For decades, we have presumed the death of hair cells and spiral ganglion neurons are the main cause of hearing loss and difficulties understanding speech in noise, but new findings suggest synapse loss may be the key contributor. Specifically, recent preclinical studies suggest that the synapses between inner hair cells and spiral ganglion neurons with low spontaneous rates and high thresholds are the most vulnerable subcellular structures, with respect to insults during aging and noise exposure. This cochlear synaptopathy can be "hidden" because this synaptic loss can occur without permanent hearing threshold shifts. This new discovery of synaptic loss opens doors to new research directions. Here, we review a number of recent studies and make suggestions in two critical future research directions. First, based on solid evidence of cochlear synaptopathy in animal models, it is time to apply molecular approaches to identify the underlying molecular mechanisms; improved understanding is necessary for developing rational, effective therapies against this cochlear synaptopathy. Second, in human studies, the data supporting cochlear synaptopathy are indirect although rapid progress has been made. To fully identify changes in function that are directly related this hidden synaptic damage, we argue that a battery of tests including both electrophysiological and behavior tests should be combined for diagnosis of "hidden hearing loss" in clinical studies. This new approach may provide a direct link between cochlear synaptopathy and perceptual difficulties.