Short Interspike Intervals and Double Discharges of Anconeus Motor Unit Action Potentials for 1 the Production of Dynamic Elbow Extensions 2 3 Short Interspike Intervals and Double Discharges 2

22 Incidence of double discharges (DDs, >100Hz) and short interspike intervals (ISIs, >50Hz 23 <100Hz) is reported to vary widely among different muscles and tasks, with a higher incidence 24 in motor unit (MU) trains of fast muscles and for the production of fast contractions in humans. 25 However, it is unclear whether human muscles with a large composition of slower motor units 26 exhibit DDs or short ISIs when activated with maximal synaptic drive, such as those required for 27 maximal velocity dynamic contractions. Thus, the purpose of this study was to determine the 28 effect of increasing peak contraction velocity on the incidence of DDs and short ISIs in the 29 anconeus muscle. Seventeen anconeus MUs in ten young males were recorded across dynamic 30 elbow extensions ranging from low submaximal velocities (16% of maximal velocity) up to 31 maximal velocities. A low incidence of DDs (4%) and short ISIs (29%) was observed among the 32 583 MU trains recorded. Despite the low incidence in individual MU trains, a majority (71% and 33 94%, respectively) of MUs exhibited at least one DD or short ISI. The number of short ISIs 34 shared no variance with MU recruitment threshold (R=0.02), but their distribution was skewed 35 toward higher peak velocities (G=-1.26) and a main effect of peak elbow extension velocity was 36 observed (p<0.05). Although a greater number of short ISIs was observed with increasing 37 velocity, the low incidence of DDs and short ISIs in the anconeus muscle is likely related to the 38 function of the anconeus as a stabilizer rather than voluntary elbow extensor torque and 39 velocity production. 40 SHORT INTERSPIKE INTERVALS AND DOUBLE DISCHARGES 3 Introduction 41 Voluntary torque production in muscles that control joint movements occurs in 42 response primarily to increases in two primary motor unit (MU) mechanisms; recruitment and 43 discharge rate (Heckman and Enoka 2012). However, in addition to these fundamental 44 properties other MU mechanisms function to amplify and optimize the resultant mechanical 45 output of the neuromuscular system (Garland and Griffin 1999; Semmler 2002; Semmler et al. 46 2004). One mechanism is the modification of interdischarge interval temporal variability in a 47 MU train to which the muscle is very responsive (Binder-Macleod and Kesar 2005). Double 48 discharge (DD), or short interspike (ISI) intervals early in a MU train of action potentials have 49 been shown to greatly increase the torque output (up to 20%) and the rate of torque 50 development (up to 50%) in animal (Binder-Macleod and Barrish 1992) and human muscle 51 (Garland and Griffin 1999; Binder-Macleod and Kesar 2005) without relying on a sustained 52 increase in the average rate of MU discharge. Studies using percutaneous electrical stimulation 53 report the optimal ISI range for generation of maximal tetanic tension is 5-10ms (Stein and 54 Parmiggiani 1979; Zajac and Young 1980a). Accordingly, recent studies of voluntary force 55 production have defined DD as the consecutive discharge of two action potentials belonging to 56 the same MU at interspike intervals of less than 5ms (>200Hz) (Van Cutsem et al. 1998; Van 57 Cutsem and Duchateau 2005) or less than 10ms (>100Hz) (Christie and Kamen 2006; 58 Mrówczyński et al. 2010). Whereas ISIs of less than 20ms (>50Hz), but greater than 10ms 59 (<100Hz), are referred to as short ISIs (Griffin et al., 1998). Either of these events is often 60 observed at, or near, the beginning of a train of MU action potentials. Many studies have 61 SHORT INTERSPIKE INTERVALS AND DOUBLE DISCHARGES 4 investigated the incidence of DDs or short ISIs (Garland and Griffin 1999; Binder-Macleod and 62 Kesar 2005), but few have explored the associated functional consequences of these short ISIs 63 during voluntary contractions in humans (Desmedt and Godaux 1978, 1979; Van Cutsem et al. 64 1998; Van Cutsem and Duchateau 2005; Christie and Kamen 2006). Of these studies, most have 65 been limited to isometric contractions (Garland and Griffin 1999). 