214
U. Wijk et al.
longer learning period and also 2 occasions per day
during the first week, which indicates that the subjects
learn even faster and within a day. The learning could
even have a “3-occasion-effect” with at least 3 h in bet-
ween occasions. In our study we used mechanotactile
stimuli, which has been proven to be easy to interpret
for sensory feedback (24). Mechanotactile is also a
more common modality to receive as sensory feedback
in daily use compared with vibrotactile or electrotactile
stimuli, which was used by Chai et al. (27). The very
good agreement between stimuli and responses in our
study indicates that it is possible to learn predefined
areas on the forearm skin that is comparable to referred
sensation in capacity to localize the predefined areas.
Chai et al. (27) show similar results, and non-soma-
totopically matched areas reached comparable levels
to the somatotopically matched areas considering re-
sponse time from the actual stimulation to the response
of the perceived stimulation, during those 3 training
days (27). Our result opens up for the possibility for
amputees without referred sensations, as well as for
congenital amputees, to learn the association and keep
it prolonged for at least 2 weeks. Compared with our
experimental learning set-up, prosthesis users would
probably wear a prosthesis with sensory feedback more
frequently and therefore get more confident with the
sensory associations.
Learning as a concept is defined as an encoding
of memory and is the process of “gradual changes in
behavior as a function of training” (31). In the dual
code theory there are separate “channels” to process
information from different senses. Therefore, multiple
senses should be used to facilitate learning, without
exposing the working memory to fatigue (32). Three
learning styles for adults are described; visual, auditory
and kinesthetic, and the best learning is achieved when
these 3 approaches are combined (33). In our study we
apply visual and kinaesthetic (sensory) information at
the same time, and in accordance with the dual code
theory and the 3 learning styles this should ease the
learning. A well-known concept in psychology and
cognitive literature is the spacing effect (34). The
spacing effect implies that practice is spread over a
period of time and the opposite is when practice is
massed at one or few close occasions. When the same
amount of time is spent practicing, learning is most
effective when spaced over time (34, 35). The memory
tends to last longer, since spaced learning keeps new
cells maintained (36). It has also been shown that the
best learning occurs when the practice intervals were
expanding over time (37). In the current study the
spacing effect was applied and the occasions were
spaced over a period of 5 weeks. In the first week the
training occasions were made twice a day, the second
www.medicaljournals.se/jrm
week the training was made once a day and there was
an interval of 1 week made respectively for the last 2
occasions. Another concept used in research for lear-
ning and memory is the testing effect. The effect in
long-term memory is better when memory tests are
made during the period of practice (38). In the present
study every learning occasion included both a pure
learning session and testing session where the subject
received feedback on the responses. This may have
been advantageous for learning.
No difference was seen in learning over time bet-
ween the sexes. However, the group of men was small,
only 6 men participated compared with 25 women,
and the lack of statistical significant difference may
be due to lack of statistical power. The results did not
show any differences between the different age groups.
The U-shape of the tactile display imitates the order
and positions of the fingers and may ease the intuitive
interpretation of the stimuli of the predefined area with
the specific finger. The middle finger was easiest to
discriminate, whereas the little finger stimulation was
most frequently mistaken, and instead associated with
being the ring finger. A possible explanation for this
is the U-shape. The stimulation for the middle finger
was applied over the flexor tendons to the fingers and
the median nerve, and some of the subjects reported a
different sensation (tingling), or a stronger sensation
of the stimulations of the area for the middle finger
compared with stimulations of the other finger areas.
The middle finger stimulation was applied in the centre
and the most distally on the forearm and might have
become a reference for the other stimulated areas which
were either on the one or the other side of the middle
finger. There was barely any misperception between
the stimulations on different side of the middle finger
(digit 1↔digit 4), (d1↔d5), (d2↔d4) and (d2↔d5),
but it was more difficult to discriminate adjacent
fingers (d1↔d2) and (d3↔d4). Nerve innervation is a
possible explanation, the 3 radial sites (d1, d2 and d3)
were applied to skin that is innervated by the median
nerve, and the 2 ulnar sites (d4 and d5) were applied
on skin innervated by the ulnar nerve.
Study limitations
Stimulation on the forearm comprised pressure from
servo motors, and it is impossible to avoid mechanical
noise. Since the speed of rotation of the servo motor
was set to be the same, when applying pressure on the
pre-defined area on the forearm, the 5 servo motors
should sound similar. However, some subjects noticed
that some servo motors could slightly differentiate in
sound, which may have affected the performance in the
progression of learning. According to dual code theory,
the involvement of more senses can facilitate learning.