Description
Scientists oft en plan experiments using litt le more than their intuitions about what needs to be accomplished next. To the seasoned researcher, this procedure is akin to navigating a familiar but vast landscape without a map and compass. Scientists have been doing this for centuries now, since science’s inception. Th is does not mean that experiments and research programs aren’t oft en thoughtfully planned and executed. Scientists routinely think long and hard about which experiments to tackle next. But litt le prepares them to systematically scan the horizons of their fi eld’s potential. Good fortune and good mentors help some fi nd successful routes among the many paths their research might take. Conversations with colleagues, journal clubs, and courses in experimental design off er some clues about how to steer research programs. But scientists won’t fi nd empirical research devoted to helping them decide how to structure a research program or what experiments to tackle next. In place of such research they will fi nd only tradition, occasional lessons from their mentors, and their intuitions.
To extend our metaphor a bit: navigating over large, richly structured geographical areas without a map and compass is especially problematic. And hopefully it won’t take much argument to show that even the current body of published neuroscience research— the terrain that experimenters must now navigate to review and plan experiments—is overwhelming. Over the past three decades neuroscientists have published more than 1.6 million articles, spanning increasingly specialized yet increasingly complex and interconnected fields and subfields. Already this is a corpus far beyond human reading capacity. Th is terrain of knowledge is rich and highly structured, but with our current resources we have barely tapped it. How many unrecognized but important conclusions lay buried within the vastness of this record? How often do we duplicate published work, or pursue dead-end experimental paths that could have been avoided had we been fully aware of the contents and implications of the full published experimental record, even in our own fields? How much published research has actually advanced our understanding of nervous
system functions? What percentage of this published research instead represents litt le more than small variations on previous findings, retracing of familiar ground in the immense landscape of the published record? If we had a tool to help us both navigate the vastness of the published record and grasp its implications, how might this tool change the course of current and foreseeable research? Even the most optimistic principal investigator, grant reviewer, or journal referee has to admit that these questions lack easy answers. Yet we can no longer aff ord to continue to conduct research at current scales—and expense—without addressing them. Th e corpus of neuroscience research against which we now ask these questions will only continue to expand.