66 The timing of MU DD and short ISIs during voluntary contractions is believed to be 67 closely related to their proposed function (Desmedt and Godaux 1978; Binder-Macleod and 68 Kesar 2005; Mrówczyński et al. 2010; Mrówczyński et al. 2011). These functions include 69 absorbing the resting slack of the muscle to enable linear isometric force production; increasing 70 series elastic stiffness for greater force transmission (Wilson and Larimer 1968; Parmiggiani and 71 Stein 1981; Binder-Macleod and Kesar 2005); or rapid summation of MU twitch tensions to 72 increase the rate of isometric torque production (Desmedt and Godaux, 1978; Van Cutsem et 73 al., 1998; Mrówczyński et al., 2010; Mrowczynski et al., 2011). Although DDs have been 74 observed periodically in long trains of MU action potentials (Bawa and Calancie 1983; Kudina 75 and Alexeeva 1992; Westad et al. 2004) and at termination of isometric force production 76 (Kudina and Churikova 1990; Kudina and Alexeeva 1992; Stephenson and Maluf 2010), they are 77 more often found at or near minimal firing rates (Bawa and Calancie 1983; Calvin and Schwindt 78 1972) when MUs are first recruited during low isometric torque contractions (Bawa and 79 Calancie 1983; Denslow 1948; Hoff and Grant 1944; Kudina 1974; Kudina and Churikova 1990), 80 or preceding fast ballistic isometric contractions (Bawa and Calancie 1983; Desmedt and 81 Godaux 1977). Similarly, short ISIs have been noted preceding moderately fast (<250°/s) 82 SHORT INTERSPIKE INTERVALS AND DOUBLE DISCHARGES 5 dynamic contractions in 97% of triceps brachii MUs recorded (Griffin et al. 1998). Furthermore, 83 increases in peak shortening velocity have been demonstrated in muscles stimulated initially 84 with high frequency stimulation indicating a role for DDs and short ISIs in maximizing 85 contraction velocity during initial shortening (MacIntosh et al. 2008). 86 Another important factor related to the incidence of DDs or short ISIs in humans is 87 motor unit type. Although equivocal (Garland and Griffin 1999), more often DDs and short ISIs 88 (<20ms ISI) are observed in higher threshold motor units in humans (Kudina 1974; Griffin et al. 89 1998), which presumably innervate predominantly type II muscle fibers with faster contractile 90 properties. The force-frequency relationships of slow and fast muscle fibers have been shown 91 to differ in animal preparations (Fitts et al. 1998), to differ between fast and slow whole human 92 muscles (Bellemare et al. 1983), and between fast and slow motor units of a single muscle 93 (Thomas et al. 1991). Generally, slower muscles reach tetanic fusion or peak forces at lower 94 stimulation frequencies than faster muscles. Therefore, muscles with slower twitch contractile 95 properties and a greater type I fiber composition likely do not require the high rates of 96 activation potentially provided by a DD to reach maximal level of torque development, and 97 ultimately maximal shortening velocities. Rather, torque production in slower muscles may 98 experience similar relative changes in mechanical output when activated with short interspike 99 intervals (short ISIs, <20ms) to that observed when faster muscles are subjected to DDs 100 (ISI<10ms). Studies of the cat hindlimb muscles, which range from slow to fast, have shown 101 that the incidence of DDs varies widely (8-95%) among motor neurons during slow locomotion 102 (Zajac and Young 1980b; Hoffer et al. 1987). To date, no study has investigated the effect of 103 SHORT INTERSPIKE INTERVALS AND DOUBLE DISCHARGES 6 DDs or short ISIs on peak shortening contraction velocity in either slow or fast human skeletal 104 muscles. Thus, the purpose of the present study was to determine the effect of increasing 105 resultant peak shortening contraction velocity on the incidence of MU DDs and short ISIs in a 106 muscle with a large composition of slow motor units. Based on previous investigations of 107 human MU behavior, we hypothesized that in the anconeus muscle, during its contribution to 108 elbow extensor movements: 1) that few MU DDs would occur but that short ISIs would be 109 more prevalent as peak elbow extension velocity was increased, and 2) that MU DDs and short 110 ISIs would occur more frequently during the first interspike interval (initiation of force 111 production) compared to all subsequent interspike intervals of a recruited MU train. 