AUTHOR=Yamada Shigeki , Aoyagi Yukihiko , Ishikawa Masatsune , Yamaguchi Makoto , Yamamoto Kazuo , Nozaki Kazuhiko
TITLE=Gait Assessment Using Three-Dimensional Acceleration of the Trunk in Idiopathic Normal Pressure Hydrocephalus
JOURNAL=Frontiers in Aging Neuroscience
VOLUME=13
YEAR=2021
URL=https://www.frontiersin.org/journals/aging-neuroscience/articles/10.3389/fnagi.2021.653964
DOI=10.3389/fnagi.2021.653964
ISSN=1663-4365
ABSTRACT=
Background: The subjective evaluation of pathological gait exhibits a low inter-rater reliability. Therefore, we developed a three-dimensional acceleration of the trunk during walking to assess the pathological gait quantitatively.
Methods: We evaluated 97 patients who underwent the cerebrospinal tap test and were diagnosed with idiopathic normal pressure hydrocephalus (iNPH) and 68 healthy elderlies. The gait features of all patients were evaluated and classified as one of the following: freezing of gait, wide-based gait, short-stepped gait, shuffling gait, instability, gait festination, difficulty in changing direction, and balance disorder in standing up. All gait features of 68 healthy elderlies were treated as normal. Trunk acceleration was recorded automatically by a smartphone placed on the umbilicus during a 15-foot walking test. Two novel indices were created. The first index was a trunk acceleration index, which was defined as (forward acceleration fluctuation) + (vertical acceleration fluctuation) – (lateral acceleration fluctuation) based on the multivariate logistics regression model, and the second index was created by multiplying the forward acceleration with the vertical acceleration. Additionally, 95% confidence ellipsoid volume of the three-dimensional accelerations was assessed.
Results: Forward and vertical acceleration fluctuations were significantly associated with the probability of an iNPH-specific pathological gait. The trunk acceleration index demonstrated the strongest association with the probability of an iNPH-specific pathological gait. The areas under the receiver-operating characteristic curves for detecting 100% probability of an iNPH-specific pathological gait were 86.9% for forward acceleration fluctuation, 88.0% for vertical acceleration fluctuation, 82.8% for lateral acceleration fluctuation, 89.0% for trunk acceleration index, 88.8% for forward × vertical acceleration fluctuation, and 87.8% for 95% confidence ellipsoid volume of the three-dimensional accelerations.
Conclusions: The probability of a pathological gait specific to iNPH is high at the trunk acceleration fluctuation, reduced in the forward and vertical directions, and increased in the lateral direction.