INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The recent refinement of a technique for recording and classifying the discharge of single motor units during muscle shortening (1) has allowed the behavior of motor units to be assessed in human respiratory muscles during quiet breathing (2). The discharge frequencies of large numbers of single motor units in the parasternal intercostals, scalenes, and diaphragm have been recorded in healthy individuals and in patients with chronic obstructive pulmonary disease (COPD) (2, 3). Whereas the usual peak frequency of discharge was 10.5 Hz in the costal diaphragm of healthy subjects, it was markedly greater at 17.9 Hz in the patients. On the other hand, the increase in discharge frequency for the parasternal intercostals was from 10.1 Hz in the healthy subjects to only 13.4 Hz in the patients, and the corresponding increase in the scalenes was from 8.5 Hz to only 11.4 Hz. Therefore, it appears that when the drive to breathe is increased as a result of a chronic pulmonary disease, the effect on inspiratory synergists is not uniform: enhanced drive preferentially increases the discharge frequency of phrenic motoneurons rather than that of other inspiratory muscles.

This preferential increase in the firing rate of diaphragmatic motor units may reflect a specific adaptation of the central controller to the chronic alteration in the mechanical properties of the lungs and chest wall due to COPD. Alternatively, it may be an expression of the usual behavior of human inspiratory motoneuron pools in response to an increased respiratory drive. The present study was designed to distinguish between these two possibilities.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Studies were performed on four healthy volunteer subjects (Table 1) who were free from respiratory disease. As it would have taken too long to study the diaphragm, the parasternal intercostals, and the scalenes in one session, subjects were studied in two sessions separated by approximately 1 wk. In the first study recordings were made from the diaphragm and scalene muscles, and in the second study recordings were made from the parasternal intercostals. Studies were performed with the subjects comfortably seated in an upright chair. Informed consent was obtained. Procedures were approved by the local institutional ethics committee and conducted according to the Declaration of Helsinki.

General Procedures

Before each study the thickness of the chest musculature at the sites of needle insertion was assessed using real-time ultrasonography (model 128XP/4; Acuson Corporation, Mountain View, CA). Measurements were made in the second intercostal space (1 to 4 cm lateral to the sternal edge) for the parasternal intercostal recordings and in the seventh or eighth intercostal space in the anterior axillary line overlying the zone of apposition of the costal diaphragm for the diaphragm recordings. This procedure helped to localize the site for needle insertion and provided an approximate depth for insertion.

Subjects breathed through a mouthpiece and volume was obtained by integration of the flow signal from the pneumotachometer. Respiratory movements were also monitored with inductance plethysmographs (Respitrace, Ardsley, NY) around the middle of the chest and abdomen, and the gains of the two signals were adjusted using the isovolume maneuver (4). Once the gains were adjusted, the sum of the change in abdominal and rib cage signals changed in proportion to changes in lung volume. The relative contributions of movement of abdomen and rib cage to production of a change in lung volume could then be assessed by measurement of the change in the individual signals expressed as a percentage of the change in the summed signal.

The mouthpiece was connected via a 3-way tap to a piece of corrugated tubing (38 mm minimal internal diameter). The dead space of the apparatus proximal to the tubing was 170 ml. End-tidal CO₂ was measured close to the mouthpiece. To vary the ventilatory drive, two lengths of tube were used, one with a volume of 700 ml and the other with a volume of 1,150 ml. Three tasks were performed with an interval of 4 to 5 min between them: quiet breathing with no tube; breathing through the "short" tube; and breathing through the "long" tube. After the initial period of quiet breathing, the size of the dead space was selected at random and was unknown to the subject. For each task an initial period of 3 to 5 min was allowed before any measurements of ventilation and discharge of motor units were undertaken. This period allowed ventilation to stabilize based on the monitored end-tidal CO₂ and tidal volume (VT). To minimize extraneous stimuli which may affect ventilation, subjects were blindfolded and listened to music through headphones.

Recording and Analysis Techniques

Recordings were made using an active Teflon-coated monopolar electrode with an exposed tip of 0.15 mm² or 0.25 mm². Recordings were referenced to a surface electrode 2 to 3 cm from the monopolar needle. Subjects were grounded via a large flexible strap on the right shoulder. For the diaphragm, the needle was inserted perpendicular to the skin in the seventh or eighth intercostal space close to the anterior axillary line at a site that was below the pleural reflection and close to the origin of costal fibers (Figure 1) (5). During insertion electromyographic (EMG) activity was monitored on a loudspeaker as the needle passed through the abdominal and intercostal muscle layers. For the parasternal intercostals, the electrode was inserted at an angle of about 45° in the second intercostal space on the right with the needle tip pointing toward the sternum.

