Figure 11.5 Example of muscle activation patterns for one female subject performing forces in a vertical arm position. The curves show the EMG level for 13 shoulder muscles when performing a 10 N force in different directions in the horizontal plane (representing attempted abduction, flexion, adduction, and extension). The 10 N force corresponded to approximately 14 per cent MVC for all force directions. The circles represent a level of 10 per cent EMG.

In the glenohumeral joint a torque due to arm elevation depends strongly on shoulder muscle forces as discussed above. Concerning the neck one study has reported that an increase in neck flexion resulting in an increase in joint extension moment from 32 per cent to 45 per cent of the maximal extension moment was performed with an unchanged mean splenius EMG activity level of 9 per cent EMGmax (Finsen, 1998). Thus, when the neck is flexed the torque in the seventh cervical-first thoracal joint is less dependent on muscle force and relatively more dependent on counteracting forces produced by passive structures such as ligaments.

The recording of working postures and external forces allows for the application of inverse dynamics for the calculation of net forces and moments. These forces and moments should be produced by the combined action of muscle, ligament and bone forces. Since there are far more muscles and muscle parts which can be independently activated than force and moment constraints, the number of possible muscle force combinations is infinite. To determine which muscle force patterns can be used for a given external shoulder load, modelling is needed. Most biomechanical modelling is based on optimisation principles by choosing the set of muscle forces which minimise a certain cost function (Högfors et al., 1995; Niemi et al., 1996; van der Helm, 1994). By using such optimisation principles only one force distribution is found for each load condition. A different approach is to use EMG for the determination of the voluntary muscle activation patterns and based on this to assess the muscle force distribution which actually is used. However, as mentioned above, EMG may provide good information on muscle activation while muscle force estimation from EMG is difficult. One reason is that many shoulder muscles are involved when performing external forces or sustaining a force-demanding posture. Therefore, the combined action of the shoulder muscles should be considered.