Keywords

1 Introduction

In recent years, the number of patients with dementia has been increasing worldwide. According to the report by the International Alzheimer’s Association, the number of patients with dementia will augment approximately from 50 million in 2018 to 152 million in 2050 [1]. In Japan, the number of people with dementia is expected to increase from approximately 4.62 million in 2012 to 6.75 million in 2025 [2]. There are many types of dementia, and the most widespread ones are cerebrovascular dementia and Alzheimer’s dementia. At present, however, the fundamental treatment for this disease has not been found yet, and only slows the progression of dementia. However, it has been reported that 14%–44% of patients with mild dementia, also referred to as mild cognitive impairment (MCI), which is a precursor of dementia, return to normal state owing to early detection and appropriate rehabilitation [3]. On average, the rate of progression from MCI to dementia is 10% per year [4]. It has also been reported that approximately 50% of patients with MCI will progress to dementia within five years if the condition remains untreated [5]. Therefore, early detection of MCI is an important issue. At present, typical screening tests for dementia include the mental statement (MMSE) and the revised version of the Hasegawa simple intelligence scale (HDS-R). MMSE consists of 11 items, including orientation to time, orientation to place, calculation, and repetition of sentences; its duration is 10 min. HDS-R includes nine items, including such factors as age, orientation to time, orientation to place, calculation, and fluency of language; its duration is five minutes. However, the problem is that the sensitivity of MMSE depends on the severity of dementia. Specifically, the rate is about 100% in the moderate dementia stage, approximately 50% in the mild dementia stage, and about 30% in the MCI stage. Moreover, there is a difficulty that the specificity of HDS-R is lower than the sensitivity. Therefore, a new screening method with a high discrimination accuracy is required. The results of the previous related studies have demonstrated the effectiveness of screening MCI patients and healthy elderly persons by using instrumental activities of daily living (IADL) based on the frequency of micro-errors (ME), which correspond to stagnation of the behavior during reaching movements. Previous studies have been focused on investigating the occurrence conditions of ME from the viewpoint of environmental factors. Moreover, the effects of visual conditions, such as color and shape of a reaching target, on the causes of ME occurrence have been clarified. Specifically, it has been reported that ME occurs when the stimulus of a reaching subject is close to the color and shape of a conspiracy stimulus displayed simultaneously. This observation suggests that the gazing behavior contributes to ME occurrence immediately before starting reaching movements.

The purpose of the present study is to investigate the frequency of ME when only the viewing angle is changed in an environment in which visual attention is controlled aiming to measure the effect of gaze behavior. The following research questions have been considered: (1) whether the frequency of ME increases with an increase in the viewing angle, (2) whether the frequency of ME varies depending on the direction of stimulation, and (3) whether ME can be avoided in the environment in which visual attention is controlled.

2 Previous Study

2.1 Virtual Kitchen Challenge (VKC) System

In the previous research, the VKC system was developed to measure the performance of VR-IADL by reproducing the IADL task environment on a tablet device using the virtual reality (VR) technology [6]. The “Lunchbox Task” presented in VKC implied composing a lunch with sandwiches and cookies made by spreading jam and peanut butter on slices of bread, a bottle with juice and executing the task of putting these three items in a lunch box. These actions could be operated by touching or dragging an object on the tablet screen. Previous studies demonstrated the presence of a correlation between VKC and IADL tasks in the real space with the number of ME occurrences [7]. Figure 1 represents the Lunchbox Task in VKC.

Fig. 1.
figure 1

Setup the lunchbox task

Full size image

2.2 Relationship Between the Visual Field and Fixation

The eye-gaze behavior is defined as the movement of the eye to obtain visual information. In ergonomics, the relationship between the visual field range and discrimination ability of the eye-gaze point is considered, as shown in Fig. 2. The range of the central vision is 1°; the recognition limit for one or several words is from 5° to 10°; and the recognition limit for symbols from is 5° to 30°. These findings suggest that there may be a difference in the number of ME occurrences in the case when the objects to be reached and interfering stimuli are presented at each visual field angle.

Fig. 2.
figure 2

Relationship between the visual field and fixation

Full size image

2.3 Shape Task

It is considered that VKC can be influenced by color, shape, meaning, function, and planning. Therefore, in the previous study, Shape Task was formulated as a task to measure the influence of the shape similarity on ME [8]. In this task, multiple interfering stimuli and the target stimulus were simultaneously displayed on the screen, and the target stimulus was supposed to be touched. Specifically, an interfering stimulus (distractor object: DO) with a high correlation coefficient was randomly placed around the target object (target object: TO) having the similar form to investigate the frequency of ME. As a result, it was concluded that ME was generated by an interfering stimulus with the form similar to that of TO [8]. The task screen is represented in Fig. 3.

Fig. 3.
figure 3

Shape task

Full size image

3 Method

3.1 Eye Task

In the task environment established for measuring ME in the previous research, TO and DO were displayed over multiple visual fields, and ME was evaluated in the visual fields with a different discrimination ability. In the present study, we formulated a task to measure the number of ME occurrences in the case when the viewing angle and stimulus position were changed in an environment with the controlled visual attention. The subject was suggested to seat down with a gaze meter (EMR -9) attached to the chin rest 35 cm away from the center of the monitor with the gaze point displayed on it. In addition to the gaze point, a fork, which was TO, and a knife, which corresponded to DO, were displayed on the monitor. Each stimulus was displayed by using the vertical and horizontal patterns at 5°, 10°, 20°, and 30° around the gaze point. The subject was requested to touch the TO as quickly as possible when the monitor was displayed while maintaining the gaze point as gaze as possible.

