Despite decades of genetic studies on late onset Alzheimer disease (LOAD), the molecular mechanisms of Alzheimer disease (AD) remain unclear. Furthermore, different cell types in the central nervous system (CNS) play distinct roles in the onset and progression of AD pathology. To better comprehend the complex etiology of AD, we used an integrative approach to build robust predictive (causal) network models which were cross-validated over multiple large human multi-omics datasets in AD. We employed a published method to delineate bulk-tissue gene expression into single cell-type gene expression and integrated clinical and pathologic traits of AD, single nucleotide variation, and deconvoluted gene expression for the construction of predictive network models for each cell type in AD. With these predictive causal models, we are able to identify and prioritize robust key drivers of the AD-associated network state. In this study, we focused on neuron-specific network models and prioritized 19 predicted key drivers modulating AD pathology. These targets were validated via shRNA knockdown in human induced pluripotent stem cell (iPSC) derived neurons (iNs), in which 10 out of the 19 neuron-related targets (JMJD6, NSF, NUDT2, YWHAZ, RBM4, DCAF12, NDRG4, STXBP1, ATP1B1, and FIBP) significantly modulated levels of amyloid-beta and/or phosphorylated tau peptides in the postmitotic iNs. Most notably, knockdown of JMJD6 significantly altered the neurotoxic ratios of Aβ to 40 and p231-tau to total tau, indicating its potential therapeutic relevance to both amyloid and tau pathology in AD. Molecular validation by RNA sequencing (RNAseq) in iNs further confirmed the network structure, showing significant enrichment in differentially expressed genes after knockdown of the validated targets. Interestingly, our network model predicts that these 10 key drivers are upstream regulators of REST and VGF, two recently identified key regulators of AD pathogenesis.