Neurodegenerative diseases, such as Alzheimer’s disease (AD) and frontotemporal dementia (FTD), are incurable, fatal, and affect millions of people worldwide. The brain is composed of several cell types, predominantly neurons, astrocytes, microglia, and oligodendrocytes, that work together to ensure people’s normal functioning. In neurodegenerative diseases, aberrant protein aggregation occurs, resulting in complex cascades that result in the development of pathological astrocytes and microglia as well as the death of particular subsets of neurons, termed vulnerable neurons, within affected brain regions. However, the molecular underpinnings of different cell types’ differential phenotypic changes in disease are not fully understood. To identify and characterize pathological reactive astrocytes in AD and vulnerable neurons in behavioral variant FTD (bvFTD), we isolated nuclei from affected regions of donor human brains and utilized single nucleus RNA sequencing to obtain the transcriptomes of hundreds of thousands of cells. We used clustering algorithms to iteratively group nuclei into transcriptionally similar groups, identifying the brain’s many cell types, and within each cell type, we identified disease-relevant subtypes. We characterized these disease-relevant subtypes using differential gene expression and gene ontology analyses to identify what genes and biological pathways are perturbed in these cells. Finally, we used multiplex immunofluorescence to confirm our RNA-based bioinformatics findings are also evident at the protein level. In AD, we found gray matter, protoplasmic astrocytes to exist within a spectrum of reactivity. This spectrum was hallmarked by mild upregulations of genes canonically known to be associated with reactive astrocytosis (inflammation and cytoskeletal components), but this reactivity spectrum was even more profoundly associated with the loss of homeostatic function, especially transcription factors, suggesting that AD reactive astrocytes’ loss of typical maintenance functions may contribute to neurodegeneration. In bvFTD, we found excitatory neurons to be more affected than inhibitory neurons, especially a group of layer 2-3 intratelencephalic neurons. Within those neurons, we found a subset of PPARG expressing neurons to be depleted in disease. Analyzing these neuron populations’ transcriptional changes in bvFTD, we identified genes and biological pathways that may confer vulnerability or resistance in disease and identified potential therapeutic targets for future investigation.