Adaptable Categorization of Hands and Tools in Prosthesis Users

Full text

Psychological Science
1 –4
© The Author(s) 2017
Reprints and permissions:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/0956797616685869
journals.sagepub.com/home/pss
Short Report
Some theories propose that tools become incorporated
into the neural representation of the hands (a process
known as tool embodiment; Maravita & Iriki, 2004).
Others suggest that conceptual body representation is
rigid and that experience with one’s own body is insuf-
ficient for adapting bodily cognition, as shown in indi-
viduals born without hands (Vannuscorps & Caramazza,
2016) and in amputees with persistent phantom hand
representation (Kikkert et al., 2016). How sharp is the
conceptual boundary between hands and tools? This
question is particularly relevant for individuals who have
lost one hand and use prosthetic hands as tools to sup-
plement their missing hand function. Although both con-
genital one-handers (i.e., amelia patients) and one-handed
amputees are encouraged to use prostheses, the former
show a greater tendency than the latter to use prosthetic
hands in daily tasks (Jang et al., 2011). One-handers have
a fully functional remaining hand (allowing them to use
handheld tools, etc.), which makes them less likely to
show semantic distortions in hand and tool representa-
tion. However, their bodies and their interactions with
their environment are fundamentally altered by their dis-
ability (Makin et al., 2013; Makin, Wilf, Schwartz, &
Zohary, 2010).
To determine how real-world experience shapes con-
ceptual categorization of hands, tools, and prostheses,
we recruited one-handers with congenital or acquired
unilateral hand loss to take part in a study involving a
priming task. We predicted that one-handers, particularly
congenital one-handers, would show more conceptual
blurring between hands and tools than control partici-
pants would, as a result of less experience with a hand
and more reliance on prostheses (which are essentially
tools) for typical hand functions. We further predicted
that individual differences in prosthesis usage would be
reflected in implicit categorization of hands, manual
tools, and prostheses.
Method
Twenty-four one-handers (12 born without a hand and
12 who lost a hand through amputation) and 21 matched
control participants performed a visual priming task in
which they verbally categorized target images of hands
and tools (Fig. 1a). On each trial, participants saw a prime
stimulus followed by a target stimulus, each of which
appeared for 32 ms (stimulus onset asynchrony = 600
ms). On baseline trials, the prime was always a scram-
bled image, and on experimental trials, the prime could
be an image of a hand, a tool, or a prosthesis. On all tri-
als, the target was either a hand or a tool. Participants
were instructed to ignore the prime and to verbally report
whether the target image was a hand or a tool. Time from
the start of the target display to voice onset was recorded
as the participants’ reaction time (RT). Participants com-
pleted 40 baselines trials and four blocks (60 trials each)
of experimental trials.
Ten different exemplars were used as prime and target
items in each category (for example stimuli, see Fig. S1 in
the Supplemental Material available online). Hand and
prosthesis images showed the side of the participants’
missing hand (one-handers) or nondominant hand (con-
trols). When possible, images were of the participant’s
own prosthesis, so daily prosthesis usage (assessed both
by asking how often participants wore their prosthesis
and by using an adapted version of the Motor Activity
Log; Makin et al., 2013) generally reflected individuals’
experience with the prime prosthesis image presented to
them (see the Supplemental Material for additional meth-
odological details).
685869PSSXXX10.1177/0956797616685869van den Heiligenberg et al.Flexible Categorization
research-article2016
Corresponding Author:
Tamar R. Makin, University College London, Institute of Cognitive
Neuroscience, 17 Queen Square, London, WC1N 3AZ, United
Kingdom
E-mail: tmakin@ucl.ac.uk
Adaptable Categorization of Hands
and Tools in Prosthesis Users
Fiona M. Z. van den Heiligenberg1, Nick Yeung2, Peter Brugger3,
Jody C. Culham4, and Tamar R. Makin1,5
1FMRIB Centre, Nuffield Department of Clinical Neurosciences, University of Oxford; 2Department of
Experimental Psychology, University of Oxford; 3Neuropsychology Unit, Department of Neurology,
University Hospital Zurich, Switzerland; 4Brain and Mind Institute, Department of Psychology,
University of Western Ontario; and 5Institute of Cognitive Neuroscience, University College London
Received 1/18/16; Revision accepted 12/3/16

