Perception, 2013, volume 42, pages 473 – 476

doi:10.1068/p7400

SHORT AND SWEET

A new method for assessing self-touch enhancement
of the foot in stroke patients with mobility problems
Rebekah C White1, Anne M Aimola Davies1,2,3
1

Department of Experimental Psychology, University of Oxford, South Parks Road,
Oxford OX1 3UD, UK; e-mail: Rebekah.White@psy.ox.ac.uk; Anne.Aimola@psy.ox.ac.uk;
2
Faculty of Philosophy and NIHR Biomedical Research Centre, University of Oxford, Oxford, UK;
3
Research School of Psychology, The Australian National University, Canberra, Australia
Received 23 October 2012, in revised form 12 March 2013
Abstract. Patient NG is the first reported case of lower-limb ‘self-touch enhancement’ following
stroke. Mobility problems prevented NG from reaching to touch her foot, thus we used a self-touch
rubber-hand paradigm to mimic the conditions of self-administered touch. With vision precluded,
NG administered stimulation to a prosthetic limb while the Examiner administered synchronous
stimulation to NG’s affected left foot. NG detected all stimulation administered with our self-touch
paradigm, whereas in the control condition (with NG not involved in administering stimulation),
NG had failed to detect one-third of Examiner-administered stimulation. When mobility problems are
a barrier to investigating self-touch enhancement, the self-touch paradigm can be used to demonstrate
residual tactile sensation following stroke.
Keywords: body representation, rubber-hand illusion, self-touch, sensation, stroke

Stroke patients with a sensory impairment may not detect touch administered by another
person but they may detect touch of the same intensity when it is self-administered. This
phenomenon, termed ‘self-touch enhancement’ (STE), has been investigated in five
studies. Self-touch enhancement can occur when the patient administers self-touch without
Examiner assistance (van Stralen et al 2011) and when the patient is guided by the Examiner
to administer touch (Valentini et al 2008; Weiskrantz and Zhang 1987; White et al 2010a,
2010b). Possible explanations for how STE occurs include: (1) proprioceptive information,
the patient infers touch based on the relative position of the patient’s administering hand
and the affected body part; (2) spatial attention, the patient’s administering hand provides a
spatial cue that directs attention toward the affected body part; (3) temporal expectation, the
action of the patient’s administering hand provides a temporal cue that focuses attention as
touch is administered. Although there is no reason to presume that STE is limited to the upper
limbs, all five previous studies have focused on the hands. Methodological ease is the most
likely explanation. Self-administering touch to the lower limbs (particularly the feet) can be
challenging, if not impossible, when stroke patients have mobility problems. We describe a
simple paradigm that overcomes these problems.
NG is a 21-year-old female, who experienced a right-hemisphere stroke resulting in
left hemiplegia. (For full details of NG’s neuropsychological assessment, see White et al
2010a, 2010b.) At 3½ months after stroke NG’s motor function had improved, but she had
problems with sensation in her left foot. Whereas individuals without sensory impairment
detect stimulations of 0.07 g intensity with a Semmes–Weinstein monofilament, NG did not
detect ⅓ of Examiner-administered stimulations of 15 g intensity to the left foot. We set out
to assess STE using a self-touch rubber-hand paradigm because NG could not comfortably
reach her left foot to administer self-touch.

474

R C White, A M Aimola Davies

In the self-administered-stimulation condition (with and without localisation cues;
see table 1), the Examiner guided NG’s right hand, which held the Semmes–Weinstein
monofilament, to administer stimulation to a prosthetic hand positioned on NG’s lap. The
Examiner administered synchronous stimulation to the dorsal surface of the toes of NG’s
left foot using a Semmes–Weinstein monofilament of 15 g intensity (see figure 1: top panel).
The procedure was demonstrated to NG prior to testing, thus NG was aware that she was
administering stimulation to a prosthetic hand and receiving stimulation to her foot. NG
wore a blindfold during testing, so that her responses were based on tactile rather than visual
information. After each trial, the Examiner asked NG whether she had detected touch to her
foot and whether she could localise the touch to a specific toe. There were three conditions—
an Examiner-administered condition and two self-administered conditions. Each condition
comprised 60 trials: 10 stimulations to each toe, and an additional 10 sham trials to ensure
reliability of responses. On sham trials in the Examiner-administered-stimulation condition,
no stimulation was administered by the Examiner to NG’s foot. On sham trials in the two
self‑administered-stimulation conditions (with or without localisation cues), no stimulation
was administered by the Examiner to NG’s foot, but NG’s right hand (guided by the Examiner)
administered stimulation to the prosthetic hand. It was important to prevent NG from noticing
which trials were sham trials (eg by the presence or absence of rustling sounds from the
Examiner’s clothing). Thus, for sham trials in all three conditions, the Examiner carried out
the actions for administering stimulation, but paused when the tip of the monofilament was
1–2 cm from NG’s foot.
Table 1. Three conditions for the assessment of self-touch enhancement.
Condition

Procedure

Task

1 Examineradministered
stimulation

Examiner administered stimulation to foot

(a) report if touch
was detected (yes/no)
(b) report where
touch was detected
(1–5; 1 = big toe,
5 = little toe)

