Discriminability of simple and complex haptic vibrations in single-cell computational and human psychophysical settings

Abstract

A multiscale, multiphysics model of the Pacinian Corpuscle (PC) was used to study the neurophysiological response to haptic vibrations in the 100-200Hz range. The computational results were compared to human psychophysical experiments, emulating the pairing of psychophysics with in vivo electrophysiology in PC research. A first assessment of this approach was made by examining the discriminability (dꞌ) of pairs of vibrotactile stimuli. The discrimination task was performed psychophysically and in silico for both one- and two-frequency stimuli. Both firing rate and inter-spike interval neural decoding schemes were used to calculate dꞌ from simulation data. Human subjects discriminated between frequencies with two components (complex stimuli) more effectively than isolated frequencies (simple stimuli), possibly due to the presence of beat frequencies in dissonant stimuli. Over a given stimulus set, in silico dꞌ values correlated well with the psychophysical data (R2 > 0.6), but when the simple and complex data were combined the model did not match the experiment (R2 < 0.1). Firing rate resulted in better predictions than inter-spike interval, and was more robust to noise. Results suggest that a single simulated PC can capture some but not all of the observed psychophysical response to a vibrotactile stimulus.