Spatial information from cutaneous and muscle receptors contributes to perceived finger position and ownership

There is little known about the physiological basis for how we know that our limbs are a part of our body. Our understanding of the mechanisms of body ownership has recently been strengthened by studies using the rubber hand illusion (RHI). A rubber hand placed in view and stroked in time with the participant's hand, which is out of view, results in the perception that the rubber hand is their real hand. This sense of ownership of the hand develops over tens of seconds. A probabilistic strategy best explains the RHI, whereby an internal body model is compared to incoming sensory feedback, resulting in the conscious perception of the rubber hand as the real hand. Specifically it predicts that there is a misattribution of sensation and ownership to an object when it resembles a real hand and is in an anatomically plausible position. Further, there needs to be temporal coincidence and integration of the incoming sensory signals that indicate that sensation is arising from the object. However, there is the possibility in the absence of temporally coincident signals that ownership could be induced by these other factors.

Héroux et al. (2013) found evidence that perceived ownership of an artificial finger can be induced in the absence of vision using information from muscle receptors. In the study the left arm was positioned 12 cm above the right arm such that they were in vertical and posterior–anterior alignment. In the critical condition the left index finger and thumb were clamped around an artificial finger. In congruence with the notion that the left hand was grasping the right index finger, a silicone clamp applied pressure to both sides of the right index finger. Synchronous passive movements of the two hands resulted in the illusion that the hands were closer together, a proprioceptive drift. Surprisingly, passive grasping without movement also resulted in a decrease in the perceived distance between the hands. This grasp illusion critically differs from the RHI paradigm as it occurs within a single modality and can create the additional illusion of self‐touch in some individuals. The mechanism underlying this illusion is fundamental to proprioception, to differentiate between touching your own finger, someone else's finger, or a finger‐like object. Furthermore, this grasp illusion might shed light on the relation between ownership and proprioceptive drift, given there is a lack of consensus within the field over whether they represent separate central processes (Abdulkarim & Ehrsson, 2016).

In a cleverly designed study recently published in The Journal of Physiology Héroux and colleagues (2018) used their grasp illusion paradigm to test if the physical resemblance of the artificial finger to a real finger influences the strength of embodiment. Two separate studies were performed, each with 30 subjects. Experiment 1 tested if the grasp illusion causes a shift in perceived finger position and an illusory sense of ownership of the artificial finger. Experiment 2 manipulated the physical characteristics (temperature, shape, texture, compliance) of the artificial finger to create incongruence with those of the real finger, to test if these properties are important for embodiment.

An important aspect of the study was the integrity of the methods used to analyse and report the data. The authors preplanned every aspect of the data analysis and did not perform any exploratory analyses. To aid the reader to evaluate whether the effects observed were consistent across subjects, all individual data points were shown in the figures. Further, in the figures and the text all summary values were reported as means with 95% confidence intervals to easily convey significant findings to the reader without having to report any P‐values.

In experiment 1 the perceived vertical spacing between the fingers reduced by ∼3 cm in 3 min, in grasp and no grasp conditions. Grasping resulted in an additional ∼5–6 cm compression compared to not grasping. Interestingly, within the 3 min of grasping, the percentage of subjects rating ownership above neutral increased from 33% to 50%.

In experiment 2 changes to the physical characteristics of the artificial finger elicited substantial changes to ownership but only subtle modulations of proprioceptive drift. Compared to the control finger, ownership ratings were reduced for the cold finger (12°C), whereas the hot finger (48°C) made no difference. There were robust reductions in perceived ownership when an odd‐shaped or rectangular finger was used. A finger with rough texture produced lower ownership ratings than control, whereas using a smooth finger or changing the compliance of the artificial finger did not change ownership relative to the control finger.

