Our understanding of past interactions between the development of human societies, the evolution of climate and associated changes in ecosystems and landscape dynamics is conditioned by the acquisition of high resolution records within which specific tracers allow us estimating variability. The recent development of a molecular biomarker (miliacin) specific of cultivated millet (Panicum miliaceum) expands the possibilities afforded by organic geochemistry applied to sedimentary archives to unravel these interactions. Within the frame of the Pygmalion project (ANR Blanc, France) we measured the concentration in miliacin of about 400 sediment samples extracted from two cores drilled in Lake le Bourget in order to reach an infra-decadal resolution for the 2000-600 BC time period that covers the Bronze Age. Miliacin is detected for the first time in both cores in sediments dated back to ca. 1700 BC, in agreement with the supposed date of introduction of P. miliaceum in the region. Miliacin concentration is low (ca. 20 ng.g-1) during the 1700-1400 BC interval and then rises to values up to 300 ng.g-1 at 850 BC before the strong decrease to 20 ng.g-1 at 750 BC imputable to the abandonment of palaffitic habitats due to a climatic deterioration at the Bronze Age/Iron Age transition. In addition to this general trend, miliacin concentration shows century-scale variations in the 1700-800 BC interval that are comparable in both cores and coincide with other data. For example, two periods of miliacin high concentrations at 950 and 850 BC coincide with high densities of dendrochronological dates acquired on wooden piles and with two periods of lake level lowering. A relationship between miliacin concentration in sediment and demography around Lake le Bourget cannot be disregarded at this stage. This exceptional record should be compared to other records affording such a high temporal resolution, which remains rare in the region.