The Department of Electronics, Information and Bioengineering at Politecnico di Milano is at the forefront of a study investigating the use of a sensorized pen for the early screening of writing difficulties in children. The research was conducted as part of the PRIN project “e-School 2.0”, coordinated by Prof. Simona Ferrante in collaboration with the University of Insubria, and involved more than 700 primary and lower secondary school students.
At the heart of the study is the THInkPen (Tele-Health Ink Pen), a patented sensorized ink pen that is used on paper just like a regular pen while collecting detailed information about the writing process.
During the study, children used the THInkPen to complete two tasks from the BVSCO-3 – Battery for the Clinical Assessment of Writing and Orthographic Competence, one of the main tools used in Italy to assess writing difficulties, dysgraphia and dysorthography.
Thanks to its integrated sensors, the pen captures digital indicators related to features such as pressure on the paper, fluidity of movement and pen tilt. The collected data were analysed using artificial intelligence algorithms to investigate their relationship with clinical scores and assess their effectiveness in identifying writing difficulties.
The results revealed significant relationships between the digital indicators and clinical scores. AI models were also able to distinguish students with writing difficulties from their peers, while explainable AI techniques helped identify the factors most strongly associated with below-average performance. The indicators also proved capable of capturing the progressive development of graphomotor skills as children grow.
The THInkPen stems from a project patented by Politecnico di Milano and the University of Milan, in collaboration with Prof. Alberto Borghese. Initially developed for the screening of neurodegenerative diseases, the technology was subsequently adapted to assess neurodevelopment, an evolution to which Linda Greta Dui, a researcher at the Department of Electronics, Information and Bioengineering, also contributed.
The study thus highlights the potential of combining sensor technology and artificial intelligence with traditional clinical tools to support earlier and more in-depth identification of writing difficulties.