Project DetailsDiagnosis of autism at an early age is an extensive area of research, as it has a massive impact on the ability to treat and aid those suffering from the syndrome. So far diagnosis has been based on professional behavioral observation, a flawed tool since it is subjective and imprecise, but also due to the fact that it is only effective at a late developmental stage (age 4-5 years).
The goal of this work is to develop a diagnostic-assist tool for classifying mice into two categories: mice with symptoms of ASD (Autism Spectrum Disorder) and mice without such symptoms, based on recordings of their squeaks.
Our work is generally made up of selection of prominent data features that can help us successfully classify the mice, and the implementation of the final classification model. The model is based on machine learning methodologies, more specifically XGBoost, which is a state-of-the-art and very popular algorithm in recent years, following tremendous success on Kaggle and other machine learning competitions. It receives features that we extracted from the dataset, which is made up of thousands of mice recordings split into single-syllable segments, and additional data such as ID, sex, inter-syllable interval duration, etc., including a manual tag of the mouse as ASD/healthy.
Finally, after optimizing the algorithm, we achieved a final classification accuracy of approximately 88%, completing, for the first time, a full pipeline for detection of autism in mice.