112 Methods 113 Interspike intervals and instantaneous MU discharge rates of the anconeus, and elbow 114 extension torque, position, and velocity were recorded during dynamic contractions of varying 115 peak velocities in ten young men (26±2y, 177.3±8.5cm, 77.7±7.0kg) free from orthopaedic, 116 neuromuscular, and cardiorespiratory limitations. Informed written consent was obtained from 117 all subjects prior to participation, and all procedures were approved according to the policies 118 and guidelines of the local Research Ethics Board for human participants and conformed to the 119 Declaration of Helsinki. 120 One to three visits (~1.5hr/visit) were required to ensure an adequate quantity and quality of 121 MU recordings. Elbow extension torque, position, and velocity measures were obtained using a 122 Biodex System 3 multi-joint dynamometer (Biodex Medical Systems, Shirley, NY, USA) with the 123 subject’s left shoulder joint flexed 90° and arm abducted 20°. The subject’s arm rested on a 124 SHORT INTERSPIKE INTERVALS AND DOUBLE DISCHARGES 7 support positioned ~10cm proximal to the olecranon process of the ulna and the forearm was 125 secured in the semi-prone position to a custom-built support attached to the Biodex lever arm. 126 The protocol has been described in detail previously (Harwood et al. 2011; Harwood and 127 Rice 2012). Briefly, the protocol began with three (~5-s) isometric elbow extension MVCs at 60° 128 elbow flexion (0° = full extension) of which the highest value was taken as the representative 129 MVC and used to determine the load (25%MVC) of the subsequent dynamic contractions. 130 Dynamic contractions were defined as those in which a predetermined load was held relatively 131 constant and the velocity was allowed to vary throughout the ROM. Five loaded (25%MVC) 132 maximal velocity elbow extensions (Vmax25) were performed over 120° ROM (starting from 120° 133 elbow flexion to 0° elbow extension). Torque and velocity output were displayed on a 134 computer screen for visual feedback during MVCs and Vmax25, respectively, and verbal 135 encouragement was provided during each maximal effort. Four target peak velocities (25, 50, 136 75, and 100%Vmax25) were calculated for each subject individually from the highest Vmax25 137 recorded. Four sets of at least five loaded (25%MVC) elbow extensions were performed at each 138 of the four target peak velocities (25%, 50%, 75%, 100%Vmax25) in a randomized order. Subjects 139 were instructed to equal the peak velocity of each submaximal elbow extension performed to a 140 horizontal cursor displayed on the computer screen that corresponded to one of the three 141 submaximal target velocities (25%, 50%, and 75%Vmax25). Additional elbow extensions were 142 performed when subjects failed to match the target peak velocities. Therefore, the number of 143 contractions performed often exceeded the number of contractions required for completion of 144 the protocol. As a result, each subject performed a brief MVC (~5-s) to verify fatigue was not 145 SHORT INTERSPIKE INTERVALS AND DOUBLE DISCHARGES 8 induced in response to the protocol. Each submaximal elbow extension was separated by ~30-s 146 rest, whereas MVCs and Vmax25 were separated by at least 2-min rest. Additional rest was 147 allotted at the request of the subject. 