Received November 5, 2007
Revised November 28, 2007
Accepted after revision March 10, 2008

¹ Auckland Univ of Technology
² Wolters Kluwer Health Adis
³ University of Auckland

^* To whom correspondence should be addressed. E-mail: simeon.cairns{at}aut.ac.nz.

	Abstract

We present a curve-fitting approach that permits quantitative comparisons of fatigue profiles obtained with different stimulation protocols in isolated slow twitch soleus and fast twitch extensor digitorum longus (EDL) muscles of mice. Profiles from our usual stimulation protocol (125 Hz for 500 ms, evoked once every second for 100 300 s) could be fitted by single-term functions (sigmoids or exponentials) but not by a double-exponential. A clearly superior fit - as confirmed by the Akaiki Information Criterion - was achieved using a double-sigmoid function. Fitting accuracy was exceptional; mean square errors were typically <1% and r2 > 0.9995. The first sigmoid (early fatigue) involved ~10% decline of isometric force to an intermediate plateau in both muscle-types; the second sigmoid (late fatigue) involved a reduction of force to a final plateau - the decline being 83% of initial in EDL and 63% of initial in soleus. The maximum slope of each sigmoid was 7 to 8 fold greater in EDL than in soleus. The general applicability of the model was tested by fitting profiles with a severe force loss arising from repeated tetanic stimulation evoked at different frequencies or rest periods, or with excitation via nerve terminals in soleus. Late fatigue, which was absent at 30 Hz, occurred earlier and to a greater extent at 125 than 50 Hz. The model captured small changes in rate of late fatigue for nerve terminal versus sarcolemmal stimulation. We conclude that a double sigmoid expression is a useful and accurate model to characterise fatigue in isolated muscle preparations.

Key Words: Muscle contraction, Muscle fatigue, Muscle fibre