- Email: [email protected]
Assessment of Breathing Patterns and Respiratory Muscle Recruitment During Singing and Speech in Quadriplegia
250
ORIGINAL ARTICLE
Assessment of Breathing Patterns and Respiratory Muscle Recruitment During Singing and Speech in Quadriplegia Jeanette Tamplin, MMus, Danny J. Brazzale, BAppSc, Jeffrey J. Pretto, BAppSc, Warren R. Ruehland, BSc, Mary Buttifant, BAppSc (SpPath), Douglas J. Brown, MBBS, David J. Berlowitz, PhD ABSTRACT. Tamplin J, Brazzale DJ, Pretto JJ, Ruehland WR, Buttifant M, Brown DJ, Berlowitz DJ. Assessment of breathing patterns and respiratory muscle recruitment during singing and speech in quadriplegia. Arch Phys Med Rehabil 2011;92:250-6.
Key Words: Quadriplegia; Rehabilitation; Respiratory muscles; Speech; Spinal cord injuries; Voice. © 2011 by the American Congress of Rehabilitation Medicine
Objectives: To explore how respiratory impairment after cervical spinal cord injury affects vocal function, and to explore muscle recruitment strategies used during vocal tasks after quadriplegia. It was hypothesized that to achieve the increased respiratory support required for singing and loud speech, people with quadriplegia use different patterns of muscle recruitment and control strategies compared with control subjects without spinal cord injury. Design: Matched, parallel-group design. Setting: Large university-affiliated public hospital. Participants: Consenting participants with motor-complete C5-7 quadriplegia (n⫽6) and able-bodied age-matched controls (n⫽6) were assessed on physiologic and voice measures during vocal tasks. Interventions: Not applicable. Main Outcome Measures: Standard respiratory function testing, surface electromyographic activity from accessory respiratory muscles, sound pressure levels during vocal tasks, the Voice Handicap Index, and the Perceptual Voice Profile. Results: The group with quadriplegia had a reduced lung capacity (vital capacity, 71% vs 102% of predicted; P⫽.028), more perceived voice problems (Voice Handicap Index score, 22.5 vs 6.5; P⫽.046), and greater recruitment of accessory respiratory muscles during both loud and soft volumes (P⫽.028) than the able-bodied controls. The group with quadriplegia also demonstrated higher accessory muscle activation in changing from soft to loud speech (P⫽.028). Conclusions: People with quadriplegia have impaired vocal ability and use different muscle recruitment strategies during speech than the able-bodied. These findings will enable us to target specific measurements of respiratory physiology for assessing functional improvements in response to formal therapeutic singing training.
ESPIRATORY DYSFUNCTION resulting from SCI is a R major cause of morbidity, mortality, and economic burden. Weak or paralyzed respiratory muscles result in reduced
From the Institute for Breathing and Sleep (Tamplin, Brazzale, Pretto, Ruehland, Brown, Berlowitz), Victorian Spinal Cord Service (Tamplin, Brown), and Voice Analysis Clinic, (Buttifant), Austin Health, Melbourne, Australia; University of Melbourne, Australia (Tamplin, Brown, Berlowitz); and Department of Respiratory and Sleep Medicine, John Hunter Hospital, NSW, Australia (Pretto). Presented to the European Respiratory Society, September 2010, Barcelona, Spain. Supported by the Victorian Neurotrauma Initiative (grant no. D066). No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated. Clinical Trial Registration Number: ANZCTRN12607000306415. Correspondence to Jeanette Tamplin, MMus, Research Music Therapist, Institute for Breathing and Sleep, Bowen Centre, Austin Hospital, 145 Studley Rd, Heidelberg, Vic 3084, e-mail: [email protected]. Reprints are not available from the author. 0003-9993/11/9202-00598$36.00/0 doi:10.1016/j.apmr.2010.10.032
Arch Phys Med Rehabil Vol 92, February 2011
1
lung volume, ineffective cough, increased number of respiratory tract infections, and reduced chest wall compliance.2,3 In addition, respiratory dysfunction can negatively affect vocal production by reducing vocal loudness and decreasing phonation length.4 In able-bodied people during quiet breathing, inspiration is performed by contraction of the diaphragm, and expiration is essentially passive.5 As ventilatory demands increase, such as during singing or speaking, accessory muscles of inspiration and expiration may be called on. Neck muscles such as the sternocleidomastoid and scalenes may provide inspiratory assistance, whereas expiration can be assisted by abdominal muscles.6 Paralysis of the intercostals in quadriplegia means that inspiration is reliant on only the movement of the diaphragm and the accessory muscles around the neck.3,7 Both sternocleidomastoid and trapezius are important inspiratory muscles in quadriplegia. Contraction of the sternocleidomastoid can elevate the upper rib cage when the head is held fixed by the upper fibers of the trapezius.8 Paralysis of the abdominal muscles in quadriplegia also means that the only avenue for active expiration (to increase expiratory length and voice volume) is provided by the clavicular head of pectoralis major.9 In other patient groups with respiratory compromise (Duchenne muscular dystrophy, spinal muscular atrophy, amyotrophic lateral sclerosis, chronic obstructive pulmonary disease), accessory muscle recruitment in the neck is common during demanding tasks such as exercise10-16 or pursed-lip breathing.17 This contrasts with the situation in quadriplegia where accessory respiratory muscles are commonly recruited at rest,18 and to an even greater degree during exertion.19 Anecdotal reports have suggested that singing may have positive effects on respiratory function in this population,20 and because accessory respiratory muscles are commonly recruited during singing,21 we postulated that therapeutic singing training may aid subjects with quadriplegia to develop effective breathing strategies.22 In addition, a professional singer with C5 quadriplegia had previously been observed by the authors to
List of Abbreviations ASIA AUC EMG SCI SPL
American Spinal Injury Association area under the curve electromyogram spinal cord injury sound pressure level
RESPIRATORY MUSCLE RECRUITMENT AFTER SPINAL CORD INJURY, Tamplin
251
Fig 1. The anatomic positions of electrode placement on the sternocleidomastoid (SCM), diaphragm (D), and trapezius (TR) muscles. Transducer bands were positioned around chest and abdomen.