148 Single MU action potential trains of the anconeus were recorded with custom made 149 insulated stainless steel fine wire electrodes (100μm, California Fine Wire Company, Grover 150 Beach, CA). Two hooked tip fine wires (15-30cm length) were passed through a 27.5 gauge 151 hypodermic needle (Becton Dickinson and Company, Franklin Lanes, NJ) and inserted into the 152 belly of the anconeus ~2-4cm distal to the space between the olecranon process of the ulna 153 and the lateral epicondyle of the humerus. To maximize the number of single MU trains 154 recorded throughout a protocol, two needles were inserted and withdrawn immediately 155 leaving the two bipolar pairs of fine wires embedded in the muscle. A common ground 156 electrode for the fine wire electrode pairs was placed over the styloid process of the radius and 157 secured with surgical tape. 158 Intramuscular EMG of the anconeus was pre-amplified (100-1000x), high-pass filtered 159 (10Hz, Neurolog, Welwyn City, England) and digitized using analog-to-digital converter 160 (Cambridge Electronics Design, Cambridge, UK) at a rate of 15kHz. Torque, position, and 161 velocity data were sampled at 100Hz. All data were stored offline for analysis where 162 intramuscular EMG signals were high-pass filtered at 100Hz to remove any movement artifact. 163 Data Analyses 164 Torque, position, and peak elbow extension velocity were determined using a custom 165 software package (Spike 2 version 7.0, CED, Cambridge, UK) for all dynamic contractions 166 SHORT INTERSPIKE INTERVALS AND DOUBLE DISCHARGES 9 included in the data analysis. Average rate of torque development (RTD) was determined 167 beginning with a 5%MVC departure from baseline torque to the attainment of the requisite 168 load and normalized to MVC. Peak velocity was expressed relative to Vmax25. For each 169 contraction in which a MU discharged either a DD or short ISI, the absolute constant error 170 [(|peak velocity – target velocity|/target velocity)*100%] was calculated. Average absolute 171 constant errors were computed for the full range of elbow extension velocities, and for each of 172 the target velocities. 173 Single MUs were identified using a template matching algorithm (Spike 2 version 7.0, 174 CED, Cambridge, UK) which amassed waveforms of sequential action potentials sharing similar 175 temporal and spatial characteristics. The ultimate determinant in deciding whether an action 176 potential belonged within a train of MU action potentials was visual inspection by an 177 experienced investigator (B.H.). Motor unit discharge times (ms) were determined for each MU 178 action potential from which interspike intervals (ISIs) and subsequently instantaneous MU 179 discharge rates were calculated between each successive MU action potential (Spike 2 version 180 7.0, Cambridge, UK). The inclusion criteria for MUs to be included in the statistical analysis 181 required that each MU: 1) fired at least five consecutive action potentials, 2) was active during 182 both the initiation phase (torque development) and movement phase of each elbow extension, 183 and 3) was consistently present during each set of dynamic contractions. 184 An instantaneous MU discharge rate (MUDR) was recorded for each ISI sequentially 185 until peak velocity was attained, and an average MUDR was calculated for each MU train. 186 Double discharges (>100Hz) and short ISIs (>50Hz to <100Hz) were not included in the 187 SHORT INTERSPIKE INTERVALS AND DOUBLE DISCHARGES 10 calculation of average MUDR. The number of instantaneous MUDRs recorded for a single MU 188 train ranged from 4 to 24, for which 24 represented the last functional discharge (preceding 189 peak contraction velocity) of the slowest firing single MU recorded. Double discharges (>100Hz, 190 or <10ms) and short ISIs (>50Hz to <100Hz, or 10ms to<20ms) were identified within each MU 191 train meeting the inclusion criteria for a dynamic contraction. Two separate analyses were 192 performed for DDs and short ISIs. The first involved determining the incidence and number of 193 DDs and short ISIs over the first 4 to 24 ISIs, which was dependent upon the length of each MU 194 action potential train. The second focused on the incidence of DDs and short ISIs for the 1 ISI, 195 because the discharge of short ISIs during the 1 ISI has been identified as a potential 196 mechanism by which RTD may be increased (Bawa and Calancie 1983; Van Cutsem et al. 1998). 