2 Self-administered
stimulation with
localisation cues

Examiner administered stimulation to foot, while
guiding NG’s right hand, which held the Semmes–
Weinstein monofilament, to administer stimulation
to the spatially analogous digit of the prosthetic
hand (eg if Examiner administered stimulation to
NG’s big toe then NG administered stimulation
to the thumb of the prosthetic hand)

as above

3 Self-administered
stimulation without
localisation cues

Examiner administered stimulation to foot, while
guiding NG’s right hand, which held the Semmes–
Weinstein monofilament, to administer stimulation
to the back of the prosthetic hand. Because NG
administered stimulation to the same location
(back of the prosthetic hand) on each trial, this
condition allowed us to rule out the possibility
that NG used subtle (leftward and rightward)
movements of her administering hand to help her
with detection and/or localisation

as above

Self-touch enhancement in stroke patients

475

Detections
  localised
  mislocalised

100
Detections/%

80
60
40
20
0

Examiner-administered
stimulation

Self-administered
stimulation with
localisation cues

Self-administered
stimulation without
localisation cues

Figure 1. Top panel: Experimental set-up for three stimulation-administration conditions. Bottom
panel: Localised and mislocalised detections in each experimental condition. Error bars represent 95%
Wilson Confidence Intervals.

Analyses were conducted with the Q' test (Michael 2007), used in single-case analyses to
test the hypothesis of equal proportions. First, we assessed for STE, that is, better detection
of self-administered stimulation compared to Examiner-administered stimulation. Second,
we assessed for differences in stimulus localisation, that is, correct report of which toe was
stimulated.
NG demonstrated STE (figure 1: bottom panel). Although NG detected only 31/50 Examineradministered stimulations, she detected 50/50 self-administered stimulations with localisation
cues, and 50/50 self-administered stimulations without localisation cues [Q'(1) = 18.61,
p < 0.0001]. Crucially, NG did not report touch on any of the sham trials. Thus, it is not that
NG reported touch each time she carried out the actions for self-administered stimulation.
Instead, NG reported touch only when self-administered stimulation coincided with the
Examiner’s stimulation of her affected left foot.
There were no significant differences found for accuracy, when localising which toe
was stimulated, in the three experimental conditions (Examiner-administered = 23/31, 74%
correctly localised; self-administered with localisation cues = 30/50, 60% correctly localised;
self-administered without localisation cues = 29/50, 58% correctly localised; all p values
> 0.196). The task of localising stimulation on the toes is notoriously difficult (Halnan and
Wright 1960) even for individuals without a sensory impairment. It is thus not surprising
that NG mislocalised a large proportion of stimulations. More important is the finding that
NG was not ‘guessing’ about the location of felt touch, and that she must have had some
information about location of stimulation because 87.76% of mislocalisations were to the
neighbouring toe.
When NG administered stimulation to a prosthetic hand while the Examiner administered
synchronous stimulation to her foot, NG may have experienced the illusion of a single selftouch event (Ramachandran and Hirstein 1997). According to a Bayesian explanation, there is

476

R C White, A M Aimola Davies

a low probability that the action of NG’s hand and the sensation on her foot could correspond
so precisely by chance. Thus it may have seemed to NG that she was administering touch
directly to her foot. Notably, this ‘illusion of self-touch’ can only be elicited if the participant
has sensation in the receptive body part (White et al 2010b).
NG is the first published case of ‘lower limb’ STE following stroke. Residual sensation
in her left foot was demonstrated with a paradigm that was designed to overcome the
mobility problems that prevented NG from touching her foot. NG administered stimulation
to a prosthetic hand (positioned on her lap) and received stimulation to her affected left foot
(positioned on the ground). Because the administering hand and affected foot were never in
contact with each other, STE cannot be explained by (1) NG using proprioceptive information
to infer touch or (2) NG’s attention being directed by spatial cues from the administering
hand. Instead, the findings support the temporal expectation hypothesis (White et al 2010b):
the action of NG’s administering hand provided a temporal cue that focused attention as
touch was administered.
References
Halnan C R E, Wright G H, 1960 “Tactile localization” Brain 83 677–700
Michael G A, 2007 “A significance test of interaction in 2 × K designs with proportions” Tutorials in
Quantitative Methods for Psychology 3 1–7
Ramachandran V S, Hirstein W, 1997 “The three laws of qualia: What neurology tells us about the
biological functions of consciousness, qualia and the self ” Journal of Consciousness Studies 4
429–458
Valentini M, Kischka U, Halligan P W, 2008 “Residual haptic sensation following stroke using
ipsilateral stimulation” Journal of Neurology, Neurosurgery, and Psychiatry 79 266–270
van Stralen H E, van Zandvoort M J, Dijkerman H C, 2011 “The role of self-touch in somatosensory
and body representation disorders after stroke” Philosophical Transactions of the Royal Society
of London B 366 3142–3152
Weiskrantz L, Zhang D, 1987 “Residual tactile sensitivity with self-directed stimulation in hemi­
anaesthesia” Journal of Neurology, Neurosurgery, and Psychiatry 50 632–634
White R C, Aimola Davies A M, Kischka U, 2010a “Errors of somatosensory localisation in a patient
with right-hemisphere stroke” Neurocase 16 238–258
White R C, Aimola Davies A M, Kischka U, Davies M, 2010b “Touch and feel? Using the rubber
hand paradigm to investigate self-touch enhancement in right-hemisphere stroke patients”
Neuropsychologia 48 26–37