The pain or neuronal memory hypothesis revolves around the often-cited experience that phantom pain resembles the pain felt before amputation (Nikolajsen et. al., 1997). The hypothesis attributes phantom limb pain to neural representation of pre-amputation pain that are formed in the somatosensory cortex and strengthened before amputation. These representations encode a particular pattern of nerve impulses associated with pain, which remain after amputation occurs and may be activated by input from other areas (Katz & Melzack, 1990). Importantly, these neural representations are not simply recollections but actually activate an output pattern that causes the individual to experience pain.
A related hypothesis is that of ‘learned paralysis’ (Ramachandran & Hirstein, 1998). This theory proposes that during the experience of paralysis prior to amputation the brain continues to send signals to the motor cortex that effect no movement in the limb. The feedback that cortical areas receive, indicating paralysis, are then learned by the neural circuitry. Following amputation, the neural circuitry indicating paralysis is maintained by the lack of sensory feedback, and results in a perceptual/cognitive experience of a painful paralysed phantom limb. Evidence that long lasting noxious input can cause changes to cortical representations in the primary somatosensory cortex has been demonstrated in sufferers of chronic back pain (Flor et al., 2006). The short video below discusses how pain may cause changes to the neural circuitry and persist after the physical experience of pain has ceased.
Consistent with the hypothesised role of preamputation pain, estimates suggest that up to 57% of subjects who had pain at or near the time of amputation reported phantom pain that resembled these sensations (Katz & Melzack, 1990). Furthermore, there is some evidence that preoperative pain blockers can reduce the subsequent effects of phantom limb pain (Jensen & Nikolajsen, 2000).
The pain memory and learned paralysis hypotheses are broadly consistent with Melzack’s Neuromatrix Theory (1999) in that they suggest a representation which remains after amputation and is activated by spontaneous inputs from elsewhere. However, these theories have only garnered limited support. Ramachandran and Hirstein (1998) argue that learned paralysis should theoretically be attenuated by visual feedback that signals to the brain that paralysis has ceased. However, their use of a mirror box paradigm to demonstrate this showed mixed efficacy with 40% patients failing to obtain relief from painful paralysed phantoms. The most obvious criticism of these theories is that they cannot account for phantom pain in those who did not have prolonged period of preamputation pain, such as those experiencing traumatic loss of limb (Weeks et al., 2010) or for those experiencing phantom pain in congenitally absent limbs (Flor et al., 2006; Weeks et al. 2010). NEXT
Phantom Limb Pain 