use innovative postures and unusual muscle recruitment patterns to achieve greater expiratory pressures for singing and voice projection. Singing training may not only help with phonation and vocal loudness, but may also assist in development of respiratory function and thereby provide beneficial effects in activities of daily living. We are unaware of any published controlled clinical trials in this area. The aim of this study was to develop measurement techniques for quantitative assessment of breathing patterns and the use of accessory respiratory muscles in quadriplegia for singing and speech. It was hypothesized that to achieve the increased respiratory support required for singing and loud speech, people with quadriplegia may use different patterns of muscle recruitment and control strategies than do control subjects without SCI. The assessment methodology used in the current study has been outlined previously.22 In the current article, we document a range of physiologic and voice measures taken during vocal tasks for people with C5-7 quadriplegia and compare these measurements with those for able-bodied matched controls. These results will provide baseline descriptive data about breathing strategies used during vocalization in quadriplegia. METHODS Participants A matched, parallel-group design was used to assess respiratory muscle recruitment for participants with C5-7 quadriplegia (n⫽6) and able-bodied matched controls (n⫽6) during vocal tasks. Potential participants with quadriplegia (⬎1 year postinjury) were recruited from the Victorian Spinal Cord Service database (Victoria, Australia). Convenience sampling was used to recruit control participants, where consenting participants with quadriplegia were invited to bring a friend of the same age, sex, and level of singing experience to act as their control. All participants were English speaking, in stable general health at the time of assessment, and able to travel for the assessments. Exclusion criteria were a history of speech disorder, tracheostomy, respiratory disease, psychiatric disorder, or neurologic impairment before the SCI. The project was approved by the institutional Human Research Ethics Committee and is registered at the Australian New Zealand Clinical Trials
Registry (www.anzctr.org.au). We certify that all applicable institutional and governmental regulations concerning the ethical use of human volunteers were followed during the course of this research. Measurements All participants participated in a single assessment session of 45 to 60 minutes, consisting of (1) completion of the Voice Handicap Index,23 (2) respiratory function tests, and (3) vocal assessment. Respiratory function tests were conducted according to the guidelines of the American Thoracic Society24,25 and modified to incorporate the limitations associated with SCI.26,27 Ventilatory function and upper airway function were assessed using maximal flow-volume loops to determine maximal inspiratory and expiratory flow rates and timed lung volumes with an EasyOne spirometer.a Static lung volumes were measured using inert-gas dilution via a P.K. Morgan M8 rolling seal spirometer with a helium analyzer.b Respiratory muscle strength was assessed by measuring maximal inspiratory pressure, maximal expiratory pressure, and sniff nasal inspiratory pressures28 using a portable MicroRPM respiratory pressure meter.c During the vocal assessment, surface electromyography, respiratory inductance plethysmography, and audio data were collected. Surface electromyography was used to measure accessory respiratory muscle function during vocal tasks. Surface EMG signals from the sternocleidomastoid, trapezius, and diaphragm muscles (fig 1) were amplified using the Compumedics S series sleep systemd and acquired using the Micro1401 data acquisition system and Spike2 software.e The respiratory inductance plethysmography data were collected via Inductotrace transducer bandsf around the participant’s chest and abdomen, which recorded relative contributions from the thoracic and abdominal compartments to ventilation during vocal tasks. Audio signals were recorded using a calibrated Ono Sokki MI-1211 Type 1 condenser microphoneg positioned at a distance of 30cm from the participant’s mouth and analyzed using an EASERA software analyzer.h Protocol For the vocal assessment, each participant was seated in a soundproofed room. An investigator directed participants Arch Phys Med Rehabil Vol 92, February 2011
252
RESPIRATORY MUSCLE RECRUITMENT AFTER SPINAL CORD INJURY, Tamplin Table 1: Demographic Information (Including Voice and Respiratory Function Data) for All Participants
2
BMI (kg/m ) VHI score FEV1 (% predicted) FVC (% predicted) MEP (% predicted) MIP (% predicted) SNIP (% predicted) FRC (% predicted) TLC (% predicted) RV (% predicted)
Quadriplegia (n⫽6)
Able-Bodied (n⫽6)
Wilcoxon Signed-Ranks Tests P
26.1 (23.4, 27.6) 22.5 (8.5, 35.8) 61 (55, 87) 71 (46, 83) 67 (56, 115) 87 (63, 98) 42 (25, 47) 69 (63, 88) 78 (73, 84) 105 (85, 126)
26.2 (22.9, 31.3) 6.5 (0.8, 10) 101 (88, 108) 102 (93, 107) 96 (88, 120) 112 (79, 121) 91 (67, 102) 64 (55, 80) 100 (91, 103) 79 (71, 89)
.753 .046 .028 .028 .028 .173 .028 .345 .028 .028
NOTE. Values are median (interquartile range) or as otherwise indicated. Abbreviations: BMI, body mass index; FEV1, forced expiratory volume in 1 second; FRC, functional residual capacity; FVC, forced vital capacity; MEP, maximum expiratory pressure; MIP, maximum inspiratory pressure; RV, residual volume; SNIP, sniff nasal inspiratory pressure; TLC, total lung capacity; VHI, Voice Handicap Index.
through a protocol consisting of phonatory exercises, standardized reading passages, and familiar songs. “Happy Birthday” was chosen as a well-known song to be spoken as well as sung to ensure linguistic control for the speech versus singing conditions. SPL analyses were conducted on the following conditions: “Rainbow Passage”29 with and without background noise, loud versus soft vowels, and speech versus singing. Background noise was delivered to each participant via headphones and was standardized to the SPL recorded during normative speaking conditions (signal-to-noise ratio, 0).
Analysis The electromyographic data were rectified and smoothed with a 0.1-second moving window.30 Three parameters were derived from the rectified and smoothed electromyographic data: background amplitude, peak amplitude, and AUC. The background amplitude is the size in microvolts of the average baseline activity between events (ie, in the absence of phonation). The peak amplitude is the largest amplitude of muscle activity during the generation of sound and was expressed as a
Fig 2. Waveform representation of raw electromyographic signals during loud and soft vocal tasks for a matched pair of participants. Control participant (left) and participant with quadriplegia (right).
Arch Phys Med Rehabil Vol 92, February 2011
253
RESPIRATORY MUSCLE RECRUITMENT AFTER SPINAL CORD INJURY, Tamplin Table 2: Median Differences in Peak Amplitude Data During Soft and Loud Vocal Tasks†. Difference in Peak Amplitude
SCM Trapezius Diaphragm
Loud versus Soft
Quadriplegia versus Control
Q: Loud – Soft
C: Loud – Soft
Loud: Q – C
Soft: Q – C
233* (104, 456) 81* (52, 121) 90 (–263, 307)
126* (3, 244) 0 (0, 8) 82 (0, 119)
543* (301, 856) 168* (150, 184) 199* (99, 301)
259* (202, 627) 90* (77, 106) 58 (–50, 632)
NOTE. Values are median (interquartile range). Abbreviations: Q, Quadriplegic; C, Control; SCM, sternocleidomastoid. *Denotes a statistically significant difference (P⬍.05) determined via the Wilcoxon signed-rank test. † Peak amplitude expressed as percentage increase from baseline.
percentage of baseline.31 To normalize the AUC for background amplitude and task completion time, background amplitude was multiplied by task completion time and subtracted from the total AUC to give an AUC measure corrected for time. An acoustic engineer extracted acoustic parameters from the vocal data (pitch, amplitude, spectral characteristics) using ADA EASERA software for analysis. Perceptual voice assessments were conducted by a speech pathologist using Oates and Russell’s Perceptual Voice Profile.32 Recordings were deidentified to ensure blinding during these analyses. Data are summarized as median (interquartile range). We used nonparametric statistics (Wilcoxon signed-rank) to analyze the results. RESULTS Demographic and Respiratory Function Data Six participants with quadriplegia (4 men) and 6 controls matched for age and sex were recruited to the study; their age range was 28 to 62 years (table 1). The 6 participants with quadriplegia had lesion levels ranging from C5 to C7 (C5, n⫽3; C6, n⫽1; C7, n⫽2). All had a complete SCI (ASIA A) apart from 1 C5 participant who was ASIA B (motor complete, sensory incomplete). As anticipated, the group of participants with quadriplegia exhibited smaller lung capacities (P⫽.028) and reduced respiratory pressures (P⫽.028). The group of participants with quadriplegia also reported significantly greater vocal impairment as measured by the Voice Handicap Index (P⫽.046). Electromyographic Results: Soft Versus Loud Speech The most significant result from the electromyographic analyses was found in peak amplitude differences during the loud versus soft vocal task. This difference is visible in the raw electromyographic data both within and between participants (fig 2). Quantitative measures derived from the integrated electromyographic data demonstrated that a significantly higher peak amplitude was recorded from participants with quadriplegia during loud and soft speech in comparison with control
participants (table 2). We also recorded a significantly higher peak amplitude (from both sternocleidomastoid and trapezius muscles) during loud speech versus soft speech by participants with quadriplegia, whereas control participants only exhibited a higher peak amplitude from the sternocleidomastoid muscle during loud speech. The only significant difference in diaphragm peak amplitude was observed between participants with quadriplegia, and control participants during loud speech. The AUC analysis of the electromyographic data demonstrated significantly higher activation of the sternocleidomastoids and diaphragm during loud (in comparison with soft) vocalization for the control group, and significantly higher activation of the trapezius for the group of participants with quadriplegia (table 3). In terms of between-group comparisons, AUC measures were significantly greater from the sternocleidomastoid during soft speech for participants with quadriplegia, and significantly greater from the diaphragm during loud speech for control participants. The appearance of “preparatory peak events” during analyses (fig 3) was an unexpected finding. These preparatory peak events were defined as the electromyographic amplitude above background amplitude before the onset of phonation, and they occurred almost exclusively in the group of participants with quadriplegia before both soft and loud vocal tasks (fig 4). Electromyographic Results: Singing Versus Speech There were no statistically significant differences in peak amplitude within or between groups in the singing versus speech task (data not shown). The AUC was significantly higher during singing than speech for all muscles measured in the control group, but only for the diaphragm in the group of participants with quadriplegia (table 4). No difference in AUC was observed between the study groups. As with the loud versus soft results, preparatory peak events occurred almost exclusively in the group of participants with quadriplegia before both speech and singing (see fig 4). The sung version of “Happy Birthday” took substantially longer than the spoken version for both groups (participants with quadriplegia: 11.4s vs 8.3s, P⫽.028; control participants: 10.3s vs 7.3s, P⫽.046).