In the present study, we measured two factors such as the viewing angle (5°, 10°, 20°, 30°) and direction (top, bottom, right, left) as independent variables; and the other ones such as the number of ME occurrences, the three-dimensional position of the sliders by using a motion sensor, and the two-dimensional position of the gaze line obtained by using a gaze meter as dependent variables. In this experiment, 11 male university students (Mean = 21.55, Standard Deviation = 1.13) were asked to listen to white noise for one minute after each experimental patterns to measure fatigue of the examinees. The experiment was conducted considering a counter balance to offset the sequential effect. The established measurement environment is demonstrated in Fig. 4, and the stimulation pattern displayed on the Eye Task monitor is represented in Fig. 5.

Fig. 4.
figure 4

Measurement environment

Full size image
Fig. 5.
figure 5

Stimulation patterns

Full size image

4 Results

4.1 Correlation Coefficient Between the Number of ME Occurrences and the Viewing Angle

The correlation coefficient r between the number of ME occurrences and the viewing angle was calculated to be r = 0.7182, indicating a positive correlation as a whole. Therefore, it was confirmed that the number of ME occurrences increased as the viewing angle became wider.

4.2 Number of Occurrences of ME

During the experiment, execution of the tasks by the participants was recorded as moving images, which were then analyzed to determine the number of ME occurrences for each task pattern. Table 1 outlines the results of the analysis on variance on the number of ME occurrences, which is a dependent variable. Table 2 provides the results of the multiple comparison test, and Fig. 6 represents a plot of the average number of ME occurrences relatively to the direction.

Table 1. Results of the analysis on variance
Full size table
Table 2. Results of the multiple comparison test
Full size table
Fig. 6.
figure 6

Plot of the mean number of ME

Full size image

4.3 Eye Movement During ME

Considering that the previous studies have demonstrated that ME can be caused by attachment to DO [8], we assume that ME does not occur in an environment in which visual attention is controlled. However, based on the results of the conducted experiment in the environment with controlled visual attention, we found that ME occurred more frequently when the subject’s gaze was not moving than when they were moving. Here, the threshold to identify for whether the subject’s gaze moved was set as 5 cm or more and less than 10 cm. Figure 7 represents the plot depicting the change in the distance from the gaze point with respect to time under the 30° condition in the case of subject 1.

Fig. 7.
figure 7

Change in the distance from the eye-gaze point under the 30° condition

Full size image

In the present study, we conducted the experiments in the environment in which visual attention was controlled, and therefore, it was important to determine whether the line of sight was moving. To determine this, we calculated the distance from the point of gaze. The calculation method for the distance is outlined below: (1) The position coordinates of the parallax-corrected line of sight are obtained by using the line of sight measuring instrument. (2) The distance d from the fixation point to the two-dimensional coordinate of the line of sight is calculated using the pythagorean theorem, as shown in Fig. 8.

Fig. 8.
figure 8

2D coordinates of the line of sight

Full size image
$$ d\, = \,\sqrt {d.x^{2} + d.y^{2} } $$
(1)

(1) Table 3 provides the results of the analysis on variance with regard to the number of eye movements when ME occurs, and Table 4 outlines the results of multiple comparison tests.

Table 3. Results of the analysis on variance
Full size table
Table 4. Results of the multiple comparison tests
Full size table

4.4 ME and Viewing Angle Considerations

The results of the multiple comparison test provided in Table 2 outline no significant difference between 20° and 30°, suggesting that there has been no variance in the number of ME occurrences in the visual field beyond 20°, as it falls within the range of peripheral vision, and that no significant difference between 20° and 30° has been observed. Therefore, it is considered that the overall correlation coefficient was affected.

4.5 Study on the Direction Based on the Fixation Point and the Number of ME Occurrences

We hypothesized that there was no lateral difference in the number of ME occurrences but there was vertical difference based on human visual field characteristics. However, the analysis on variance presented in Table 1 demonstrated no significant difference in the direction of ME occurrences in the case when the fixation point was used as a reference. Investigations on the attention and accuracy have been conducted in the previous studies, and it has been clarified that a difference in the accuracy rate when the number of stimuli that responded vertically to the fixation point increases [9]. These result suggest that there was no difference in direction in this study due to the small number of stimuli.

4.6 Study on ME and Eye Movement

The results of the analysis on variance presented in Table 3 indicated a significant difference in the viewing angle; however, the results of the multiple comparison test presented in Table 4 demonstrated no difference between the levels. These results indicate that with an increase on the viewing angle, the number of eye movements increases as well.

5 Conclusion・Future Prospects

In the present study, we examined the following three research questions: (1) whether the number of ME events increases with an increase in the viewing angle; (2) whether the number of ME events varies depending on the location of a stimulus; and (3) whether ME can be avoided in an environment in which visual attention is controlled. Although no significant difference was observed in terms of the direction, a considerable variation was noted in the visual field angle, and a positive correlation was indicated between the number of occurrences of ME and the visual field angle. Based on these results, the relationship between the gazing behavior and ME was clarified. In the present study, the attention was fully focused on the gazing point; however, as a future prospect, we plan to conduct a comparative study for the case of moving the eyes, the similar experiment as presented in this study considering MCI patients, and an experiment with an increasing number of stimuli.