Article Figures & Data
Figures
- Fig. 1.
Schematic illustration of the stimulation protocols. Top: stimulation rates and patterns used for stationary activation tests;middle and bottom: stimulation patterns of 2 separate protocols used for transient activation tests. A1: schematic illustration of mean stimulation rates used in 27 consecutive measurements using concurrent stimulation of multi-motor unit (MMU) filaments (see methods) (see also Hulliger et al. 2001); the thick vertical lines (a–e) identify the tests, for which both combined multi-channel and up to 10 single channel stimulation trials were carried out. B1: individual stimulation profiles of the 10 channels of test c (in A1) illustrating the inter-stimulus interval (ISI) variability of individual pulse trains (see also text and Fig. 4). Middle: background components of individual transient stimulation profiles;bottom: instantaneous rate displays of individual stimulation profiles, with pseudo-random Gaussian noise added to the background components of middle (see text). A, 2and 3: imitation of a pure recruitment strategy, featuring staggered activation of filaments at a relatively high rate (28/s) but without subsequent increase of activation rate; the 4 profiles were used for activation of MMU filaments. B, 2 and 3: imitation of a combined recruitment and rate modulation strategy, featuring staggered activation of MUs at a low rate (7/s) combined with a subsequent systematic increase in activation rate. The 10-channel pattern of B3 was used for activation of single motor units (MUs; see Fig. 8). A reduced version (channels 1, 4, 7, and 10) was used for trials with activation of MMU filaments (see Fig. 6).
- Fig. 2.
Controlled waveform interaction to evaluate the effect of action potential overlap and signal cancellation. Top: illustration of selected examples of stimulus timing; 2 MMU filaments were activated at the same constant rate (20/s), while time shifts between the 2 were systematically varied between −20 and 20 ms (see methods). Examples were selected to illustrate interactions between AP segments with the same or opposite phase polarities (see also text).Middle: MMU action potentials (MMUAPs) evoked by separate stimulation of 2 ventral root filaments. Bottom: waveform interaction on combined activation of the 2 filaments. See also text.
- Fig. 3.
Signal cancellation due to AP overlap and interaction. Averaged rectified electromyogram (AEMG) calculated over 2-s segments of recorded and synthesized EMG generated by activation of 2 MMU filaments. A: data from the pair of Fig. 2.B: data from all 5 pairs of filaments, for which waveform interaction tests were performed, averaged after separate calculation of AEMG values for each pair and each time shift value.Top: AEMG values as a function of the time shift between stimulus pulse trains (see Fig. 2) during combined electrical stimulation of the 2 filaments. Bottom: AEMG values of EMG traces that were synthesized from the records obtained during separate stimulation of each filament (see Fig. 2,middle). Thick lines, synthesized data; thin lines, data recorded during combined activation (added for comparison; same traces as in top).
- Fig. 4.
Comparison of recorded and synthesized EMG during stationary activation of 10 MMU filaments. A: stimulation protocol, featuring separate and combined activation of 10 filaments at intermediate rates, ranging from 17/s to 31/s (cf. also C, and Fig.1 A, test d). B: surface EMG recorded during combined electrical stimulation. C: MMUAP trains evoked by separate stimulation of each filament. D: EMG synthesized off-line by algebraic summation of the individual MUAP trains shown in C. Note, in C, the absence of a clear correlation between MMUAP amplitude and filament size (tetanic force), as the filaments were ordered according to their size (channel 1, smallest; channel 10, largest). See also text.
- Fig. 5.
Quantitative analysis of the magnitude of recorded and synthesized EMG for stationary activation of 10 MMU filaments. A: dependence of AEMG on ensemble activation rate; thick line (C), recorded during combined stimulation; thick broken line (Σ), synthesized from individual MMUAP trains (recorded separately); thin line, sum of rectified MUAP trains. Shaded area, the amount of signal loss attributable to signal cancellation. For each trace, AEMG was normalized to the maximum recorded level during combined stimulation. Note the residual nonlinearity in the sum of rectified MUAP trains (seediscussion). B: power function regression analysis of pooled AEMG values. Filled squares, recorded EMG; open triangles, synthesized EMG; measurements at (from 3 different 10-channel sets) each of 5 EAR values. Power functions were fitted separately to the pooled data sets (recorded, synthesized). Note the close agreement of the fitted functions: thick line, recorded EMG; broken line, synthesized EMG; thin solid lines, 80% confidence limits calculated from population regression coefficients and variance (see text and methods).
