Tactile stimulus location is automatically transformed from somatotopic into external spatial coordinates, rendering information about the location of touch in three-dimensional space. This process is referred to as tactile remapping. Whereas remapping seems to occur automatically for the hands and feet, the fingers may constitute an exception in that some studies have implied purely somatotopic coding of touch to the fingers. When participants judge the order of two tactile stimuli, they often err when the stimulated body parts (usually the two hands) are crossed, presumably because somatotopic and external coordinates are in conflict in crossed postures. Using this task, we investigated, first, whether the fingers are unlike other limbs with regard to spatial coding, by testing whether crossing effects, indicative of external coding, were observable when stimulating two fingers, either on the same or on different hands. Second, we investigated the interaction of hand and finger posture in tactile localization of finger stimuli. Crossing effects emerged when fingers and hands were crossed, suggesting external coding for all body parts. Crossing effects were larger when both hand and finger were located in the hemifield opposite to their body side, and smaller when only hand or finger lay in the opposite hemifield. We suggest that tactile location is estimated by integrating the external location of all relevant body parts, here of a finger and its belonging hand, and that such integrative coding may represent a general principle for body part processing as well as for tool use.

Different reference frames appear to be relevant for tactile spatial coding. When participants give temporal order judgments (TOJ) of two tactile stimuli, one on each hand, performance declines when the hands are crossed. This effect is attributed to a conflict between anatomical and external location codes: hand crossing places the anatomically right hand into the left side of external space. However, hand crossing alone does not specify the anchor of the external reference frame, such as gaze, trunk, or the stimulated limb. Experiments that used explicit localization responses, such as pointing to tactile stimuli rather than crossing manipulations, have consistently implicated gaze-centered coding for touch. To test whether crossing effects can be explained by gaze-centered coding alone, participants made TOJ while the position of the hands was manipulated relative to gaze and trunk. The two hands either lay on different sides of space relative to gaze or trunk, or they both lay on one side of the respective space. In the latter posture, one hand was on its "regular side of space" despite hand crossing, thus reducing overall conflict between anatomical and external codes. TOJ crossing effects were significantly reduced when the hands were both located on the same side of space relative to gaze, indicating gaze-centered coding. Evidence for trunk-centered coding was tentative, with an effect in reaction time but not in accuracy. These results link paradigms that use explicit localization and TOJ, and corroborate the relevance of gaze-related coding for touch. Yet, gaze and trunk-centered coding did not account for the total size of crossing effects, suggesting that tactile localization relies on additional, possibly limb-centered, reference frames. Thus, tactile location appears to be estimated by integrating multiple anatomical and external reference frames.