Can EEG Be Adopted as a Neuroscience Reference for Assessing Software Programmers’ Cognitive Load?

Abstract

An emergent research area in software engineering and software reliability is the use of wearable biosensors to monitor the cognitive state of software developers during software development tasks. The goal is to gather physiologic manifestations that can be linked to error-prone scenarios related to programmers’ cognitive states. In this paper we investigate whether electroencephalography (EEG) can be applied to accurately identify programmers’ cognitive load associated with the comprehension of code with different complexity levels. Therefore, a controlled experiment involving 26 programmers was carried. We found that features related to Theta, Alpha, and Beta brain waves have the highest discriminative power, allowing the identification of code lines and demanding higher mental effort. The EEG results reveal evidence of mental effort saturation as code complexity increases. Conversely, the classic software complexity metrics do not accurately represent the mental effort involved in code comprehension. Finally, EEG is proposed as a reference, in particular, the combination of EEG with eye tracking information allows for an accurate identification of code lines that correspond to peaks of cognitive load, providing a reference to help in the future evaluation of the space and time accuracy of programmers’ cognitive state monitored using wearable devices compatible with software development activities.

Quality Evaluation of Modern Code Reviews Through Intelligent Biometric Program Comprehension

“…In other words, the ANS biomarkers could capture the cognitive load induced by the code comprehension and review as well as the EEG biomarkers could do. This result aligns with what Medeiros et al [82] have achieved, showing that EEG can assess programmers' cognitive load accurately. Still, its intrusiveness compared to ANS signals is less prevalent in practice and natural software development environments.…”

“…Although plenty of papers show that HRV and Pupillography can be used to measure cognitive load (e.g., [18][19] [20]) we want to ensure further that they can be applied to intellectually complex and abstract tasks such as code reviews. Medeiros et al [82] showed that EEG biomarkers could be used to finetune or validate the results obtained with ANS signals, which can be acquired using non-intrusive wearable devices.…”

“…A standard preprocessing pipeline was performed to guarantee a reasonable signal-to-noise ratio of the neural signals for the remaining analysis. The preprocessing comprised different steps (see [82] for more details): (i) reduction of MR-induced EEG artifacts (gradient artifact and pulse artifact) due to the nature of the experimental protocol designed; (ii) standard pipeline where filtering, noisy channels interpolation, re-referencing, and blind source separation steps were performed to remove additional typical EEG artifacts (i.e., noise in the signal due to power lines and other equipment).…”

Ensemble Deep Neural Network for Automatic Classification of EEG Independent Components

“…2) BASE Dataset: BASE dataset is composed by 2,340 ICs obtained from EEG signals recorded during code comprehension in a cognitive load assessment study from BASE project [53], [54]. This dataset contains short-term EEG signals, acquired from 30 subjects (24 males and 6 females, with ages ranging from 19 to 42 years), with an average duration of 9.45±3.92 minutes.…”

Software Bug Detection Causes a Shift From Bottom-Up to Top-Down Effective Connectivity Involving the Insula Within the Error-Monitoring Network

“…This region is an important part of the salience network, which not only drives attention toward target stimuli but also relates to decision-making and error-monitoring ( Droutman et al, 2015 ; Lamichhane et al, 2016 ; Bastin et al, 2017 ; Billeke et al, 2020 ). Our previous study suggests that the insula is activated for challenging deep-level analysis of code inspection and demanding error-monitoring ( Castelhano et al, 2019 ; Medeiros et al, 2021 ). Moreover, the use of mathematical- or reading-related skills to source-code understanding has been a matter of recent studies ( Prat et al, 2020 ; Castelhano et al, 2021 ).…”