The MAIJU Wearable Jumpsuit Hides Motion Trackers for ML-Powered Motor Development Monitoring

Using laminated pockets to hold motion-tracking pucks, the MAIJU wearable could offer in-home medical monitoring.

Researchers working at Helsinki Children's Hospital's BABA Center have developed a smart jumpsuit, designed to track toddlers' motor development — by hiding multiple wireless movement sensors in the fabric.

"The development of the MAIJU [Motor Assessment of Infants with a Jumpsuit] wearable required a technical breakthrough in the development of machine learning algorithms for this purpose," claims project lead Manu Airaksinen, PhD. "This was achieved by combining a new kind of motility description with state-of-the-art deep learning solutions."

The soft, washable MAIJU jumpsuit could help doctors keep better track of infants' motor development through machine learning. (📹: University of Helsinki)

The MAIJU project tracked infants aged between five and 19 months, each wearing the specially-designed jumpsuit with its four motion-tracking pucks — waterproof Movesense devices with Bluetooth Low Energy (BLE) connectivity, tucked in laminated pockets on each limb — during spontaneous play, while a camera provided a visual record.

Using the camera footage to label the movement data, the team developed a new motility description scheme to train a machine-learning algorithm for posture and movement recognition — hitting, the team claims, an accuracy level equivalent to visual assessment by a trained expert, but entirely automatically.

"Our research shows that it is very possible to assess the motor development of an infant outside of a hospital or special laboratory setting," says Sampsa Vanhatalo, professor of physiology and BABA Center lead. "A particular advantage of the MAIJU methodology is the fact that it allows us to carry out developmental assessments in the natural environment of the child, such as a home or daycare.

"Our methods can be automatized and scaled up for very wide use. It is also possible that our technology could be adapted for developing wearable solutions to help other patient groups, such as older children or even elderly people."

The data captured by the system could be used, the team suggests, to track motor development and offer early warnings of potential motor or neurological difficulties — as well as help to support and monitor the efficacy of various therapies.

"Methods of this kind are urgently needed to support the research and novel therapeutic innovations of early neurological development," claims Leena Haataja, professor of child neurology and one of the team working on the project, of the MAIJU wearable. "It [is] important to encourage children to move as naturally and as much as possible in everyday life situation."

The team's work has been published under open-access terms in the journal Communications Medicine.

Gareth Halfacree
Freelance journalist, technical author, hacker, tinkerer, erstwhile sysadmin. For hire:
Latest articles
Sponsored articles
Related articles
Latest articles
Read more
Related articles