Geographical/geometric indicators like azimuth provide a real-time and low-cost method of measuring spatial performance during navigation, especially in view of their accessibility on mobile phones in the form of a compass or GPS. This study aimed to investigate azimuth-based assessment of spatial orientation performance and its potential in diagnosing cognitive problems. Methods: This was a descriptive survey, and included multivariate logistic regression and multi-layer neural network analysis. We measured the spatial orientation performance of participants using an azimuth-based compass. Their demographic data, including age, gender, years of driving experience, field of study, and cognitive health status were collected. The statistical population consisted of 52 females and 48 males, 18 of whom had experienced cognitive problems in their lives. The participants were from different ethnic backgrounds living in the US, and in the age range of 20-85. The census method was then applied. Multivariate data analysis was conducted to illustrate the effectiveness of each feature variable on spatial orientation performance. Logistic regression was run by fitting a logit function, and multi-layer neural network analysis was developed and evaluated to predict the risk of cognitive problems based on spatial orientation performance and four other features under study. Results: Multivariate analysis showed that participants’ age (P=0.005), years of driving experience (p <0.001), and cognitive problems (p <0.001) contributed to predicting spatial orientation performance. Those who had experienced cognitive problems deviated 13.50 degrees from the destination with increasing age. The accuracy of the fitted logit model by NN analysis over training process was 0.8, indicating that the model can predict 8 out of 10 cases accurately. Conclusion: According to the findings, azimuth-based assessment of performance and other demographic data could be an appropriate means of determining individuals’ spatial orientation ability, especially when performed on larger and more homogeneous groups.
|Journal||Interdisciplinary Journal of Virtual Learning in Medical Sciences|
|Publication status||Published - 2 Nov 2021|