Tuesday, 30 September 2014

Enacting a Phantom

Ken Pepper
This post is by Ken Pepper, who recently submitted his PhD at the University of York.

Amputees often feel 'phantom' sensations emanating from their missing limb (for a review, see Giummarra et al 2007). This entry discusses the role of action and perception in the constitution of these physically absent yet phenomenally present body parts. I urge the view that phantoms are to some extent enactive – they are constituted by active perceptual engagement with the world (see e.g. Noë 2004).

Impressed by the way in which a blind man localised sensations at the tip of his cane, Head and Homes (1911) hypothesised that his brain must update its representation of bodily posture on the fly and treat the cane as part of his arm. It turns out that they were correct; neural representations of limb locations are highly adaptable and continually modified by vision, touch, and kinaesthesia. Experiments on macaques reveal that while using a rake to retrieve food, the receptive field of neurons in somatosensory cortex – the location of the brain's inner model of the body – extends beyond the boundary of the monkeys' arm to incorporate the area occupied by the rake (Iriki, Tanaka & Iwamura1996).

Similar effects have been observed in human tool use also (Maravita and Iriki 2004, Maravita, Spence, and Driver 2003). While using a grasping tool (like the device used by park keepers to collect litter) to estimate the size of an object, the brain treats the haptic signal as originating from the tip of the tool rather than the hand (Takahashi, Diedrichsen & Watt 2009). Following this kind of tool use subjects are temporarily prone to overestimate the distance between tactile sensations on their arm, suggesting that for a short time their brains continue to represent their arm length as extended (Cardinali et al 2009).
The flexibility of the brain's representation of the body strongly supports the view that the neural representations underlying it are action-oriented. Action-oriented representations are minimally contentful sub-personal states which are poised to guide behaviour in a contextually sensitive way (Clark 1997: 49). In other words, action-oriented representations depend for their content on what the subject is currently perceiving and doing, and this content is 'wide' as opposed to 'narrow' –what they signify depends on the environment in which they are embedded rather than intrinsic properties of the system in which they inhere (see Putnam1975; Burge 1979).
Phantom limbs owe to the preservation of neural representations of absent body parts. If these neural representations are action-oriented, then phantom limbs should also be amenable to contexts of perception and action. And indeed they are. Though some patients experience their phantoms as immobile many report being able to move them intentionally. These movements can constrain the movement of surviving body parts in the same way as real limbs (Franz and Ramachandran 1998).
One patient born without arms feels two short phantoms which gesticulate in conversation, but slump motionlessly while walking (Ramachandran and Blakeslee 2005 :40–2). Immobile phantoms can be revived using mirrors to create the appearance of the subject's intact hand taking the place of their missing hand, thereby fooling the patient's brain into attributing the outcome of their intention to act to the perceived movement, and can be used to ease spasms felt in the missing hand (Ramachandran, Rogers-Ramachandran and Cobb 1995).
This newly acquired felt movement is transient, and is lost without visual perception of the superimposed limb (Ramachandran and Blakeslee 2005: 53). Conversely, phantom movement can be eliminated through 'learned paralysis', where the lack of visually perceived movement nullifies the subject's act intentions (Ramachandran and Blakeslee 2005: 46). Amputees who use prostheses tend not to feel phantom pains, but come to feel their phantom as being embodied by the prosthetic limb (Giummarra et al 2007 pp.222–3), and it is thought by some doctors that phantoms are essential to successful use of a prosthetic limb (Sacks 1985: 71). Skilled users of prostheses can even gain the ability to scratch phantom itches by scratching their prosthetic limb (Giummarra et al 2008: 154).
The surviving neural representations of absent body parts depend heavily for their content on perception and action. So although the experience of phantoms owes initially to preserved neural representations of missing body parts, the content of these representations, and what shows up in the subject's experience on their basis, is determined by the subject's active engagement with the world. Hence the malleable nature of phantoms. Far from being passive, private mental episodes, phantom limbs are a largely enactive phenomenon.

1 comment:

  1. So what if itches are actually some realignment method for the phenominal model? The itches are 'model checks', to update the overall model (which changes as a creature evolves - though that takes a long time to happen, the modeling system itself is very old as well)


Comments are moderated.