In the early era of muscle research it was assumed that activation is uniform within muscles. However, recent research revealed regional differences in activation within several muscles, which has been linked to region-specific muscle adaptations to stimuli such as pain, fatigue, and training. Non-uniform within- and between-muscle distribution of hamstring injuries may be linked to non-uniform within- and between-muscle distribution of muscle activity. However, there is a lack of information about regional differences in hamstring muscle activity during hamstring exercises, and in high injury-risk movements such as high-speed running. Therefore, the aim of this thesis was to examine within- and between-muscle differences in biceps femoris and semitendinosus electromyography (EMG) activity using high-density surface EMG in typical hamstring exercises and at different running speeds. Our results showed that in most examined exercises, biceps femoris activity was lowest in the proximal region, which is the region where most hamstring injuries occur. The magnitude of regional differences was exercise-specific. In semitendinosus, some exercises presented uniform EMG activity proximo-distally, while others activated the middle and/or proximal region(s) the most. Generally, knee-oriented movements resulted in larger regional differences than hip-oriented movements in both muscles. Between-muscle differences were exercise-specific, whereby some knee-oriented movements activated semitendinosus preferentially, while hip extension performed on a roman chair activated biceps femoris more than semitendinosus. It was also found that performing Nordic hamstring exercise with hips flexed to 90° shows higher knee flexion torque and lower hamstring EMG activity as compared to performing Nordic hamstring exercise with neutral hips. In running, large inter-individual variability was observed in regional and muscle-specific EMG activity patterns, which were qualitatively consistent across a range of running speeds. These results suggest that a single pair of EMG electrodes placed over biceps femoris or semitendinosus does not accurately describe the activity of the whole muscle. Instead, spatially robust methods such as high-density EMG should be used to comprehensively describe hamstring activity. The sources and consequences of variations in regional activity patterns between exercises and between individuals should be further examined.