The toolish hand illusion: embodiment of a tool based on similarity with the hand

Abstract

A tool can function as a body part yet not feel like one: Putting down a fork after dinner does not feel like losing a hand. However, studies show fake body-parts are embodied and experienced as parts of oneself. Typically, embodiment illusions have only been reported when the fake body-part visually resembles the real one. Here we reveal that participants can experience an illusion that a mechanical grabber, which looks scarcely like a hand, is part of their body. We found changes in three signatures of embodiment: the real hand's perceived location, the feeling that the grabber belonged to the body, and autonomic responses to visible threats to the grabber. These findings show that artificial objects can become embodied even though they bear little visual resemblance to the hand.

Figure 1

Study summary. Experiment 1 (N = 16) aimed to assess whether embodiment for a tool can be induced with an RHI-like setup. As a control, we used a balloon as in Ma and Hommel, 2015. We measured brushing-induced proprioceptive drift of the index finger and subjective feelings of embodiment using a questionnaire. Experiment 2 (N = 40) tested tool embodiment in a larger sample and investigated the role of functional use of the tool. The same measures as in Experiment 1 were used. Experiment 3 (N = 32) investigated the physiological correlates of embodiment, measuring skin conductance responses after threat to the embodied tool.

Figure 2

Perceived index finger position. In experiment 1, after 2 min of synchronous brushing, participants (N = 16) localised their index finger drifted toward the tool. The same participants did not show significant drift after synchronous brushing of the balloon. Error bars indicate 95% C.I.

Figure 3

In Experiment 2, participants (N = 40) perceived their index finger drifted toward the tool only after synchronous (upper panel), but not after asynchronous (lower panel), brushing. Error bars indicate 95% C.I.

Figure 4

In Experiment 2, participants (N = 40) showed higher agreement for statement regarding changes in perceived touch location (Q1, Q2 and Q3 – light blue bars) and tool embodiment (Q5, Q6 and Q14 – yellow bars) after synchronous stimulation only. Error bars indicate 95% C. I.

Figure 5

In Experiment 3, participants (N = 32) showed a Galvanic Skin Response to the threat of the tool after it was brushed synchronously, but not asynchronously, with their own index finger. Error bars indicate 95% C.I.