🚨 New preprint on arXiv from @OxWearables @pixl_oxford ! Can vision-language models (VLMs) help automatically annotate physical activity in large real-world wearable datasets (⌚️+📷, 🇬🇧 + 🇨🇳). 📄 https://t.co/08Klw511JV 🧵1/7

JOB ALERT: Please disseminate to any ECR's in your networks that may be interested in joining our team to assist in procuring large validation datasets for the wider research community. 2 posts available:

Of particular interest to @ProPASSconsort @M_Stamatakis @ismpb_org @ISPAH @F_Schwendinger @PC_Dempsey

Here, we analysed data from 3,138 UK Biobank participants with repeated accelerometer measurements to assess the long-term reproducibility of 9 PA & sleep phenotypes. We also illustrated how associations of these with disease risk may have been substantially underestimated. 2/2

𝐂𝐚𝐧 𝐚 𝐬𝐢𝐧𝐠𝐥𝐞 𝟕-𝐝𝐚𝐲 𝐚𝐜𝐜𝐞𝐥𝐞𝐫𝐨𝐦𝐞𝐭𝐞𝐫 𝐦𝐞𝐚𝐬𝐮𝐫𝐞𝐦𝐞𝐧𝐭 𝐜𝐚𝐩𝐭𝐮𝐫𝐞 𝐡𝐚𝐛𝐢𝐭𝐮𝐚𝐥 𝐩𝐡𝐲𝐬𝐢𝐜𝐚𝐥 𝐚𝐜𝐭𝐢𝐯𝐢𝐭𝐲 𝐚𝐧𝐝 𝐬𝐥𝐞𝐞𝐩 𝐩𝐚𝐭𝐭𝐞𝐫𝐧𝐬?⁣⁣ Preprint: https://t.co/YfxXTG85e0⁣ ⁣ Great work by @CharilaosZisou. Feedback welcome! 1/2

Today we were delighted to host @M_Stamatakis at @Oxford_NDPH who presented his work over the last few years investigating how micro-patterns of physical activity are associated with beneficial health outcomes.

It's been 1 week since we welcomed 32 researchers from around the world to Oxford to learn about machine learning of wearables data in large scale biomedical studies. We can't wait to see how this knowledge is used in future work! https://t.co/MzIeIG7L26

Thanks for your inspiring contribution to our reproducible machine learning course!

Great analysis by @KoffmanLil77098 and @StrictlyStat on step counting algorithms. Our StepCount package is available on Github for all researchers to use freely:

Interested in learning more about wearables in large-scale biomedical studies? We're running a residential short course from 22 - 26 September at Oxford Includes: inspirational speakers, great tutors, and hands-on data analysis. https://t.co/5tB4QKfHvo

I had a great time presenting and discussing my poster with others at #ICAMPAM2024 yesterday. OxWEARS will produce the largest annotated validation dataset in the world, and will join other datasets made publicly available by the @OxWearables group at @Oxford_NDPH @UniofOxford

Members of our @Oxwearables group will be delivering 4 oral presentations, 1 workshop, 1 symposium and 1 poster at next weeks #ICAMPAM2024 conference organised by @ismpb_org We are all looking forward to coming and sharing the work of our group.

Congratulations Rosemary!

For @ismpb_org (and other) researchers interested in using @uk_biobank wearable sensor data, we have a really exciting workshop at the #ICAMPAM2024 conference with @DrAJLewandowski, @uk_biobank @AlainaShreves, @theNCI https://t.co/NpsURvVuRC

StepCount can be used for Axivity, Actigraph, GENEActiv data, or a custom .csv file. The repository can be found on Github with detailed instructions on how to use it: https://t.co/ecxOIqXro5 2/2

Just published: Performance of our open-source, transparent method for deriving steps from wrist-worn accelerometry 'StepCount'. https://t.co/PBYUr3Tt7n 1/2

Deadline 10th May. Please do get in touch if interested!

The fantastic Shing Chan keeps implementing new features into the @OxWearables stepcount package! Along with validated steps, the package now outputs peak cadence values. 6-minute video tutorial here:

New article! Read Self-supervised learning for human activity recognition using 700,000 person-days of wearable data https://t.co/lHkTu9AEdT in npj Digital Med

Happy to announce our foundation model for wearables published at npj Digital Medicine today. This model sets a new standard in #Wearables, significantly outperforming human activity recognition benchmarks in diverse conditions. 🚀 https://t.co/RfIRY5UMTz