Novel Methods for Handwriting-based Automated Grading of Degree of Handedness

Hand preference and degree of handedness are two different aspects of human behavior which are often confused to be one. While hand preference is affected by various natural, environmental, and socio-cultural causes, degree of handedness reflects the inherent capability of a human gained by nature which can be modified with training. In this paper, we present two novel methods to quantify the degree of handedness for the first time using handwriting features of dominant and non-dominant hand on three categories of subjects- “Unidextrous”, “Partially-Unidextrous”, and “Ambidextrous”. Methods: Time, static, and dynamic variables of handwriting signal were used as the features for quantification of degree of handedness. Davies Bouldin Score as a statistical method and Neural Network accuracy-based 4-point scale as an automated method for this quantification were presented and compared with the well-known degree of handedness assessment questionnaires from Edinburgh Inventory (EI). Results: Davies Bouldin Score and 4-point scale from Neural Network were found to be in accordance with the EI questionnaires. 4-point scale was preferred over Davies Bouldin score as a robust grading mechanism. Conclusion: The presented methods can be used as an assessment tool for quantifying degree of handedness with multiple applications in i) determining the feasibility of switching hand preference under societal norms ii) neuro-rehabilitation strategies iii) neurological disorder diagnosis iv) study of brain lateralization and activation levels v) forensics vi) sports applications vii) human behavioral assessments.