noise ratio of the EMG signals, a 7 cm tubing was used, which is shorter than the 15 cm typically used for adults.17 This adjustment was made primarily due to the lower tidal volume in children, as longer tubing could lead to an increased risk of dead space ventilation. Additionally, the choice to use the right for recording EMGdi is intended to minimize interference from the ECG. If the EMGdi signal of the right diaphragm could not be obtained, EMGpara can be used as a substitute.30,31 Our data showed that ΔEMGpara could also distinguish asthmatic children from the healthy, although its AUC is significantly smaller than that of ΔEMGdi.
Our previous data found that EMGdi expressed as a function of tidal volume (ΔEMGdi/VT) has higher discriminative power during the challenge test than ΔEMGdi alone in adults, although it failed to reach statistical significance.17 However, there was an opposite trend in children’s data, which showed the ROC area from ΔEMGdi was higher than that from ΔEMGdi/VT. This may be because mask leakage is more common in children than in adults. In addition, tidal volume in children is smaller than that of adults, so slight fluctuations are more likely to cause data bias. Therefore, extra caution should be taken if using a mask to record airflow in children.
Maarsing et al’s study demonstrated that even when asthmatic children exhibited distinct clinical symptoms—including wheezing, persistent cough, prolonged expiration, or increased respiratory rate—at the end of the challenge test, no significant differences in EMGdi or EMGpara values were observed between these symptom groups.32 However, in our study, when asthmatic children exhibited two distinct response patterns during BDR—improvement in FEV1 and improvement in wheezing symptoms—significant differences were observed in their corresponding EMGdi and EMGpara values (see Figure 2). This may be related to the varying degrees of improvement in airway resistance during BDR under these two response patterns. However, the more pronounced changes in EMGdi and EMGpara following bronchodilator administration in the Asthma-U group suggest that surface respiratory muscle EMG may be useful in distinguishing patients with asthma from those without, particularly in individuals who are unable to perform spirometry.
Our study has several limitations. The sample size was relatively small, especially in the asthma group. In addition, not all participants were able to complete spirometry, including FEV1, which limited our ability to compare changes in respiratory EMG with FEV1 across all subjects. Finally, we did not compare respiratory EMG changes with airway resistance measured by body plethysmography, which could have offered additional insights into airway dynamics.
In conclusion, surface respiratory muscle EMG signal is feasible in children even for those who are unable to perform FEV1 during BDR. EMGdi and EMGpara, rather than those combined with VT, could be useful to assess the BDR and differentiate uncontrolled asthma from healthy in children during BDR.
Data Sharing Statement
The original data are available upon request. Please contact the corresponding author at [email protected] to request access.
Statement of Ethics
This study protocol was reviewed and approved by the ethics committee of the first affiliated Hospital of Guangzhou Medical University, approval number [2020 K-123], and written informed consent was obtained from participants’ parent to participate in the study.
Author Contributions
He, Li, and Liu are co-first authors who contributed equally to the study design, data acquisition, data interpretation, and manuscript writing. Luo supervised the study and contributed to the conception, design, execution, data interpretation, and manuscript preparation. All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.
Funding
This study was funded by National Natural Science Foundation of China (Grant Number: 82070090), Independent Project of State Key Laboratory of Respiratory Disease (Grant Number: SKLRD-Z-202413), Guangzhou Municipal Health Commission Major Technical Project Grant (Project Number: 2023C-GX02), the Clinical and Epidemiological Research Project of State Key Laboratory of Respiratory Disease (Grant No. SKLRD-L-202401).
Disclosure
The authors have no conflicts of interest to declare for this work.
References
1. GINA. Global strategy for asthma management and prevention; 2022 [Updated May 2022]. Available from: www.ginasthma.org.
2. Dundas I, Mckenzie S. Spirometry in the diagnosis of asthma in children. Curr Opin Pulm Med. 2006;12(1):28–33. doi:10.1097/01.mcp.0000199809.42011.58
3. Pellegrino R, Viegi G, Brusasco V, et al. Interpretative strategies for lung function tests. Eur Respir J. 2005;26(5):948–968. doi:10.1183/09031936.05.00035205
4. Nadel JA, Tierney DF. Effect of a previous deep inspiration on airway resistance in man. J Appl Physiol. 1961;16:717–719. doi:10.1152/jappl.1961.16.4.717
5. Pellegrino R, Sterk PJ, Sont JK, et al. Assessing the effect of deep inhalation on airway calibre: a novel approach to lung function in bronchial asthma and COPD. Eur Respir J. 1998;12(5):1219–1227. doi:10.1183/0903.1936.98.12051219
6. Henry C, Boucher M, Boulay M-È, et al. Oscillometry with or without spirometry for methacholine testing. Physiol Rep. 2025;13(10):e70387. doi:10.14814/phy2.70387
