Components of gait in people with and without mild cognitive impairment

Background: Several objective gait parameters are associated with cognitive impairment, but there is limited
knowledge of gait models in people with mild cognitive impairment (MCI).
Research question: How can 18 objective gait characteristics be used to define different components of gait in
people with MCI (with suspected incipient neurocognitive disorder) and cognitively unimpaired people (CU),
respectively?
Methods: Spatiotemporal gait data were collected by using an electronic walkway (GAITRite®), i.e. assessments
in comfortable gait speed. Using cross-sectional gait data, two principal component analyses (PCA) were performed
(varimax rotation) to define different components of gait in people with MCI (n = 114) and CU (n = 219),
respectively, from the BioFINDER-2 study.
Results: Both PCAs produced four components, here called Variability, Pace/Stability, Rhythm and Asymmetry.
Total variance explained was 81.0% (MCI) versus 80.3% (CU). The Variability component explained the largest
amount of variance (about 25%) in both groups. The highest loading gait parameter was the same for both
groups in three out of four components, i.e. step velocity variability (Variability), mean step length (Pace/Stability)
and mean step time (Rhythm). In the asymmetry component, stance time asymmetry (MCI) and swing
time asymmetry (CU) loaded the highest.
Significance: The gait components seem similar in people with and without MCI, although there were some
differences. This study may aid the identification of gait variables that represent different components of gait.
Gait parameters such as step velocity variability, mean step length, mean step time as well as swing and stance
time asymmetry could serve as interesting core variables of different gait components in future research in people
with MCI (with suspected incipient neurocognitive disorder) and CU. However, the selection of gait variables
depends on the purpose. It needs to be noted that assessment of variability measures requires more advanced
technology than is usually used in the clinic.