Emotion Recognition
Emotion Recognition tests whether participants can identify the emotion expressed in facial images. Researchers upload up to 50 images and label each with one of five emotions. Participants are shown a random subset of 10 images and must select the matching emotion.
ActivitySpec: lamp.emotion_recognition
Cognitive domain: Social cognition
Configuration
Researchers must provide a dataset of up to 50 facial images and label each with the corresponding emotion. Default emotions: happiness, sadness, fear, anger, and neutral. Custom emotion labels beyond these five can be added in the activity settings.
A random subset of up to 10 images is selected and shuffled each session. If fewer than 10 images are uploaded, all images will be shown in random order.
Sample Instructions
"In this task you will be presented with up to 10 facial images and asked to identify the emotion expressed by each one from a list of 5 emotions."
Usage
The participant sees one image at a time, selects the emotion they believe matches, and taps "Save" to proceed. The assessment cycles through up to 10 randomly selected images.
Scoring
Scoring is based on whether the user correctly identifies the emotion expressed in each image.
Screenshots
Data
temporal_slices
| Field | Type | Description |
|---|---|---|
item | number | Trial number (1-indexed, up to 10) |
value | string | Selected emotion: "Happiness", "Sadness", "Fear", "Anger", "Neutral" |
type | boolean | true = correct, false = incorrect |
duration | number | Response time (ms) |
level | string | Correct emotion text label (the expected answer for this trial) |
Cortex Features
No Cortex features are currently available for this activity.
View in Portal | Python SDK | API Reference