Non-invasive imaging techniques have been developed to detect subtle metabolic changes in 2D cultures and engineered 3D brain-like tissues. Using two-photon images based on endogenous contrast, metrics based on optical redox ratio, NADH-bound fraction and levels of mitochondria clustering can be used to identify differences between neurons and astrocytes. In particular, stress responses within these cell populations can be detected by observing changes in the relative levels of redox ratios. By quantitatively using this non-destructive, high-resolution approach for detecting neurodegeneration-related metabolic derangement (e.g., oxidative stress) at the cellular level, an improved understanding of  mechanisms can be achieved and provide new approaches for diagnosing and treating neurodegenerative disease.