Table 3: Median Differences in AUC Data During Soft and Loud Vocal Tasks. Loud versus Soft Difference in AUC
SCM Trapezius Diaphragm
Quadriplegia versus Control
Q: Loud – Soft
C: Loud – Soft
Loud: Q – C
Soft: Q – C
2 (–6, 24) 7* (4, 20) 3 (–2, 8)
6* (2, 8) 0 (–3, 1) 3* (1, 11)
18 (–3, 50) 7 (1, 18) 17* (9, 51)
13* (3, 38) –1 (–9, 7) 21 (3, 69)
NOTE. Values are median (interquartile range). Abbreviations: Q, Quadriplegic; C, Control; SCM, sternocleidomastoid. *Denotes a statistically significant difference (P⬍0.05) determined via the Wilcoxon signed- rank test.
Arch Phys Med Rehabil Vol 92, February 2011
254
RESPIRATORY MUSCLE RECRUITMENT AFTER SPINAL CORD INJURY, Tamplin
Fig 3. Waveform representation of rectified and smoothed electromyographic signals during loud and soft vocal tasks for a matched participant with quadriplegia (right) and control participant (left) pair.
Voice Analysis Results No statistically significant differences in SPL were found between participants with quadriplegia and control participants for any of these vocal tasks. All participants demonstrated a statistically significant within-subject increase in speech projection as anticipated according to the Lombard effect33 (tendency to increase vocal intensity in response to increased background noise) when subjected to noise in the headphones. Dynamic range was determined by comparing the SPL difference between impulsive loud and soft vowels. Collectively, control participants had a wider dynamic range (median, 18dBA) in comparison with participants with quadriplegia (median, 15dBA), although this difference was not statistically significant (P⫽.345). Control participants were able to sustain a vowel for a longer period (median, 16.4s) than participants with quadriplegia (median, 15.2s). This difference was not significant (P⫽.753), and both the control group and the par-
Fig 4. Percentage of anticipatory peak events occurring before soft and loud vocal tasks and speech and singing tasks. Abbreviation: SCM, sternocleidomastoid.
Arch Phys Med Rehabil Vol 92, February 2011
ticipants with quadriplegia displayed marked variability between subjects (range, 5–24.9s and 5.2–19.9s, respectively). Both groups demonstrated an increased SPL when singing in comparison with speaking the lyrics to “Happy Birthday” (median 4dBA increase for both the control group and participants with quadriplegia); however, this difference was not statistically significant (P⫽.715). The perceptual voice analyses confirmed the findings of the acoustic analyses above, suggesting that voice quality was within normative limits for all participants. However, 4 participants with quadriplegia were rated with mild to moderately impaired voice quality in the areas of breathiness, strain, and roughness. DISCUSSION Muscle Recruitment Patterns This study found that people with quadriplegia recruited accessory respiratory muscles (sternocleidomastoid and trapezius) to increase vocal intensity more than able-bodied control participants did during speech. The peak electromyographic values were significantly higher across all the accessory muscles in participants with quadriplegia compared with their able-bodied counterparts. In addition, the peak electromyographic amplitude was higher in the group of participants with quadriplegia during more demanding (loud vs soft) vocal tasks (statistically significant in trapezius only). In the control group, there was a statistically significant increase in sternocleidomastoid activity when increasing from soft to loud. These data are consistent with the premise that people with quadriplegia have a limited amount of “muscle reserve” that they can call on, and thus recruit their available muscle capacity in a burst of activation. Able-bodied people may potentially use a similar amount of muscle activation spread more evenly over the course of vocalization. Examination of the shape of both the raw
255
RESPIRATORY MUSCLE RECRUITMENT AFTER SPINAL CORD INJURY, Tamplin Table 4: Median Differences in AUC Data During Speech and Singing Tasks Singing versus Speech Difference in AUC
SCM Trapezius Diaphragm
Quadriplegia versus Control
Quadriplegia
Control
Speech
Singing
–4.9 (–12.1, 1) 4.4 (1.2, 8.3) 20.8* (10.1, 58.7)
4* (–3.8, 7.8) 18.8* (–8, 36.2) 14.9* (5, 30.6)
18.7 (–11.5, 40.7) 7.3 (–12.7, 29.2) 32.8 (–8, 71.6)
6.6 (–11.7, 71.1) –18.7 (–23.2, 6.6) 40.3 (0.9, 93.2)
NOTE. Values are median (interquartile range). Abbreviation: SCM, sternocleidomastoid. *Denotes a statistically significant difference (P⬍.05) determined via the Wilcoxon signed-rank test.
(see fig 2) and smoothed (see fig 3) EMG and other respiratory traces would suggest this was the case. The significantly greater sternocleidomastoid activation demonstrated in all electromyographic measures by participants with quadriplegia during soft speech suggests that they have to work harder even to achieve vocalization at a low volume. In addition, the unexpected occurrence and high incidence of preparatory peak events (see fig 4) before both loud and soft vocal tasks (and singing) supports this notion that people with quadriplegia may require more intense activation of their accessory respiratory muscles than able-bodied controls both in preparation for, and during vocalization. In contrast with the loud versus soft task, no difference in peak electromyographic amplitude was observed in either group during singing versus speech. The control group increased the AUC in all muscles when singing, but only the diaphragm AUC increased for the group of participants with quadriplegia. There were no between-group differences observed for either singing or speech tasks. We observed a larger AUC in trapezius and diaphragm muscles within both participant groups during singing (see table 4). These results are consistent with previous research examining muscle activation by able-bodied participants during singing. Pettersen et al34 and Pettersen and Westgaard35 reported that during singing, the trapezius contributes to the compression of the upper thorax, thus serving as an accessory muscle of expiration. In the participants with quadriplegia, it is possible that the lack of an increase in activation for the higher effort task (singing) occurred because they were already recruiting most of their available muscle capacity during the speech task and thus had nothing further to recruit. However, considerable effort from the participants with quadriplegia was observable through the high incidence of preparatory muscle contractions. Further controlled trials are required to replicate these results and to evaluate the clinical significance of these differences. Respiratory Function and Voice Assessments The group of participants with quadriplegia scored significantly lower than their matched controls on most respiratory measures, fitting with the pattern of respiratory impairment reported after SCI in previous research.2,3 However, the reported perception of impaired voice production by these participants (as measured by the Voice Handicap Index) is an interesting and novel finding of this study. The lack of a statistically significant difference in SPL between singing and speech is interesting and was contrary to what was expected. Although there was a 4-dBA increase for both groups, this increase was not statistically significant, which may suggest that the study was underpowered. This result may also have occurred because “Happy Birthday” is ingrained in most people as a song and as such was not representative of normal running speech. The use of rhythmic emphasis and stress when speaking the lyrics increased the
