Research Papers

Long Term Consistency of Handwriting Grip Kinetics in Adults

[+] Author and Article Information
Bassma Ghali

Bloorview Research Institute,
Holland Bloorview Kids Rehabilitation Hospital,
Institute of Biomaterials and
Biomedical Engineering,
University of Toronto,
Toronto, ON M4G 1R8, Canada
e-mail: bassma.ghali@utoronto.ca

Khondaker A. Mamun

Department of Computer Science
and Engineering,
Ahsanullah University of Science and Technology,
Dhaka 1208, Bangladesh
e-mail: k.mamun@ieee.org

Tom Chau

Bloorview Research Institute,
Holland Bloorview Kids Rehabilitation Hospital,
Institute of Biomaterials and
Biomedical Engineering,
University of Toronto,
Toronto, ON M4G 1R8, Canada
e-mail: tom.chau@utoronto.ca

1Corresponding author.

Contributed by the Bioengineering Division of ASME for publication in the JOURNAL OF BIOMECHANICAL ENGINEERING. Manuscript received June 19, 2013; final manuscript received January 15, 2014; accepted manuscript posted February 6, 2014; published online March 24, 2014. Assoc. Editor: Zong-Ming Li.

J Biomech Eng 136(4), 041005 (Mar 24, 2014) (7 pages) Paper No: BIO-13-1269; doi: 10.1115/1.4026641 History: Received June 19, 2013; Revised January 15, 2014; Accepted February 06, 2014

While there is growing interest in clinical applications of handwriting grip kinetics, the consistency of these forces over time is not well-understood at present. In this study, we investigated the short- and long-term intra-participant consistency and inter-participant differences in grip kinetics associated with adult signature writing. Grip data were collected from 20 adult participants using a digitizing tablet and an instrumented pen. The first phase of data collection occurred over 10 separate days within a three week period. To ascertain long-term consistency, a second phase of data collection followed, one day per month over several months. In both phases, data were collected three times a day. After pre-processing and feature extraction, nonparametric statistical tests were used to compare the within-participant grip force variation between the two phases. Participant classification based on grip force features was used to determine the relative magnitude of inter-participant versus intra-participant differences. The misclassification rate for the longitudinal data were used as an indication of long term kinetic consistency. Intra-participant analysis revealed significant changes in grip kinetic features between the two phases for many participants. However, the misclassification rate, on average, remained stable, despite different demarcations of training, and testing data. This finding suggests that while signature writing grip forces may evolve over time, inter-participant kinetic differences consistently exceeds within-participant force changes in the long-term. These results bear implications on the collection, modeling and interpretation of grip kinetics in clinical applications.

Copyright © 2014 by ASME
Your Session has timed out. Please sign back in to continue.



Grahic Jump Location
Fig. 1

Data collection instrumentation set-up. A close-up of the instrumented pen is shown above, highlighting the section of the force sensor array of interest. The full array consisted of 6 columns with 16 sensors in each column; however, only 4 columns of 8 sensors each were needed to cover the pen barrel.

Grahic Jump Location
Fig. 2

Box plots of the three grip force features based on phase 1 and phase 2 data. For each participant, the first box represents the phase 1 feature distribution while the second box is the phase 2 distribution of the same feature. NCC = normalized correlation coefficient; TAF = total average force; TFIQR = total force interquartile range.

Grahic Jump Location
Fig. 3

Heat maps showing percentage of participants exhibiting significant kinetic differences between phases 1 and 2. The left subplot presents the results for the “all signatures” case while the right subplot presents the results based on the first five signatures. Each heat map shows the percentage of participants exhibiting a significant difference for each grip force feature (horizontal axis NCC = normalized correlation coefficient; TAF = total average force; TFIQR = total force interquartile range) using each statistical test (vertical axis RS = Wilcoxon rank-sum test; KS = Kolmogorov-Simrnov test; AB = Ansari-Bradley test).

Grahic Jump Location
Fig. 4

Feature distributions for phase 1 (left) and phase 2 (right) signatures. Data from participant 13 are shown. Circles denote feature vectors from authentic signatures from participant 13. X’s denote feature vectors from other participants. The dark diagonal line is the separating plane determined by linear discriminant analysis. (TFIQR = total force interquartile range; TAF = total average force; NCC = normalized correlation coefficient)

Grahic Jump Location
Fig. 5

Misclassification rates for phases 1 and 2. Asterisks indicate that the participant has a significant difference between phase 1 and phase 2 MCRs. The dashed lines show the average MCRs across all 18 participants.




Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In