# Depressive Tendency Biases Ensemble Perception of Emotional Faces

## TL;DR

People with depressive tendencies perceive the average emotion of multiple faces differently when consciously judging them, but not when doing so automatically.

## Contribution

This study reveals that depressive tendencies bias explicit, but not implicit, ensemble perception of emotional faces.

## Key findings

- High depressive symptom individuals showed a bias toward anger in explicit emotion averaging tasks with longer exposure times.
- No significant difference in automatic ensemble emotion perception was found between high and low depressive symptom groups.
- The bias in explicit perception disappeared when exposure time was shortened to 50 ms.

## Abstract

Previous studies have shown that individuals with depression have an impaired ability in encoding single facial expressions. However, little is known about how depressive tendencies—subclinical emotional distress that may progress to clinical depression—affect the perception of the average emotion of multiple faces. To address this question, the current study investigated whether depressive tendencies affect explicit or implicit ensemble perception of emotion. In Study 1, participants viewed sets of four emotionally varying faces (ranging from angry to happy) for 2000 or 50 ms, then judged if a subsequent test face was angrier than the average emotion of the preceding set. Results showed that the high depressive symptom (HDS) group had a point of subjective equality (PSE) more biased toward anger compared to the low depressive symptom (LDS) group when exposure time was 2000 ms. However, this difference disappeared when the time was shortened to 50 ms. In Study 2, we assessed the automatic perception of ensemble emotion by requiring participants to judge whether a probe face was a member of the preceding set, a task that does not explicitly demand averaging. Results indicated that the HDS and LDS groups had a similar likelihood of misidentifying the set mean as a member under both 2000 and 50 ms conditions, indicating a comparable automatic coding of ensemble emotion. Together, the current research demonstrates that depressive tendencies can bias ensemble coding for emotional faces at explicit level but not at implicit level.