overall SPL and thus may have contributed to the lack of a significant difference between speaking and singing. Study Limitations Several methodological considerations must be addressed when assessing our conclusions. A significant limitation of this study is the small sample size; however, we have been conservative accordingly in our use of nonparametric analyses. The relationship between electrical activity (EMG) and muscle contraction is complex and cannot be directly assessed. Thus, there is no way to confirm that the increased electromyographic activity observed in participants does in fact correlate with increased muscle force. Surface electromyographic recordings of respiratory muscle activity have previously been reported in able-bodied subjects during speech36 and singing.34,35 Although there is no consistent, universal relationship between muscle force and dynamic electromyographic data, it has been demonstrated that the integrated EMG is linearly correlated with isometric force under static conditions.37 One method for quantifying surface electromyographic measurements involves the measurement of a maximal muscle contraction and then representing the data as a percentage of this maximal value.38 We did not use this method because it is difficult to perform a task requiring absolute maximal effort of respiratory muscles, accessory muscles, or both, and so we could not reliably achieve a maximal contraction of the muscles assessed. In addition, it was not possible to isolate diaphragm activity from abdominal muscle activity in the able-bodied subjects using the surface electromyographic electrodes. This expiratory muscle activity of the abdominals in the able-bodied controls may have masked any differences between the groups, particularly during the more intensive (loud speech and singing) tasks. A further potentially confounding factor in the electromyographic analysis is that we could not guarantee that electromyographic surface electrode placement was precisely the same in all participants. Differences in electrode position may explain some of the variability between participants. However, given that most of our analyses involved assessment of changes in measurements within each subject, this issue is thought to be of reduced importance. Finally, the use of convenience sampling ensured wellmatched control participants, but also introduced bias and limits generalization of the findings, as control participants were not randomly drawn from the general population. CONCLUSIONS In summary, the results of this study showed that the participants with quadriplegia exhibited the expected respiratory impairments and a self-perception of vocal impairment. Participants with quadriplegia appeared to recruit accessory respiratory muscles more when speaking and singing (particularly at louder volumes) than their matched able-bodied controls. Participants with quadriplegia also made unexpected, preparatory respiratory efforts that were not observed in their able-bodied Arch Phys Med Rehabil Vol 92, February 2011
256
RESPIRATORY MUSCLE RECRUITMENT AFTER SPINAL CORD INJURY, Tamplin
peers. The role of these preparatory efforts is unclear and warrants further investigation. The participants with quadriplegia could speak and sing adequately (and at comparable intensity levels to their controls), but appeared to use unusual muscle recruitment techniques to compensate for respiratory impairments. Future research should examine whether there may be a role for further development and training of these atypical muscle recruitment techniques to enhance respiratory and vocal function for people with quadriplegia. Acknowledgments: We thank Darren Tardio (Vipac Engineers) for his technical assistance, and Matthew Absalom-Wong for his artistic expertise. References 1. DeVivo MJ, Krause S, Lammertse DP. Recent trends in mortality and causes of death among persons with spinal cord injury. Arch Phys Med Rehabil 1999;80:1411-9. 2. Brown R, DiMarco AF, Hoit JD, Garshick E. Respiratory dysfunction and management in spinal cord injury. Respir Care 2006;52:853-68. 3. Chen CF, Lien IN, Wu MC. Respiratory function in patients with spinal cord injuries: effects of posture. Paraplegia 1990;28:81-6. 4. MacBean N, Ward E, Murdoch BE, Cahill L, Salley M, Geraghty T. Characteristics of speech following cervical spinal cord injury. J Med Speech Lang Pathol 2006;14:167-84. 5. Bunch M. Dynamics of the singing voice. 2nd ed. New York: Springer-Verlag Wien; 1993. 6. Shaffer TD, Wolfson MR, Bhutani VK. Respiratory muscle function, assessment and training. Phys Ther 1981;61:1711-23. 7. Frisbie JH. Breathing pattern in tetraplegic patients. Spinal Cord 2002;40:424-5. 8. Winslow C, Rozovsky J. Effect of spinal cord injury on the respiratory system. Am J Phys Med Rehabil 2003;82:803-14. 9. Estenne M, Knoop C, Van Vaerenbergh J, Heilporn A, De Troyer A. The effect of pectoralis muscle training in tetraplegic subjects. Am Rev Respir Dis 1989;139:1218-22. 10. DePalo VA, McCool FD. Respiratory muscle evaluation of the patient with neuromuscular disease. Semin Respir Crit Care Med 2002;22:201-10. 11. Wohlgemuth M, van der Kooi EL, van Kesteren RG, van der Maarel SM, Padberg GW. Ventilatory support in facioscapulohumeral muscular dystrophy. Neurology 2004;63:176-8. 12. Misuri G, Lanini B, Gigliotti F, et al. Mechanism of CO2 retention in patients with neuromuscular disease. Chest 2000;117:447-53. 13. Delgado H, Braun S, Skatrud B, Reddan W, Pegelow D. Chest wall and abdominal motion during exercise in patients with chronic obstructive pulmonary disease. Am Rev Respir Dis 1982; 126:200-5. 14. Howard RS, Wiles CM, Loh L. Respiratory complications and their management in motor neuron disease. Brain 1989;112:1155-70. 15. Breslin EH. Breathing retraining in chronic obstructive pulmonary disease. J Cardiopulm Rehabil 1995;15:25-33. 16. Epstein SK, Celli BR, Williams J, Tarpy S, Roa J, Shannon T. Ventilatory response to arm elevation. Its determinants and use in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1995;152:211-6. 17. Breslin EH. The pattern of respiratory muscle recruitment during pursed-lip breathing. Chest 1992;101:75-8. 18. De Troyer A, Peche R, Yernault JC, Estenne M. Neck muscle activity in patients with severe chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1994;150:41-7. 19. Fujiwara T, Hara Y, Chino N. Expiratory function in complete tetraplegics: study of spirometry, maximal expiratory pressure, and muscle activity of pectoralis major and latissimus dorsi muscles. Am J Phys Med Rehabil 1999;78:464-73.
Arch Phys Med Rehabil Vol 92, February 2011
20. Batavia AI, Batavia M. Karaoke for quads: a new application of an old recreation with potential therapeutic benefits for people with disabilities. Disabil Rehabil 2003;25:297-300. 21. Mendes AP, Brown WS, Sapienza C, Rothman HB. Effects of vocal training on respiratory kinematics during singing tasks. Folia Phoniatr Logop 2006;58:363-77. 22. Tamplin J. The link between singing and respiratory health for people with quadriplegia. Aust J Music Ther 2009;20(special issue):45-55. 23. Jacobson BH, Johnson A, Grywalski C, et al. The Voice Handicap Index (VHI): development and validation. Am J Speech Lang Pathol 1997;6:66-70. 24. ATS/ERS statement on respiratory muscle testing. Am J Respir Crit Care Med 2002;166:518-624. 25. Miller MR, Hankinson JL, Brusasco V, et al. Standardisation of spirometry. Eur Respir J 2005;26:319-38. 26. Ashba J, Garshick E, Tun CG, et al. Spirometry—acceptability and reproducibility in spinal cord injured patients. J Am Paraplegia Soc 1993;16:197-203. 27. Kelley A, Garshick E, Gross ER, Lieberman SL, Tun CG, Brown R. Spirometry testing standards in spinal cord injury. Chest 2003; 123:725-30. 28. Black LF, Hyatt RE. Maximal static respiratory pressures in generalized neuromuscular disease. Am Rev Respir Dis 1971;103:641-50. 29. Fairbanks G. Voice and articulation drillbook. 2nd ed. New York: Harper and Row; 1960. p 196. 30. Gamet D, Fokapu O. Electromyography. In: Wnek GE, Bowlin GL, editors. Encyclopedia of biomaterials and biomedical engineering. London: Informa Healthcare; 2007. p 956-67. 31. Eckert DJ, McEvoy RD, George KE, Thomson KJ, Catcheside PG. Genioglossus reflex inhibition to upper-airway negativepressure stimuli during wakefulness and sleep in healthy males. J Physiol 2007;581:1193-1205. 32. Oates JM, Russell A. A sound judgement. In: CD-ROM package. Melbourne: La Trobe University; 1997. 33. Winkworth AL, Davis PJ. Speech breathing and the Lombard effect. J Speech Hear Res 1997;40:159-69. 34. Pettersen V, Bjorkoy K, Torp H, Westgaard RH. Neck and shoulder muscle activity and thorax movement in singing and speaking tasks with variation in vocal loudness and pitch. J Voice 2005; 19:623-34. 35. Pettersen V, Westgaard RH. The activity patterns of neck muscles in professional classical singing. J Voice 2005;19:238-51. 36. MacFarland DH, Smith A. Surface recordings of respiratory muscle activity during speech: some preliminary findings. J Speech Hear Res 1989;32:657-67. 37. Wootten ME, Kadaba MP, Cochran GV. Dynamic electromyography. I. Numerical representation using principal component analysis. J Orthop Res 1990;8:247-58. 38. Mezzanotte WS, Tangel DJ, White DP. Waking genioglossal electromyogram in sleep apnea patients versus normal controls (a neuromuscular compensatory mechanism). J Clin Invest 1992;89:1571-9. a. b. c. d. e. f. g. h.
Suppliers OPS Medical, LLC, 8055 Ritchie Hwy, Ste 103, Pasadena, MD 21122. P.K. Morgan Ltd, Chatham, Kent, UK. Micro Medical Ltd, PO Box 6, Rochester, Kent ME1 2AZ, England. Compumedics Sleep Pty Ltd, 1 Marine Parade, Abbotsford, Victoria, 3067, Australia. Cambridge Electronic Design Ltd, 4 Science Park, Milton Rd, Cambridge, Cambridgeshire, CB4 0FE, England. Ambulatory Monitoring, Inc, 731 Saw Mill River Rd, Ardsley, NY 10502-0609. VIPAC Engineers & Scientists Ltd, Victorian Technology Centre, 275 Normanby Rd, Port Melbourne, Victoria, 3207, Australia. SDA Software Design Ahnert GmbH, Arkonastrae 45-49, 13189, Berlin, Germany.