View larger version (21K): [in this window] [in a new window]   Figure 1.   Approximate sites for needle EMG recording from the diaphragm, parasternal intercostals, and scalenes. Subjects breathed through a mouthpiece under control conditions and with added dead space consisting of a short or long tube. Subjects also listened to music through headphones and were blindfolded. Airflow was measured with a pneumotachometer and integrated to give volume. Rib cage and abdominal movements were measured with calibrated inductance bands around the chest and abdomen.

The usual site was 2 to 3 cm from the sternal edge. EMG activity was monitored as the needle was advanced through pectoralis major into the parasternal intercostal muscle. If advanced further, expiratory activity from triangularis sterni was detected. For the scalene muscles, the needle was inserted about 1 to 2 cm above the clavicle in the posterior triangle of the neck. The location was defined by palpation during voluntary inspiratory maneuvers. Prior to recordings from the diaphragm and parasternal intercostal muscles, a small injection of local anesthetic was made at the site of needle insertion (1 to 1.5 ml of lignocaine 2% with adrenalin). This minimized any discomfort associated with the procedures. Once a suitable recording site had been secured in the relevant muscle, 5 to 10 breaths were monitored before manipulation of the electrode to a new recording site. In any subject an average of 10 sites was studied for each muscle under each recording condition. By monitoring the end-tidal CO₂, respiratory movements, and VT signal, we were able to analyze recordings made under stable ventilatory conditions.

The EMG signals were amplified (1,000 to 10,000 times), filtered (53 Hz to 3 kHz), and played on loudspeakers. The activity of motor units were also monitored continuously by one of the experimenters on an oscilloscope (5 ms/division). All data were stored on video tape using pulse code modulation (Vetter PCM, model 3000; A.R. Vetter Co., Rebersburg, PA) and on computer using a commercial spike-analysis system (Spike2 with 1401 plus interface, Cambridge Electronic Design, Cambridge, UK). The pulse code modulator sampled each input at 22 kHz and prior testing indicated that this was adequate to encompass the highest frequencies observed with the type of monopolar electrode used in this study (J. B. Leeper and S. C. Gandevia, unpublished observations).

The EMG data were reanalyzed in two stages. First, from the computer-stored data, all possible spikes from single motor units were extracted based on a threshold crossing. In the second stage, all spikes were manually sorted using individual "templates" based on their size and detailed morphology. Details of the customized software have been provided previously (2, 6). The software allowed updating of the mean shape of motor units stored in each template, review of the frequency plot for each single motor unit, simultaneous display of volume and movement signals, superimposition of all spikes from a motor unit, and superimposition and summation of selected templates from individual motor units (so that simultaneous activity in units could be determined). Although time-consuming, this method allows the discharge frequency of up to 4 single motor units to be measured reliably from each recording (2, 3). The main analysis was performed on the mean discharge frequency at the peak rate of firing during inspiration. In addition, the onset discharge frequencies were noted along with the presence of tonic firing throughout inspiration and expiration. Data from units that discharged sporadically with less than 3 interspike intervals during inspiration were excluded.

Statistical Analyses

Both the discharge frequencies of single motor units and the various ventilatory parameters were assessed by analyses of variance, which took into account ventilatory condition (i.e., control, short and long tube), subject, and muscle. When significant changes were found, post hoc analysis was conducted. Statistical significance was set at the 5% level. Unless indicated, values are given as mean ± SD in the text and shown as mean ± SEM in the figures.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ventilatory Responses

All subjects showed the expected ventilatory responses to breathing through an increased dead space. Two levels of increased ventilation were produced by partial rebreathing through the short tube and the long tube (Figure 2). End-tidal CO₂ rose from 5.2 ± 0.2% during quiet breathing through the mouth piece and associated dead space (170 ml) to 5.7 ± 0.3% with the short tube and to 6.2 ± 0.5% with the long tube (p < 0.05). Ventilation increased approximately threefold from an average of 12 ± 2.2 L/min to 38 ± 7 L/min with the larger dead space (p < 0.001), and VT increased from 0.8 ± 0.1 L to 2.1 ± 0.2 L (p < 0.001). There was also a small but significant decrease in inspiratory time from 1.7 ± 0.4 s to 1.5 ± 0.2 s (p < 0.01).

View larger version (14K): [in this window] [in a new window]   Figure 2.   Mean data (± SEM) for the group of subjects for ventilatory variables under three conditions: quiet breathing ("control"), and breathing through a short and long tube. (A) VT; (B) inspiratory time (TI); (C ) minute ventilation ( E); (D) end-tidal CO2.

In the three tasks, the relative contribution of rib cage and diaphragm movement to volume generation was estimated with inductance bands that measured changes in rib cage and abdominal circumference (see METHODS). Based on data from all studies, in the control task 57 ± 14% of the VT was contributed by rib cage movement, and this was the same as when breathing through the long tube (57 ± 11%).