197 Recruitment thresholds of MUs were also determined for MU action potential trains meeting 198 the inclusion criteria for both analyses as the relative force (%MVC) at which the first MU action 199 potential discharged. 200 Statistical Analysis 201 To investigate the difference between the expected and observed frequencies of DD 202 and short ISIs, chi square tests were performed across all elbow extension velocities, and for 203 each range of peak elbow extension velocity separately. Separate frequency histograms were 204 generated for DDs and short ISIs relative to peak elbow extension velocity and MU recruitment 205 threshold. Skewness (G) was determined for each histogram, where a value of zero indicates a 206 symmetrical distribution, a positive value indicates a leftward skewed distribution, and a 207 negative value indicates a rightward skewed distribution. 208 SHORT INTERSPIKE INTERVALS AND DOUBLE DISCHARGES 11 Analytical statistics were performed for short ISIs over the first 24 ISIs because only one 209 MU recorded elicited more than one DD over the first 24 ISIs and the incidence of short ISIs and 210 DDs during the 1 ISI is a binary variable. Shapiro-Wilks tests of normality determined that the 211 number of short ISIs over the first 24 ISIs was not normally distributed relative to peak elbow 212 extension velocity (P<0.05) or MU recruitment threshold (p<0.05). Therefore, Kruskal-Wallis 213 ANOVAs were used to determine whether a main effect of peak elbow extension velocity or 214 MU recruitment threshold was present. Due to unequal variances between the average 215 incidence of DDs and short ISIs for the four peak elbow extension velocity ranges (<25, 25-<50, 216 50-<75, and ≥75%Vmax25), and and MU recruitment threshold ranges (<10, 10-<15, 15-<20, and 217 ≥20%MVC), Welch’s t-tests were used to test whether there was a difference between any two 218 ranges upon determination of a main effect. Regression analyses were performed and 219 coefficients of determination (R) were calculated to demonstrate the amount of shared 220 variance between the number short ISIs and MU recruitment threshold; between the number 221 of short ISIs and RTD; and between the number of short ISIs and peak contraction velocity. 222 Results 223 A total of 583 MU trains were recorded from 17 distinct MUs active throughout the 224 whole protocol in 10 subjects (1-3 MUs per subject). Representative data from one MU for 225 which a short ISI was recorded at 75% Vmax25 (left panel), but not 25%Vmax25 (right panel) is 226 shown in Figure 1. The distribution, average recruitment thresholds, and discharge rates of the 227 recorded MUs in addition to contractile properties of the elbow extensions from which these 228 MUs were recorded are presented in Table 1. The average absolute constant error for 229 SHORT INTERSPIKE INTERVALS AND DOUBLE DISCHARGES 12 contractions across the full range of velocities was 9.3±8.1%, wherein the ability of subjects to 230 equal the target velocity was greater at higher velocities (15.6±8.5% at 25%Vmax25 vs. 3.9±4.4% 231 at 100%Vmax25). Of the 17 MUs recorded, 12 MUs (71%) exhibited DDs and 16 (94%) exhibited 232 short ISIs. The number of short ISIs over the first 24 ISIs was 171, of which 88 (15% of MU 233 action potential trains) were recorded during the 1 ISI. The number of DDs recorded over the 234 first 24 ISIs was 21. Similarly, the number of DDs recorded for the 1 ISI was 20, which 235 corresponded to ~4% of the total MU action potential trains recorded. A detailed description of 236 the incidence of DDs (>100Hz) and short ISIs (>50<100Hz) in each MU is provided in Table 2. 