ORIGINAL ARTICLE
Assessment of Breathing Patterns and Respiratory Muscle Recruitment During Singing and Speech in Quadriplegia Jeanette Tamplin, MMus, Danny J. Brazzale, BAppSc, Jeffrey J. Pretto, BAppSc, Warren R. Ruehland, BSc, Mary Buttifant, BAppSc (SpPath), Douglas J. Brown, MBBS, David J. Berlowitz, PhD ABSTRACT. Tamplin J, Brazzale DJ, Pretto JJ, Ruehland WR, Buttifant M, Brown DJ, Berlowitz DJ. Assessment of breathing patterns and respiratory muscle recruitment during singing and speech in quadriplegia. Arch Phys Med Rehabil 2011;92:250-6.
Key Words: Quadriplegia; Rehabilitation; Respiratory muscles; Speech; Spinal cord injuries; Voice. © 2011 by the American Congress of Rehabilitation Medicine
Objectives: To explore how respiratory impairment after cervical spinal cord injury affects vocal function, and to explore muscle recruitment strategies used during vocal tasks after quadriplegia. It was hypothesized that to achieve the increased respiratory support required for singing and loud speech, people with quadriplegia use different patterns of muscle recruitment and control strategies compared with control subjects without spinal cord injury. Design: Matched, parallel-group design. Setting: Large university-affiliated public hospital. Participants: Consenting participants with motor-complete C5-7 quadriplegia (n⫽6) and able-bodied age-matched controls (n⫽6) were assessed on physiologic and voice measures during vocal tasks. Interventions: Not applicable. Main Outcome Measures: Standard respiratory function testing, surface electromyographic activity from accessory respiratory muscles, sound pressure levels during vocal tasks, the Voice Handicap Index, and the Perceptual Voice Profile. Results: The group with quadriplegia had a reduced lung capacity (vital capacity, 71% vs 102% of predicted; P⫽.028), more perceived voice problems (Voice Handicap Index score, 22.5 vs 6.5; P⫽.046), and greater recruitment of accessory respiratory muscles during both loud and soft volumes (P⫽.028) than the able-bodied controls. The group with quadriplegia also demonstrated higher accessory muscle activation in changing from soft to loud speech (P⫽.028). Conclusions: People with quadriplegia have impaired vocal ability and use different muscle recruitment strategies during speech than the able-bodied. These findings will enable us to target specific measurements of respiratory physiology for assessing functional improvements in response to formal therapeutic singing training.
ESPIRATORY DYSFUNCTION resulting from SCI is a R major cause of morbidity, mortality, and economic burden. Weak or paralyzed respiratory muscles result in reduced
From the Institute for Breathing and Sleep (Tamplin, Brazzale, Pretto, Ruehland, Brown, Berlowitz), Victorian Spinal Cord Service (Tamplin, Brown), and Voice Analysis Clinic, (Buttifant), Austin Health, Melbourne, Australia; University of Melbourne, Australia (Tamplin, Brown, Berlowitz); and Department of Respiratory and Sleep Medicine, John Hunter Hospital, NSW, Australia (Pretto). Presented to the European Respiratory Society, September 2010, Barcelona, Spain. Supported by the Victorian Neurotrauma Initiative (grant no. D066). No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated. Clinical Trial Registration Number: ANZCTRN12607000306415. Correspondence to Jeanette Tamplin, MMus, Research Music Therapist, Institute for Breathing and Sleep, Bowen Centre, Austin Hospital, 145 Studley Rd, Heidelberg, Vic 3084, e-mail: [email protected]. Reprints are not available from the author. 0003-9993/11/9202-00598$36.00/0 doi:10.1016/j.apmr.2010.10.032
Arch Phys Med Rehabil Vol 92, February 2011
1
lung volume, ineffective cough, increased number of respiratory tract infections, and reduced chest wall compliance.2,3 In addition, respiratory dysfunction can negatively affect vocal production by reducing vocal loudness and decreasing phonation length.4 In able-bodied people during quiet breathing, inspiration is performed by contraction of the diaphragm, and expiration is essentially passive.5 As ventilatory demands increase, such as during singing or speaking, accessory muscles of inspiration and expiration may be called on. Neck muscles such as the sternocleidomastoid and scalenes may provide inspiratory assistance, whereas expiration can be assisted by abdominal muscles.6 Paralysis of the intercostals in quadriplegia means that inspiration is reliant on only the movement of the diaphragm and the accessory muscles around the neck.3,7 Both sternocleidomastoid and trapezius are important inspiratory muscles in quadriplegia. Contraction of the sternocleidomastoid can elevate the upper rib cage when the head is held fixed by the upper fibers of the trapezius.8 Paralysis of the abdominal muscles in quadriplegia also means that the only avenue for active expiration (to increase expiratory length and voice volume) is provided by the clavicular head of pectoralis major.9 In other patient groups with respiratory compromise (Duchenne muscular dystrophy, spinal muscular atrophy, amyotrophic lateral sclerosis, chronic obstructive pulmonary disease), accessory muscle recruitment in the neck is common during demanding tasks such as exercise10-16 or pursed-lip breathing.17 This contrasts with the situation in quadriplegia where accessory respiratory muscles are commonly recruited at rest,18 and to an even greater degree during exertion.19 Anecdotal reports have suggested that singing may have positive effects on respiratory function in this population,20 and because accessory respiratory muscles are commonly recruited during singing,21 we postulated that therapeutic singing training may aid subjects with quadriplegia to develop effective breathing strategies.22 In addition, a professional singer with C5 quadriplegia had previously been observed by the authors to
List of Abbreviations ASIA AUC EMG SCI SPL
American Spinal Injury Association area under the curve electromyogram spinal cord injury sound pressure level
RESPIRATORY MUSCLE RECRUITMENT AFTER SPINAL CORD INJURY, Tamplin
251
Fig 1. The anatomic positions of electrode placement on the sternocleidomastoid (SCM), diaphragm (D), and trapezius (TR) muscles. Transducer bands were positioned around chest and abdomen.
