Text this: Leveraging transparent ontology learning to refine constructs in neuroscience