237 Chi square (X) tests showed that the observed frequency of both DDs and short ISIs was less 238 than the expected frequency for the full range of elbow extension velocities investigated and 239 within each velocity range (Table 3). The distribution of these data relative to each target 240 velocity range and MU recruitment threshold for all ISIs and for the 1 ISI is presented in 241 Figures 2 and 3, respectively. The distributions of peak contraction velocities at which DDs and 242 short ISIs were recorded were skewed toward higher velocities (G = -0.33 and-0.51, 243 respectively), and the distribution of MU recruitment thresholds at which short ISIs were 244 recorded was skewed toward lower MU recruitment thresholds (G = 0.84) when all 24 ISIs were 245 considered (Figure 2A, 2B). However, the distribution of MU recruitment thresholds in which 246 DDs were observed was normally distributed (G=-0.09) about a mean and standard deviation of 247 11.3±7.2%MVC (0.2-25.6%MVC, Figure 2B). When considering only the 1 ISI, the peak 248 contraction velocities in which a DD or short ISI were observed were skewed (G = -0.29 and 249 1.26, respectively) toward higher velocities (Figure 3A). However, DDs were only observed in 250 SHORT INTERSPIKE INTERVALS AND DOUBLE DISCHARGES 13 higher threshold MUs (>10%MVC) during the 1 ISI and the distribution of short ISIs during the 251 1 ISI was bimodal with a greater number of short ISIs being recorded in both the highest and 252 lowest MU recruitment threshold ranges (Figure 3B). 253 Regression analyses demonstrated no shared variance between MU recruitment 254 threshold and the number of short ISIs recorded (R=0.02, p=0.14) when all ISIs were 255 considered. However, a greater number of short ISIs were recorded during contractions that 256 exhibited higher RTDs (R=0.24, p<0.05) and greater peak velocities (R=0.14, p<0.05). 257 Moreover, the Kruskal-Wallis ANOVA revealed a main effect of target velocity for short ISIs 258 (X(3) =10.42, p<0.05). Welch’s t-tests demonstrated fewer short ISIs at <25, 25-<50, and 50259 <75%Vmax25 than at ≥75%Vmax25 (p<0.05, Δ=1.25-6.86), and fewer short ISIs at <25 and 25260 <50%Vmax25 than at 50-<75%Vmax25 (p<0.05, Δ=0.52 and 1.86, respectively, Figure 2A). 261 Discussion 262 Anconeus MUs tracked during the production of a range of elbow extension velocities 263 (16-100%Vmax25) in the present study exhibited very few DDs (4% of MU action potential trains, 264 1.2±1.1) and short ISIs (29% of MU action potential trains, 10.1±12.3) across the first 24 ISIs of 265 the MU trains recorded. Even fewer short ISIs (15% of MU action potential trains, 5.2±5.1) 266 were observed when only the 1 ISI interval of MU action potential trains was considered 267 (Table 2). Despite the relatively low incidence of DDs and short ISIs within any individual MU 268 action potential train in this study, overall 71% and 94% of the recorded anconeus MUs 269 exhibited at least one DD or short ISI, respectively. The number of MUs discharging short ISIs or 270 DDs is reported to range widely (0-97%) across human studies during mainly isometric tasks, 271 SHORT INTERSPIKE INTERVALS AND DOUBLE DISCHARGES 14 and is believed to be more frequent in fast contracting muscles (Garland and Griffin 1999). 272 However, MU studies of the slow soleus and intercostal muscles (Johnson et al. 1973; Mizuno 273 and Secher 1989) and the relatively faster biceps brachii (Denslow 1948; Kukulka and Clamann 274 1981) have reported low numbers of DDs (0-23% of MUs) (Kukulka and Clamann 1981; Kudina 275 and Alexeeva 1992; Whitelaw and Watson 1992) indicating that the relationship between 276 contractile properties of the skeletal muscle and incidence of DDs is equivocal. In general, the 277 results of the present study provide support for low incidence of DDs in slow human skeletal 278 muscle. 