In this work, statistical learning approaches are used to detect biomarkers for neurodegenerative diseases (NDs). NDs are becoming increasingly prevalent as populations age, making understanding of disease and identification of biomarkers progressively important for facilitating early diagnosis and the screening of individuals for clinical trials. Advancements in gene expression profiling has enabled the exploration of disease biomarkers at an unprecedented scale. The work presented here demonstrates the value of gene expression data in understanding the underlying processes and detection of biomarkers of NDs. The value of novel approaches to previously collected -omics data is shown and it is demonstrated that new therapeutic targets can be identified. Additionally, the importance of meta-analysis to improve power of multiple small studies is demonstrated. The value of blood transcriptomics data is shown in applications to researching NDs to understand underlying processes using network analysis and a novel hub detection method. Finally, after demonstrating the value of blood gene expression data for investigating NDs, a combination of feature selection and classification algorithms were used to identify novel accurate biomarker signatures for the diagnosis and prognosis of Parkinson’s disease (PD) and Alzheimer’s disease (AD). Additionally, the use of feature pools based on previous knowledge of disease and the viability of neural networks in dimensionality reduction and biomarker detection is demonstrated and discussed. In summary, gene expression data is shown to be valuable for the investigation of ND and novel gene biomarker signatures for the diagnosis and prognosis of PD and AD.