Raman spectroscopy is an optical technique that can identify molecules in a label-free manner, and is therefore heavily investigated in various areas ranging from biomedical engineering to materials science. Probe-based Raman spectroscopy can perform minimally invasive chemical analysis, and thus has potential as a real-time diagnostic tool during surgery. In this study, Raman experimentation was calibrated by examining the Raman shifts with respect to the concentrations of chemical substances. Raman signal characteristics, targeted for normal mice and cerebral tissues of the 5xFAD dementia mutant model with accumulated amyloid beta plaques, were measured and analyzed to explore the possibility of diagnosis of Alzheimer's disease. The application to the diagnosis of dementia was cross-validated by measuring Raman signals of amyloid beta. The results suggest the potential of Raman spectroscopy as a diagnostic tool that may be useful in various areas of application.