cMIND AI Web Tool Usability Study (Hong Kong)

Overview

This study is testing a new web-based tool that uses artificial intelligence (AI) to help older adults in Hong Kong check how healthy their Cantonese-style meals are for brain health. The tool is based on the cMIND diet, a Chinese-adapted version of a known healthy eating pattern that may support memory and thinking skills. Participants will use the web app to take photos of their usual meals for at least 10 days over two weeks. The AI will automatically identify ingredients and give a score showing how well the meal follows the cMIND diet. The study will also ask participants to complete a short questionnaire and a brief interview to find out how easy and useful the tool is for older adults. The purpose of this small study is to see whether the AI tool is user-friendly and acceptable for older people. Results will help improve the tool for future use to support healthy ageing and brain health.

Mild cognitive impairment (MCI) is common among older adults and can progress to dementia. Diet plays an important role in brain health. The cMIND diet is a culturally adapted Chinese version of the Mediterranean-DASH diet, designed to support cognitive function. However, many older adults find it difficult to track their adherence to this diet using traditional methods. This study is developing and testing a simple web-based AI tool to help older adults in Hong Kong monitor their Cantonese-style meals. Users take photos of their meals (such as dim sum, stir-fries, or congee) using the web app. The AI automatically identifies ingredients in mixed dishes and calculates a cMIND adherence score (0-12), giving immediate personalised feedback on how well the meal supports brain health. The main part of the study is a small usability and acceptability test. We will recruit 20 community-dwelling older adults aged 60 years and above who regularly eat Cantonese-style meals. Participants should not have a self-reported diagnosis of dementia or Alzheimer's disease, or other psychiatric/medical conditions that would interfere with participation or valid outcome assessment. Eligible participants will receive a 20-minute training session on how to use the web tool. They will then use the app to photograph their usual meals for at least 10 days over a two-week period, without changing their normal eating habits. At the end of the two weeks, participants will complete a short online questionnaire about the ease of use and usefulness of the tool. They will also take part in one short individual interview (about 20 minutes, audio-recorded) to share their experiences and suggestions. This low-risk study aims to understand whether older adults find the AI tool easy and acceptable to use in daily life. The results will help improve the prototype for future larger studies. All data will be kept strictly confidential, and ethics approval has been obtained from the Hong Kong Metropolitan University Research Ethics Committee (Reference: HE-FRSE/2026/08).