Motor and Neural Adaptation during an Eight-Week Writing
Training with the Non-Preferred Hand
Manuel Bange
1
, Gaëtan André
2
and Diana Henz
1
1
Johannes Gutenberg University, Institute for Training and Movement Science, Mainz, Germany
2
Toulouse, France
1 OBJECTIVES
Drawing on the work of Haken et al. (1985) and
Hollerbachs oscillatory handwriting model (1981),
Athenes et al. (2004) showed that graphomotor skills
are governed by nonlinear dynamic coupling of two
(nearly) orthogonal oscillators, measured as relative
phase (RP). Other studies evaluated the degree of
automaticity by means of kinematic parameters, e.g.
the number of velocity inversions (NIV) per stroke
(Mai and Marquardt, 1998). Electrophysiological and
neuroimaging studies showed that motor training
produced altered activity in task-related brain-areas,
in both early and later stages (Patel et al., 2013; Bar
and DeSouza, 2016). In the present study, we
investigated the development of RP, velocity, and
NIV of the letter “e” during eight weeks of training to
write with the non-preferred hand, as well as changes
in spectral maps of the cortex in three of the nine
sessions. By applying an exploratory longitudinal
single case study design to the task, we hope to gather
new insights about individual motor and neural
adaptation. We hypothesized that writing velocity and
automaticity increase with training, while RP
becomes more stable. Furthermore, subjects should
show enhanced neural activity in task related regions.
Areas disengaging in later stages could play a role
during early learning.
2 METHODS
Five adult, right-handed participants performed eight
weeks (3x30 minutes/week) of unsupervised
differential training (Schöllhorn et al., 2015) to
improve their left-handed writing. Motor adaptation
was tested before the intervention and after every
week (9 sessions). Subjects received eight sets of four
letters on a screen placed in front of them. Every set
was presented for 15 seconds and consisted of either
one “e” or “m” and three random letters, which had
to be written on a sheet of paper attached to a graphics
tablet (Wacom Intous 3, 542*318mm, 2540 dpi,
200Hz). Kinematic data was recorded and kernel
filtered with the software CS (Marquardt and Mai,
1994). All “e” were further analysed.
Average stroke velocity was calculated by CS.
Corresponding NIV was calculated as an average of
acceleration zero crossings of all vertical strokes of
the letter. Mean values of velocity and NIV of every
session were compared qualitatively. Continuous RP
for was calculated with the Hilbert transform (Danna
et al., 2012) and combined for early, mid, and late
stages (three sessions each). As KS and Shapiro Wilk
tests (Razali and Wah, 2011) showed that RP-data
was not normally distributed, we compared the
standard deviation (SD) of the stages.
For EEG acquisition, 19 electrodes were placed
according to the international 10-20 system (Jasper,
1958), which recorded cortical activity at 1024 Hz
(Brain Quick, Micromed; SystemPlus Evolution) for
three conditions (rest 1, task, rest 2) during session 1,
5 and 9. Data was bandpass (0.8 Hz, 99 Hz) and notch
filtered (50 Hz, 43 Hz) with the Matlab EEGLAB
toolbox (Delorme and Makeig, 2004). Artifacts were
removed by visual inspection and independent
component analysis. We analysed theta (4-8 Hz) and
gamma (30-99 Hz) bands by spectral mapping for
every participant.
3 RESULTS
Individual mean velocity is plotted in fig. 1a for all
sessions. Courses proceed differently and show fluc-
tuations for every participant. One commonality was
a decline of writing velocity after the first week for
four subjects, which was followed by an increasing
trend by three subjects.
The NIV-courses show fluctuations as well (fig.
1b). Three subjects had increased values even after
eight weeks of training.
Fig. 1c presents the standard deviation of RP for
every participant in the early, mid and late stage of
Bange, M., André, G. and Henz, D.
Motor and Neural Adaptation during an Eight-Week Writing Training with the Non-Preferred Hand.
