Background and aim:The role of thoraco-lumbar muscles is important in spinal column stability. Following fatigue due to constant activity, these muscles encounter with variable control pattern and variations of median and mean frequencies are little in relation to torque of force. However these parameters are useful in the literatures, but due to complexity of neuromuscular interaction and variety of motor control, signal processing can determine a wide range of changes and measurements.
Signal processing nonlinear techniques exploit in biologic signals. Variables of nonlinear techniques are recurrence, determinism, entropy and so on.
The purpose of this study was to determine changes of entropy as nonlinear parameter in comparison with linear parameter and applicability of entropic measurements of the erector spinae muscles during fatigue.
Materials and methods: Ten healthy women and 6 women with nonspecific low back pain (NSLBP) with a range of 20-30 years old patticipated in this study. Surface electromyography of isometric activities recorded from trunk (T12), lumbar (L3) and biceps femoris muscles during modified Sorenson isometric fatigue test. Median and mean frequency and also nonlinear parameters such as entropy and trend measured in one second of muscles activities before and onset of fatigue.
Results: Following fatigue in healthy group, median and mean frequencies reduced at a range of 12-20% (p<0.05). This decrement in LBP group was little (4 - 20%, p<0.05). Entropy increased 120-200% and trend reduced 800-2000% in normal subjects (p<0.05), whereas subjects with LBP indicated increment of entropy 65-220% and decrement of trend 240-500% (p<0.05). Before and after fatigue there was a significance difference between two groups in entropy parameter (p<0.05), whereas median and mean frequencies differences were not significant.
Conclusion:Following static positions, fatigue occurred in three levels of above muscles particularly at lumbar region. Traditional fatigue indicators showed good differences, but percentage of variability was low in comparison to nonlinear parameters. It suggests that nonlinear variables especially entropy are more sensitive than traditional indicators and can explain these stochastic behaviors