This thesis concerns a particular relationship between music practice, computer technology and the importance of rhythm as a vehicle of musical expression. The intention is to explore new technology to allow for computer-mediated improvisation of rhythmic forms that are derived from a consideration of West African drumming. In developing the thesis an analysis of personal practice is used to reveal the role of music technology and the musical aims implicit behind its adoption. The exploration of rhythm is seen as important both compositionally and as a means of mediating collaborative musical expression, and can be understood as the exploration of a particular form of complexity. In considering its importance, the idea of music as 'embodied action' is discussed and music is considered as a form of knowledge and communication equal to spoken language but different from it. These considerations form the background for a discussion of the musical potential that may lie within certain key developments in artificial life, situated robotics, and the computer modelling of perception. These technologies are looked at for their ability to recognize rhythmic complexity and to be able to suggest subtle adaptations of complex rhythmic forms. A proposal of an 'adaptive rhythm synthesis' is put forward. In its consideration, the rhythmic structures and improvisational styles found in West African music are seen to pose particular challenges for computer modelling and a range of possible solutions is explored. A consideration is also given to the nature of machine 'autonomy'. It is suggested that the progress may lie in some combination of these solutions. Finally, some thoughts about the relationship between art, music and science are discussed.