The amputation of a limb is accompanied by alterations in body representation such as a changed perception of the missing limb. Amputees in general (Nico et al. 2004) and especially those experiencing PLP (Reinersmann et al. 2010) show an increase in reaction time when asked to mentally rotate a hand representing their missing limb. This delayed response has been viewed as indicative of a dysfunction in the processing of the body schema, which refers to more implicit aspects of body representation compared to the body image, which has been viewed as an explicit, conscious percept. As implicit and explicit aspects of body representation are dynamic and closely interwoven and cannot always be separated, we prefer the term body representation. Other basic processes involved in body representation, however, seem to be less affected. Thus, amputees can be induced to perceive a rubber hand as belonging to their body, when the seen rubber hand is synchronously touched together with the hidden residual limb (Ehrsson et al. 2008). The underlying neuronal principles in amputees (Schmalzl et al. 2014) appear to be similar to those in healthy controls (Ehrsson et al. 2004). No significant differences between amputees with and without PLP related to the responsiveness to this so-called rubber hand illusion have been found (Ehrsson et al. 2008). The intact ability of amputees with PLP to perceive an artificial limb as belonging to their body might be useful for the treatment of PLP. For example, functional prosthesis use has been found to be negatively associated with both dysfunctional cortical reorganization and PLP intensity (Lotze et al. 1999; Weiss et al. 1999), indicating that restoring the amputee’s body integrity might alleviate or prevent PLP. There is evidence that the perception of ownership of the prosthetic device plays a mediating role in this process, with more intense ownership experience of the prosthesis being associated with lower levels of PLP (Kern et al. 2009). Similarly, sensory feedback from the prosthesis (Dietrich et al. 2012) might enhance perceived ownership. However, amputees who report a telescopic distortion of their phantom stated less often the occurrence of ownership sensations for the prosthesis (Giummarra et al. 2010), suggesting that the presence of a telescope might influence the effectiveness of PLP treatment as well. In line with this, Foell et al. (2014) reported that the presence of a telescope is an important predictor for the effectiveness of mirror therapy. During mirror therapy, patients perform movements with their intact limb in front of a mirror and have to mentally combine the seen movements with the self-executed movements of the phantom. Amputees who experienced a telescope failed to relate the seen movement to the felt movement and did not benefit from the mirror intervention (Foell et al. 2014). Since the sensation of a telescope is associated with similar reorganizational processes like PLP (Grüsser et al. 2001), this finding highlights the importance of body perception and its neural correlates for the treatment of PLP.

It has been proposed that “somatosensory memories” of non-painful and painful sensations in the missing limb play a role in phantom limb awareness and PLP (Katz and Melzack 1990; Katz 1992; Flor et al. 2006). Anderson-Barnes et al. (2009) proposed that “proprioceptive memories” could explain sensations in a phantom limb and that there might exist learned associations between proprioceptive memories and pain perceived before amputation. Before amputation, pain in the affected limb is common, for example, due to a tumor, vascular disease, or injury. Katz and Melzack (1990) suggested that these types of pain are encoded and can later be triggered, for example, by peripheral input from the residual limb, and experienced as PLP. Support for this comes from retrospective reports showing a relationship between reported memories referring to the phase before or during amputation and later phantom sensations (Katz and Melzack 1990; Giummarra 2011a). In Katz and Melzack’s (1990) study, almost 60 % of amputees who reported some kind of pain before the amputation also reported that painful sensations continued or recurred in the phantom limb. It has been proposed that neural plastic changes following long-term nociceptive input can be seen as a neuronal mechanism underlying pain memories (Flor 2003, 2008). Maladaptive plasticity associated with PLP might be more severe if chronic pain precedes the amputation. The concept of pain memories and pain prior to the amputation is closely linked to the question whether the formation of pain memories and subsequent PLP can be prevented if nociceptive input is blocked before amputation, for example, by the use of anesthetic drugs. However, the evidence on this is controversial (Ypsilantis and Tang 2010; Jensen and Nikolajsen 2000). In a prospective study, Jensen et al. (1985) and Nikolajsen et al. (1997) showed that PLP during the first 6 months but not long-term PLP was predicted by pain before amputation. However, usually long-standing pain was not taken into account.

