# Artificial vibrotactile feedback elicits neural correlates of sense of agency

## TL;DR

This study shows that artificial vibrotactile feedback can create a sense of control over movements, which is important for assistive devices like prosthetics.

## Contribution

The study introduces vibrotactile feedback as a novel method to elicit neural correlates of sense of agency in assistive technologies.

## Key findings

- Vibrotactile feedback significantly modulated the N100 ERP component, indicating a neural correlate of sense of agency.
- Vibrotactile feedback enhanced implicit neural measures of agency, while auditory feedback had the opposite effect.
- Implicit neural measures were more sensitive to self-generated outcomes than explicit or behavioral measures.

## Abstract

The Sense of Agency (SoA) refers to the subjective experience of having control over our own actions and their outcomes. SoA is experienced when there is a match between the predicted and actual sensory outcomes of an intended motor action. Thus, receiving sensory feedback related to motor actions is essential for the formation of the SoA. As a consequence, individuals with sensory loss due to neuropathies or limb impairments might experience reduced SoA. Therefore, incorporating artificial sensory feedback in assistive technologies (e.g. prosthetics, exoskeletons, sensory substitution systems) is crucial for patients to properly experience SoA and thereby enhance the effective use of such technologies. In this study, with the use of neuro-cognitive measures, we validated the effect of a vibrotactile sensory feedback system on SoA for lower-limb movements.

Healthy participants (n = 25) performed a stepping task while receiving vibrotactile sensory feedback and, in a control condition, auditory feedback. We employed SoA measures to validate the sensory feedback system, in particular the N100 ERP component of the electroencephalography (EEG) signal as an implicit neural measure of SoA, temporal interval estimation tasks as an implicit behavioural measure, and SoA questionnaires as an explicit measure. Statistical analyses, including repeated-measures ANOVA and t-tests, were conducted to examine the effects of externally generated versus self-generated stimuli on N100 modulation, temporal interval estimations, and SoA questionnaire results.

Our results showed a significant modulation of the N100 component for self-generated, relative to externally generated, sensory feedback. The direction of the effect was opposite for vibrotactile and auditory sensory feedback, with enhancement in vibrotactile and attenuation in auditory modality. In contrast, no notable effects were found in interval estimation tasks or explicit agency measures, pointing to a possible dissociation between implicit and explicit measures of SoA and indicating higher sensitivity of implicit neural measures.

Sensory processing of the vibrotactile sensory feedback is sensitive to the difference between outcomes that are actively generated compared to passively received, as indexed by the modulation in amplitude of the N100 component. This is an indication that vibrotactile sensory feedback is capable of eliciting sense of agency. Additionally, we found that modulation of sensory processing is more sensitive than temporal interval estimates to self- vs. externally-generated outcomes. Therefore, we suggest using it as a metric for the design and validation of assistive technologies.