StudentSADD: Mobile Depression and Suicidal Ideation Screening of College Students during the Coronavirus Pandemic

ML Tlachac, Ricardo Flores, Miranda Reisch, Rimsha Kayastha, Nina Taurich, Veronica Melican, Connor Bruneau, Hunter Caouette, Ermal Toto, Elke Rundensteiner

Keywords: mobile health, mental health assessment, transfer learning, digital phenotype

TL;DR: We present the StudentSADD dataset with machine learning model baselines for mental illness screening.

Abstract: The growing prevalence of depression and suicidal ideation among college students is alarming, with the Coronavirus pandemic further highlighting the need for universal mental illness screening technology. While traditional screening questionnaires are too burdensome to achieve universal screening in this population, data collected through mobile applications has the potential to identify at-risk students. However, knowing the modalities that students are willing to share and that contain strong screening capabilities is critical for developing such mental illness screening technology. Thus, we deployed a mobile application to over 300 students during the pandemic to collect the Student Suicidal Ideation and Depression Detection (StudentSADD) dataset. Overall, students were most willing to share text responses, unscripted voice recordings, and scripted voice recordings. To provide baselines, we trained machine learning and deep learning methods on these modalities to screen for depression and suicidal ideation. The novel StudentSADD dataset is a valuable resource for developing mobile mental illness screening technologies.