Alzheimer’s disease (AD) is the leading form of dementia and one of the leading causes of death in the United States. Key hallmarks include cognitive decline, neuronal network dysfunction, amyloid-beta (Aβ) deposits, and aberrant microglia activation1. Neuronal dysfunction in AD manifests as interneuron dysfunction, network hyperexcitability, and deficits in neuronal oscillations. While interneurons have been implicated as key drivers of neuronal network dysfunction, the contribution of microglia to these deficits remains poorly understood. Microglia are critical regulators of neuronal function, capable of sensing and directly influencing neuronal activity2-4. Evidence that microglial-mediated inflammatory responses to Aβ contribute to cognitive decline5 implicates microglia over-activation as a potential driver of disease progression.In our first study, we investigated the hypothesis that neuroinflammation serves as a key driver of AD-related neuronal dysfunction. Using a model of microglia nuclear factor kappa B (NF-κB) hyperactivation (IKKβ-CA), we identified robust transcriptomic changes affecting both microglia and neurons. Through simultaneous local field potential (LFP) and single-unit recordings in the posterior parietal cortex (PPC), we uncovered deficits in pyramidal cell firing that disrupt spike-phase coupling and reduce theta and gamma oscillation power, contributing to spatial learning and memory deficits. By crossing these mice with newer, humanized App knock-in (KI) mice (AppNLGF), we discovered that microglia NF-κB signaling and Aβ pathology have distinct yet convergent effects on neuronal network synchronization.
Our second study addresses the critical need for more behavioral paradigms and modeling approaches that assess goal-directed navigation in ethologically-relevant environments. Studying these behaviors can provide valuable insights into the behavioral mechanisms underlying cognitive disorders, such as AD. We thus developed a novel labyrinth maze that mirrors naturalistic foraging, whereby mice can freely navigate from their home cage through a series of interconnected decision points along an optimal path to reach rewards. This design specifically engages cognitive mapping, planning, and decision-making processes essential for goal-directed behavior. To quantify these complex strategies, we created a hierarchical probabilistic modeling framework—Cognitive Mapping of Planning Actions with State Spaces (CoMPASS)—that decodes both locomotor states (level 1) and higher-order goal-oriented states (level 2) from movement patterns. While wild-type mice rapidly learned the goal-directed nature of the task, humanized App-KI mice (AppSAA) exhibited overt navigational deficits to the target and search efficiency. CoMPASS analysis revealed that AppSAA mice exhibited specific deficits in active surveillance state (level 1) at decision nodes and reduced engagement in goal-oriented behaviors (level 2). Wireless EEG recordings in the PPC of wild-type mice validated our framework by demonstrating that successful goal-directed navigation is characterized by increased gamma synchronization, particularly during surveillance at decision nodes. This behavioral and computational platform provides a sensitive approach for detecting subtle alterations in goal-directed navigation in mouse models of AD.
Together, these complementary approaches provide novel insights into how microglial activation disrupts neuronal network function and offer sensitive methodologies for assessing complex cognitive processes in AD models, advancing our understanding of the fundamental mechanisms leading to cognitive decline in Alzheimer’s disease.
