Paired silver/silver chloride surface electrodesc placed 2 cm apart were used to record from pectoralis major, upper trapezius, and middle deltoid. Intramuscular hook-wire electrodes prepared in the laboratory in accordance with Basmajian and DeLuca (1985) were inserted into rhomboid major, lower trapezius, infraspinatus, supraspinatus, subscapularis, check details teres major, latissimus dorsi, and serratus anterior in that sequence using a 23 gauge needle as a cannula. Insertion sites of the indwelling electrodes were in accordance with the recommendations of Kabada and colleagues (1992) for subscapularis, and Geiringer (1994) for all remaining muscles. Correct

electrode placement, in the majority of muscles examined, was confirmed by comparing see more the signals during submaximal contractions expected to generate high levels of activity in the target muscle, to contractions expected to produce low activity in the target

muscle or to activate surrounding muscles into which the intramuscular electrode may have been inserted incorrectly. Because of the difficulty in distinguishing between rhomboid major and lower trapezius using this method, intramuscular electrodes were inserted into these muscles using an ultrasonically guided insertion techniqued. Following insertion of the indwelling electrodes, the shoulder was moved passively to determine the extent of wire excursion through the skin during the abduction range of movement required for the testing procedure. Allowing for this excursion, all wires were then looped and taped to the skin to prevent accidental removal and to reduce movement artefact during the testing procedure. A large surface ground electrodee was placed over the spine and acromion of the scapula of the opposite shoulder Thymidine kinase (Figure 1). The EMG signals were amplified and filteredf (gain = 100, bandpass between 10 Hz and 1 kHz) before transferring to a personal computer with

a 16 bit analog to digital converterg at a sampling rate of 2564 Hzh. Electromyographic signals were high pass filteredi, rectified, and low pass filteredj. These values were then expressed as a percentage of the maximum value of the filtered electromyographic signal generated for each muscle during the Shoulder Normalisation Tests. Mean electromyographic data for each muscle for each participant were calculated at each test position and each load by averaging a 1-sec sample from the two trials conducted. Group mean (SD) electromyographic data were subsequently calculated. A 3-factor, repeated measures ANOVA was performed to compare the levels of electromyographic activity across the 11 muscles, 3 angles, and 4 loadsk. Statistical significance was set at p < 0.05. Tukey post hoc analysis with pairwise comparisons was used to identify specific differences when significant ANOVA results were obtained. Fifteen people participated in the study.