A substantial gap exists between the impoverished level of description of experiment events, tasks, and conditions required by current data archiving standards and the more robust level required for event-related analysis (and mega-analysis) of data across neuroimaging studies. The HED-3G framework, a standardized methodology for creating machine-actionable tag annotations, provides a practical pathway for closing this gap. Annotators use tools to select tags from a simple, hierarchically structured base vocabulary to describe the sensory presentations, participant actions, tasks, environment, and design variables. Tools validate these annotations and automatically expand and assemble them, producing machine-actionable data annotations that can provide an enabling infrastructure for data search, collection, analysis, and interpretation. HED-3G also supports development of library schema vocabularies by and for specialized research communities. HED-3G offers an easily usable front-facing annotation interface whose results can support sophisticated analysis and also be connected (on the back end) to formal ontologies and related tools. A key challenge is to motivate data authors to perform adequate annotation before archiving and/or sharing their data. The additional downstream functionality provided by HED annotation is designed to encourage data authors to consider not only their own immediate uses for the data but also its potential for broader use and impact.