Author's personal copy The science of research: The principles underlying the discovery of cognitive and other biological mechanisms Alcino J. Silva * Department of Neurobiology, UCLA, 695 Charles Young Drive South, Los Angeles, CA 90095-1761, United States Department of Psychiatry, UCLA, 695 Charles Young Drive South, Los Angeles, CA 90095-1761, United States Department of Psychology, UCLA, 695 Charles Young Drive South, Los Angeles, CA 90095-1761, United States Brain Research Institute, UCLA, 695 Charles Young Drive South, Los Angeles, CA 90095-1761, United States Abstract Studies of cognitive function include a wide spectrum of disciplines, with very diverse theoretical and practical frameworks. For exam- ple, in Behavioral Neuroscience cognitive mechanisms are mostly inferred from loss of function (lesion) experiments while in Cognitive Neuroscience these mechanisms are commonly deduced from brain activation patterns. Although neuroscientists acknowledge the limi- tations of deriving conclusions using a limited scope of approaches, there are no systematically studied, objective and explicit criteria for what is required to test a given hypothesis of cognitive function. This problem plagues every discipline in science: scientific research lacks objective, systematic studies that validate the principles underlying even its most elemental practices. For example, scientists decide what experiments are best suited to test key ideas in their field, which hypotheses have sufficient supporting evidence and which require further investigation, which studies are important and which are not, based on intuitions derived from experience, implicit principles learned from mentors and colleagues, traditions in their fields, etc. Philosophers have made numerous attempts to articulate and frame the prin- ciples that guide research and innovation, but these speculative ideas have remained untested and have had a minimal impact on the work of scientists. Here, I propose the development of methods for systematically and objectively studying and improving the modus operandi of research and development. This effort (the science of scientific research or S2) will benefit all aspects of science, from education of young scientists to research, publishing and funding, since it will provide explicit and systematically tested frameworks for practices in science. To illus- trate its goals, I will introduce a hypothesis (the Convergent Four) derived from experimental practices common in molecular and cellular biology. This S2 hypothesis proposes that there are at least four fundamentally distinct strategies that scientists can use to test the con- nection between two phenomena of interest (A and B), and that to establish a compelling connection between A and B it is crucial to develop independently confirmed lines of convergent evidence in each of these four categories. The four categories include negative alter- ation (decrease probability of A or p(A) and determine p(B)), positive alteration (increase p(A) and determine p(B)), non-intervention (examine whether A precedes B) and integration (develop ideas about how to get from A to B and integrate those ideas with other avail- able information about A and B). I will discuss both strategies to test this hypothesis and its implications for studies of cognitive function. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Science of research; Cognition; Plasticity; Memory; Learning 1. Introduction In the last 100 years science has grown at a vertiginous pace and it has changed every aspect of the world around us. In the process, scientific discovery has also been trans- formed from an activity that attracted a few people to an 0928-4257/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.jphysparis.2007.11.002 * Address: Department of Neurobiology, UCLA, 695 Charles Young Drive South, Los Angeles, CA 90095-1761, United States. Tel.: +1 310 7946609; fax: +1 310 794 7088. E-mail address: silvaa@ucla.edu www.elsevier.com/locate/jphysparis Available online at www.sciencedirect.com Journal of Physiology - Paris 101 (2007) 203–213