Single Motor Unit Recordings

Typical data for two breaths during quiet breathing and while breathing through the long tube are shown for one subject in Figure 3. Under control conditions, a total of 44 motor units were sampled for the diaphragm (mean 11 units for each of the four subjects, range 9 to 13), 66 motor units for the parasternal intercostals (mean 17 units, range 12 to 19), and 60 motor units for the scalene muscles (mean 15, range 8 to 19). Overall, 170 units were recorded under control conditions, 172 units when breathing with the small added dead space and 174 units with the larger dead space.

View larger version (19K): [in this window] [in a new window]   Figure 3.   (A) Sample of data from one experimental session recording from diaphragm and scalene muscles during quiet breathing ("control," left panels) and with increased ventilation produced by a large dead space ("long tube," right panels). Discharge frequency and volume shown for two breaths in each condition. Typical recordings of one unit from the diaphragm (above) and scalenes (below). Action potentials from the single units are superimposed at right of the traces. (B) Data from another experimental session in the same subject under similar conditions except that recordings were made from the parasternal intercostal muscle (second interspace). Traces arranged as in panel A. Horizontal dashed lines in the frequency plots on right (in panels A and B) mark the mean peak firing rate under control conditions.

All but four of the 516 units increased their discharge during inspiration with few (46 units, 9%) having a tonic discharge that persisted throughout expiration. Figure 4 shows histograms of the firing frequencies of all units recorded from each muscle under the three conditions. The discharge frequencies under control conditions were highest in the diaphragm and lowest in the scalenes (p < 0.01). The discharge frequencies of motor units in the diaphragm increased from 11.0 ± 2.7 Hz to 17.7 ± 3.3 Hz between the control task and breathing with the long tube (p < 0.001). There was a smaller but significant increase for the parasternal intercostal muscles (from 10.0 ± 1.6 Hz to 11.9 ± 1.9 Hz, p < 0.001). These increases were more marked as the size of the added dead space increased (diaphragm, p < 0.001; parasternals, p < 0.02). On the other hand, although the increase in frequency for the scalene muscles (from 8.7 ± 1.8 Hz to 9.5 ± 2.1 Hz) was significant (p < 0.05), this effect was not graded with the size of the added dead space. The behavior of the motor units in the three different muscles was significantly different across the three tasks (p < 0.001).

View larger version (22K): [in this window] [in a new window]   Figure 4.   Histogram of discharge frequencies for all units in the diaphragm (left column), parasternal intercostals (middle column), and scalene muscles (right column). Data for quiet breathing ("control," top row), ventilation through the small tube (middle row) and through the long tube (lower row). Bin width 0.5 Hz. Mean values are marked on each histogram.

Figure 5 shows data from each subject for the three muscles and tasks. The main result obtained when data from all motor units were pooled is also evident in data from single subjects. Thus, there is a prominent increase in the discharge frequency of diaphragmatic motor units as ventilatory drive increased.

View larger version (25K): [in this window] [in a new window]   Figure 5.   Discharge frequencies for four subjects shown as mean ± SEM under control conditions (left panels), ventilation through short tube (middle panels) and through long tube (right panels). Diaphragm, open bars; parasternal intercostals, hatched bars; and scalenes, filled bars.

Tonically firing units were only found in the scalene and parasternal intercostal muscles. In the scalenes, 29 tonic units were sampled (11, 13, and 5 for control, short tube, and long tubes, respectively). In the parasternal intercostals, there were 17 tonic units (13, 2, and 2 for control, short tube, and long tube, respectively). Most of the tonic units were recorded from Subject 1 (62% and 53%, for scalenes and parasternal intercostals, respectively). In contrast to the behavior of the whole population of units, there were no significant differences in the firing frequencies of tonic units between the three protocols or between the two muscles. For the scalenes the mean frequencies were 10.1, 10.8, and 10.5 Hz for the control, short tube, and long tube, respectively, and for the parasternal intercostals 10.7, 8.7, and 11.7 Hz, respectively. There were also no differences in the firing frequencies at end-expiration for these units (3.9, 4.5, and 3.6 Hz, respectively, for the scalenes, and 4.9, 5.5, and 4.9 Hz for the parasternal intercostals). One unit in the scalenes (Subject 1) and 3 units in the parasternal intercostals (Subject 4) had a tonic discharge which was not related to respiration. The frequencies for these units were 6.8 Hz in the scalenes and a mean of 12.3 Hz in the parasternal intercostals.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present results show that when respiratory drive is increased in humans, the firing rate of inspiratory motor units is elevated (see also Reference 3 for humans, and Reference 7 for cats). However, the firing rate does not increase equally for each inspiratory motoneuron pool. Additional dead space, which provides a chemical stimulus to breathe, increased ventilation about threefold and this was associated with an increase in the firing rate of motor units of 60% for the costal diaphragm, but only 19% for the parasternal intercostals, and 9% for the scalene muscles. In human subjects these muscles are always active during inspiration, even when VT is very small (8). The discussion focuses on the technical validity of the results and on some of their implications.