In Extended Abstracts (icSPORTS 2016), pages 11-13
Copyright
c
2016 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
11
Figure 1: (a) Average writing velocity and (b) averaged NIV-values for the letter “e” for every subject on every session. (c)
SD of relative phase for all five subjects during early, mid, and late stages of the intervention. (d) Spectral mapping of theta
and gamma activity of subject TK in sessions 1,5 and 9. Blue = low, green = medium, red = high spectral power.
the intervention. It shows that the variation decreased.
This was observed in the histograms (not presented
here) as well, where distributions became narrower
between 60° and 90°.
Cortical spectral activity of theta and gamma
bands is presented for TK in fig. 1d. Three Subjects
(SM, MD, TK) showed enhanced theta power under
the task condition on the whole cortex. TK showed
lower increases during later stages, especially in
frontal areas. Frontal theta power increase was less
for three other subjects in session 9. SB demonstrated
increased theta power over the frontal lobes in
sessions 1 and 5, as well as in parietal and occipital
lobes in sessions 5 and 9. TJ showed increased power
mainly on the right hemisphere during sessions 1 and
5, and reduced power in both lateral motor cortices in
session 9.
Under rest 2 condition, four participants (MD, SB,
SM, TJ) exhibited reduced theta activity over the
frontal cortex during session 1, but not during other
sessions. Slight increases in parietal and motor areas
were found for four subjects (TK, MD, SM, SB) in
session 5. TK matched baseline activity in rest 2
during session 9, while two participants showed
regional decreases (SM frontal, TJ parietal and
central), and two other participants increases (SB
frontal, MD parietal and occipital).
With some exceptions gamma power of the whole
cortex was enhanced in all subjects in the task
condition. Exceptions were SB in session 1 (only
parietal and occipital increase), SM in session 1 (only
frontal increase) and TJ in sessions 5 and 9 (only
parietal and occipital increase). Rest 2 gamma power
remained above baseline for TK and SB in all three
sessions and MD in session 9. It dropped below or at
baseline for SM and TJ (and MD in session 1).
4 DISCUSSION
The time courses of NIV and velocity were unlike
usual learning curves, which are often smoothed by
averaging over many trials and subjects (Ritter and
Schooler, 2002). This was unexpected, as we thought
that a training-related increase of speed and decrease
of NIV would be clearer. While TK and SM showed
the tendency to increase writing speed during the
intervention, we cannot distinguish between natural
fluctuations and progress for MD, SB and TJ. The
common decline from session 1 to 2 might reflect a
shift from speed to shape constraints. Lower writing
speed facilitates visual feedback for movement
correction, which is utilized during early grapho-
motor learning (Danna and Velay, 2015).
The number of velocity inversions can be seen as
a marker for corrective movements (Mai and
Marquardt, 1998). Higher Values indicate using more
corrective movements which is associated with
icSPORTS 2016 - 4th International Congress on Sport Sciences Research and Technology Support
12
increased feedback control. SM and MD increased
their NIV and therefore seemed to focus more on
shape constraints during the intervention. TK and SB
decreased their NIV and possibly relied less on feed-
back during later stages. The Variance in partly
contradicting NIV and velocity values can be
explained as subjects are seeking the optimal solution
for a speed-accuracy trade-off during learning.
A clear progress was visible for RP, however.
Lower SD indicates the formation of increasingly
stable attractors, which supports the idea that writing
can be modelled by updating RP, amplitude and
frequency on a piecewise manner to spare neural
resources (Andre et al., 2014). Semi-permanent stable
RP would support automation as well. However, well
learned letters exhibit more than one stable phase
angle and are therefore not normally distributed. The
transition between those angles is of special interest
and can be further investigated by means of pattern or
time series analyses.
We planned to reveal cortical adaptations by
examining spectral maps for all participants. Perfetti
et al. (2011) demonstrated that enhanced gamma
activity in right parietal regions is associated with
initial learning. We found that all subjects had
increased power in this area, as well as in other areas
in the task condition. Gamma activity remained
enhanced shortly after the task, indicating that
memory formation processes were still active.
