AI Identifies Subtle Indicators of Depression in Students’ Facial Expressions

Subthreshold depression (StD) is a milder form of depression that doesn’t meet the criteria for a clinical diagnosis, yet serves as an early warning sign of potential future mental health problems. It’s like standing at the brink of more severe depression.

Individuals with depression often display reduced facial expressions, and previous studies have used facial cues to identify conditions such as anxiety and stress. However, the challenge remains: it’s still unclear whether Subthreshold Depression (StD) influences how people express or interpret emotions through facial expressions.

Researchers are now investigating whether slight changes in facial behavior could serve as early indicators of Subthreshold Depression (StD), potentially allowing intervention before it develops into a more severe condition.

Waseda University Researchers Use AI to Uncover Early Facial Indicators of Depression

At Waseda University, Associate Professor Eriko Sugimori and doctoral student Mayu Yamaguchi examined facial data from Japanese university students, using artificial intelligence to explore how subtle facial cues might reveal early signs of depression.

Their study revealed specific patterns of facial muscle movement associated with depressive symptoms, even among individuals without a clinical diagnosis. This suggests that AI could play a valuable role in identifying early signs of emotional distress, offering a promising approach to preventative mental health care.

Sugimori explained, “With growing concerns about mental health, I was interested in examining how subtle non-verbal cues—like facial expressions—affect social perceptions and reveal underlying mental states through AI-driven facial analysis.”

AI Analyzes Student Self-Introductions to Link Facial Cues with Perceived Emotions

The researchers had 64 Japanese students record brief self-introduction videos. A separate group of students then watched these clips and rated the speakers on traits such as friendliness, expressiveness, and naturalness. At the same time, an AI program called OpenFace 2.0 examined subtle facial muscle movements in the videos.

The findings revealed a clear pattern: students showing mild depressive symptoms (StD) were perceived as less friendly and expressive—but not unnatural or insincere. In other words, StD doesn’t create a negative impression; it simply dulls positive emotional expressions.

By using AI to analyze the videos, the researchers identified subtle facial movements—like slight eyebrow raises, eye widening, and mouth stretching—that were more frequent in students with StD. These micro-expressions were strongly linked to higher depression scores, even though they were too subtle for most people to consciously detect.

Importantly, the study centered on Japanese students, where cultural norms influence how emotions are expressed. This emphasizes the significance of cultural context when interpreting facial cues.

Sugimori explained, “Our innovative method—using brief self-introduction videos combined with automated facial analysis—can be adapted for mental health screening in schools, universities, and workplaces.”

This technique holds promise for integration into mental health technologies, digital health platforms, and employee wellness programs, offering an efficient way to track psychological well-being.

Sugimori concluded, “Overall, our research introduces an accessible, non-invasive AI-based facial analysis tool for the early detection of depression—before clinical symptoms appear—allowing for timely intervention and care.”

---

Read the original article on: TechXplore

Read more: Kissing Your Spouse Could Transmit Anxiety and Depression, Study Suggests