Author(s): GUO, L.Y., LIN, J.1, MAI, J.2, HSU, M.1, YAN, C.2, Institution: KAOHSIUNG MEDICAL UNIVERSITY, Country: TAIWAN, Abstract-ID: 1515

INTRODUCTION:
Cervical discomfort could lead to the transmission of abnormal signals, disrupting neck proprioception and potentially triggering dizziness. In previous studies, the smooth pursuit neck torsion test [1] had been used for the assessment of oculomotor control mechanism. Eye tracking systems were often associated with high costs [2]. This study aimed to assess the accuracy and reliability of the low-cost Tobii 5C eye tracker in gaze tasks.
METHODS:
The study involved five male participants with an average age of 23 ± 1 years old. None of the participants wore glasses. Subjects were instructed to look at a red dot on the screen at a distance of 68 cm with their head stationary. After calibration, a red dot appeared at five positions: upper left (UL), upper right (UR), neutral (N), lower left (LL), and lower right (LR) on the screen, each displayed for five seconds. Participants were instructed to gaze at these targets in 20 repeated sessions. The definition of accuracy is the visual angle between the target and gaze point. The data processing removed blinks and saccades using I-VT filtering [4]. The accuracy and intraclass correlation coefficient (ICC) of the Tobii 5C were assessed. The learning effect analyzed by comparing the accuracy of the first 10 trials to the next 10 trials. All statistical analyses were performed using SPSS.20. Significance level was set at p<0.05.
RESULTS:
The results showed average accuracy between 0.66° and 1.29°, consistent with previous research. Notably, performance at the central position was found to be the most stable. Our study found no significant differences between participants in their accuracy following horizontal (H) and vertical (V) movements for each gaze point location: UL(H: 1.33° ± 0.21°, V: 1.30° ± 0.20°); UR(H: 1.38° ± 0.16°, V: 1.28° ± 0.22°); N( H: 1.12° ± 0.29°, V: 1.29° ± 0.20°); LL( H: 1.19° ± 0.24°, V: 1.23° ± 0.24°); LR(H:1.37° ± 0.21°, V: 1.37° ± 0.20°). The ICC values, ranging from 0.20 to 0.75, indicated a moderate effect. Higher ICC values, particularly for positions N and LR, suggest increased stability when the gaze position is closer to the eye tracker. With 20 trials per participant to observe ocular learning effects, the difference was not statistically significant, but there was a trend toward improved accuracy with increasing number of trials.
CONCLUSION:
The study assessed the accuracy of the Tobii 5C eye tracker and found it to be acceptable. Slightly oculomotor learning effects were observed over 10 trials. This study also confirms the reliability of the device for gaze tracking. However, further research is necessary to determine its ability to accurately capture pursuit movements.
REFERENCES:
1. Majcen Rosker Z. et al., Diagnostics (Basel) 2021
2. Ban S. et al., Biosensors (Basel) 2022