The science behind Sense.

Validating OCOsense Smart Glasses in a Three-Week Home-based Study: Assessing Detection of Eating, Food Identification and the Use of Haptic Feedback to Aid Behaviour Modification

Padmanabhan A., Stankoski S., Panchevski F., Armitage R. M., Cox S., Baert C., Sazdov B., Darkovska M., Kiprijanovska I., Fatoorechi M., Sofronievski B., Indovska E., Nikodinovska A., Kulebanova S., Cleal A., Gjoreski M., Sevenoaks T., Nduka C., Gjoreski H., Yeomans M. R. Appetite, Vol. 218, 2026, 108382. ISSN 0195-6663. https://doi.org/10.1016/j.appet.2025.108382

This study looked at whether Sense glasses can understand how people eat in real life, and help them gently change their habits.

Participants wore the glasses during their normal day for several weeks. The glasses were able to detect when people were eating with high accuracy, without needing constant manual input.

The system also tested simple feedback, such as:

• reminders to slow down eating
• prompts to chew more
• encouragement to pause between bites

Over time, people who followed these prompts began to eat more slowly, chew more per bite, and take longer breaks between bites — all behaviours linked to better digestion, blood sugar control, and long-term health.

The study shows that everyday eating habits can be measured passively, that these habits are often hard to notice on your own, and that small, real-time nudges can help shift behaviour in a meaningful way.

Sensing the future of eating behaviour measurement: validation of OCOsense glasses to detect chewing behaviour

Armitage R. M., Padmanabhana A., Harvey T., Baert C., Gjoreski H., Stanoski S., Sazdov B., Nduka C. & Yeomans M. R. Appetite, Vol. 213, Supplement, 2025, 108064. ISSN 0195-6663. doi:10.1016/j.appet.2025.108064

This study tested whether Sense glasses can accurately measure how people eat — specifically how they chew — compared to traditional lab methods.

Participants ate a normal breakfast while:

• being recorded and analysed by researchers
• wearing the glasses at the same time

The results showed that the glasses were highly accurate:

• they measured chewing patterns almost identically to expert human analysis
• chewing rate and number of chews closely matched lab results
• they correctly identified most eating vs non-eating moments

In other words, the glasses can capture the same type of detailed eating behaviour data as a lab study, but without needing cameras, manual coding, or artificial conditions.

Pilot study to reduce chewing and eating rates using haptic feedback from the OCOsense glasses

Baert C., Sazdov B., Stankoski S., Gjoreski H., Nduka C. and Jordan C. Appetite, Vol. 213, 2025, 108056. ISSN 0195-6663. doi:10.1016/j.appet.2025.108056

This study explored whether Sense glasses can help people slow down how they eat using gentle, real-time feedback.

Participants first ate normally to establish their usual eating pace. A week later, they ate the same meal again, but this time the glasses gave subtle vibrations when they were chewing too quickly.

The results showed that with this feedback:

• people chewed more slowly
• they also reduced their overall eating speed
• importantly, they did this without eating less food

Participants found the system easy to use and acceptable in practice.

Controlled and Real-Life Investigation of Optical Tracking Sensors in Smart Glasses for Monitoring Eating Behavior Using Deep Learning: Cross-Sectional Study

Stankoski S., Kiprijanovska I., Gjoreski M., Panchevski F., Sazdov B., Sofronievski B., Cleal A., Fatoorechi M., Nduka C., Gjoreski H. JMIR Mhealth Uhealth 2024;12:e59469. DOI: 10.2196/59469

This study looked at whether Sense glasses can accurately detect eating and chewing in real life, without relying on food diaries or memory.

The glasses use small sensors to track subtle facial movements and distinguish between:

• eating
• speaking
• other movements like teeth clenching

The results showed that the system can reliably detect when you’re eating, measure chewing behaviour in detail, and do this both in controlled settings and in everyday life.

Importantly, it doesn’t just detect that you’re eating — it can analyse how you eat, including chewing patterns and pace.

FedMMA-HAR: Federated Learning for Human Activity Recognition With Missing Modalities Using Head-Worn Wearables

Gobbetti A., Gjoreski M., Gjoreski H., Lane N. & Langheinrich M. IEEE Pervasive Computing, Vol. 23, No. 04, pp. 40–49, Oct.–Dec. 2024. doi:10.1109/MPRV.2024.3475473

This study explored how wearable devices like Sense glasses can use data to understand behaviour, while also protecting user privacy.

Typically, systems like this rely on continuous data collection, which can raise privacy concerns. This research looked at a different approach, where:

• data can stay on the user’s device
• only necessary insights are shared
• the system can still work even if some data is missing

The results showed that it’s possible to build robust, accurate systems without needing perfect or complete data all the time.