Pain catastrophizing is an exaggerated, negative orientation towards pain and has been found to predict chronic pain and impairment (Flor et al. 1993; Sullivan et al. 1995; Linton and Shaw 2011) as well as a negative outcome (Linton and Shaw 2011; Wertli et al. 2014). In several studies, catastrophizing has been shown to be significantly positively correlated to the magnitude of PLP in amputees (Hill 1993; Hill et al. 1995; Jensen et al. 2002; Hanley et al. 2004; Richardson et al. 2007; Vase et al. 2011, 2012). Jensen et al. (2002) and Hanley et al. (2004) showed that catastrophizing 1 month after amputation was correlated with concurrent PLP and depression. However, high catastrophizing was associated with an improvement of PLP and depression at 6 months or 2 years later. This finding seems contradictory. However, these results might be related to regression to the mean: due to the high correlation of PLP and depression with catastrophizing at the first time point, subjects with high values in catastrophizing show high values in depression and PLP as well, leaving subjects little chance to further aggravate. Hence, the lagged relationships in these two studies should not be interpreted in a way that catastrophizing predicts improvement but rather that the initial magnitude of pain needs to be taken into account as well. Richardson et al. (2007) showed that catastrophizing before the amputation predicted PLP 6 months after the amputation such that a high degree of catastrophizing was associated with more PLP. Catastrophizing has also been examined with respect to coping with disability, showing that catastrophizing predicted physical and psychosocial disability in amputees (Whyte and Carroll 2004). Vase et al. (2011) showed that catastrophizing accounted for 35 % of the variance found in PLP even after statistically controlling for depression and anxiety. Moreover, catastrophizing also correlated with wind-up-like pain, elicited by pinpricks at the residual limb. The wind-up test is a dynamic pain measure in which moderately painful stimuli of the same intensity are repetitively presented at the same site. Usually, stimuli are perceived as increasingly painful. This measure is seen as an indicator of amplification of peripheral nociceptive input. The authors assumed that both catastrophizing and wind-up interact and contribute to PLP and that, given the trait-like nature of catastrophizing and the fact that it precedes PLP, catastrophizing might lead to wind-up. In another study (Vase et al. 2012), electroencephalography was used to record cortical responses to noxious and non-noxious stimuli presented at the affected and non-affected limb. For the affected side, there was a correlation between catastrophizing and the root mean square power of the N/P135 dipole, which was located in the area of secondary somatosensory cortex, known to play a role in attentional processes. The authors interpret this finding as an indicator that catastrophizing relates to hypervigilant attention for noxious and non-noxious stimuli.

Pain coping strategies describe various ways to “deal” with pain after it has been attended to and interpreted (appraised) as being a threat (Rosenstiel and Keefe 1983; Linton and Shaw 2011) and can be divided into cognitive and behavioral strategies. Examples of cognitive coping strategies are distracting attention from a sensation or reinterpreting pain (Hill 1993). Behavioral coping refers to strategies like increasing or decreasing social or physical activity or seeking social or medical support (Hill 1993; Linton and Shaw 2011). Coping with PLP was first systematically been studied by Hill (1993) who used the Coping Strategies Questionnaire in 60 male amputees with PLP. A principal component analysis in the amputee sample revealed a factor structure that was similar to the one originally discovered for chronic back pain patients (Rosenstiel and Keefe 1983). Three main components were found which the authors called “cognitive coping,” “helplessness,” and “pain denial.” They explained about 20 % of the variance in both PLP and psychological distress. An analysis of subscales revealed that catastrophizing was by far the most powerful factor accounting for most of the variance explained by the “helplessness” factor. The authors concluded that PLP sufferers use a limited amount of coping strategies that help to alleviate distress and pain and that “successful” coping rather means not to catastrophize. In another study, Hill et al. (1995) found that catastrophizing explained 26 % of the variance in pain as opposed to other strategies that only explained 3 %. Whyte and Niven (2011b) examined 89 amputees with pain diaries assessing PLP and coping strategies. Unlike other studies, strategies were captured in a free format without standardized questions. Diary entries were made once per hour for 1 week. The participants used a limited number of strategies falling into the categories of distraction, relaxation, seeking support, exercise, manipulation of the residual limb, and drug or alcohol use. Interestingly, none of the reported strategies turned out to be effective in reducing PLP. This study confirms that PLP sufferers have few effective coping strategies.