This study shows that artificial neural networks can be trained to provide high resolution, high accuracy models of multivariate radiobiological data collected by flow cytometry.

Nine faunal teeth from Layer 20 of El Castillo cave in Cantabrian Spain were dated using electron spin resonance (ESR). Two teeth were rejected due to inconsistent sub-sample ages, while the remaining teeth yielded a mean age that is consistent with the stratigraphic expectations: 42.7±3.5. Uncertainty in the external γ dose rate results in a potential systematic uncertainty of ±6.4 ka that should affect all samples equally. The results provide independent confirmation of previously-reported 14 C ages for Layer 20.