Assessing the quality of supplementary sensory feedback using a crossmodal congruency task

Abstract

Peripheral nerve interfaces show promise in making prosthetic limbs more biomimetic and ultimately more intuitive and useful for patients. However, approaches to assess these emerging technologies are limited in their scope and the insight they provide. When outfitting a prosthesis with a new feedback system it would be helpful to quantify its physiological correspondence, i.e. how well the experimental feedback mimics the perceived feedback in an intact limb. Here we present an approach to quantify physiological correspondence using a modified crossmodal congruency task. We trained 60 able-bodied subjects to control a bypass prosthesis under different feedback conditions and training durations. We find that the crossmodal congruency effect (CCE) score is sensitive to changes in feedback modality (multi-way ANOVA; F(2,48) = 6.02, p<0.05). After extended training, the CCE score increased as the spatial separation between expected and perceived feedback decreased (unpaired t-test, p<0.05). We present a model that can quantitatively estimate physiological correspondence given the CCE result and the measured spatial separation of the feedback. This quantification approach gives researchers a tool to assess an aspect of emerging augmented feedback systems that is not measurable with current motor assessments.