Thermal phantom sensations in arm amputees and what it means for future prosthetics

Comment on: Iberite, F., Muheim, J., Akouissi, O., Gallo, S., Rognini, G., Morosato, F., Clerc, A., Kalff, M., Gruppioni, E., Micera, S., & Shokur, S. (2023). Restoration of natural thermal sensation in upper-limb amputees. Science.2023;380(6646):731-735.doi:10.1126/Science.adf6121

For a long time, the focus of prosthetic development for amputees was on designing robotic devices to restore motor functions. In the last decade, several groups, including ours, have aimed to enhance prosthetics with sensory feedback. A new generation of bionic limbs allows amputees to perceive objects’ shapes, sizes, and textures [1]. By using neural implants that can interface with nerves, it is possible to target sensory fibers. Electrical stimulation of these fibers has been shown to consistently induce tactile sensations. In a pivotal study, Valle and colleagues proposed several encoding strategies to improve the naturalness of sensation [2]. They found that by (bio)mimicking the stimulation frequency of mechanoreception at the skin, the user’s phenomenological experience was improved: participants rated the tactile sensation produced by these biomimetic stimulation encodings as more “natural.” But, in the quest to restore the rich palette of sensory feedback for prosthetic users, one sensory modality has often been neglected: thermal sensation.

Adding thermal feedback for prosthetic users can allow the perception of nuances of haptic sensation, for example, the detection of objects’ materials based on their thermal properties (e.g. copper at room temperature cools down the skin more than plastic). But even more important is the presence of thermal sensation for the social aspect of touch and pleasantness. Adding thermal sensation to bionic hands could significantly improve users’ embodiment of their prostheses. In a recent work [3], we presented a finding that allowed us to implement a simple yet effective way to restore natural thermal sensation in upper limb amputees. We describe here our findings and the future steps that could boost the development of bionic hands with integrated thermal feedback.

We performed detailed cartography of the residuum of 27 upper limb amputees using a custom-made thermode (Metaphysiks, Switzerland) set at three temperatures: cool (25°C), neutral (32°C), and warm (37°C). We found that thermal stimulation of some of the spots of the residual arm triggered a thermal sensation in their phantom hand. Participants reported all sorts of sensations in their phantom hand; one participant said, “I feel the thermal sensation is irradiating into my missing hand.” In another case, when a very cold stimulation was applied to the residuum, the participant said, “My phantom finger seems frozen.” Overall, in 17 out of the 27 participants, we found that at least one spot on the residuum was projected in the phantom hand. Importantly, we found that thermal phantom sensations were stable over time (the thermal stimulation of a given point in the residuum projected to the same point in the phantom hand over several days) and that they were phenomenologically similar to the thermal sensations on the intact hand. Indeed, participants reported that thermal stimulation on their phantom hand and thermal stimulation on the mirrored spot on the intact hand produced a “matching” sensation both in terms of temperature and position.

Exploiting these findings, we developed a wearable device that can integrate existing prosthetic hands. The system, called the MiniTouch, mediates the thermal feedback monitored at the fingertip of a prosthetic to a Peltier element that stays in contact with the skin in one of the projecting spots on the residuum (Figure 1). An active warming element in the sensor ensures that the prosthetic finger maintains a baseline temperature close to the human skin ($\sim$ 32°) and mimics the thermal flow during contact with a thermally conductive material (e.g. when touching copper). Blindfolded amputee participants wearing the MiniTouch could detect the temperature of various objects in real-time and discriminate between different materials.

Figure 1. Transradial amputee testing the MiniTouch system. The MiniTouch’s thermode is placed on the residual arm of the participant. The experimenter places the sensor over different objects with various temperatures (here, a cup with cold water). The participant reports the thermal sensation in the phantom limb produced by this stimulus. During the tests, the participant was blindfolded. The image is printed with the permission of the participant and photographer.

But how could this finding be used for future prosthetics? We discuss the technical and scientific breakthroughs that could boost thermal feedback development in future prosthetics. So far, we have provided thermal feedback passively: the thermode was placed on one of the thermal projected spots, and the experimenter used the sensor to touch different objects. Tests with the MiniTouch integrated inside the prosthetic socket are necessary to investigate whether participants can functionally use thermal sensation in close-loop tasks (e.g. grasping or scanning an object) and see if the thermal sensations are maintained with long periods of use. Energy efficiency and heat extraction are some of the challenges of this integration.

While we found phantom thermal sensations in most participants, it is unclear why some amputees have these sensations and others do not. At this stage, we can only speculate. The type of accident and the surgical procedure following the amputation might influence the presence of phantom sensations; for example, amputation due to electrocution could necessitate a skin graft, which could change the sensory perception. Another factor might be the possible cortical reorganization following the amputation [4]. Ultrasound analysis of the nerves in the stump and comparative fMRI studies of amputees with and without thermal phantom sensation are necessary to elude this point.

The absence of phantom thermal sensation poses a question: How can we help those who lack it? One solution is thermal remapping, which could involve mapping thermal sensitivity to a different body part. A study by [5] compared different body parts and found the abdomen to be a promising spot for detecting thermal stimuli. This approach could lead to a personalized solution for all prosthetic users.

Finally, while participants with severe phantom limb pain were excluded from the current study, we had several cases of individuals with low levels of pain describing the thermal sensation as “reducing the discomfort of their phantom limb.” This is not completely surprising, as the presence of sensory feedback in prosthetics has been shown to have a positive effect on reducing phantom limb pain [6]. Building on this idea, adding thermal feedback to lower limb prosthetics might be an elegant solution for this critical clinical challenge. In conclusion, restoring thermal sensation in amputees could add a new dimension to the already rich field of sensory feedback for prosthetic users.