Increased activity over the whole cortex indicates
engagement of a wide-spread network during grapho-
motor learning. High-density EEG could help
localizing involved areas with a better resolution.
Wong et al. (2014) revealed that theta and gamma
activities in the frontal cortex are having a negative
relationship with task familiarity. In this line, we
revealed lower frontal theta activities in session 9 for
four subjects, while theta power was increased in
other areas (especially occipital) during the task. This
finding supports the idea that subjects need less
attention with higher task familiarity.
Our aim was to gather insight about motor and
neural parameters during eight weeks of learning. We
found inter-individual differences for both, which
could reflect using different strategies or learning
with different speeds. Additionally, we discovered
common features for RP as well as gamma and theta
activities. Future studies could correlate behavioural
with high-density EEG data (e.g. NIV with frontal
activity) to reveal coherent adaptations or possible
strategy-related differences in the neural network.
REFERENCES
André, G., Kostrubiec, V., Buisson, J.-C., Albaret, J.-M., &
Zanone, P.-G. (2014). A parsimonious oscillatory
model of handwriting. Biological cybernetics, 108(3),
321336.
Athenes, S., Sallagoïty, I., Zanone, P.-G., & Albaret, J.-M.
(2004). Evaluating the coordination dynamics of
handwriting. Human movement science, 23(5), 621
641.
Bar, R. J., & DeSouza, J. F. X. (2016). Tracking plasticity:
effects of long-term rehearsal in expert dancers
encoding music to movement. PloS one, 11(1),
e0147731.
Danna, J., & Velay, J.-L. (2015). Basic and supplementary
sensory feedback in handwriting. Frontiers in
psychology, 6.
Delorme, A., & Makeig, S. (2004). EEGLAB: an open
source toolbox for analysis of single-trial EEG
dynamics including independent component analysis.
Journal of neuroscience methods, 134(1), 921.
Haken, H., Kelso, J. S., & Bunz, H. (1985). A theoretical
model of phase transitions in human hand movements.
Biological cybernetics, 51(5), 347356.
Hollerbach, J. M. (1981). An oscillation theory of
handwriting. Biological cybernetics, 39(2), 139156.
Jasper, H. (1958). Report of the committee on methods of
clinical examination in electroencephalography.
Electroencephalogr Clin Neurophysiol, 10, 370375.
Mai, N., & Marquardt, C. (1998). Registrierung und
Analyse von Schreibbewegungen: Fragen an den
Schreibunterricht: na.
Marquardt, C., & Mai, N. (1994). A computational
procedure for movement analysis in handwriting.
Journal of neuroscience methods, 52(1), 3945.
Patel, R., Spreng, R. N., & Turner, G. R. (2013). Functional
brain changes following cognitive and motor skills
training a quantitative meta-analysis. Neurorehabilita-
tion and neural repair, 27(3), 187199.
Perfetti, B., Moisello, C., Landsness, E. C., Kvint, S.,
Lanzafame, S., Onofrj, M., et al. (2011). Modulation of
gamma and theta spectral amplitude and phase
synchronization is associated with the development of
visuo-motor learning. The Journal of Neuroscience,
31(41), 1481014819.
Razali, N. M., & Wah, Y. B. (2011). Power comparisons of
shapiro-wilk, kolmogorov-smirnov, lilliefors and
anderson-darling tests. Journal of statistical modeling
and analytics, 2(1), 2133.
Ritter, F. E., & Schooler, L. J. (2001). The learning curve.
International encyclopedia of the social and behavioral
sciences, 13, 86028605.
Schöllhorn, W. I., Eekhoff, A., & Hegen, P. (2015).
Systemdynamik und differenzielles Lernen.
Sportwissenschaft, 45(3), 127137.
Wong, S. W. H., Chan, R. H. M., & Mak, J. N. (2014).
Spectral modulation of frontal EEG during motor skill
acquisition: a mobile EEG study. International Journal
of Psychophysiology, 91(1), 1621.
Motor and Neural Adaptation during an Eight-Week Writing Training with the Non-Preferred Hand
13