The immune system is composed of hundreds of highly-specialized cell types that collaboratively orchestrate an efficient response to pathogens and damage. Central to immune function, B lymphocytes participate in both the fast but non-specific innate, and persistent adaptive immune responses by sensing conserved pathogen-associated molecular patterns such as bacterial or viral CpG DNA as well as pathogen-specific patterns recognized by uniquely generated B-cell receptors. Upon activation, B cells undergo rapid expansion in number, deal with the threat by carrying out specific effector functions, and eventually die by programmed cell death or become long-lived memory cells. As a result, B-cell dynamics dictate vaccine efficiency, while aberrant proliferation and/or survival is the hallmark of autoimmune disorders, immune deficiency, and cancer. Decades of Nobel-worthy studies have characterized the key molecular players, cellular behaviors, and population dynamics of B cells, but the implicit heterogeneity and multi-scale nature of the B-cell response pose fundamental challenges to meaningful interpretation in specific contexts. A multi-scale understanding has only recently become possible with the advent of single-cell assays and the advancement of computational methods. To better-understand how individual cells orchestrate the population response, we developed CFSE flow cytometry deconvolution of cell populations, time-lapse cell tracking, and agent-based multi-scale computational modeling methods which we combined with single-cell and traditional biochemical assays and literature mining to develop a mechanistic understanding of the B cell immune response from the molecular pathways governing NFκB signaling, growth, cell-cycling, and apoptosis to cellular behavior and ultimately the population dynamics. We find that 1)the population behavior is best explained by individual B cells making decisions to either grow and divide, or die 2)that NFκB signaling serves as a central enforcer of B cell decision making by promoting division and survival and 3)that a multi-scale model can accurately predict population behavior with a lower dose of the stimulus, when NFκB cRel missing, and when pretreated with the drug rapamycin. The methods and models developed as part of this dissertation serve as predictive frameworks for future hypothesis-driven discovery and model-driven analysis, enabling meaningful interpretation of patient data, and drug target prediction across biological scales.