Abstract: This study was designed to contribute to research on the relative roles and contributions of afference, proprioception, and efference to touch-based (or haptic) texture perception. We sought to determine whether theoretical accounts of roughness constancy, in active and passive perception and across broad velocity ranges, would be supported when spatial density gradients were explored. We created textures with sinusoidal spatial gradients that participants explored with sinusoidal velocity directions: positive gradients resulted in densities that increased in the region where exploration velocities were largest while negative gradients counteracted the velocity increases and decreases with decreases and increases in spatial density, respectively. While these gradients could all be detected reasonably well, we found a slight shift in discriminability in favour of positive gradients and found that surfaces with no gradient at all were more likely to be judged to have a positive gradient. This finding runs counter to the roughness constancy models. Counterintuitively, we found that providing the participants with feedback as to their accuracy had a negative impact on gradients. The spatial gradients were of different magnitudes and in different their accuracy and confidence and we found no evidence of perceptual learning in the task. We consider the data relative to existing models and accounts of haptic texture perception.

Key words: touch, texture perception, haptic discrimination