Role for Astrocytes in Complex Cognition: An Evolutionary, Behavioral and Computational Perspective

The human brain has more computational power than modern supercomputers by utilizing only a few signaling systems, rendering it the most efficient computational device ever created. Evolution, though not prospective, has iteratively constructed neural networks of increasing complexity that have ultimately endowed humans with extraordinary cognitive abilities. Through an evolutionary lens, there is a clear correlation between organismal sophistication and network signaling complexity Astrocytes, a non-neuronal cell that makes up approximately half the cells in the human brain, show dramatic increases in morphological and signaling complexity in higher-order species. This co-evolution is likely not coincidental but rather represents the development of sophisticated communication hubs that enhance network computation. Notably, system xc- (Sxc), a cystine-glutamate antiporter expressed predominantly on astrocytes, appears almost exclusively in vertebrate species. Additionally, the neuropeptide pituitary adenylate cyclase activating polypeptide (PACAP) signaling system experienced rapid functional enhancement in early mammals. These parallel evolutionary trends in signaling complexity warrant investigation into potential mechanistic connections between these network-enhancing developments. In this dissertation, I employ an interdisciplinary approach integrating molecular evolution, neuroscience, and computational engineering to uncover astrocytic involvement in higher-order cognition. I demonstrate that the neuropeptide PACAP attenuates drug-seeking behavior through pathway- and cell-specific mechanisms, revealing Sxc as a critical regulatory mediator for PACAP-induced recruitment of astrocytes within glutamate signaling networks. To understand the specific role that Sxc plays in controlling drug-seeking, I dissect this complex behavior into its constitutive parts, categorizing them as evolutionarily conserved and phylogenetic recent cognitive functions. These findings suggest that Sxc plays a discrete role in behaviors requisite of complex cognitive signaling and displays a temporal-specific reliance in learning acquisition but not maintenance. Based on these findings, I propose a novel hypothesis for cognitive evolution: coordinated co-evolution of molecular proteins within glutamate signaling networks drove the expansion of cognitive complexity. Collectively, this work introduces a paradigm-shifting approach for understanding network communication with transformative implications for both molecular medicine and artificial intelligence.