Reading comprehension based on visualization of eye tracking and EEG data
- PDF / 1,278,066 Bytes
- 3 Pages / 595 x 842 pts (A4) Page_size
- 49 Downloads / 227 Views
. MOOP .
November 2020, Vol. 63 214101:1–214101:3 https://doi.org/10.1007/s11432-019-1466-7
Reading comprehension based on visualization of eye tracking and EEG data Shiwei CHENG* , Yilin HU, Jing FAN & Qianjing WEI School of Computer Science and Technology, Zhejiang University of Technology, Hangzhou 310023, China Received 25 February 2019/Revised 9 May 2019/Accepted 26 July 2019/Published online 10 October 2020 Citation Cheng S W, Hu Y L, Fan J, et al. Reading comprehension based on visualization of eye tracking and EEG data. Sci China Inf Sci, 2020, 63(11): 214101, https://doi.org/10.1007/s11432-019-1466-7
Using eye tracking technology can help us understand reading behavior. Furthermore, sharing the teacher’s eye tracking features resulted in improving the students’ comprehension of the same reading material [1]. On the other hand, a user’s intention can be analyzed by physiological data, such as electroencephalogram (EEG) [2]. EEG is closely related to human cognition [3]. Recently, researchers have tried to use EEG-based engagement measures to augment learning activities. The BRAVO system constantly analyzes users’ brain activity, and estimates their attention and meditation levels, and presents users with learning material that only results in high engagement [4]. FOCUS is an EEG augmented reading system that monitors a child’s engagement level in real time, and it provides contextual brain computer interaction (BCI) training sessions to improve a child’s reading engagement [5]. This study proposed an approach to serve novice readers, i.e., students, and recorded eye tracking and EEG data of the teacher and then converted the raw data into visualized measures. During the reading process, the students adjusted their reading patterns according to their teachers’ visualization, and improved reading comprehension. Eye tracking related measures. How fast a paragraph is read and how many times the person read the same paragraph are the most discriminative features for measuring comprehension [6]. We defined three measures: reading speed for a single area of interest (AOI), reading time for each AOI,
switching frequency between two AOIs, and denote p as a time threshold in the AOI, and q as a time threshold between two AOIs (the values of p and q are set empirically) [1]. EEG related measures. Reading engagement has been referred to as general intent on reading and writing, the capacity to focus on text meaning and avoid distractions, and the state of immersion in the narrative [7]. We define reading engagement based on EEG measures and denote E as the value of reading engagement when an individual user is reading a specific AOI. It is calculated as follows [5]: β E= , (1) α+θ where α, β, θ represents the amplitude of the alpha wave rhythm, beta wave rhythm, and theta wave rhythm, respectively. E is able to identify changes in engagement related to external stimuli (e.g., AOI). To avoid differences across all individuals, we calculate a normalized value Enorm (0 to 1) as E − Emin Enorm = , (2) Emax − Emin where
Data Loading...
