The traditional method in the development of drugs for neurodegenerative diseases, such as Alzheimer's or Parkinson's disease, necessitates the use of various animal models to assess complex behavioral changes. This process is not only time-consuming but also prone to misjudgments due to interspecies behavioral differences. In contrast, neural activity data gathered by head-mounted microscopy platforms represent integrated information that is relatively straightforward to interpret and exhibits low variability between species. This makes it an excellent resource for the development of corresponding artificial intelligence algorithms.We injected 6-Hydroxydopamine (6-OHDA) or adeno-associated virus (AAV)- mediated overexpression of alpha-synuclein into the nucleus accumbens, and collected calcium imaging and corresponding behavioral changes from these Parkinson's disease mouse models. We found that compared to the sham control group, both different Parkinson's disease mouse models produced a unique pattern of neural activity. Upon repeated verification of this pattern's association with Parkinson's disease, we discovered that this neural activity pattern does not appear in other animal models (such as those for pain, opioid drug treatments, etc.). Further use of this neural activity pattern to establish an artificial intelligence algorithm may provide an opportunity to create a rapid screening platform for the development of Parkinson's disease drugs.