use innovative postures and unusual muscle recruitment patterns to achieve greater expiratory pressures for singing and voice projection. Singing training may not only help with phonation and vocal loudness, but may also assist in development of respiratory function and thereby provide beneficial effects in activities of daily living. We are unaware of any published controlled clinical trials in this area. The aim of this study was to develop measurement techniques for quantitative assessment of breathing patterns and the use of accessory respiratory muscles in quadriplegia for singing and speech. It was hypothesized that to achieve the increased respiratory support required for singing and loud speech, people with quadriplegia may use different patterns of muscle recruitment and control strategies than do control subjects without SCI. The assessment methodology used in the current study has been outlined previously.22 In the current article, we document a range of physiologic and voice measures taken during vocal tasks for people with C5-7 quadriplegia and compare these measurements with those for able-bodied matched controls. These results will provide baseline descriptive data about breathing strategies used during vocalization in quadriplegia. METHODS Participants A matched, parallel-group design was used to assess respiratory muscle recruitment for participants with C5-7 quadriplegia (n⫽6) and able-bodied matched controls (n⫽6) during vocal tasks. Potential participants with quadriplegia (⬎1 year postinjury) were recruited from the Victorian Spinal Cord Service database (Victoria, Australia). Convenience sampling was used to recruit control participants, where consenting participants with quadriplegia were invited to bring a friend of the same age, sex, and level of singing experience to act as their control. All participants were English speaking, in stable general health at the time of assessment, and able to travel for the assessments. Exclusion criteria were a history of speech disorder, tracheostomy, respiratory disease, psychiatric disorder, or neurologic impairment before the SCI. The project was approved by the institutional Human Research Ethics Committee and is registered at the Australian New Zealand Clinical Trials
Registry (www.anzctr.org.au). We certify that all applicable institutional and governmental regulations concerning the ethical use of human volunteers were followed during the course of this research. Measurements All participants participated in a single assessment session of 45 to 60 minutes, consisting of (1) completion of the Voice Handicap Index,23 (2) respiratory function tests, and (3) vocal assessment. Respiratory function tests were conducted according to the guidelines of the American Thoracic Society24,25 and modified to incorporate the limitations associated with SCI.26,27 Ventilatory function and upper airway function were assessed using maximal flow-volume loops to determine maximal inspiratory and expiratory flow rates and timed lung volumes with an EasyOne spirometer.a Static lung volumes were measured using inert-gas dilution via a P.K. Morgan M8 rolling seal spirometer with a helium analyzer.b Respiratory muscle strength was assessed by measuring maximal inspiratory pressure, maximal expiratory pressure, and sniff nasal inspiratory pressures28 using a portable MicroRPM respiratory pressure meter.c During the vocal assessment, surface electromyography, respiratory inductance plethysmography, and audio data were collected. Surface electromyography was used to measure accessory respiratory muscle function during vocal tasks. Surface EMG signals from the sternocleidomastoid, trapezius, and diaphragm muscles (fig 1) were amplified using the Compumedics S series sleep systemd and acquired using the Micro1401 data acquisition system and Spike2 software.e The respiratory inductance plethysmography data were collected via Inductotrace transducer bandsf around the participant’s chest and abdomen, which recorded relative contributions from the thoracic and abdominal compartments to ventilation during vocal tasks. Audio signals were recorded using a calibrated Ono Sokki MI-1211 Type 1 condenser microphoneg positioned at a distance of 30cm from the participant’s mouth and analyzed using an EASERA software analyzer.h Protocol For the vocal assessment, each participant was seated in a soundproofed room. An investigator directed participants Arch Phys Med Rehabil Vol 92, February 2011
252
RESPIRATORY MUSCLE RECRUITMENT AFTER SPINAL CORD INJURY, Tamplin Table 1: Demographic Information (Including Voice and Respiratory Function Data) for All Participants
2
BMI (kg/m ) VHI score FEV1 (% predicted) FVC (% predicted) MEP (% predicted) MIP (% predicted) SNIP (% predicted) FRC (% predicted) TLC (% predicted) RV (% predicted)
Quadriplegia (n⫽6)
Able-Bodied (n⫽6)
Wilcoxon Signed-Ranks Tests P
26.1 (23.4, 27.6) 22.5 (8.5, 35.8) 61 (55, 87) 71 (46, 83) 67 (56, 115) 87 (63, 98) 42 (25, 47) 69 (63, 88) 78 (73, 84) 105 (85, 126)
26.2 (22.9, 31.3) 6.5 (0.8, 10) 101 (88, 108) 102 (93, 107) 96 (88, 120) 112 (79, 121) 91 (67, 102) 64 (55, 80) 100 (91, 103) 79 (71, 89)
.753 .046 .028 .028 .028 .173 .028 .345 .028 .028
NOTE. Values are median (interquartile range) or as otherwise indicated. Abbreviations: BMI, body mass index; FEV1, forced expiratory volume in 1 second; FRC, functional residual capacity; FVC, forced vital capacity; MEP, maximum expiratory pressure; MIP, maximum inspiratory pressure; RV, residual volume; SNIP, sniff nasal inspiratory pressure; TLC, total lung capacity; VHI, Voice Handicap Index.
through a protocol consisting of phonatory exercises, standardized reading passages, and familiar songs. “Happy Birthday” was chosen as a well-known song to be spoken as well as sung to ensure linguistic control for the speech versus singing conditions. SPL analyses were conducted on the following conditions: “Rainbow Passage”29 with and without background noise, loud versus soft vowels, and speech versus singing. Background noise was delivered to each participant via headphones and was standardized to the SPL recorded during normative speaking conditions (signal-to-noise ratio, 0).
Analysis The electromyographic data were rectified and smoothed with a 0.1-second moving window.30 Three parameters were derived from the rectified and smoothed electromyographic data: background amplitude, peak amplitude, and AUC. The background amplitude is the size in microvolts of the average baseline activity between events (ie, in the absence of phonation). The peak amplitude is the largest amplitude of muscle activity during the generation of sound and was expressed as a
Fig 2. Waveform representation of raw electromyographic signals during loud and soft vocal tasks for a matched pair of participants. Control participant (left) and participant with quadriplegia (right).
Arch Phys Med Rehabil Vol 92, February 2011
253
RESPIRATORY MUSCLE RECRUITMENT AFTER SPINAL CORD INJURY, Tamplin Table 2: Median Differences in Peak Amplitude Data During Soft and Loud Vocal Tasks†. Difference in Peak Amplitude
SCM Trapezius Diaphragm
Loud versus Soft
Quadriplegia versus Control
Q: Loud – Soft
C: Loud – Soft
Loud: Q – C
Soft: Q – C
233* (104, 456) 81* (52, 121) 90 (–263, 307)
126* (3, 244) 0 (0, 8) 82 (0, 119)
543* (301, 856) 168* (150, 184) 199* (99, 301)
259* (202, 627) 90* (77, 106) 58 (–50, 632)
NOTE. Values are median (interquartile range). Abbreviations: Q, Quadriplegic; C, Control; SCM, sternocleidomastoid. *Denotes a statistically significant difference (P⬍.05) determined via the Wilcoxon signed-rank test. † Peak amplitude expressed as percentage increase from baseline.
percentage of baseline.31 To normalize the AUC for background amplitude and task completion time, background amplitude was multiplied by task completion time and subtracted from the total AUC to give an AUC measure corrected for time. An acoustic engineer extracted acoustic parameters from the vocal data (pitch, amplitude, spectral characteristics) using ADA EASERA software for analysis. Perceptual voice assessments were conducted by a speech pathologist using Oates and Russell’s Perceptual Voice Profile.32 Recordings were deidentified to ensure blinding during these analyses. Data are summarized as median (interquartile range). We used nonparametric statistics (Wilcoxon signed-rank) to analyze the results. RESULTS Demographic and Respiratory Function Data Six participants with quadriplegia (4 men) and 6 controls matched for age and sex were recruited to the study; their age range was 28 to 62 years (table 1). The 6 participants with quadriplegia had lesion levels ranging from C5 to C7 (C5, n⫽3; C6, n⫽1; C7, n⫽2). All had a complete SCI (ASIA A) apart from 1 C5 participant who was ASIA B (motor complete, sensory incomplete). As anticipated, the group of participants with quadriplegia exhibited smaller lung capacities (P⫽.028) and reduced respiratory pressures (P⫽.028). The group of participants with quadriplegia also reported significantly greater vocal impairment as measured by the Voice Handicap Index (P⫽.046). Electromyographic Results: Soft Versus Loud Speech The most significant result from the electromyographic analyses was found in peak amplitude differences during the loud versus soft vocal task. This difference is visible in the raw electromyographic data both within and between participants (fig 2). Quantitative measures derived from the integrated electromyographic data demonstrated that a significantly higher peak amplitude was recorded from participants with quadriplegia during loud and soft speech in comparison with control
participants (table 2). We also recorded a significantly higher peak amplitude (from both sternocleidomastoid and trapezius muscles) during loud speech versus soft speech by participants with quadriplegia, whereas control participants only exhibited a higher peak amplitude from the sternocleidomastoid muscle during loud speech. The only significant difference in diaphragm peak amplitude was observed between participants with quadriplegia, and control participants during loud speech. The AUC analysis of the electromyographic data demonstrated significantly higher activation of the sternocleidomastoids and diaphragm during loud (in comparison with soft) vocalization for the control group, and significantly higher activation of the trapezius for the group of participants with quadriplegia (table 3). In terms of between-group comparisons, AUC measures were significantly greater from the sternocleidomastoid during soft speech for participants with quadriplegia, and significantly greater from the diaphragm during loud speech for control participants. The appearance of “preparatory peak events” during analyses (fig 3) was an unexpected finding. These preparatory peak events were defined as the electromyographic amplitude above background amplitude before the onset of phonation, and they occurred almost exclusively in the group of participants with quadriplegia before both soft and loud vocal tasks (fig 4). Electromyographic Results: Singing Versus Speech There were no statistically significant differences in peak amplitude within or between groups in the singing versus speech task (data not shown). The AUC was significantly higher during singing than speech for all muscles measured in the control group, but only for the diaphragm in the group of participants with quadriplegia (table 4). No difference in AUC was observed between the study groups. As with the loud versus soft results, preparatory peak events occurred almost exclusively in the group of participants with quadriplegia before both speech and singing (see fig 4). The sung version of “Happy Birthday” took substantially longer than the spoken version for both groups (participants with quadriplegia: 11.4s vs 8.3s, P⫽.028; control participants: 10.3s vs 7.3s, P⫽.046).
