In people suffering brain metastasis, cognitive impairment has been traditionally thought to result from the mass effect of the tumor growth. Recently, we teamed with the group of Manuel Valiente (CNIO), to discover that different experimental models of brain metastasis actually co-opt microcircuit activity in a very specific manner (Sanchez-Aguilera et al., Cancer Cell 2023). We recorded LFP activity in the vicinity of brain metastases from various primary sources and oncogenomic profiles. Using data science and machine learning strategies, we found that a model trained in LFP signals surrounding the three different brain metastasis can help to disambiguate between subtypes when confronted with recordings from a different metastatic line, and to anticipate brain metastases early in advance. These alterations were linked to changes in the density of GABAergic inhibitory synapses and specific transcriptional programs. Our results pave the way to investigate the functional signatures of the crosstalk between cancer cells and the surrounding neuronal circuit as a potential translational biomarker.