Emotion-Sensing Smart Wristband LumiBand: Reading Your Real Mood Through Skin Electrical Signals

LumiBand, released by NEC's Biomedical Research Laboratory, is a wearable device that recognizes a user's emotional state in real time through skin electrical signals. The wristband contains a micro-sensor that continuously collects skin conductance, heart rate, and temperature changes at the user's wrist, and uses an on-device AI model to infer the emotional state, distinguishing between anxiety, pleasure, calm, focus, or sadness.

The project grew out of long-running observations in the field of affective computing. Traditional questionnaires and self-reports struggle to capture subtle changes in emotion, while skin electrical signals, as a direct indicator of the sympathetic nervous system, reflect emotional fluctuations within milliseconds. The NEC team set out to miniaturize and consumerize this professional psychophysiological measurement device, so that ordinary users could also understand their real emotions.

The core technology is a skin electrical sensor and an on-device inference chip built into the wristband. The sensor samples skin conductance 32 times per second, and together with optical heart rate and temperature sensors, forms a multimodal signal stream. The on-device chip runs a lightweight neural network with about 80 million parameters, performing emotion classification locally in real time, without uploading any biometric data to the cloud.

For clinical validation, NEC partnered with the University of Tokyo Faculty of Medicine in a nine-month study, recruiting 420 volunteers who wore LumiBand in different life scenarios, including work, socializing, exercise, and sleep. Volunteers simultaneously filled out emotion scales as a reference standard. Results show that LumiBand achieved 89 percent accuracy in identifying anxiety, pleasure, and calm, and 82 percent for the focus state. This precision is sufficient to support consumer-facing emotion management applications.

For applications, LumiBand is particularly suited to scenarios where emotional regulation is critical. Meditation app Calm and Japanese telecom operator NTT Docomo have both announced that they will integrate LumiBand into their respective health services. Calm will push personalized meditation courses based on the user's real-time emotion. NTT Docomo is using LumiBand in an employee mental health pilot, covering 12,000 staff members inside the group.

On the commercialization path, LumiBand will be sold as a consumer electronics product at about 48,000 yen (around 340 dollars) per unit. The companion LumiCare cloud service is available as a monthly subscription at 980 yen. NEC has also launched a B2B version called LumiBand Pro, priced at 120,000 yen, and has signed contracts with several large Japanese companies. Shinichi Kawabata, who heads NEC's Biomedical Research Laboratory, said the product's design goal is to keep all emotion data on the user's side and avoid third-party misuse.

Critics point out that privacy risks for emotion data cannot be ignored. Even if LumiBand data is not uploaded to the cloud, third-party apps calling its API could still access the user's emotional information. NEC responded that all third-party calls must be explicitly authorized by the user, and the API can only return aggregated emotion trends, not the raw biological signals.