When we initiated the present investigation, we attempted to follow the discharge of the same motor units during a gradual increase in chemical respiratory drive. However, several minutes were necessary to achieve significant increases in ventilation, and small displacements of the recording electrodes could not be avoided. Consequently, we elected to sample from many sites within individual inspiratory muscles during repetitions of the required task (i.e., breathing in any given condition). Recording sites were held for about 40 s before the needle was moved to a different region of the muscle. Auditory feedback of the needle EMG was available to the experimenter, but not the subject, during the studies. Data were obtained without discomfort or long-term problems in four subjects during quiet breathing and when chemical drive had increased ventilation up to threefold. We are confident about the validity of the sample of single motor units because every recorded spike has been re-sorted manually (several times) using interactive software specifically designed for this purpose (2, 6). Analysis was only undertaken when the measures of ventilation were stable as assessed both on- and off-line. Furthermore, the technique has been applied to patients in whom respiratory drive was increased such that rates of motor unit discharge were similar to those observed here with the largest dead space (2, 3). The technique has also been shown to be sensitive to smaller changes in drive (11). In that study the firing frequencies of single motor units in the diaphragm changed proportionally to inspiratory flow and volume when the same units were followed during various volume-tracking tasks (mean firing frequency ranged from 7.8 to 11.0 Hz). Finally, in the present study, large numbers of motor units were sampled from each muscle at each level of ventilation.

The approach of sampling from many single motor units was first introduced by Bellemare and colleagues (12) to determine their rates of discharge during maximal voluntary contractions of limb muscles under isometric conditions. Here we have applied the method to follow the discharge of units during shortening of the relevant muscle. This approach was possible even for the diaphragm and one reason for this may have been that recordings were made close to the costal insertion of fibers so that movement of the electrode tip was minimized. Even during "isometric" contractions, there is significant shortening of muscle fibers (13), so that length changes themselves do not preclude successful recordings.

The disproportionate increase in the discharge rate of diaphragmatic motor units compared with scalene and parasternal motor units as respiratory drive increased does not necessarily mean that the work output of the diaphragm also increased disproportionately. This is because an increase in muscle work may be achieved by an increase in discharge rates of motor units combined with recruitment of additional units. No attempt was made to measure the recruitment of new motor units in the present study, but one observation suggests that the three inspiratory muscles may use a different balance between frequency modulation and recruitment at the level of the motoneuronal pool. Expansion of the rib cage remained in proportion to the displacement of the abdomen throughout the increases in respiratory drive caused by the added dead space. Inspiratory expansion of the rib cage compartment of the chest wall occurs mostly by activation of the parasternal intercostals and scalenes, whereas abdominal expansion reflects mostly the activation of the diaphragm. The proportionate expansion of the rib cage and abdominal components would be most easily achieved if the force generated by all the various inspiratory muscles acting on the rib cage and abdomen increased proportionately with respiratory drive. Thus, it is possible that there was a bias toward recruitment in the scalene and parasternal muscles. For limb muscles, it is well established that when drive increases, some muscles increase force predominantly by modulation of the frequency of recruited units with most units recruited at low force levels (e.g., intrinsic hand muscles), whereas for other muscles there is progressive recruitment of new units throughout the contraction (e.g., proximal arm muscles; 16, 17). One interesting corollary is that if the diaphragm uses a strategy of frequency modulation, at least for the levels of drive studied here, its behavior matches that of intrinsic muscles, which generate small forces, rather than that of large proximal muscles.

Whether the drive to respiration is increased by an external dead space added acutely in healthy subjects or over many years by COPD, there is a similarity in the firing behavior of inspiratory motoneurons, particularly those innervating the diaphragm. In healthy individuals breathing quietly, the final firing frequency of motor units in the three inspiratory muscles is about 9 to 10 Hz, but in patients with severe COPD the frequencies are 17.9, 13.4, and 11.4 Hz for the diaphragm, parasternal intercostal, and scalene muscles, respectively (2, 3). In the present study, when respiratory drive was increased, the final frequencies were 17.7, 11.9, and 9.5 Hz, respectively. This suggests that the markedly elevated firing rates of diaphragmatic motor units compared with those in other inspiratory muscles in the patients reflect the usual pattern of motor output when respiratory drive increases in normal subjects. The motoneuron pools of the scalenes and parasternal intercostals were activated slightly more in the patients than the control subjects, but the significance of this is unclear.