279 The anconeus muscle provided an attractive model to investigate the incidence of DDs 280 and short ISIs during the production of maximal velocity contractions because of its common 281 innvervation with the main elbow extensor, the triceps brachii (Fornalski et al. 2003, Hwang et 282 al. 2004). Together they function to extend the elbow, and have similar relative changes in 283 fascicle length as a function of elbow joint range of motion (Buchanan et al. 1986; Murray et al. 284 2000; Harwood et al. 2010; Pereira 2012). Furthermore, the anconeus is activated throughout 285 all contractile intensities and velocities of elbow extension (Le Bozec et al. 1980a, 1980b; 286 Harwood et al. 2011). Based on these shared properties of the anconeus and triceps brachii, it 287 is assumed that a similar level of synaptic drive activated these muscles at the various elbow 288 extension velocities investigated. However, the anconeus possesses fewer MUs (Stevens et al. 289 2013) than other upper limb muscles tested (Power et al. 2012), even those of a similar size 290 (Boe et al. 2005), and the range of anconeus MU recruitment thresholds is low (Le Bozec and 291 Maton 1987, Harwood and Rice 2012) in comparison to other large force-producing muscles 292 SHORT INTERSPIKE INTERVALS AND DOUBLE DISCHARGES 15 (Seki and Narusawa 1996; Griffin et al. 1998). Indeed, these features may have lessened the 293 probability that the number of DDs or short ISIs would differ among anconeus MUs of different 294 recruitment thresholds. Furthermore, during high velocity contractions, the MU recruitment 295 threshold range of the anconeus is compressed (Harwood and Rice 2012) contributing to the 296 lower probability that differences between high and low threshold MUs would be detected. 297 Despite these factors, intramuscular EMG signals in the anconeus exhibit high signal clarity in 298 comparison to the larger triceps brachii (Davidson and Rice 2010; Harwood et al. 2011) which 299 facilitates tracking of anconeus MUs across repeated contractions, at high forces, and velocities 300 (Harwood et al. 2011, 2012; Harwood and Rice 2012). 301 Short ISIs in the first 24 ISIs of MU action potential trains in the present study were 302 skewed toward higher anconeus MU recruitment thresholds, but the number of short ISIs 303 recorded within a MU train shared no variance (R0.02, p=0.14) with MU recruitment 304 thresholds. Moreover, the distribution of short ISIs relative to recruitment threshold during the 305 1 ISI of MU action potential trains was bimodal. These findings indicate no relationship 306 between MU recruitment thresholds and the number of short ISIs within the pool of anconeus 307 MUs studied. One previous study of the triceps brachii during moderate velocity (<250°/s) 308 elbow extensions (Griffin et al. 1998), observed more short ISIs in high than low threshold MUs. 309 In that study (Griffin et al. 1998), MU recruitment threshold range of the triceps brachii was 310 broader (0-40%MVC) than that recorded for the anconeus in the present study (0-24%MVC) 311 and in one other study (0-26%) (Harwood et al. 2012). However, it must be noted that the 312 requisite torque limited the range over which MU recruitment thresholds could be recorded in 313 SHORT INTERSPIKE INTERVALS AND DOUBLE DISCHARGES 16 the present study. The moderation of MU recruitment threshold range compression during 314 movements, especially at high velocities (Harwood and Rice 2012), likely occurs as MUs 315 otherwise inactive during low to moderate submaximal isometric contractions are recruited at 316 these faster velocities. The additional recruitment of these higher threshold MUs serves to 317 increase rates of torque development, concurrently extending the breadth of the compressed 318 range over which anconeus MUs are recruited. Thus, the low incidence of DDs (4%) and short 319 ISIs (15%) observed during the 1 ISI of MU action potential trains in the present study cannot 320 be attributed to MU recruitment threshold differences within the anconeus muscle. 