ePoster

Increase in dimensionality and sparsification of neural activity over development across diverse cortical areas

Lorenzo Butti, Nathaniel Powel, Bettina Hein, Deyue Kong, Jonas Elpelt, Haleigh Mulholland, Gordon Smith, Matthias Kaschube
Bernstein Conference 2024(2024)
Goethe University, Frankfurt, Germany

Conference

Bernstein Conference 2024

Goethe University, Frankfurt, Germany

Resources

Authors & Affiliations

Lorenzo Butti, Nathaniel Powel, Bettina Hein, Deyue Kong, Jonas Elpelt, Haleigh Mulholland, Gordon Smith, Matthias Kaschube

Abstract

Interacting with the environment to process sensory information, generate perceptions, and shape goal-directed behavior engages neural networks in brain areas with highly varied representations, ranging from unimodal sensory cortices to higher-order association areas. These networks form and undergo considerable and potentially area-specific refinement over development. However, recent work suggests a much greater degree of commonality in the functional organization of networks across such diverse cortical areas during early development. It is currently unknown whether this initially common structure undergoes an equally common developmental trajectory across areas and what mechanisms might guide this maturation. Here we examine the development of network organization across diverse cortical regions using in vivo calcium imaging of spontaneous activity in the ferret together with in silico network models. We find that across different cortical regions (including sensory and association areas), networks exhibit a highly similar pattern of changes over development: with activity in local populations becoming increasingly sparse, less correlated, and higher dimensional over an approximately 3 week developmental period spanning eye opening and the transition to predominantly externally-driven sensory activity. Strikingly, we find that both a modular functional organization and millimeter-scale correlated networks—both prominent features of spontaneous activity in all areas during early development—exhibit a similar weakening but remain present in all cortical areas examined. Prior work has shown that local excitation / lateral inhibition (LE/LI) network models can reproduce the modular structure of activity in the developing cortex. Here we show that the changes we observe in vivo can be accounted for by a developmental strengthening of inhibition within the network. Thus, these highly conserved changes in network organization across diverse cortical areas indicate a common trajectory of network refinement and suggest a common mechanism potentially supporting an expansion of representational capacity throughout the developing cortex.

Unique ID: bernstein-24/increase-dimensionality-sparsification-af924ef8