Please activate JavaScript!
Please install Adobe Flash Player, click here for download

uni'wissen 02-2015_ENG

other hand, we need the expert knowledge of psychologists to develop our systems because they’re made for people.” In cooperation with the Dresden University Medical Center, Van Laerhoven has been working since 2014 on a study with 24 test subjects suffering from bipolar disorders who have agreed to wear a sensor on their wrist day and night. The device consists of a battery, acceleration, light, and temperature sensors, and a memory card that records 100 data sets per second over the course of two weeks. Every day it records a total of approximately 25 million data sets. The data on arm and hand movements provide information on the wearer’s actual activities, because even bicycle riding and sleeping are associated with characteristic arm movements and postures. “Since an astounding amount of activities can be identified from typical hand movements, the sensor can reconstruct an entire daily routine,” explains Van Laerhoven. “Yoga exercises, for instance, are very easy to analyze, because they always involve particular arm movements.” Smoking is also associated with a certain posture of the wrist and the equally typical repetitive movement with which the cigarette is raised to the mouth. Combined with the data from the included light and tempera- ture sensors, this data can be used by the soft- ware to create a daily and nightly activity log. Nonstop Operation Once the patients have put the sensor on their wrist, they don’t need to do anything more: they don’t need to turn it off or recharge it, because the battery has a life of several weeks. This also means that there aren’t any gaps in the data – as long as the test subjects don’t take off the sensor again. The compact size and light weight of the sensor are also helpful: “The patients confirm that they didn’t even notice the tiny device any- more after a short time. This makes the recorded movements very realistic.” That is also the sensor’s main advantage over video recordings: For one thing, it is not possible to record all of a patient’s activities nonstop with a video camera. In addition, uni wissen 02 2015 The patient’s day includes a bike ride to the of- fice, three cups of coffee, regular games of tennis, yoga to wind down, and peaceful sleep at night – but several days later he finds himself canceling the sport, smoking cigarettes with his coffee, and sleeping longer but less peacefully. The daily routine of patients with bipolar disorders can change in the transitions from manic and depressive phases. This is information the psycho- logists treating the illness need to know. A detailed record of the patient’s activities enables them to analyze these transitions, determine which activities were stopped and which were started, and to help their patients better. Up to now, psychologists have had to rely on their patients’ journal entries to reconstruct the transition between these phases. This type of diagnosis and analysis depends for its success on the self-perception and reliability of the patient, and it is thus not always possible for the therapist to create a precise record of the patient’s activities in this way. Prof. Dr. Kristof Van Laerhoven from the Faculty of Engineering of the University of Freiburg has found a new means of recording activities without interruption: The computer scientist and his team are building tiny sensors that can be worn like wristwatches and capture all movements of the arm and hand. A special computer program then identifies which motion patterns stand for which activities and compiles a detailed weekly summary. “We want to develop a diagnostic instrument to support psychologists in their work,” explains Van Laerhoven. The Freiburg researcher sees a close link between his field of research and psychology: “In computer science and microsystems engineering we are concerned with developing systems that can collect and manage data. That is an important tool for empirical research in psychology. On the Yoga with a wearable sensor: The tiny wrist-worn device records the typical motion patterns for this activity. Photo: Thomas Kunz “An astounding amount of activities can be identified from typical hand movements.” 21 uni wissen 022015