2 van den Heiligenberg et al.
Results
We first examined the prime-target congruency effect on
trials with hand and tool primes in one-handers and con-
trol participants. RTs for controls were slower when the
prime and target were congruent (i.e., from the same
category; e.g., hand prime and hand target) than when
they were incongruent (i.e., from different categories;
e.g., hand prime and tool target), which indicates same-
category prime interference in target processing (Boy &
Sumner, 2010; see also Vainio, 2011, for negative stimu-
lus-response compatibility for hand images). Next, a 2
(prime) × 2 (target) × 2 (group) repeated measures analy-
sis of variance revealed a significant three-way interac-
tion, F(1, 43) = 5.37, p = .025; this indicates that, unlike
the results for control participants, the priming effect was
absent in one-handers (for further analysis, see Results
and Fig. S2 in the Supplemental Material). In control par-
ticipants, we found a significant Prime × Target interaction,
F(1, 20) = 11.24, p = .003. Further, planned comparisons
using paired-samples t tests revealed a significant RT
difference between congruent and incongruent trials for
bothhand targets, t(20) = 2.19, p = .041, d = 0.175, and
tool targets, t(20) = 3.31, p = .003, d = 0.261. However, in
one-handers, there was no significant Prime × Target
interaction (F < 1), which suggests that the conceptual
hand-tool category boundary is blurred in one-handers,
compared with control participants.
We calculated each participant’s congruency effect by
subtracting mean RTs on incongruent trials from mean
RTs on congruent trials. Given our finding that control
participants’ RTs were slower on congruent than on
incongruent trials, a greater congruency effect reflects
greater dissociation between hands and tools. Although
one-handers did not show a significant congruency
effect, there was evidence that their categorization behav-
ior was modulated by their case histories, specifically the
age at which one-handed amputees lost their hand and
their habitual prosthesis usage. We found a significant
correlation between age at hand loss and congruency
effect, r(10) = .65, p = .022 (Fig. 1b): Hand loss earlier in
life related to weaker congruency effects, whereas ampu-
tees who lost a hand later in life (and therefore had more
experience with the now-missing hand) showed greater
congruency effects. These findings suggest that the con-
ceptual distinction between hands and tools develops
through experience with natural hands. We also found
that one-handers who used their prosthesis more tended
to show weaker congruency effects than those who
used prostheses less frequently—correlation between
Baseline Trials
Experimental Trials
Scrambled
Prime
Prosthesis
Prime
Tool
Prime
Hand
Prime
Hand Target Tool Target
600-ms SOA
32 ms 32 ms
a
Age at Hand Loss (years)
Hand-Tool Congruency Effect (ms)
4540 N/A
150
100
50
–50
–100
–150
0
35302520150
b
One-Handed Amputees
Controls
Congenital One-Handers
Fig. 1. Experimental procedure and results. On each trial (a), participants saw a 32-ms prime stimulus followed by a 32-ms target stimulus, with a
600-ms stimulus onset asynchrony (SOA). Participants were asked to make a speeded, forced-choice verbal response as to whether the target was a
hand or a tool. Baseline trials differed from experimental trials only in that the prime was a scrambled image rather than an image of a hand, tool,
or prosthesis. Reaction times on incongruent trials (in which prime and target stimuli were from different categories) were subtracted from reaction
times on congruent trials (in which prime and target stimuli were from the same category) to calculate the hand-tool congruency effect. The scatter
plot (b; with best-fitting regression line) shows the mean hand-tool congruency effect as a function of age at hand loss for congenital one-handers
and one-handed amputees, along with the mean congruency effect for control participants.