- Fig. 6.
Comparison of recorded and synthesized EMG during transient activation of MMU filaments. A: stimulation profiles of the protocol imitating recruitment of rate modulation of 4 activation channels. B: EMG recorded during combined stimulation of 4 MMU filaments. C: MMUAP trains recorded during separate stimulation of each of the 4 filaments, using the same stimulation profiles (shown in A) as used during combined stimulation (B). D: synthesized EMG obtained by algebraic summation of the 4 MMUAP trains (C).
- Fig. 7.
Quantitative analysis of the magnitude of recorded and synthesized EMG for transient activation of 4 MMU filaments, using the combined recruitment and rate modulation protocol. A: dependence of AEMG on ensemble activation rate; thick line (C), recorded during combined stimulation; thick broken line (Σ), synthesized from individual MMUAP trains (recorded separately); thin solid line, sum of rectified MMUAP trains. Shaded area, the amount of signal loss attributable to signal cancellation. For each trace, AEMG was normalized to the maximum recorded level during combined stimulation.B: power function regression analysis of pooled AEMG values. Power functions were fitted separately to the pooled data sets (recorded, synthesized). Note the close agreement of the fitted functions: thick line, recorded EMG; broken line, synthesized EMG; thin solid lines, 95% confidence limits calculated from population regression coefficients and variance (see text). The underlying stimulation patterns and examples of EMG signals are shown in Fig. 6.
- Fig. 8.
Comparison of the surface EMG produced by combined stimulation and algebraic summation of single MUs. Top: the 10 single MU, transient recruitment and rate modulation protocol.Middle: qualitative assessment (single experiment) of the EMG superimposed from the combined stimulation and algebraic summation conditions for an entire data sweep. Bottom: 2 selected segments of the superimposed traces with increased temporal resolution (thick black line, combined stimulation; white thin line, algebraic summation). These segments are highlighted in themiddle row (width exaggerated for clarity). Note the qualitative similarity even of individual peaks and troughs of the combined-activation and synthesized signals. See also text.
- Fig. 9.
Linear regression analysis to compare recorded with synthesized EMG signals during combined activation of 10 single MUs. A: scatter plots of pairs of corresponding data points (recorded, synthesized) sampled at the same point in time. Dashed line inA1, reference with unitary slope. B: fitted linear regression lines. Top: data of the example illustrated in Fig. 8. Bottom: data from all 10 10-channel sets superimposed. Since EMGs were sampled and synthesized at a resolution of 2,000/s, superposition of 10 12-s episodes would have yielded 240,000 data points in A2; therefore only each third data point was plotted in A. Note the remarkable degree of correlation and the extremely small number of outlier points in A2. In line with this, the fitted regression lines (10) in B2 cannot all be recognized individually.
- Fig. 10.
Quantitative analysis of the magnitude of recorded and synthesized EMG for transient activation of 10 single MU filaments. Activation pattern imitating recruitment and rate modulation (see Fig. 8 A).A: dependence of averaged rectified EMG (AEMG) on ensemble activation rate; thick line (C), recorded during combined stimulation; thick broken line (Σ), synthesized from individual MMUAP trains (recorded separately); thin solid line, sum of rectified MUAP trains. Shaded area, the amount of signal loss attributable to signal cancellation. For each trace, AEMG was normalized to the maximum recorded level during combined stimulation. B: quadratic regression analysis of pooled AEMG values. Data from 10 different sets of single MUs. Quadratic functions were fitted separately to the pooled data sets (recorded, synthesized). Note the close agreement of the fitted functions: thick line, recorded EMG; broken line, synthesized EMG; thin solid lines, 95% confidence limits calculated from population regression coefficients and variance (see text andmethods).
StumbleUpon