321 Although the incidence of MU DD varies across different muscles (Garland and Griffin 322 1999), the low percentage of DDs in anconeus MU action potential trains relative to other 323 upper limb muscles investigated (Kukulka and Clamann 1981) is likely related to the proposed 324 functions of DDs. Double discharges have been shown to either facilitate high rates of torque 325 development (Desmedt and Godaux 1978; Van Cutsem et al. 1998) or to lessen the slack of 326 elastic elements in the muscle (Binder-Macleod and Kesar 2005), and to increase series elastic 327 stiffness for greater force transmission (Parmiggiani and Stein 1981). The anconeus is 328 responsible for less than 15% of the elbow extension torque (Zhang and Nuber 2000) and 329 therefore it is unlikely the resultant rate of torque development is influenced greatly by the 330 anconeus especially considering its contractile speed (Le Bozec and Maton 1987), size (Hwang 331 et al. 2004; Pereira 2012) and mechanical arrangement for elbow extension torque (Pereira 332 2012). It is also unlikely that a great deal of slack accumulates in the anconeus. The superficial 333 anconeus tendon spans ~70% of the muscle length, but the length of the tendon in series with 334 SHORT INTERSPIKE INTERVALS AND DOUBLE DISCHARGES 17 anconeus muscle fibers is very short in comparison to other muscles (Keener et al. 2010; 335 Pereira 2012). Moreover, the anconeus is active at all elbow joint angles (Le Bozec et al. 1980a) 336 and thus slackening of its elastic elements would likely only accumulate in a hyperextended 337 position when the anconeus is inactive. Thus, a role for DDs in reducing series elastic slack in 338 the anconeus would be unnecessary and seems highly unlikely. 339 Some (Bawa and Calancie 1983) have proposed that short ISIs preceding ballistic (high 340 rate of force development) isometric contractions require ‘massive synaptic inputs’ in order to 341 discharge action potentials in the ‘secondary firing range’ (Granit et al. 1966; Calvin and 342 Schwindt 1972) to achieve requisite rates of torque development. Synaptic drive to the 343 anconeus motor neuron pool was likely high in the present study because average MU 344 discharge rates at 75% and 100%Vmax25 (28.7 and 39.0Hz, respectively) were higher than during 345 isometric MVC (23.8Hz) in the anconeus in an identical protocol (Harwood et al. 2011). 346 Furthermore, muscle-dependent motor cortical excitability and spinal reflex studies show 347 excitability to be higher during shortening than isometric contractions (Duchateau and Enoka 348 2008; Duclay et al. 2011). Despite a high level of synaptic drive, a low percentage of DDs and 349 short ISIs were observed during the 1 ISI of anconeus MU action potential trains (3% and 16% 350 of action potential trains, respectively), yet the number of anconeus MUs (94%) for which at 351 least one short ISIs were recorded in the 1 ISI was similar to the number of triceps brachii MUs 352 (97%) exhibiting short ISIs during similar, but less intense, contractions (Griffin et al. 1998). 353 However, it must be noted that resampling bias as a consequence of recording single MUs 354 across repeated contractions in this study and that of Griffin et al. (1998) likely increased the 355 SHORT INTERSPIKE INTERVALS AND DOUBLE DISCHARGES 18 probability that a MU would fire at least one short ISI. In view of this limitation, these findings 356 indicate that regardless of the architectural, histochemical, and electrophysiological differences 357 among muscles, a high level of synaptic drive is likely sufficient to elicit DDs in a majority of 358 MUs. 359 It was hypothesized that fewer DDs would be observed in the anconeus due to potential 360 diminished responsiveness to high frequency stimulation as indicated by the force-frequency 361 relationships of human skeletal muscle exhibiting slower contractile speeds and greater type I 362 fiber composition. The incidence of DDs in anconeus MU action potential trains was observed 363 to be much less (~20%) than that reported for two other upper limb muscles; the adductor 364 pollicis and biceps brachii (Kukulka and Clamann 1981). The type I fiber composition and 365 contractile speed of these two upper limb muscles studied differed greatly, with the anconeus 366 (67% type I, 80ms HRT) situated between the adductor pollicis (~80% type I, 131ms half 367 relaxation time (HRT) ) and biceps brachii (~45% type I, 67ms HRT) (Johnson et al. 1973; Round 368 et al. 1984; Le Bozec and Maton 1987; Dalton et al. 2010). Moreover, the contractile speed and 369 type I fiber composition of the anconeus and triceps brachii vary greatly (Le Bozec and Maton 37