Bjo ¨ rn Sondell*
Lars Nyberg
Staffan Eriksson
Department of Community
Medicine and Rehabilitation
Umeå University
Umeå, Sweden
*Correspondence to
bjorn.sondell@germed.umu.se
Bjo ¨ rn Engstro ¨m
Anders Backman
Kenneth Holmlund
VRlab
Umeå University
Umeå, Sweden
Go ¨ sta Bucht
Lillemor Lundin-Olsson
Department of Community
Medicine and Rehabilitation
Umeå University
Umeå, Sweden
Presence, Vol. 14, No. 2, April 2005, 191–197
© 2005 by the Massachusetts Institute of Technology
Altered Walking Pattern in a
Virtual Environment
Abstract
Falls and fractures among elderly persons constitute a major health problem. Many
falls occur while walking and falls that occur during turning often result in a fracture.
Methods aimed at understanding the complex mechanisms involved in walking
should therefore assess tested individuals during walks and turns. In order to iden-
tify persons at risk and take the correct preventive measures, it is important to find
methods that quantify movements as the tested persons are processing multisen-
sory input. In a clinical setting this is sometimes difficult to achieve in a controlled
manner, since tests are difficult to set exactly the same from one time to another.
Using a virtual environment (VE) and a tracker system, conditions such as light,
sound, events, body movements, and room size can be controlled and measured.
Tests in VE can therefore be identically reproduced over and over again to evaluate
if a person can withstand changing outer demands at any given moment. In order
to perform quantitative measures 8 persons (21–74 years) were tested in immer-
sive virtual reality. The VE was a corridor in which expected and unexpected events
could be produced. Events studied were doors swinging open in front of the sub-
jects during a walk and a virtual tilting of the environment. Trackers were used for
collecting and analyzing the movement data. Our results show that the system was
well tolerated among the subjects and that there was a clear tendency that the
system could generate fall tendency among the subjects. There was also a differ-
ence among the subjects regarding walking strategies when subjected to the various
events.
1 Introduction
Falls and fall-related injuries pose a serious threat to common health and
are the most common cause of injury in old age (Sjo ¨gren & Bjo ¨rnstig, 1989;
Nyberg, 1996; Binder, 2002). One in three community-living persons aged
65+ fall every year, many of them more than once, and almost half of these
falls lead to an injury. Most falls occur during walking (Jensen, Lundin-Olsson,
Nyberg, & Gustafson, 2002). One out of twenty falls leads to a fracture (Ti-
netti, Speechley, & Ginter, 1988; Luukinen, Koski, Laippala, & Kivela ¨, 1997).
If a fall occurs during turning it more often results in a fracture (Cummings et
al., 1995). In Sweden (population 9,000,000), about 18,000 persons suffer a
hip fracture every year, and this number will increase as the aging population
increases (Socialstyrelsen, 2003). A hip fracture imposes limitations to the af-
fected individual and is a common reason for moving to a residential care facil-
ity (Sernbo & Johnell, 1993; Tinetti & Williams, 1997). This type of fracture
Sondell et al. 191