Principal dynamic mode analysis of a spider mechanoreceptor action potentials

Abstract

The nonlinear dynamics of a spider mechanoreceptor are examined using the principal dynamic mode (PDM) methodology. The cuticilar sense organ of an adult Cupiennius Salei spider was stimulated with a pseudorandom Gaussian process spectrally bound at 400 Hz and the resulting action potentials were recorded. Data analysis reveals three PDMs that describe the system dynamics, based on the second order Volterra kernel, which is estimated from the recorded input/output datasets. The first PDM exhibits high-pass behavior, illustrating the importance of the speed of the slit displacement to the mechanoreceptor, while the other two PDMs exhibit bandpass behavior. The probability of firing an action potential is represented by a static multiple-input nonlinearity that receives the values of the convolution of each mode with the pseudorandom input as its inputs. The probability of firing function exhibits asymmetric behavior with respect to its arguments, suggesting directional dependence of the mechanoreceptor response on the PDM outputs. Trigger regions for a probability threshold value of 0.1 are also presented.