Summary: Depression’s earliest signs can be hard to spot, but a new study shows AI can detect them in subtle facial movements. Japanese students with subthreshold depression were perceived as less friendly and expressive by peers, despite not seeming nervous or fake.
AI analysis revealed specific eye and mouth muscle activity patterns that aligned with depression scores. The approach may one day enable accessible screening in schools, workplaces, and digital health platforms.
Key Facts
- Peer Ratings: Students with subthreshold depression appeared less expressive, likable, and friendly.
- AI Detection: Micro-movements in eyes and mouth strongly correlated with depression scores.
- Early Screening Potential: The method offers a non-invasive way to detect depression before clinical symptoms appear.
Source: Waseda University
Depression is one of the most common mental health challenges, but its early signs are often overlooked. It is often linked to reduced facial expressivity.
However, whether mild depression or subthreshold depression (StD) (a mild state of depressive symptoms that does not meet the criteria for diagnosis but is a risk factor for developing depression) is associated with changes in facial expressions remains unknown.
In light of this, Associate Professor Eriko Sugimori and doctoral student Mayu Yamaguchi from the Faculty of Human Sciences, Waseda University, Japan, have now investigated changes in facial expression in Japanese undergraduates using facial data and artificial intelligence.
The study was published in the journal of Scientific Reports on 21 August 2025.
“As concerns around mental well-being have been rising, I wanted to explore how subtle non-verbal cues, such as facial expressions, shape social impressions and reflect mental health using artificial intelligence-based facial analysis,” says Sugimori.
The researchers asked 64 Japanese university students to record short self-introduction videos. Another group of 63 students then rated how expressive, friendly, natural, or likeable the speakers appeared. At the same time, the team used OpenFace 2.0, an artificial intelligence system that tracks micro-movements in facial muscles, to analyze the same videos.
The results revealed a consistent pattern. Students who reported subthreshold depressive symptoms were rated by their peers as less friendly, expressive, and likable. Interestingly, they were not judged as more stiff, fake, or nervous. This suggests that StD does not make people appear overtly negative but rather tones down their positive expressivity.
Artificial intelligence analysis revealed specific patterns of eye and mouth movements, such as the inner brow raiser, upper lid raiser, lip stretcher, and mouth-opening actions that were more frequent in participants with StD. These subtle muscle movements were strongly linked to depression scores, even though they were too fine for untrained observers to pick up on.
The researchers note that their study was conducted with Japanese students, an important consideration given that cultural norms influence how people express emotions.
“Our novel approach of short self-introduction videos and automated facial expression analysis can be applied to screen and detect mental health in schools, universities, and workplaces,” says Sugimori.
The proposed approach could be used in mental health technology, digital health platforms, or employee wellness programs to monitor psychological well-being efficiently.
“Overall, our study provides a novel, accessible, and non-invasive artificial intelligence-based facial analysis tool for early detection of depression (before the appearance of clinical symptoms), enabling early interventions and timely care of mental health,” concludes Sugimori.
About this AI and depression research news
Author: Armand Aponte
Source: Waseda University
Contact: Armand Aponte – Waseda University
Image: The image is credited to Neuroscience News
Original Research: Open access.
“Subthreshold depression is associated with altered facial expression and impression formation via subjective ratings and action unit analysis” by Eriko Sugimori et al. Scientific Reports
Abstract
Subthreshold depression is associated with altered facial expression and impression formation via subjective ratings and action unit analysis
Depression is often linked to reduced facial expressivity and to biases in recognizing others’ emotions. Whether subthreshold depression (StD)—a putative prodromal stage—shows comparable alterations remains unclear.
We recorded 10‑second self‑introduction videos of Japanese undergraduates (ratees; n = 64) and obtained subjective impression ratings from a separate group (raters; n = 63).
Both groups completed the Beck Depression Inventory‑II (BDI‑II). Raters’ depressive tendency was not associated with their impression ratings (p > .10, Benjamini–Hochberg corrected).
In contrast, ratees with StD (BDI‑II = 11–20) received significantly lower scores on positive items—expressive, natural, friendly, likeable—than healthy ratees (BDI‑II = 1–10; partial η² = 0.18–0.70).
Automated analysis with OpenFace 2.0 showed higher presence/intensity of AU01 (Inner Brow Raiser), AU05 (Upper Lid Raiser), AU20 (Lip Stretcher), AU25/26/28 (mouth‑opening AUs) in StD faces; five of these AUs correlated with BDI‑II after false‑discovery‑rate correction (q < 0.05).
Subthreshold depression was associated with muted positive expressivity and distinct eye‑ and mouth‑movement patterns, but did not bias observers’ first‑impression judgements. The observed AU signatures may aid early identification of individuals at risk for clinical depression.
Together, our findings suggest that subthreshold depression is associated with alterations in facial expressivity, particularly in positive expressions, while not significantly influencing how others perceive those expressions.
