Publish By:Beatrix Vereijken
Blog by: Dr. Wei Zhang, Dr. Anisoara Ionescu and Prof. Kamiar Aminian. Laboratory of Movement Analysis and Measurements, EPFL Zurich, Switzerland.
Life expectancy is higher than ever, making it crucial that we empower older adults to take care of their own health and function as much and as long as possible. Physical activity plays an important role in healthy living. Thanks to the advancement of welfare technologies, such as mobile health (mHealth) technology, we are one step further in delivering services to assist older adults to stay physically active and adopt healthy behaviour in their daily life.
What is physical activity and how can it be measured by a smart phone or smart watch?
Physical activity[1]is any bodily movement that works your muscles and requires more energy than resting. Walking, running, cycling but also house cleaning and gardening are examples of physical activities performed in daily life. Physical activity can be quantified by its duration and intensity by monitoring activity by inertial motion sensors, typically accelerometers and gyroscopes, which are embedded in a portable device, such as a smart mobile phone or a smart watch. For example, the intensity of the movement is indicated by the magnitude of the accelerometer signals. A flat accelerometer signal at around 0 indicates that a person remains inactive (Figure 1).
Figure 1. Postures and movements can be identified based on the accelerometer signal features.
With advanced pattern recognition techniques, the types of different activities can be distinguished, thereby allowing to detect activities the person is performing or to estimate the energy consumed by performing the activities or the number of steps taken during a day.
What is physical behaviour and how can it be measured by mHealth technologies?
Physical behaviour is an umbrella term, which includes the behaviour of a person in terms of body postures, movements, and daily activities in his/her own environment[2]. In recent years, multi-modal sensing and sensor fusion techniques have enabled new developments in mHealth. Beyond the different activities identified by the motion sensors, more information provided by the smart phones and smart watches can be used to deduce the context in which the physical activities are performed. For example, a mHealth system can tell whether a person has walked outdoors and the distance walked by sampling the GPS sensor data. Such data provides additional insight into different contexts of the same types of physical activities (indoors or outdoors). Moreover, this context information can be used to deduce other aspects of a person’s behaviour, such as social interaction with the outside world.
How to quantify one’s physical behaviour?
There are many different ways to describe physical activity and behaviour. Some typical examples are percentage of time spent on different postures and movements in a day, energy expenditure in the form of burnt calories, the number of steps or the distance walked, etc. This information is usually illustrated using a pie chart or a list of numerical values. In PreventIT, we developed a new way to register and illustrate the physical behaviour that incorporates the multi-dimensional nature of physical activity in the daily environment by using a colour-coded barcode. The colour changes from cold (dark blue) to hot (dark red) when the person is active and thus the barcode changes from a lower (that is, more passive) to a higher (that is, more active) state. The barcode illustrates each posture or movement bout in a given state based on its type, duration and intensity in a time sequence. A physical activity bout is assigned to a higher state than a sedentary bout. The longer the physical activity lasts or the higher the intensity is, the higher the barcode state will be and hence, the ‘hotter’ the colour is in the barcode. The barcode maps physical activities with useful details to a time sequence of the monitored period. It directly shows the distribution of different physical activities and the dynamic change between the postures and movements. From the barcode example in Figure 2, we can tell the difference between the behaviours of the two persons. Person A is a patient suffering from chronic pain, whose barcode shows mostly bluish colours, which indicates that the person spent most of the time being inactive and had several sustained sedentary periods. On the other hand, the age-matched healthy person B’s barcode is filled with more reddish states, where he/she was engaging in high intensity physical activity.
Figure 2. Two examples of barcode illustration of physical activity of a chronic pain patient (A) and an age-matched health person (B) [3]
From Figure 2, we can also tell that the healthy person B had more bouts of physical activities performed with various length and intensity throughout the monitored period, reflecting a more varied or complex pattern throughout his/her day. In contrast, chronic pain patient A had fewer physical activity bouts with little variation (few colours), reflecting a rather simple day. The barcode suggests that the chronic pain patient was not able to dynamically alternate between various body movements or activities, probably because of pain or other factors such as fear of movement and activity avoidance3.
In order to have a numerical signature of the pattern illustrated by the barcode, the entropy of the barcode can be computed to quantify the level of complexity of the physical activity of the monitored period. The entropy, which is based on the Lempel-Ziv algorithm, captures unique sub-patterns or combinations of postures and movements encoded in the barcode. The more sub-patterns and combinations found, the higher the entropy is. It means that the more variety or alternation between postures and movements the person performs, the more complex the person’s physical behaviour is.
Adopt a healthy behaviour by engaging in physical activity in everyday life
Reduced variation in behaviour or complexity does not only appear in patients suffering from chronic conditions. With ageing, reduced complexity in physical behaviour reflects functional decline (Figure 3). Reduced complexity at old age implies that the person might be in a down-spiral circle. This circle leads the person to reduced adaptability and being vulnerable to the fluctuations and perturbations in the daily living environment.
Figure 3. Complexity of physical behaviour reduces at high age.
To assist older adults in maintaining their physical function as long as possible, PreventIT proposes to add extra strength and balance challenges to daily activities, break up sedentary periods, and increase length and intensity of active periods. For example, one suggestion can be to break long sitting periods and take a walk; while walking, trying to take a longer path and increasing the speed on the way to the destination. Even such small changes will fill up a day with more variety in physical activities (Figure 4) and thereby allow the person to adopt and maintain a complex physical behaviour. The higher complexity one manages to maintain at old age, the better equipped one is to meet daily challenges.
Figure 4. Increase the complexity by breaking up long sitting and walking longer and faster
[1] Paraschiv-Ionescu A, Perruchoud C, Buchser E, Aminian K. Barcoding Human Physical Activity to Assess Chronic Pain Conditions. PLOS ONE. 2012 Feb 23;7(2):e32239.
[2] J. B. J. Bussmann and R. J. G. van den Berg-Emons, “To total amount of activity and beyond: perspectives on measuring physical behavior,” Front. Psychol., vol. 4, no. JUL, pp. 1–6, 2013.
[3] Paraschiv-Ionescu A, Perruchoud C, Buchser E, Aminian K. Barcoding Human Physical Activity to Assess Chronic Pain Conditions. PLOS ONE. 2012 Feb 23;7(2):e32239.