Animals move their eyes, head or limbs in simple directed actions, but also coordinate these actions in more complex, specific and adaptive ways during locomotion. This high-performance control of locomotion depends on a precise knowledge of the configuration, speed, and direction of the body. Yet, how distributed neural circuits in the brain estimate self-movement remains unclear. Although locomotion generates visual signals, vision alone is inadequate to accurately estimate self-movement because retinal signals are confounded by changes in gaze, and by the movement of objects in the world. Here we will discuss our attempts to test the idea that an accurate internal estimate of the direction of the head and/or body movements during walking involves the integration of visual and non-visual information generated in conjunction with locomotion.