Young adults are a primary high-risk group for depression, yet due to factors like healthcare resources and stigma associated with the disease, most sufferers do not receive effective treatment.

In music therapy, music plays a key role in activating autobiographical memories. If the music selection is inappropriate, it may trigger negative memories, potentially worsening the patient's condition. Choosing music materials based on the characteristics of autobiographical memory activation related to positive emotional valence that could improve treatment outcomes.

Therefore, this project aims to build a music stimulus library targeted at young adults, using machine learning algorithms to analyse the features of autobiographical memory activation and cognitive characteristics associated with positive emotional valence in young individuals with depression. This will help in forming standardized treatment materials and protocols for music-based digital therapy and lay the groundwork for subsequent research into the neural mechanisms involved.

​