Australasian Science: Australia's authority on science since 1938

Dead Hands and Phantoms

Credit: iStockphoto

Credit: iStockphoto

By Lee Walsh, Janet Taylor and Simon Gandevia

Recent studies have highlighted how central signals in the brain can change our sensation of the position and movement of joints, and how phantom limbs form when sensory information is lost.

You have probably woken at night occasionally with a “dead” hand or arm due to compression or stretching of the nerves going down your arm. The hand can be both paralysed (it won’t move when you try to move it) and anaesthetised (you can touch it with your other hand, but it feels numb and “dead”).

Logically you might expect to feel nothing – or the absence of an arm – when your brain receives no input from sensory nerves. Instead, you feel the presence of a real “dead” hand.

Similarly, you may have had dental anaesthesia. The anaesthetised tongue, jaw or lips again do not disappear, but remain in your consciousness, commonly feeling larger than usual.

What is going on? Awareness of a phantom body part develops rapidly and seamlessly in our brain without any cautionary clue that the brain has “created” something. It suggests that the brain maintains some representation of the body that does not rely on ongoing sensory input, so that a hand, for example, can be perceived despite the lack of sensory input from the hand to signal its existence.

The creation of a phantom happens in less common circumstances, such as when a limb has been amputated. The clinical descriptions of such cases are varied and even bizarre, probably because some nerves to the removed limb, or their central connections, are not completely “silent” or the nerves are connected to damaged peripheral structures.

Recently we used anaesthesia and paralysis of the arm to generate a phantom arm in healthy volunteers. We wanted to know whether a phantom arm would move in the direction that you willed if you tried to move it.

This is a fundamentally important question in understanding proprioception, the “sixth” sense that tells us where the parts of our body are and whether they are moving. While we know that sensory signals from receptors in the limb muscles and the skin around joints can generate sensations of altered position and movement of parts of the body, it was uncertain whether the brain’s command to move also plays a role in these sensations. Previous results in healthy subjects had suggested this command did nothing, but clinical anecdotes suggested otherwise.

To settle the question, we simulated a phantom hand in healthy subjects whose right arm was rested in a frame that allowed the experimenter to move the wrist. A blood pressure cuff on the upper arm was inflated to block the blood supply – after about 40 minutes the arm beyond the cuff was anaesthetised and paralysed.

Before inflation of the cuff, subjects accurately signalled the angle of the wrist. But when subjects had a dead hand and perceived a phantom, the position of the phantom did not change when the real hand was moved passively.

The crucial test was whether the phantom hand appeared to move or change position when subjects tried to push their hand in one direction or the other with about 30% maximal strength. The effort produced no movement as the arm was paralysed, and no sensory feedback as it was anaesthetised. However, all subjects perceived, and signalled, that their hand “moved” about 20° in the same direction as it was willed to go. Thus the brain’s “command to move” had itself influenced sensory parts of the brain to generate the altered proprioceptive sensation.

Further work showed that the arm did not “snap” from one position to another, but it moved smoothly, influenced by the exact size and duration of the effort. By abolishing peripheral information about the real position of the joint, we unmasked a role for the central command signals.

During the 40 minutes while we waited for the arm to be fully anaesthetised and paralysed, subjects generally reported that they felt their fingers and wrist become flexed. These reports prompted us to perform new research to track the perceptual evolution of the experimental phantom hand. Was there a “default” posture occupied by the phantom, perhaps genetically determined, or could the final posture be changed by the history of evolving sensory loss?

To examine this, subjects were studied over 2 days. On one day, the wrist and fingers were comfortably flexed and kept in that posture throughout. On the other day, the hand was kept with the wrist and fingers extended. Subjects repeatedly showed us the perceived postures of wrist and finger joints using an artist’s articulated hand.

After about 40 minutes of cuff inflation, when there was no useful sensory information coming to the brain from the arm and hand, the subject felt that the evolved phantom was not in the same posture that the hand had started in. When the subject’s actual wrist and fingers were flexed, the phantom was in an extended posture. Conversely, when the subject’s wrist and fingers were in an extended posture, the phantom was flexed. The final phantom was in a different posture in the two conditions: surprisingly, it was more flexed after the hand was held out extended. Thus there is no default posture for a phantom limb.

We speculate that the progression of the block produced by the blood pressure cuff gradually changes the sensory input so that the loss of dominant signals of flexion is interpreted as the phantom wrist and fingers gradually extending, with the opposite occurring when dominant signals of extension are lost. This progressive loss of sensory input is unlikely to mimic a traumatic amputation.

When people perceive a phantom hand, even though the sensation is unusual they do not doubt that it is their own hand. Something about the brain’s representation of a hand is retained.

But this sense of “ownership” is labile. In the well-known “rubber-hand” illusion, if you see a rubber hand on the table in front of you being stroked with a brush and at exactly the same time we gently stroke the same part of your own hand, you soon believe that the rubber hand is your own. This transfer of your bodily “ownership” to an inanimate artificial hand relies on signals from skin receptors evoked by brushing. The transfer is so complete that a threat to stab the rubber hand elicits a physiological “ouch” response.

While this manipulation of ownership requires signals from the skin elicited by the brushing, we wondered if signals from muscle stretch receptors could also generate the same transfer. Indeed, they can. If your real index finger is moved passively while you can’t see it but can see a magician’s finger moving in time with your finger, this causes you to “own” the false finger. When you are asked the location of your own finger, you think that it is much closer to the magician’s finger than it really is.

Importantly, this transfer even occurs when the skin of the finger is blocked with a local anaesthetic. This leaves the receptors in muscles that move the finger as the only source of information about passive finger motion.

Skin receptors and muscle stretch receptors have one thing in common. Your brain can only receive signals from them if something happens to your own body. This is different to vision or hearing, which can tell you about something touching or moving someone else’s hand. It seems that it is the coincidence of sensory inputs, one of which can only come from your body, that tells you that something you can see or hear actually belongs to you.

These studies have highlighted several areas about conscious awareness of our own bodies when we are at rest and when we are making movements. Although we have a definite sense of what belongs to our body, this can be manipulated by signals from the skin and muscles. This sense is labile – it can be transferred to an inanimate object that is external to our own body.

Better understanding these disturbances better is important for two reasons. First, we need to know how information about proprioception is used in conscious awareness of our body and the control of our movements and posture. Second, we need to understand how disturbances to these systems arise. As examples, in schizophrenia the sense of control of voluntary actions can be altered, while after a stroke the affected limbs may be neglected or thought not belong to their owner.

Lee Walsh completed his PhD on proprioception at Neuroscience Research Australia under the supervision of Janet Taylor and Simon Gandevia.