
Mascha Pauelsen. Photo: Karin Sjöberg
30 November 2020
Uses AI to examine balance
The system involved in human balance is a complex one. To be able to investigate it properly, Mascha Pauelsen uses artificial intelligence, AI. – I have used machine learning, which can analyse many variables simultaneously, she says.
Mascha Pauelsen, PhD Student in Physiotherapy at Luleå University of Technology, is interested in in the things that influence our balance.
– I want to investigate the relationship between balance and fear of falling as we age. There are many things that can influence our balance, for example sight, sensitivity in our feet and the fear to fall. Most of all it is about the collaboration between many factors, she says.
She wants to understand the mechanisms behind it all to be able to develop interventions that work. There is good evidence for training and improving balance, but interventions for fear of falling are not as successful.
– After the treatment period ends, the fear gets worse again. We would be able to prevent that if we really understood the phenomenon, says Mascha Pauelsen.
Collected a lot of data
Earlier, she invited people aged 70 and over to the Human Movement Science Lab at Luleå University of Technology. They performed a myriad of activities, some of them while standing on a force plate and wearing reflective markers for motion capture. She was able to collect a lot of data this way.
In collaboration with researchers from the Department of Computer Science, Electrical and Space Engineering, she then analysed all the gathered information by using machine learning methods.
– I used existing software and code, but when you apply those to new areas you may have to adapt the code a little, as was the case here, says Mascha Pauelsen.
AI analyses the different data in an unending number of combinations.
– It can discover things you would otherwise never see and can perform very complex analyses. It can combine and analyse all the variables and estimate how they work together in a way that a human mind cant.
Learn to process
Mascha Pauelsen now wishes she had access to more data.
Machine learning processes the data and iteration after iteration it learn to sort and process the information in more and more advanced ways.
– Today, its difficult due to national and international law and regulations on how we are allowed to handle human data. To reach the size of data that is preferable, you have to share and reuse data, but that is not something that is allowed when dealing with human health-related data, says Mascha Pauelsen.
She points out that the issue is gaining political awareness and it’s being investigated, and laws and regulations might change in the near future. Mascha Pauelsen is of the opinion that with access to larger data, care and rehabilitation concerning a persons balance could improve.
– By using machine learning this way, we might be able to identify the variables that are the biggest cause of fear of falling.
Published:
Updated: