What is an adaptive pattern of brain activity? It depends on one’s environment

• In Event: Adaptation, not deficit: Contextual and environmental influences of SES on neurodevelopment and self-regulation

Abstract

Prior research indicates that certain patterns of resting state functional connectivity are more adaptive—for example, associated with better cognitive test scores. However, most study samples are skewed towards higher-socioeconomic status individuals—and what is adaptive for one population may not be for another. Here, we probe this assumption for one such association.

One pattern of functional connectivity thought to be adaptive is an anti-correlation between lateral frontoparietal network (LFPN; supports executive functions), and Default Mode Network (DMN; supports internally-directed thought; Figure 1A). Lower LFPN-DMN connectivity has been linked to higher cognitive test performance, leading to a view that it in order to focus on a cognitively demanding task, the LFPN must operate independently of the DMN. However, most studies are based on non-representative samples of individuals from higher-socioeconomic status backgrounds. Children living in poverty are at the greatest risk of low performance on cognitive tests, yet we know little about the neural underpinnings of success for them.

In a pre-registered study, we analyzed resting-state fMRI data from 6839 children ages 9-10 years. For children above the federal poverty line (N=5805), we replicated the prior finding: better cognitive performance was related to weaker connectivity between the lateral frontoparietal and default mode networks, B=-1.41, p=0.002. However, for children living in poverty, this relation trended in the opposite direction, B=2.11; p=0.060, and the interaction was significant, p=0.003 (Figure 1b). Several tests confirmed the reliability of this dissociation, including robust regression to account for potential outliers, bootstrapping to test the extent to which the effect found for the children below poverty would be found in a larger sample of children living above poverty, and repeating analyses restricted to children with the least head motion.

Follow-up cross-validated predictive analyses revealed that the relation between LFPN-DMN connectivity and test performance varied systematically depending on children’s environments. For children living in dangerous neighborhoods, for example, more positive LFPN-DMN connectivity was linked to better test performance; for children living in safe neighborhoods, this relation was in the expected, negative, direction (Figure 2).

The results of our study suggest that “optimal” patterns of brain development depend in part on the external pressures that children face (https://www.biorxiv.org/content/10.1101/2020.05.29.124297v1). These results go against the idea that different patterns of brain activation for children living in poverty necessarily imply a deficit, and instead point to the ways in which children adapt to various constraints. Still, it is unclear whether patterns of brain development that appear adaptive in middle childhood will continue to be adaptive over time. Thus, an important next step will be to follow these children longitudinally to see how LFPN-DMN connectivity and its relation with cognitive test performance changes across adolescence. Altogether, these results highlight the substantial variability of experiences of children living in poverty, who are often conceptualized as a homogenous group and compared to higher-SES children. Moreover, they suggest that our field’s assumptions about generalizability of brain-behavior relations are not necessarily correct, highlighting the need for more diverse representation in developmental cognitive neuroscience.

Authors

• Monica E Ellwood-Lowe, University of California - Berkeley
  Presenting Author

• Susan Whitfield-Gabrieli, Northeastern University
  Non-Presenting Author

• Silvia Bunge, University of California - Berkeley
  Non-Presenting Author