Transactions of Japanese Society for Medical and Biological Engineering
Online ISSN : 1881-4379
Print ISSN : 1347-443X
ISSN-L : 1347-443X
Classification of Elderly People with Mild Cognitive Impairment by Quantitative Assessment of Motor Sequence Learning Memory
Author information

2022 Volume 60 Issue 2-3 Pages 68-75


Mild cognitive impairment (MCI) is a precursor stage to dementia. It is necessary to understand the appropriate condition of MCI to prevent the progression of dementia. For MCI diagnosis, there is a study report on memory feature measured by a device, in addition to paper-based assessments. Motor sequence learning memory, a trait of memory feature could be used as an indicator to distinguish MCI. The purpose of this study was to develop a new method to distinguish MCI from an unimpaired condition. The data obtained from proposed method are intended to be applied to a decision tree approach to establish a classification algorithm by clarifying the variables required for classification. Sixty-seven subjects were examined in the study:a group of healthy young people, a group of healthy elderly people, and a group of people with MCI. The motor sequence learning memory was quantitatively evaluated with a visual tracking exercise using hand dexterity movement with the help of a grasping force control training device (iWakka). Sine waves were combined to create a synthetic wave for evaluation tasks. For the evaluation parameters, the difference between the target value and atonal value was used, as well as assignment accuracy and its distributed standard deviation. MCI was differentiated based on these variables. The result showed that the value obtained from iWakka was significantly lower in the elderly people with MCI compared with the other groups. Furthermore, the classification algorithm was created using a decision tree classification. The evaluation variables adopted were mAGF8 and the learning rate. In addition, the decision tree classifier was validated in terms of the area under the curve (0.79), sensitivity (0.66), specificity (0.92). From these results, it can be concluded that the proposed method with iWakka is highly likely to provide useful information when classifying MCI.

Content from these authors
© 2022 Japanese Society for Medical and Biological Engineering
Previous article Next article