Table 3: Median Differences in AUC Data During Soft and Loud Vocal Tasks. Loud versus Soft Difference in AUC
SCM Trapezius Diaphragm
Quadriplegia versus Control
Q: Loud – Soft
C: Loud – Soft
Loud: Q – C
Soft: Q – C
2 (–6, 24) 7* (4, 20) 3 (–2, 8)
6* (2, 8) 0 (–3, 1) 3* (1, 11)
18 (–3, 50) 7 (1, 18) 17* (9, 51)
13* (3, 38) –1 (–9, 7) 21 (3, 69)
NOTE. Values are median (interquartile range). Abbreviations: Q, Quadriplegic; C, Control; SCM, sternocleidomastoid. *Denotes a statistically significant difference (P⬍0.05) determined via the Wilcoxon signed- rank test.
Arch Phys Med Rehabil Vol 92, February 2011
254
RESPIRATORY MUSCLE RECRUITMENT AFTER SPINAL CORD INJURY, Tamplin
Fig 3. Waveform representation of rectified and smoothed electromyographic signals during loud and soft vocal tasks for a matched participant with quadriplegia (right) and control participant (left) pair.
Voice Analysis Results No statistically significant differences in SPL were found between participants with quadriplegia and control participants for any of these vocal tasks. All participants demonstrated a statistically significant within-subject increase in speech projection as anticipated according to the Lombard effect33 (tendency to increase vocal intensity in response to increased background noise) when subjected to noise in the headphones. Dynamic range was determined by comparing the SPL difference between impulsive loud and soft vowels. Collectively, control participants had a wider dynamic range (median, 18dBA) in comparison with participants with quadriplegia (median, 15dBA), although this difference was not statistically significant (P⫽.345). Control participants were able to sustain a vowel for a longer period (median, 16.4s) than participants with quadriplegia (median, 15.2s). This difference was not significant (P⫽.753), and both the control group and the par-
Fig 4. Percentage of anticipatory peak events occurring before soft and loud vocal tasks and speech and singing tasks. Abbreviation: SCM, sternocleidomastoid.
Arch Phys Med Rehabil Vol 92, February 2011
ticipants with quadriplegia displayed marked variability between subjects (range, 5–24.9s and 5.2–19.9s, respectively). Both groups demonstrated an increased SPL when singing in comparison with speaking the lyrics to “Happy Birthday” (median 4dBA increase for both the control group and participants with quadriplegia); however, this difference was not statistically significant (P⫽.715). The perceptual voice analyses confirmed the findings of the acoustic analyses above, suggesting that voice quality was within normative limits for all participants. However, 4 participants with quadriplegia were rated with mild to moderately impaired voice quality in the areas of breathiness, strain, and roughness. DISCUSSION Muscle Recruitment Patterns This study found that people with quadriplegia recruited accessory respiratory muscles (sternocleidomastoid and trapezius) to increase vocal intensity more than able-bodied control participants did during speech. The peak electromyographic values were significantly higher across all the accessory muscles in participants with quadriplegia compared with their able-bodied counterparts. In addition, the peak electromyographic amplitude was higher in the group of participants with quadriplegia during more demanding (loud vs soft) vocal tasks (statistically significant in trapezius only). In the control group, there was a statistically significant increase in sternocleidomastoid activity when increasing from soft to loud. These data are consistent with the premise that people with quadriplegia have a limited amount of “muscle reserve” that they can call on, and thus recruit their available muscle capacity in a burst of activation. Able-bodied people may potentially use a similar amount of muscle activation spread more evenly over the course of vocalization. Examination of the shape of both the raw
255
RESPIRATORY MUSCLE RECRUITMENT AFTER SPINAL CORD INJURY, Tamplin Table 4: Median Differences in AUC Data During Speech and Singing Tasks Singing versus Speech Difference in AUC
SCM Trapezius Diaphragm
Quadriplegia versus Control
Quadriplegia
Control
Speech
Singing
–4.9 (–12.1, 1) 4.4 (1.2, 8.3) 20.8* (10.1, 58.7)
4* (–3.8, 7.8) 18.8* (–8, 36.2) 14.9* (5, 30.6)
18.7 (–11.5, 40.7) 7.3 (–12.7, 29.2) 32.8 (–8, 71.6)
6.6 (–11.7, 71.1) –18.7 (–23.2, 6.6) 40.3 (0.9, 93.2)
NOTE. Values are median (interquartile range). Abbreviation: SCM, sternocleidomastoid. *Denotes a statistically significant difference (P⬍.05) determined via the Wilcoxon signed-rank test.
(see fig 2) and smoothed (see fig 3) EMG and other respiratory traces would suggest this was the case. The significantly greater sternocleidomastoid activation demonstrated in all electromyographic measures by participants with quadriplegia during soft speech suggests that they have to work harder even to achieve vocalization at a low volume. In addition, the unexpected occurrence and high incidence of preparatory peak events (see fig 4) before both loud and soft vocal tasks (and singing) supports this notion that people with quadriplegia may require more intense activation of their accessory respiratory muscles than able-bodied controls both in preparation for, and during vocalization. In contrast with the loud versus soft task, no difference in peak electromyographic amplitude was observed in either group during singing versus speech. The control group increased the AUC in all muscles when singing, but only the diaphragm AUC increased for the group of participants with quadriplegia. There were no between-group differences observed for either singing or speech tasks. We observed a larger AUC in trapezius and diaphragm muscles within both participant groups during singing (see table 4). These results are consistent with previous research examining muscle activation by able-bodied participants during singing. Pettersen et al34 and Pettersen and Westgaard35 reported that during singing, the trapezius contributes to the compression of the upper thorax, thus serving as an accessory muscle of expiration. In the participants with quadriplegia, it is possible that the lack of an increase in activation for the higher effort task (singing) occurred because they were already recruiting most of their available muscle capacity during the speech task and thus had nothing further to recruit. However, considerable effort from the participants with quadriplegia was observable through the high incidence of preparatory muscle contractions. Further controlled trials are required to replicate these results and to evaluate the clinical significance of these differences. Respiratory Function and Voice Assessments The group of participants with quadriplegia scored significantly lower than their matched controls on most respiratory measures, fitting with the pattern of respiratory impairment reported after SCI in previous research.2,3 However, the reported perception of impaired voice production by these participants (as measured by the Voice Handicap Index) is an interesting and novel finding of this study. The lack of a statistically significant difference in SPL between singing and speech is interesting and was contrary to what was expected. Although there was a 4-dBA increase for both groups, this increase was not statistically significant, which may suggest that the study was underpowered. This result may also have occurred because “Happy Birthday” is ingrained in most people as a song and as such was not representative of normal running speech. The use of rhythmic emphasis and stress when speaking the lyrics increased the
