Multivoxel Pattern Analysis Does Not Provide Evidence to Support the Existence of Basic Emotions

Area of Science:

• Neuroscience
• Cognitive Science
• Psychology

Background:

• Saarimaki et al. (2015) claimed to identify neural "fingerprints" for basic emotions (anger, fear, disgust, happiness, sadness, surprise) using multivariate pattern analysis.
• This interpretation has been questioned due to statistical and conceptual mischaracterizations of the findings.
• Essentialism, the belief in fixed category essences, may hinder understanding of the neural basis of emotion.

Purpose of the Study:

• To critically evaluate the interpretation of Saarimaki et al.'s (2015) findings regarding neural correlates of basic emotions.
• To argue that the identified patterns represent statistical summaries, not essential neural fingerprints.
• To propose that the results align better with a constructed emotion theory than with basic emotion theory.

Main Methods:

• Re-analysis and conceptual critique of multivariate pattern analysis findings from Saarimaki et al. (2015).
• Application of statistical principles to differentiate between abstract summaries and essentialist "fingerprints".
• Comparison of findings with criteria for basic emotion theory versus constructed emotion theory.

Main Results:

• Statistical patterns identified are abstract summaries of variable instances, not unique neural essences or "fingerprints".
• The findings do not meet the criteria for basic emotion theory.
• The results provide empirical support for a constructed emotion theory, emphasizing population variability.

Conclusions:

• Saarimaki et al.'s findings, while methodologically elegant, are better explained by a constructed emotion framework.
• The study underscores the limitations of essentialist thinking in understanding the neural basis of emotion.
• Embracing population thinking is crucial for advancing emotion research and understanding neural variability.