The dynamics of touch sensing studied in a mass-spring-damper model of the skin

It has been shown that the skin can provide highly resolvable, dynamic tactile information to the central nervous system. However, currently available skin models do not provide a matching level of dynamic complexity. Motivated by recent observations that everyday interactions create a diversity of widespread travelling waves of multiple overlaid frequencies in the skin, we here model the skin as a 3D-distributed mass-spring-damper model. Shear forces across each spring were reported back as separate sources of information, on which we performed information content analysis using principal component analysis. We found that a wide range of settings of spring constants, dampening coefficients and baseline tension resulted in highly resolvable dynamic information even for simple skin-object interactions. Optimization showed that there were some settings that were more beneficial for a higher temporal resolution, i.e. where multiple independent interactions could be more easily resolved temporally. Whereas even a single sensor reporting a skin shear force with infinite precision by itself can achieve infinite resolution, biological sensors are noisy. We therefore also analyzed the resolution of force direction in the dynamic skin model, when their simulated signal-to-noise ratio was varied. We conclude that biological skin due to its inherent dynamics can afford a low spatial resolution of sensors (subsampling) while still maintaining a very high resolution for detecting skin-object interaction dynamics, and that biological evolution moreover due to this construct likely has been free to play around with a variety of mechanical skin parameters and sensor densities without significantly compromising this resolution.