Exploring the cognition during program comprehension based on EEG activities

• Main Author: 廖宜志
• Other Authors: 林育慈
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/15700781570560949399

碩士 === 國立臺灣師範大學 === 資訊教育研究所 === 102 === Computer programming is a critical skill in computer science, which is a complex cognitive activity involving programming knowledge, retrieval, logical thinking, and problem solving. However, it is challenging for most novice programmers. Much existing research have been trying to find the reasons why learners have difficulties in learning programming, and most research studied learners’ programming by observing their behavior by the methods of interview, thinking aloud, or the questionnaire study. But these methods lack objective evidences about participants’ inner cognition. Instead of observing learners’ outer behavior, this research investigated learners’ inner cognition during program comprehension by the EEG techniques. The differences of EEG oscillations between low and high-proformance students were discussed to find how they thought in different ways while tracing and comprehending the programs. In the experiment, thirty-three nudergrauate students were asked to wear the EEG system for recording EEG activities while reading two programs. This research also utilized the eye-tracker to record the participants’ eye movements to facilitate understanding the EEG activities. Finally, interview and questionnaire were used to figure out the situation of the participants’ comprehension was found from the experiment results that high-proformance students showed significantly higher mean power of theta, lower alpha, upper alpha, and lower beta. Theta band activity is related to the encoding and retrieval of episodic memory, and the relational decoding of different types of information in working memory. The high-performance students had higher theta power, which might imply that they involved more in the transformation of current information and in the encoding and retrieving information (e.g., the variable values) in episodic memory. The lower alpha power (8-9Hz) is related to attentional processes. The high-performance students had higher lower alpha power, which indicates that they had better memory and attention performance. The high-performance students with higher upper alpha power means they were faster in retrieving related programming knowledge while comprehending the program. Besides, the high-performance students showed higher beta power while reading key statements of the program. Beta power (13-30Hz) is related to top-down attentional mechanism. It implies that the high-performance students could focus on the Beacons (key statements of the program) and trace the program in more iv logical and strategical manner. This research also found the low-performance students had worse performance on tracing programs with more complex construct, such as combination of ‘if statement’ and ‘for loop’, ‘nested for loop’, and ‘large structure of the recursive function’, due to their limited working memory capacity and attention resources. Therefore, this research suggests that it is necessary to provide visualization tools to help students grasp the program logic and process complex information to assist comprehending programs while learning.