Geneticists use olfactory conditioning in Drosophila to identify learning genes; however, little is known about how these genes are integrated into short-term memory (STM) pathways. Here, we investigated the hypothesis that the STM evidence base is weak. We performed systematic review and meta-analysis of the field. Using metrics to quantify variation between discovery articles and follow-up studies, we found that seven genes were both highly replicated, and highly reproducible. However, ∼80% of STM genes have never been replicated. While only a few studies investigated interactions, the reviewed genes could account for >1000% memory. This large summed effect size could indicate irreproducibility, many shared pathways, or that current assay protocols lack the specificity needed to identify core plasticity genes. Mechanistic theories of memory will require the convergence of evidence from system, circuit, cellular, molecular, and genetic experiments; systematic data synthesis is an essential tool for integrated neuroscience.

遗传学家在果蝇中使用嗅觉调节识别学习基因；但是，关于这些基因如何整合到短期记忆（STM）途径中的知识还很少。在这里，我们调查了STM证据基础薄弱的假设。我们对该领域进行了系统的审查和荟萃分析。使用度量标准量化发现文章和后续研究之间的差异，我们发现七个基因都高度复制且可高度再现。但是，约80％的STM基因从未被复制过。虽然只有很少的研究调查了相互作用，但综述的基因可能占了> 1000％的记忆。如此大的总效应大小可能表明不可重复性，许多共享途径，或者当前的检测方案缺乏鉴定核心可塑性基因所需的特异性。记忆的力学理论将需要来自系统的证据的融合，电路，细胞，分子和遗传实验；系统的数据合成是集成神经科学的重要工具。