Using neurophysiological signals that reflect cognitive or affective state: six recommendations to avoid common...

Area of Science:

• Neuroscience
• Machine Learning
• Psychology
• Human Factors

Background:

• Estimating cognitive or affective states from neurophysiological signals is interdisciplinary, requiring expertise in multiple fields.
• This complexity makes rigorous research and accurate evaluation challenging.
• There is a need for clear guidelines to improve the quality of studies and applications in this area.

Purpose of the Study:

• To identify and summarize common pitfalls in research using neurophysiological signals for state estimation.
• To provide actionable recommendations for researchers and readers to avoid these pitfalls.
• To support the development of high-quality studies and well-validated applications.

Main Methods:

• Review and synthesis of common challenges in the field.
• Identification of critical aspects for successful state estimation.
• Guidance on defining states, relevant neurophysiological processes, and confounding factors.
• Recommendations on classification analyses and the added value of neurophysiological measures.

Main Results:

• Detailed discussion of pitfalls related to state definition, neurophysiological underpinnings, and confounding variables.
• Insights into avoiding inadvertent "cheating" in classification analyses.
• Guidance on understanding the basis of successful state estimation.
• Emphasis on the practical utility and added value of neurophysiological measures in applications.

Conclusions:

• Adherence to the provided recommendations can enhance the quality of research in neurophysiological state estimation.
• This guidance aims to foster the creation of robust and validated applications.
• The paper serves as a valuable resource for the scientific community to improve study design and application development.