Machine Learning Reveals The Effect of Maternal Age on The Mouse Pre-Implantation Embryo Developmental Timing

Abstract

SummaryIn recent years, many women have delayed childbearing, thus increasing the necessity for assisted reproductive technology (ART) for older women1–3. Despite advances in ART4, its success rate in advanced-age women is still very low3,4. As time-lapse imaging became available, morphological features of the developing pre-implantation embryo, in-vitro, are heavily used to assess its potency5–9. Timing of embryo cleavage is also an important factor that correlates with blastocyst formation and pregnancy rates8,10–14. Yet, our understanding of the interplay between embryos’ morphology, viability, and maternal age is limited, as manual approaches to infer embryo morphokinetics are time-consuming, subjective, and prone to errors. Machine learning15–18 was recently harnessed to predict embryo developmental potential19,20, however, with limited success.Here, we develop an artificial intelligence (AI) platform that infers the embryos’ developmental stage and captures tens of morphological properties and developmental dynamics. We show that developmental timing is the most informative and predictive morphokinetic property, particularly for embryos from maternally aged females. Analyzing the timing distributions reveals that viable embryos are confined into an age-independent temporal corridor while non-viable embryos deviate from it towards slower transition times. Yet, the deviation of non-viable embryos from the temporal corridor is age-dependent. Furthermore, there is a significant correlation between consecutive developmental stages transition times that diminishes in maternally old embryos. Overall, our results suggest that maternally old embryos’ most apparent morphokinetic property is the loss of temporal regulation. Our results and platform pave the way for a more accurate, maternally-age-dependent, assisted reproductive technology.