Critically, Héroux et al. show that somatosensory information is sufficient to embody an object. The authors propose that the brain compares the current sensory input to the internally stored body model to generate conscious perception of the most likely representation of the body. In their study, the cutaneous and proprioceptive signals from each of the finger and thumb suggested that a finger‐like object was being held, and the index finger received cutaneous signals that were consistent with it being grasped by a finger and thumb. Interestingly there was a strong instantaneous effect of the grasp illusion, unlike the RHI, which tends to develop over tens of seconds (40–60 s, Rohde et al. 2011). The strength of the grasp illusion also increased with time, in agreement with a probabilistic model, whereby evidence accumulated with time that the artificial finger is in fact the real finger.

On the basis of the physical characteristics evoking changes in ownership but not perceived limb position, the authors consider the possibility that perceived limb position and body ownership are likely served by different processes. This hypothesis is consistent with the finding that simulating proprioceptive drift during the induction of the RHI, by manipulating the perceived position of the real hand through mechanical displacement at a speed below movement detection threshold, does not dampen the feeling of ownership of the rubber hand (Abdulkarim & Ehrsson, 2016).

Héroux et al. suggest that the cold finger was less readily embodied as cold temperatures alter the integration of incoming cutaneous signals more than heat. An alternative hypothesis is that the congruence of the sensory signals from the two hands could modulate the strength of ownership of the artificial finger. Maximal congruence would be achieved when the difference in temperature of the grasping hand to the artificial finger is consistent with the difference in temperature of the clamp on the index finger. For instance, congruent signals are given if the left hand feels a cold finger‐like object, and the clamps touching the right index finger are hot. Still, this does not explain why the hot finger was no less embodied than the control finger. A possibility is that an unpleasant object is less likely to be perceived as belonging to the self, as cold is more unpleasant than heat (Greenspan et al. 2003). This idea is also consistent with the reduced ownership ratings for the rough finger.

The authors propose that the illusory ownership of the artificial finger does not rely on any temporal inputs and is purely derived from spatial information that is congruent with it being the real finger. However, there is the possibility that there is an internal temporal signal available that in conjunction with this spatial information led to a sense of ownership of the finger. This temporal signal could arise from visceral information (interoception), most likely the heartbeat. In the current study, the fingertips have potential (albeit weak) temporally correlated information from the heartbeat, in conjunction with cutaneous information that they are touching a finger‐like object. Importantly, the resting heart rate is similar to the 1 Hz brushing frequency used to optimise ownership in many RHI studies. To show that interoceptive information can modulate conscious ownership, Suzuki et al. (2013) created a variant of the RHI in which the colour of a visually displayed hand was changed in synchrony with the heartbeat. In addition, they found that people that were more aware of their heartbeat showed a stronger illusion of ownership. As in Suzuki et al. (2013), the grasp illusion could arise from integration of external stimuli and interoceptive sensory signals. To test if temporal inputs in the fingertips from the pulse contribute to the grasp illusion, one could apply a pressure cuff to one or both of the arms to create an ischaemic block, and test for the immediate effect of the grasp illusion before numbness sets in.

As the grasp illusion provides the illusory sensation of self‐touch, it has clinical relevance for those with abnormal perception of self‐generated sensations arising from movement. For example, when such a patient grasps their own finger they could feel that someone else is grasping their finger. These errors in the process of the attenuation of self‐generated sensations occur in conditions such as Parkinson's disease and schizophrenia. Therefore, the grasp illusion might provide a useful tool to study these patients.

In conclusion, Héroux et al. have provided important evidence regarding the role of cutaneous and muscle receptors in the perceived position of the fingers and to update the internal model of the body. In conjunction with previous literature, they show that the brain considers different sensory information, depending on the context, to come to the most likely representation of the body. Although their study suggested that static signals from external stimuli explain the grasp illusion, the possibility remains that this is contingent upon contributions from each finger of temporally correlated interoceptive signals. In addition they show that the physical characteristics of the artificial finger affects its embodiment. The grasp illusion provides an exciting paradigm for increasing our understanding of how the hands work together and the physiological basis of body ownership. Furthermore it has potential as a tool to improve our understanding of conditions of abnormal perception of self‐generated sensory perception signals.

Additional information

Competing interests

None declared.