Neural Representation of Self: Independent or Interdependent?

Post by Lani Cupo

What's the science?

One path to better understanding how humans incorporate complex social information in relation to themselves is through the conceptualization of self. The degree to which the self is defined as independent or interdependent with others is known as self-construal, and can differ greatly across environments and cultural backgrounds. Previous research with anatomical and functional magnetic resonance imaging (fMRI) implicates a wide range of brain regions in different representations of self, however it is an open question how connectivity between different brain regions contributes to self-construal. This week in NeuroImage, Shi and colleagues used machine learning to conduct a hypothesis-free, whole-brain analysis of neural connectivity patterns to investigate the representation of self-construal in the connectivity between different brain regions.

How did they do it?

Participants comprised 307 students from Tsinghua University in Beijing - of relevance as the cultural upbringing of individuals drastically impacts self-construal. First, participants completed a questionnaire to assess their self-construal, with higher scores indicating orientation towards independence and lower scores indicating orientation towards interdependence. Then, they underwent fMRI scans for about 8.5 minutes in order to capture the activity of various brain regions. After parcellating the brain into regions based on a predefined brain atlas, the authors correlated fluctuations in the fMRI signal from across all regions to construct functional connectivity matrices describing how similar patterns of activity were between regions. In order to measure whole-brain connectivity, they used five matrices: left-brain (connectivity within the left hemisphere), right-brain (connectivity within the right hemisphere), global (left and right brain added together), interactive (connectivity between the left and right hemisphere), and asymmetric (the absolute value of right brain subtracted from left, where higher values indicate more asymmetry). The authors then trained machine learning algorithms to use data from the five connectivity matrices to predict participants’ scores on the questionnaire, determining whether they exhibited more independent or interdependent self-construal. This allowed the authors to compare which matrix most helped the model differentiate participants. Then, they verified their results using a different questionnaire in the same population. Finally, they identified the 200 connections that most contributed to model performance and checked whether these connections corresponded to known brain networks involved in cognitive functioning, including the default mode network (DMN), executive control network (ECN), and salience networks (SN).

What did they find?

Of all the matrices, the asymmetric one best predicted independent and interdependent self-construal among the participants. These results contribute to a hypothesis that there may be hemisphere specialization underlying self-referential thought. Mapping self-construal orientation onto networks revealed that independence correlated with increased connectivity between brain regions in the DMN and ECN, brain networks involved in both self-referential thought and goal-directed activity, respectively. The authors posit that these results may substantiate a theoretical idea that independent self-construal reflects more personal, affective aspects of self. Meanwhile, there was an different pattern in interdependence, with DMN-ECN connectivity showing a left dominance and DMN-SN connectivity showing a right dominance. The SN is a network involved in integrating cognitive, sensory and emotional information. While the interpretation of these results is still unclear, the authors note one hypothesis is that those with interdependent self-construals may be more sensitive to socially-relevant stimuli around them.

What's the impact?

This study is the first to quantitatively examine patterns of whole-brain functional connectivity underlying self-construal, linking the patterns to previously established neural networks. The authors’ findings suggest that the representation of the self as independent or interdependent is represented in functional connectivity, especially in the relative involvement of the two hemispheres. Overall the results of their study could present a way of understanding how social cultural contexts affect the neural activity underlying the conception of self.

Shi et al. The divided brain: Functional brain asymmetry underlying self-construal. NeuroImage (2021).Access the original scientific publication here.