7. Burity EF, Pereira CAC, Rizzo JA, et al. Reference values for spirometry in preschool children. J Pediatr. 2013;89(4):374–380. doi:10.1016/j.jped.2013.01.002
8. Dom S, Weyler JJ, Droste JHJ, et al. Determinants of baseline lung function and bronchodilator response in 4-year-old children. Eur Respir J. 2014;44(2):371–381. doi:10.1183/09031936.00100813
9. Danvers L, Lo D, Gaillard EA. The role of objective tests to support a diagnosis of asthma in children. Paediatr Respir Rev. 2020;33:52–57. doi:10.1016/j.prrv.2019.02.001
10. Siersted HC, Boldsen J, Hansen HS, et al. Population based study of risk factors for underdiagnosis of asthma in adolescence: odense schoolchild study. BMJ. 1998;316(7132):651–5;discussion655–6. doi:10.1136/bmj.316.7132.651
11. Looijmans-van DAI, van Luijn K, Verheij T. Overdiagnosis of asthma in children in primary care: a retrospective analysis. Br J Gen Pract. 2016;66(644):e152–7. doi:10.3399/bjgp16X683965
12. Starren ES, Roberts NJ, Tahir M, et al. A centralised respiratory diagnostic service for primary care: a 4-year audit. Prim Care Respir J. 2012;21(2):180–186. doi:10.4104/pcrj.2012.00013
13. Boucher M, Dufour-Mailhot A, Tremblay-Pitre S, et al. In mice of both sexes, repeated contractions of smooth muscle in vivo greatly enhance the response of peripheral airways to methacholine. Respir Physiol Neurobiol. 2022;304:103938. doi:10.1016/j.resp.2022.103938
14. Boucher M, Henry C, Khadangi F, et al. Effects of airway smooth muscle contraction and inflammation on lung tissue compliance. Am J Physiol Lung Cell Mol Physiol. 2022;322(2):L294–L304. doi:10.1152/ajplung.00384.2021
15. Gazzola M, Lortie K, Henry C, et al. Airway smooth muscle tone increases airway responsiveness in healthy young adults. Am J Physiol Lung Cell Mol Physiol. 2017;312(3):L348–L357. doi:10.1152/ajplung.00400.2016
16. Henry C, Boucher M, Boulay M-È, et al. The cumulative effect of methacholine on large and small airways when deep inspirations are avoided. Respirology. 2023;28(3):226–235. doi:10.1111/resp.14387
17. He BT, Luo Y-M, Wang L-S, et al. Identifying bronchoconstriction from the ratio of diaphragm EMG to tidal volume. Respir Physiol Neurobiol. 2021;291:103692. doi:10.1016/j.resp.2021.103692
18. Wang L, Wu S, He B, et al. Exercise-induced bronchoconstriction assessed by a ratio of surface diaphragm EMG to tidal volume. Physiol Rep. 2023;11(21):e15860. doi:10.14814/phy2.15860
19. Chen R, Collins S, Remtulla H, et al. Phrenic nerve conduction study in normal subjects. Muscle Nerve. 1995;18(3):330–335. doi:10.1002/mus.880180311
20. Luo YM, Polkey MI, Johnson LC, et al. Diaphragm EMG measured by cervical magnetic and electrical phrenic nerve stimulation. J Appl Physiol. 1998;85(6):2089–2099. doi:10.1152/jappl.1998.85.6.2089
21. Duiverman ML, van Eykern LA, Vennik PW, et al. Reproducibility and responsiveness of a noninvasive EMG technique of the respiratory muscles in COPD patients and in healthy subjects. J Appl Physiol. 2004;96(5):1723–1729. doi:10.1152/japplphysiol.00914.2003
22. Graham BL, Steenbruggen I, Miller MR, et al. Standardization of Spirometry 2019 Update. An Official American Thoracic Society and European Respiratory Society Technical Statement. Am J Respir Crit Care Med. 2019;200(8):e70–e88. doi:10.1164/rccm.201908-1590ST
23. Luo YM, Tang J, Jolley C, et al. Distinguishing obstructive from central sleep apnea events: diaphragm electromyogram and esophageal pressure compared. Chest. 2009;135(5):1133–1141. doi:10.1378/chest.08-1695
24. Qin YY, Steier J, Jolley C, et al. Efficiency of neural drive during exercise in patients with COPD and healthy subjects. Chest. 2010;138(6):1309–1315. doi:10.1378/chest.09-2824
25. Luo YM, He B-T, Wu Y-X, et al. Neural respiratory drive and ventilation in patients with chronic obstructive pulmonary disease during sleep. Am J Respir Crit Care Med. 2014;190(2):227–229. doi:10.1164/rccm.201402-0302LE
26. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44(3):837–845. doi:10.2307/2531595
27. Crenesse D, Berlioz M, Bourrier T, et al. Spirometry in children aged 3 to 5 years: reliability of forced expiratory maneuvers. Pediatr Pulmonol. 2001;32(1):56–61. doi:10.1002/ppul.1089
28. Lee JX, Ryan M, Mukharesh L, et al. Comparison of home-based spirometry and oscillometry measurements in school-age children with bronchopulmonary dysplasia. Pediatr Pulmonol. 2024;59(10):2589–2596. doi:10.1002/ppul.27072
29. Dacha S, Janssens L, Rodrigues A, et al. Comparison between manual and (semi-)automated analyses of esophageal diaphragm electromyography during endurance cycling in patients with COPD. Front Physiol. 2019;10:885. doi:10.3389/fphys.2019.00885
30. Reilly CC, Jolley CJ, Ward K, et al. Neural respiratory drive measured during inspiratory threshold loading and acute hypercapnia in healthy individuals. Exp Physiol. 2013;98(7):1190–1198. doi:10.1113/expphysiol.2012.071415
31. Lin L, Guan L, Wu W, et al. Correlation of surface respiratory electromyography with esophageal diaphragm electromyography. Respir Physiol Neurobiol. 2019;259:45–52. doi:10.1016/j.resp.2018.07.004
32. Maarsingh EJ, van Eykern LA, Sprikkelman AB, et al. Histamine induced airway response in pre-school children assessed by a non-invasive EMG technique. Respir Med. 2004;98(4):363–372. doi:10.1016/j.rmed.2003.10.014