Transcribed CGG repeat expansions in the 5’ UTR of FMR1 cause Fragile X-Associated Tremor/Ataxia Syndrome (FXTAS). CGG repeat RNA interacts with RNA binding proteins to impair their functions. FMR1 CGG repeats trigger repeat-associated non-AUG initiated (“RAN”) translation of toxic proteins - including the polyglycine protein FMRpolyG - that trigger neuronal intranuclear inclusion formation. FXTAS patients exhibit widespread neuroinflammation and white matter abnormalities. However, the mechanism by which CGG repeats elicit these downstream pathologies remains unclear. To explore this problem, we used (CGG)341 repeat KI mice, CGG transgenic mice, and AAV mediated expression of CGG repeats in different sequence contexts along with patient derived iNeuronal models to assess the relative contributions of FMRpolyG and CGG repeat RNA to neuroinflammation, neuropathology and neurodegeneration. Expression of RAN translation competent CGG repeats via both AAV mediated delivery and in transgenic mice elicit widespread neuronal intranuclear inclusions that stain positive for FMRpolyG and P62. These inclusions correlate with widespread gliosis, microglial activation, behavioral phenotypes, and Purkinje cell loss. In contrast, neither AAV mediated expression of CGG repeats with reduced FMRpolyG production or FMRpolyG expression from a non-repetitive sequence elicited significant behavioral phenotypes or gliosis, despite inclusion formation still occurring in the “FMRpolyG-only” mice. CGG repeat associated behavioral phenotypes, neuroinflammation and inclusion formation in both AAV and CGG repeat transgenic mice were markedly reduced in PKR -/- mice, whose loss reduces both repeat triggered stress response activation and RAN translation. In CGG KI mice, neuroinflammation occurs before the appearance of intranuclear inclusions, suggesting that neuroinflammation is an early event in the pathogenic cascade. In Summary, synergy between FMRpolyG production and repeat RNA expression induces neuroinflammation and neurodegeneration in mouse models of FXTAS and genetically targeting PKR suppresses these phenotypes, suggesting that PKR as a potential target for future therapy development in this currently untreatable condition.