Computing Reviews

Decoding the representation of code in the brain:an fMRI study of code review and expertise
Floyd B., Santander T., Weimer W.  ICSE 2017 (Proceedings of the 39th International Conference on Software Engineering, Buenos Aires, Argentina, May 20-28, 2017)175-186,2017.Type:Proceedings
Date Reviewed: 10/20/17

This paper presents an excellent example of an inter-disciplinary study to understand how the human brain processes common software engineering tasks by using medical imaging.

Subjectivity is an inherent property of many software engineering tasks such as code comprehension and review. Despite many studies exploring human aspects in this domain, the present knowledge about human brain reactions during such cognitive tasks and their implications is limited. This study uses functional magnetic resonance imaging (fMRI) to understand source code comprehension and review. It presents a controlled experiment on 29 participants to explore the relationship among tasks and human expertise within brain regions.

Each participant went through a detailed session of MRI setup where three types of stimuli were presented--code comprehension, code review, and prose review. The respondents were asked to assess the presented task and answer it as early as possible. During this procedure, the authors observed and recorded their brain activity in terms of blood-oxygen level dependent (BOLD) response.

The study reveals that, on average, programming and natural language tasks use different parts of the human mind. However, interestingly, as a person gains expertise in programming, the mind starts treating programming languages much more like natural language.

There are many things to learn from this study. Given that this is only the second study that uses medical imaging to understand human brain responses while performing software engineering tasks, the authors have done a good job explaining the intricacies of the experiment in detail. Furthermore, the study explains the applied data collection and processing method in sufficient detail.

Such studies open up a new direction of research. There are many open research questions that are interesting for both researchers and practitioners related to software engineering and corresponding human aspects. I sincerely hope to see more in this direction despite the high cost and involved difficulties.

Reviewer:  Tushar Sharma Review #: CR145600 (1712-0810)

Reproduction in whole or in part without permission is prohibited.   Copyright 2024 ComputingReviews.com™
Terms of Use
| Privacy Policy