Motor and Neural Adaptation during an Eight-Week Writing

Training with the Non-Preferred Hand

Manuel Bange

, Gaëtan André

and Diana Henz

Johannes Gutenberg University, Institute for Training and Movement Science, Mainz, Germany

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

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

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),

321–336.

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), 9–21.

Haken, H., Kelso, J. S., & Bunz, H. (1985). A theoretical

model of phase transitions in human hand movements.

Biological cybernetics, 51(5), 347–356.

Hollerbach, J. M. (1981). An oscillation theory of

handwriting. Biological cybernetics, 39(2), 139–156.

Jasper, H. (1958). Report of the committee on methods of

clinical examination in electroencephalography.

Electroencephalogr Clin Neurophysiol, 10, 370–375.

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), 39–45.

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), 187–199.

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), 14810–14819.

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), 21–33.

Ritter, F. E., & Schooler, L. J. (2001). The learning curve.

International encyclopedia of the social and behavioral

sciences, 13, 8602–8605.

Schöllhorn, W. I., Eekhoff, A., & Hegen, P. (2015).

Systemdynamik und differenzielles Lernen.

Sportwissenschaft, 45(3), 127–137.

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), 16–21.

Motor and Neural Adaptation during an Eight-Week Writing Training with the Non-Preferred Hand