Detecting Autism and Epilepsy through Infant Crying Signals: Key Features for Identification

Abstract

Crying is an important means of communication for babies during infancy, from birth to three months of age”. “At this stage of their development, infants are almost entirely dependent on the people who care for them. As a result, crying plays an important role for the survival, health, and development of the child”. This paper explores the relationship between autism, epilepsy, and crying patterns in newborn infants. The objective is to investigate the potential of analyzing acoustic features of infant cries as a non-invasive method for early detection and diagnosis of these neurodevelopmental disorders. By utilizing machine learning algorithms, a predictive model can be developed to differentiate between typical cries and those indicative of autism or epilepsy. To conduct this research, a specific dataset comprising crying sounds from newborns diagnosed with autism, epilepsy, and typically developing infants is utilized. Acoustic features, including pitch, intensity, and duration, are extracted from the cry recordings. These features are then used to train machine learning models, such as support vector machines (SVMs) or artificial neural networks (ANNs), enabling the classification of cries into distinct categories. The developed model's accuracy and generalizability are assessed using cross-validation techniques and evaluation metrics such as sensitivity, specificity, and the area under the receiver operating characteristic curve (AUC-ROC). Additionally, the model's reliability and applicability in real-world scenarios are validated using an independent dataset.