When you run your finger over a horizontal surface and feel a bump, what processes lead to the perception of the bump? It could be that perception of the bump is based on the position of the finger, which rises and then falls as it goes over the bump. Alternatively, perception of the bump could be based on horizontal force acting on the finger that resists and then assists lateral motion of the finger as it goes up and then down the bump, respectively. Testing these hypotheses was achieved using a simple robotic interface that could simulate horizontal forces to resist and then assist the finger, as though it were going over a bump, critically while maintaining the finger in the horizontal plane at all times. Surprisingly, the percept was that of a bump, indicating that the force cues were important.
Another key question is how haptic information is combined with other information such as visual inputs to form a single percept. When we pick up a viewed object, we receive both haptic and visual information about its width. An optimal estimator, which aims to estimate the object's width with the smallest error (that is, optimally), would combine these two sources of information using a weighted average where the weighting of each source depends on its reliability. By attaching the ends of two lightweight robots to the tip of the index finger and thumb, it is possible to simulate objects and control the width experienced haptically. By combining this robotic setup with a virtual-reality visual display, a recent study created objects with different haptic and visual widths and was able to determine the weighting given to each source of information in the perception of width, weightings that matched the predictions of an optimal estimator.
To examine the first stages of tactile processing, very fine control over the tactile input is required. By developing a robotic interface that can precisely control the position and orientation of a tactile probe over time, and combining tactile stimulation with microneurography (recordings from tactile afferent nerves), it has been possible to show in humans that the timing of first spikes from tactile primary afferents carry a great deal of information about properties of the object being contacted, such as curvature and friction.