Flexible Categorization 3
congruency effect and prosthesis usage: r(22) = −.38, p =
.068—such that the hand-tool category boundary
(reflected in the congruency effect) tended to blur with
the regularity of prosthesis usage.
Theories of tool embodiment state that prosthesis
usage should result in categorization of the prosthesis as
a hand (Murray, 2008). In the final set of analyses, we
assessed the degree to which prosthesis primes affected
responses to hands and tools as a function of prosthesis
experience. Given that categorical similarity resulted in
slower responses for congruent prime-target pairs than
for incongruent prime-target pairs, slowing of RTs for
prosthesis primes can be taken to reflect the conceptual
similarity between prostheses and hands or tools. To
investigate this, we ran a backwards regression analysis
on RTs for prosthesis-hand trials using age at hand loss,
years since hand loss, prosthesis usage, congruency
effect size, and mean RT on baseline trials as predictors.
The final model for hand-target trials, F(2, 21) = 35.08,
p< .001, R = .88, adjusted R2 = .75, included age at hand
loss, β = −0.29, t(23) = −2.72, p = .013, and baseline RT,
β = 0.78, t(23) = 7.28, p < .001 (see Fig. S3 in the Supple-
mental Material). This analysis revealed that people who
lost their hand earlier in life showed greater conceptual
similarity between the prosthesis and hands. We also ran
a backwards regression analysis on RTs for prosthesis-tool
trials using the same parameters as for the previous set of
backwards regressions. The final model, F(2, 21) = 42.48,
p < .001, R = .90, adjusted R2 = .78, included prosthesis
usage, β = 0.24, t(23) = 2.39, p = .026, and baseline RT, β =
0.83, t(23) = 8.36, p < .001 (Fig. S4 in the Supplemental
Material). This analysis showed that, in opposition to the
previous regression, the conceptual relationship between
prostheses and tools was best predicted by prosthesis usage,
with participants who used their prostheses more showing
greater conceptual similarity between prostheses and tools.
Conclusion
Together, our findings demonstrate that categorization of
hands and tools in one-handers depends on both prior
experience with a natural hand before amputation and
later artificial-hand usage. Specifically, dissociation bet-
ween hands and tools (exemplified by the congruency
effect) depends on the degree of experience with that
hand. Moreover, the representation of prostheses as hands
and tools depends on daily life experience. Given the
relatively limited semantic-category deficit but profoundly
changed body experience resulting from hand loss, we
suggest that the adaptable conceptual relationship
between hands, tools, and prostheses is embodied. Never-
theless, because high-level lexico-semantic processing may
implicitly depend on body representation (Rueschemeyer,
Pfeiffer, & Bekkering, 2010), further studies are necessary
to elucidate the underlying process.
Action Editor
Jamin Halberstadt served as action editor for this article.
Author Contributions
All authors contributed to the development of the study con-
cept. F. M. Z. van den Heiligenberg, N. Yeung, and T. R. Makin
designed the study. Testing and data collection were performed
by F. M. Z. van den Heiligenberg. F. M. Z. van den Heiligenberg
and T. R. Makin analyzed and interpreted the data. All authors
contributed to drafting the manuscript and provided critical
revisions. All authors approved the final version of the manu-
script for submission.
Acknowledgments
We thank our participants and Opcare for their help and Scott
Macdonald for assistance with data collection.
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interest with
respect to their authorship or the publication of this article.
Funding
This work was supported by the Cogito Foundation and by the
Wellcome Trust and the Royal Society (Grant No. 104128/Z/14/Z).
Supplemental Material
Additional supporting information can be found at http://journals
.sagepub.com/doi/suppl/10.1177/0956797616685869
Open Practices
All data and materials have been made publicly available via
the Open Science Framework and can be accessed at https://
osf.io/ck22q/. The complete Open Practices Disclosure for this
article can be found at http://journals.sagepub.com/doi/
suppl/10.1177/0956797616685869. This article has received
badges for Open Data and Open Materials. More information
about the Open Practices badges can be found at http://www
.psychologicalscience.org/publications/badges.
References
Boy, F., & Sumner, P. (2010). Tight coupling between posi-
tive and reversed priming in the masked prime paradigm.
Journal of Experimental Psychology: Human Perception
and Performance, 36, 892–905. doi:10.1037/a0017173
Jang, C. H., Yang, H. S., Yang, H. E., Lee, S. Y., Kwon, J. W.,
Yun, B. D., . . . Jeong, H. W. (2011). A survey on activi-
ties of daily living and occupations of upper extremity
amputees. Annals of Rehabilitation Medicine, 35, 907–921.
doi:10.5535/arm.2011.35.6.907

4 van den Heiligenberg et al.
Kikkert, S., Kolasinski, J., Jbabdi, S., Tracey, I., Beckmann,
C. F., Johansen-Berg, H., & Makin, T. R. (2016). Revealing
the neural fingerprints of a missing hand. eLife, 5, Article
e15292. doi:10.7554/eLife.15292
Makin, T. R., Cramer, A. O., Scholz, J., Hahamy, A., Henderson
Slater, D., Tracey, I., & Johansen-Berg, H. (2013).
Deprivation-related and use-dependent plasticity go hand
in hand. eLife, 2, Article e01273. doi:10.7554/eLife.01273
Makin, T. R., Wilf, M., Schwartz, I., & Zohary, E. (2010). Amputees
“neglect” the space near their missing hand. Psychological
Science, 21, 55–57. doi:10.1177/0956797609354739
Maravita, A., & Iriki, A. (2004). Tools for the body (schema).
Trends in Cognitive Sciences, 8, 79–86. doi:10.1016/j
.tics.2003.12.008
Murray, C. D. (2008). Embodiment and prosthetics. In P. Gallagher,
D. Desmond, & M. MacLachlan (Eds.), Psychoprosthetics (pp.
119–130). London, England: Springer.
Rueschemeyer, S.-A., Pfeiffer, C., & Bekkering, H. (2010). Body
schematics: On the role of the body schema in embodied
lexical–semantic representations. Neuropsychologia, 48,
774–781.
Vainio, L. (2011). Negative stimulus-response compatibility
observed with a briefly displayed image of a hand. Brain
and Cognition, 77, 382–390. doi:10.1016/j.bandc.2011.09.007
Vannuscorps, G., & Caramazza, A. (2016). Typical action
perception and interpretation without motor simulation.
Proceedings of the National Academy of Sciences, USA, 113,
86–91.