This thesis examines the movement cues underlying human kinesthesia, using a combination of psychophysical methods and an analyticrrl framework for math<'1l1aticHI and statistical modelling of functional aspects of underlying physiological sensory mechanisms. Using an innovative psychophysical paradigm, the first experiment examines whether people are more sensitive to the occurrence of movement than to movement direction. Contrary to previous investigations, the findings are that the underlying sensitivity for movement detection and direction discrimination is the same. An analytical tool for investigating what are the sensory cues for movement detection is developed as a physiologically plausible model of movement perception. The model accounts for performance in terms of displacement and velocity sensitive directionally tuned channels and probability summation over time. Statistical analysis of optimized model parameters indicates that the dominant cue for movement detection is velocity and that the contribution of displacement information is insignificant. An extended version of a previous movement detection model based on temporal integration is also applied to data from experiment l. Derived estimates of this model's parameters suggest that displacement information does make a significant contribution to movement detection. The conflict between the two models is resolved in favour of the probability summation model, based on statistical comparisons regarding model accuracy and parsimony, and an analysis of the physiological plausibility of underlying assumptions. A second experiment investigates whether manipulation of movement acceleration affects movement detection. The probability summation model is extended to account for processing of acceleration. Estimates of the model's parameters reveal that acceleration information does not contribute significantly to velocity based movement detection.