overall SPL and thus may have contributed to the lack of a significant difference between speaking and singing. Study Limitations Several methodological considerations must be addressed when assessing our conclusions. A significant limitation of this study is the small sample size; however, we have been conservative accordingly in our use of nonparametric analyses. The relationship between electrical activity (EMG) and muscle contraction is complex and cannot be directly assessed. Thus, there is no way to confirm that the increased electromyographic activity observed in participants does in fact correlate with increased muscle force. Surface electromyographic recordings of respiratory muscle activity have previously been reported in able-bodied subjects during speech36 and singing.34,35 Although there is no consistent, universal relationship between muscle force and dynamic electromyographic data, it has been demonstrated that the integrated EMG is linearly correlated with isometric force under static conditions.37 One method for quantifying surface electromyographic measurements involves the measurement of a maximal muscle contraction and then representing the data as a percentage of this maximal value.38 We did not use this method because it is difficult to perform a task requiring absolute maximal effort of respiratory muscles, accessory muscles, or both, and so we could not reliably achieve a maximal contraction of the muscles assessed. In addition, it was not possible to isolate diaphragm activity from abdominal muscle activity in the able-bodied subjects using the surface electromyographic electrodes. This expiratory muscle activity of the abdominals in the able-bodied controls may have masked any differences between the groups, particularly during the more intensive (loud speech and singing) tasks. A further potentially confounding factor in the electromyographic analysis is that we could not guarantee that electromyographic surface electrode placement was precisely the same in all participants. Differences in electrode position may explain some of the variability between participants. However, given that most of our analyses involved assessment of changes in measurements within each subject, this issue is thought to be of reduced importance. Finally, the use of convenience sampling ensured wellmatched control participants, but also introduced bias and limits generalization of the findings, as control participants were not randomly drawn from the general population. CONCLUSIONS In summary, the results of this study showed that the participants with quadriplegia exhibited the expected respiratory impairments and a self-perception of vocal impairment. Participants with quadriplegia appeared to recruit accessory respiratory muscles more when speaking and singing (particularly at louder volumes) than their matched able-bodied controls. Participants with quadriplegia also made unexpected, preparatory respiratory efforts that were not observed in their able-bodied Arch Phys Med Rehabil Vol 92, February 2011
256
RESPIRATORY MUSCLE RECRUITMENT AFTER SPINAL CORD INJURY, Tamplin
peers. The role of these preparatory efforts is unclear and warrants further investigation. The participants with quadriplegia could speak and sing adequately (and at comparable intensity levels to their controls), but appeared to use unusual muscle recruitment techniques to compensate for respiratory impairments. Future research should examine whether there may be a role for further development and training of these atypical muscle recruitment techniques to enhance respiratory and vocal function for people with quadriplegia. Acknowledgments: We thank Darren Tardio (Vipac Engineers) for his technical assistance, and Matthew Absalom-Wong for his artistic expertise. References 1. DeVivo MJ, Krause S, Lammertse DP. Recent trends in mortality and causes of death among persons with spinal cord injury. Arch Phys Med Rehabil 1999;80:1411-9. 2. Brown R, DiMarco AF, Hoit JD, Garshick E. Respiratory dysfunction and management in spinal cord injury. Respir Care 2006;52:853-68. 3. Chen CF, Lien IN, Wu MC. Respiratory function in patients with spinal cord injuries: effects of posture. Paraplegia 1990;28:81-6. 4. MacBean N, Ward E, Murdoch BE, Cahill L, Salley M, Geraghty T. Characteristics of speech following cervical spinal cord injury. J Med Speech Lang Pathol 2006;14:167-84. 5. Bunch M. Dynamics of the singing voice. 2nd ed. New York: Springer-Verlag Wien; 1993. 6. Shaffer TD, Wolfson MR, Bhutani VK. Respiratory muscle function, assessment and training. Phys Ther 1981;61:1711-23. 7. Frisbie JH. Breathing pattern in tetraplegic patients. Spinal Cord 2002;40:424-5. 8. Winslow C, Rozovsky J. Effect of spinal cord injury on the respiratory system. Am J Phys Med Rehabil 2003;82:803-14. 9. Estenne M, Knoop C, Van Vaerenbergh J, Heilporn A, De Troyer A. The effect of pectoralis muscle training in tetraplegic subjects. Am Rev Respir Dis 1989;139:1218-22. 10. DePalo VA, McCool FD. Respiratory muscle evaluation of the patient with neuromuscular disease. Semin Respir Crit Care Med 2002;22:201-10. 11. Wohlgemuth M, van der Kooi EL, van Kesteren RG, van der Maarel SM, Padberg GW. Ventilatory support in facioscapulohumeral muscular dystrophy. Neurology 2004;63:176-8. 12. Misuri G, Lanini B, Gigliotti F, et al. Mechanism of CO2 retention in patients with neuromuscular disease. Chest 2000;117:447-53. 13. Delgado H, Braun S, Skatrud B, Reddan W, Pegelow D. Chest wall and abdominal motion during exercise in patients with chronic obstructive pulmonary disease. Am Rev Respir Dis 1982; 126:200-5. 14. Howard RS, Wiles CM, Loh L. Respiratory complications and their management in motor neuron disease. Brain 1989;112:1155-70. 15. Breslin EH. Breathing retraining in chronic obstructive pulmonary disease. J Cardiopulm Rehabil 1995;15:25-33. 16. Epstein SK, Celli BR, Williams J, Tarpy S, Roa J, Shannon T. Ventilatory response to arm elevation. Its determinants and use in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1995;152:211-6. 17. Breslin EH. The pattern of respiratory muscle recruitment during pursed-lip breathing. Chest 1992;101:75-8. 18. De Troyer A, Peche R, Yernault JC, Estenne M. Neck muscle activity in patients with severe chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1994;150:41-7. 19. Fujiwara T, Hara Y, Chino N. Expiratory function in complete tetraplegics: study of spirometry, maximal expiratory pressure, and muscle activity of pectoralis major and latissimus dorsi muscles. Am J Phys Med Rehabil 1999;78:464-73.
Arch Phys Med Rehabil Vol 92, February 2011
20. Batavia AI, Batavia M. Karaoke for quads: a new application of an old recreation with potential therapeutic benefits for people with disabilities. Disabil Rehabil 2003;25:297-300. 21. Mendes AP, Brown WS, Sapienza C, Rothman HB. Effects of vocal training on respiratory kinematics during singing tasks. Folia Phoniatr Logop 2006;58:363-77. 22. Tamplin J. The link between singing and respiratory health for people with quadriplegia. Aust J Music Ther 2009;20(special issue):45-55. 23. Jacobson BH, Johnson A, Grywalski C, et al. The Voice Handicap Index (VHI): development and validation. Am J Speech Lang Pathol 1997;6:66-70. 24. ATS/ERS statement on respiratory muscle testing. Am J Respir Crit Care Med 2002;166:518-624. 25. Miller MR, Hankinson JL, Brusasco V, et al. Standardisation of spirometry. Eur Respir J 2005;26:319-38. 26. Ashba J, Garshick E, Tun CG, et al. Spirometry—acceptability and reproducibility in spinal cord injured patients. J Am Paraplegia Soc 1993;16:197-203. 27. Kelley A, Garshick E, Gross ER, Lieberman SL, Tun CG, Brown R. Spirometry testing standards in spinal cord injury. Chest 2003; 123:725-30. 28. Black LF, Hyatt RE. Maximal static respiratory pressures in generalized neuromuscular disease. Am Rev Respir Dis 1971;103:641-50. 29. Fairbanks G. Voice and articulation drillbook. 2nd ed. New York: Harper and Row; 1960. p 196. 30. Gamet D, Fokapu O. Electromyography. In: Wnek GE, Bowlin GL, editors. Encyclopedia of biomaterials and biomedical engineering. London: Informa Healthcare; 2007. p 956-67. 31. Eckert DJ, McEvoy RD, George KE, Thomson KJ, Catcheside PG. Genioglossus reflex inhibition to upper-airway negativepressure stimuli during wakefulness and sleep in healthy males. J Physiol 2007;581:1193-1205. 32. Oates JM, Russell A. A sound judgement. In: CD-ROM package. Melbourne: La Trobe University; 1997. 33. Winkworth AL, Davis PJ. Speech breathing and the Lombard effect. J Speech Hear Res 1997;40:159-69. 34. Pettersen V, Bjorkoy K, Torp H, Westgaard RH. Neck and shoulder muscle activity and thorax movement in singing and speaking tasks with variation in vocal loudness and pitch. J Voice 2005; 19:623-34. 35. Pettersen V, Westgaard RH. The activity patterns of neck muscles in professional classical singing. J Voice 2005;19:238-51. 36. MacFarland DH, Smith A. Surface recordings of respiratory muscle activity during speech: some preliminary findings. J Speech Hear Res 1989;32:657-67. 37. Wootten ME, Kadaba MP, Cochran GV. Dynamic electromyography. I. Numerical representation using principal component analysis. J Orthop Res 1990;8:247-58. 38. Mezzanotte WS, Tangel DJ, White DP. Waking genioglossal electromyogram in sleep apnea patients versus normal controls (a neuromuscular compensatory mechanism). J Clin Invest 1992;89:1571-9. a. b. c. d. e. f. g. h.
Suppliers OPS Medical, LLC, 8055 Ritchie Hwy, Ste 103, Pasadena, MD 21122. P.K. Morgan Ltd, Chatham, Kent, UK. Micro Medical Ltd, PO Box 6, Rochester, Kent ME1 2AZ, England. Compumedics Sleep Pty Ltd, 1 Marine Parade, Abbotsford, Victoria, 3067, Australia. Cambridge Electronic Design Ltd, 4 Science Park, Milton Rd, Cambridge, Cambridgeshire, CB4 0FE, England. Ambulatory Monitoring, Inc, 731 Saw Mill River Rd, Ardsley, NY 10502-0609. VIPAC Engineers & Scientists Ltd, Victorian Technology Centre, 275 Normanby Rd, Port Melbourne, Victoria, 3207, Australia. SDA Software Design Ahnert GmbH, Arkonastrae 45-49, 13189, Berlin, Germany.