Influences of Stochasticity on Macroscopic Brain Development: New Insights from Generative Modelling

By Dr. Sofia Carozza

Traditionally, the puzzle of brain development has been framed as a balance between two factors: genetic information and environmental inputs. Together, these are thought to shape the emergence of brain networks from infancy to adulthood. But another significant (and often overlooked) piece is stochasticity—essentially, the role of probabilistic processes in shaping neural outcomes.

The human brain doesn’t develop in a purely deterministic way. There is an inherent unpredictability woven into the developmental process at every level, from molecular events like gene expression to cellular mechanisms like axonal outgrowth. Even with similar genetic and environmental conditions, stochastic variations in these processes can produce meaningful differences in brain organization between individuals.

While this variability might seem like just noise or a liability, it may actually have an adaptive role, especially for children growing up in unpredictable or resource-scarce environments. In such contexts, heightened stochastic contributions to development could make the brain more flexible, more able to respond to the challenges that arise.

Our work set out to examine these questions: what contributions does stochasticity make to the organization of macroscopic brain networks? Could it confer resilience, particularly in adverse conditions?

To obtain new insights on these questions, we turned to generative network modeling, a computational method of simulating the emergence of whole-brain networks. Starting with a scaffold derived from neonatal data from the Developing Human Connectome Project, we simulated over ten million brain networks. The simulations mimicked how the brain’s white matter tracts might form under varying degrees of constraint—ranging from highly deterministic to more stochastic scenarios. This revealed how different levels of randomness affect the brain’s overall organization and robustness to perturbation.

We found that stochasticity contributes significantly to variability in brain network organization. Under conditions where developmental constraints were weaker, the brain networks showed more variation (or multifinality) in their final structure and displayed increased resilience to disruptions. Thus, while more stochastic development can lead to atypical outcomes, it may also support the brain’s adaptability to challenging environments.

We then tested these computational insights empirically, by leveraging data from neurodiverse children participating in a study at the Centre for Attention, Learning, and Memory (CALM). Here, we found that children from lower socioeconomic backgrounds showed brain networks that were better approximated by models incorporating more stochasticity.

We therefore put forward the adaptive stochasticity hypothesis: the proposal that increased stochasticity might be an advantageous feature of the developing brain when a child is navigating unpredictable or resource-scarce environments.

The hypothesis generates several testable predictions that could guide future research. Among others, these include: (1) that children in unpredictable environments will show greater variability in their brain network organization, (2) that their brains will show topological” or organizational signatures of heightened stochasticity, like lower segregation, and (3) that lower constraints on brain network organization will correlate with a higher incidence of both neurodevelopmental conditions and certain cognitive skills that may be useful in high-stress situations.

Advancing this framework could transform how we think about brain development, particularly amid adversity. Variability in development has often been viewed as solely an individual risk factor, but our findings suggest a more nuanced perspective—one that considers its potential adaptiveness in unpredictable contexts. Future longitudinal studies could shed light on how this adaptation unfolds, helping us better support children’s inherent capacities for resilience and growth.