Gender and muscle differences in EMG amplitude and median frequency, and variability during maximal voluntary contractions of the quadriceps femoris.

The purpose of this study was to evaluate gender and muscle differences in electromyographic (EMG) amplitude and median frequency mean and standard deviation during maximal voluntary contractions of the quadriceps femoris. Thirty recreationally active volunteers were assessed for isometric EMG activity of the vastus medialis (VM), vastus lateralis (VL), and rectus femoris (RF) muscles during three 5-s maximal isometric voluntary contractions (MVCs). Median frequency of the three muscles was assessed through a power spectral analysis (fast Fourier transformation, Hanning window processing, 512 points). The power spectral analysis was performed during the middle 3 s of each contraction over 11 consecutive, 512 ms epochs overlapping each other by half their length (256 ms). The median frequency (F(med)) for each of the 11 windows was determined for each muscle. The mean and standard deviation of the F(med) across the 11 overlapping windows were then calculated for each contraction and muscle. EMG amplitude was determined by calculating the root mean square (RMS-50 ms time constant) over the same contraction period for each muscle. The mean amplitude and standard deviation about the mean value were then determined. A three-factor ANOVA with repeated measures was performed on the calculated F(med) mean and standard deviation values, and RMS standard deviations, to assess any gender, muscle, or trial differences, or interactions. A two-factor (gender by muscle) ANOVA with repeated measures was performed on the RMS mean amplitude for each muscle. Intraclass correlation coefficients (ICCs-2,1), standard errors of measurement (SEMs), and associated 95% confidence intervals were then calculated for maximal quadriceps torque and F(med) for each muscle. The results from this study demonstrated that the VL muscle displayed significantly higher F(med) values than the RF and VM muscles. The RF muscle showed significantly higher F(med) values (mean of 11 overlapping windows) than the VM muscle. Intrasession reliability was found to be high for the calculated mean values (ICC=0.85-0.96), but was shown to be low for variability (ICC=0.13-0.45). The major findings of this study support the notion that the EMG signal is "quasi-random" in nature, as demonstrated by the reproducible F(med) means and unreliable variability.