The Sweet Lab was the first to demonstrate that psychosis within individuals with Alzheimer disease aggregates in families, indicating genetic causality. Subsequent studies have estimated the heritability of psychosis in Alzheimer disease as 61%. These findings, now replicated across multiple cohorts, have served as the foundation for a series of ongoing studies to identify the genetic sources of psychosis risk in Alzheimer disease. Our focus has largely been on studies of common variation in case-control designs. While we have identified a number of intriguing initial candidate genes, definitive genetic findings have yet to emerge. Nevertheless, establishing that psychosis in Alzheimer disease is heritable indicates that it therefore is likely to arise from a distinct underlying neurobiology, and not as just a non-specific consequence of neurodegeneration. We are currently testing the hypothesis that psychosis in Alzheimer disease results from an excess vulnerability of cortical synapses to dysfunction or elimination, either due to an enhanced neurodegenerative drive by pathologic amyloid beta and tau protein, and/or due to an independent, additive vulnerability selective for individuals with psychosis. We have found some support for both processes. For example, we found that the intraneuronal concentrations of hyperphosphorylated tau protein were elevated in psychosis in Alzheimer disease. This observation was then confirmed using phospho-tau specific ELISAs, which further indicated that multiple phosphorylation sites were involved. We have also found reductions in levels of kalirin in prefrontal cortex of psychotic Alzheimer disease patients. Kalirin is a psychosis risk gene and a key mediator of synaptic development and maintenance, whose mechanism of action includes activating post-synaptic phosphorylation signaling cascades.