Monitoring of Cognitive Functioning by Measuring Reaction Times with Wearable Devices Burcu Cinaz, Bert Arnrich, Gerhard Tröster ETH Zurich, Electronics Laboratory, Gloriastrasse 35, 8092 Zurich, Switzerland burcu.cinaz@ife.ee.ethz.ch Abstract—Reaction time (RT) tests are known as simple and sensitive cognitive tests. A drawback of existing RT tests is that they require the full attention of a test person which prohibits the measurement of cognitive functioning during daily routine tasks. In this contribution we present our first steps in designing and evaluating reaction time tests which can be operated throughout everyday life by means of wearable devices. In a feasibility study we induce changes in reaction times by applying cognitive load in 5 test subjects. We compare the obtained wearable reaction times with desktop-based reaction time tests. We show that relative changes in the mean duration and the variability of reaction times are similar for both desktop-based and wearable reaction time test. We conclude that wearable reaction time tests seems feasible to measure changes in reaction times and hence would allow the measurement of cognitive functioning throughout everyday life. Keywords-reaction time; response time; wearable; dementia; alzheimer; reaction time experiments I. INTRODUCTION Cognitive decline is commonly considered to be a consequence of typical aging processes. However, cognitive impairment may turn into Alzheimer’s disease or dementia [3]. Early detection of cognitive decline in the elderly would provide the opportunity to start with the treatments early in the disease process. For diagnosing cognitive decline there exist many different techniques such as neuropsychological tests, neuroradiological techniques, and laboratory testing [2,11]. Promising alternatives are computerized assessments of cognitive functioning which are also suited to early detect changes. In comparison to traditional neuropsychological assessment methods, computerized tests offer benefits such as cost reduction and time savings. An extensive review about computer-based cognitive tests for detecting cognitive decline can be found in [13]. A simple and sensitive cognitive test is the reaction time (RT) test which is defined as a measure of how rapidly information can be processed and a response to it can be activated [10]. In other words, it is the elapsed time between a stimulus and the individual's response to it. According to Jensen [8], RT tests are well suited for practical application in health care since in comparison to conventional psychometric techniques, RT tests offer a high sensitivity for detecting variation in cognitive efficiency and they can be virtually unlimited repeated. The main drawback of RT tests is the requirement of the full attention of the subject, i.e. the subject has to interrupt his daily routine for several minutes in order to perform the task on the computer. Our goal is to develop reaction time tests which can be operated throughout everyday life by means of wearable devices. An important step in the development is to ensure that wearable reaction time tests are suitable to measure changes in reaction times similar to desktop-based approaches. In this contribution we present our first results in operating a reaction time test by hand movements. We induce changes in the duration and variability of reaction times by applying cognitive load to test subjects. In the following we propose and evaluate a setup to measure the response to a stimulus by recognizing certain hand movements of a subject with a 3D accelerometer. We compare the obtained wearable reaction times with desktop-based reaction time tests in two experimental conditions: (i) single-task in which the subject has to respond to a target stimulus, and (ii) dual-task in which the subject has to solve a cognitive task in parallel to the single-task. In the following we first we present related work. Next, we describe our methods including experimental setup and data acquisition. Afterwards, we present the comparison of the wearable reaction time test with a desktop-based approach. Finally, we summarize our paper and provide an outlook for future work. II. RELATED WORK There exist mainly three kinds of reaction time tests [9]: simple, recognition and choice reaction. Simple reaction time tests consist of one stimulus and one response. For instance the subject has to press a button as soon as the letter "X" appears at a pre-defined position or as soon as a light or sound appears. In recognition reaction time tests, subject has to respond to one stimuli type and ignore other stimuli types. This is sometimes called as "go/no-go" reaction time task. Recognition of a particular sound or symbol belongs to this category. Lastly, choice reaction time tests include multiple stimuli and multiple responses. The subject has to respond to each stimulus with a corresponding response e.g. pressing a key whenever a corresponding letter appears on the screen. A detailed series of recommendations on how to conduct experiments using reaction times and how to analyze data can be found in [8,10,12]. PervasiveHealth 2011, May 23-26, Dublin, Republic of Ireland Copyright © 2012 ICST DOI 10.4108/icst.pervasivehealth.2011.246012