AI-Powered Handwriting Analysis Shows Promise in Early Detection of Dyslexia and Dysgraphia Among Young Children

A new study led by researchers at the University at Buffalo explores the potential of AI-powered handwriting analysis as an early detection tool for dyslexia and dysgraphia among young children. The research, published in the journal SN Computer Science, seeks to enhance existing screening methods, which, while effective, can be expensive, time-consuming, and often focus on one condition at a time. Early identification of these neurodevelopmental disorders is crucial to ensuring that children receive the necessary support, preventing adverse effects on their learning and socio-emotional development. The project is part of the National AI Institute for Exceptional Education, a UB-led research initiative aimed at developing AI systems to identify and assist young children with speech and language processing disorders. Key participants include Venu Govindaraju, Ph.D., a SUNY Distinguished Professor in the Department of Computer Science and Engineering, and a team of UB computer scientists. Background and Methodology Several decades ago, Govindaraju and his colleagues pioneered the use of machine learning, natural language processing, and other AI techniques to analyze handwriting. The U.S. Postal Service and other organizations still benefit from these advancements, which automate mail sorting. Now, the team is applying similar concepts to create a framework for identifying spelling issues, poor letter formation, and organizational writing problems—common indicators of dyslexia and dysgraphia. Dysgraphia, characterized by difficulties in physical writing, has been easier to detect using AI due to its observable traits. Dyslexia, primarily a reading and speech disorder, is more challenging to identify through handwriting but can show signs in spelling and other writing behaviors. The shortage of handwriting samples from children poses another significant hurdle. Collaboration and Data Collection To overcome these challenges, the UB researchers collaborated with teachers, speech-language pathologists, and occupational therapists to ensure their AI models are practical in real-world settings. They also partnered with Abbie Olszewski, Ph.D., an associate professor in literacy studies at the University of Nevada, Reno. Olszewski co-developed the Dysgraphia and Dyslexia Behavioral Indicator Checklist (DDBIC), which identifies 17 behavioral cues related to these disorders. The team collected writing samples from kindergarten to 5th-grade students at an elementary school in Reno. This data collection was approved by an ethics board and kept anonymous to protect student privacy. The samples will be used to validate the DDBIC tool, train AI models to perform the screening, and assess the models' effectiveness compared to human administrators. AI Models and Their Applications The study describes how the AI models can be used to: Identify Spelling Issues: Detect patterns and errors in spelling that might indicate dyslexia. Analyze Letter Formation: Assess the quality of letter formation, which can reveal dysgraphia. Evaluate Writing Organization: Examine how well students organize their writing, another critical indicator. Summarize Findings: Combine the outputs from different models to provide a comprehensive assessment. This integrated approach aims to streamline the screening process, making it more efficient and accessible, particularly in underserved areas where professional resources are limited. By automating the initial stages of diagnosis, AI could potentially alleviate the strain on speech-language pathologists and occupational therapists, enhancing the overall educational support system. Potential Impact and Industry Insights Industry insiders have hailed this research as a significant step toward leveraging technology for inclusive education. The development of AI-enhanced tools for early dyslexia and dysgraphia screening demonstrates a commitment to using AI for public good, providing valuable resources to those who need them most. The team's emphasis on user-centric design and ethical considerations ensures that these tools are not only technologically advanced but also responsible and practical. Venu Govindaraju, the study's corresponding author, emphasizes the importance of early intervention. "Our ultimate goal is to make these tools widely available and ensure that they can be effectively integrated into educational settings, supporting teachers and therapists in their efforts to help children," he says. Sahana Rangasrinivasan, a co-author and Ph.D. student at UB, adds, "It is crucial to consider the perspectives of educators and healthcare professionals to ensure our AI models are both accurate and user-friendly in real-world scenarios." The National AI Institute for Exceptional Education, which spearheads this research, is dedicated to advancing the field of AI in special education. Sumi Suresh, Ph.D., a visiting scholar at UB, highlights the institute's mission: "This work showcases the potential of AI to provide targeted and timely assistance, ultimately improving the lives of children with neurodevelopmental disorders." Bharat Jayarman, Ph.D., director of the Amrita Institute of Advanced Research and a professor emeritus at UB, and Srirangaraj Setlur, a principal research scientist at the UB Center for Unified Biometrics and Sensors, also contributed to the study, bringing expertise in biometrics and sensor technology to refine the AI models. Overall, this study represents a promising intersection of technology and education, offering hope for more accessible and effective early intervention strategies. The development of these AI tools could revolutionize the way we screen and support children with dyslexia and dysgraphia, ensuring they get the help they need to thrive academically and socially.