An investigation on statistical models for EMG-force relationship of anterior tibialis muscle in dorsiflexion

Document Type


Publication details

Lu, Z, Huang, LP, Tian, Q, Brooks, LO & Zhou, S 2005, 'An investigation on statistical models for EMG-force relationship of anterior tibialis muscle in dorsiflexion', paper presented to 4th Annual Conference of the Society of Chinese Scholars on Exercise Physiology and Fitness: Health and Scientific Training of Elite Athletes, Taiwan, 16-20 December.


The relationship between surface EMG and force generation has been arguably used as an indicator of neuromuscular adaptation to strength training. This relationship could be either linear or non-linear, depending upon the muscle, task and contraction intensity under investigation. No consensus has been achieved on how to statistically define variations in this relationship in response to interventions, particularly if it is non-linear. The primary aim of this study was to establish a suitable statistical model that can be used to evaluate EMG-force relationship of the tibialis anterior muscle in dorsiflexion, and the response of this relationship to different types of strength training and testing. Sixty healthy men (age 19.22±1.23yr) were randomly allocated into one of the four groups: electrical acupuncture (EA), isokinetic resistance training (RT), electrical acupuncture combined with resistance training (ER), and control (CG). The EA, RT and ER groups trained right leg three sessions per week, for four weeks. Before and after training, all subjects were tested for isometric and isotonic strength, and integrated EMG (iEMG) of the anterior tibialis at given percentages (10-100%) of maximal isometric dorsiflexion, and when lifting given weights (1.5 to 45 kg) in isotonic test, for both limbs. Results indicated that the distribution of original iEMG data was skewed, therefore transformation to natural logarithm was applied before using R Gui statistical package for further analysis. A full factorial linear mixed effect model was found to be suitable for description of the EMG-force relationship in both the isotonic and isometric tests. Significant 3-way interactions were detected for group*training*limb (p