Regression-based indices of fatigue in paraplegics' electrically stimulated quadriceps

We studied muscle fatigue in 19 paraplegic patients during an isometric sustained contraction. Both quadriceps (n = 38) were maximally stimulated for 126 s. We defined a nonlinear equation with four coefficients to fit the data and extract fatigue indices carrying information on the amount (asymptotic values, torque coordinate at the inflection point, total decrement from initial value) and the rate of torque decrease (maximum slope, time constant, time coordinate at the inflection point, time at which the torque reached 50% of initial value). Most of the indices reported in the literature provide information limited to the amount of torque decrease. Three of the coefficients we defined could be graphically estimated. The goodness of fit was very high (r2 = 0.993 +/- 0.001) but none of the indices provided information on both the amount and the rate of decrease. Major differences were encountered among the patients. Therefore, some of the computed indices varied to a great extent. We studied the curves exhibiting the same fatigue indices and defined a simple general fatigue index able to distinguish from muscles having different resistance to fatigue (asymptotic value x the time constant). This index is carrying both types of information and is expressed in relative torque x second (range from 1.81 to 22.82).