Abstract
It is common to encounter the belief that neuroscience holds promise for advancing education practice—a belief that is predicated on the assumption that neuroscientific findings can be scaled up to inform our understanding of behavior in complex education settings. In this article, we argue that this belief is not just far-fetched but misdirected. Although we acknowledge the value of neuroscience for understanding brain mechanisms, we argue that it is largely unnecessary for the development of effective learning interventions. We demonstrate how neuroscience findings have failed to generalize to classroom contexts by highlighting the recent popularity and failed results from brain-training research. We end by providing two recommendations for how future researchers and policy makers should address neuroscience and its potential for education applications.
Keywords
It is common to encounter the belief that advances in technologies that allow us to measure activity in the brain will lead directly to interventions that improve learning. Implied by this belief is the idea that education stands to benefit from the explosion of research and data generated within neuroscience. These ideas are further propelled by large-scale initiatives such as the Brain Research Through Advancing Innovative Neurotechnologies (BRAIN) initiative. However, the idea that neuroscience can have a direct impact in the classroom is a bit far-fetched. This notion was first pointed out by Bruer (1997) in his article “Education and the Brain: A Bridge too Far,” elaborated on in subsequent work (e.g., Bruer, 2006), and recently echoed by a variety of researchers (e.g., Anderson & Della Sala, 2012; Bowers, 2016; Marcus, 2014; for an alternative view, see Howard-Jones et al., 2016). Bruer’s thesis is simple: The brain is far too complex and we know far too little about how it works for this knowledge to be useful for education. What we learn from neuroscience may well be useful for understanding cognitive phenomena, but it is the latter and not the former that offers promise for education practices.
In January 2015, a group of scientists from education, developmental psychology, cognitive psychology, and neuroscience convened a workshop at the White House in Washington, DC. The overarching goal of the workshop was to revisit the question of whether (and how) neuroscience might drive education practice. Has the field of neuroscience matured to the point at which it can serve as the genesis for identifying neuroscience-based education interventions? Strikingly, although there was discussion of the potential for neuroscience to inform education practice, not one of the participants was able to identify an intervention that originated from neuroscientific findings. Despite more than two decades of research and massive investment in the brain sciences from the National Science Foundation, the National Institutes of Health, and the European Research Council, Bruer’s original critique still rings true, but why?
According to Bruer (1997, 2006), the gap between neuroscience and education is too large to bridge without intermediate landing spots. These landing spots lie in psychology, and in particular cognitive and social psychology. The challenge for developing successful education interventions is not in understanding brain mechanisms; it is in understanding the behaviors that the brain manifests in complex learning environments. Understanding behavior is necessary and sufficient for guiding the development of education interventions; understanding the brain is not. To illustrate this point, consider the recent popularity of brain training, variably known as cognitive training or working memory (WM) training. The basic premise underlying brain training is the idea that neurological processes can be changed through extensive practice on cognitive tasks, a principle often referred to as brain plasticity. In early work (e.g., Merzenich et al., 1996), researchers argued that principles of brain plasticity could be exploited to develop interventions that could address learning deficiencies (e.g., language-learning impairments) by targeting specific neurological processes related to the learning deficiency. The concept of brain plasticity dates back long before the modern field of neuroscience to at least William James (1892), who discussed the concept in relation to experience-based change (what James referred to as the formation of a habit). There is little doubt that extensive practice on a particular task leads to neurological change. For example, several studies have shown that practice on a task such as the n-back task, which is believed to rely heavily on WM skills, can lead to changes in white-matter integrity, gray-matter density, and blood-oxygen-level-dependent signal-activation patterns (e.g., Olesen, Westerberg, & Klingberg, 2004; Takeuchi, Taki, & Kawashima, 2010). However, the real question for education researchers is whether the observed changes in brain structure are generalizable, leading to improvements in the types of complex learning that take place in classroom settings. Sadly, in regard to brain training, the answer to this question seems to be “no” (Roberts et al., 2016; Strong, Torgerson, Torgerson, & Hulme, 2011).
Brain training is emblematic of the gulf between basic neuroscience and education, wherein seemingly groundbreaking neuroscience findings (e.g., brain plasticity, synaptogenesis, pruning) simply do not scale up to practical education interventions. Other examples in which neuroscience has failed to deliver practical advice for education include early work emphasizing the importance of critical periods for learning (Begley, 1996; Hirsch, 1996) and proposals to use biomarkers as predictors of targeted interventions (e.g., Supekar et al., 2013). The importance of critical periods ultimately proved to be restricted primarily to first-language learning and vision (Bruer, 2006), whereas the use of biomarkers for targeted interventions is simply impractical (Bishop, 2000). The above examples can be contrasted with another phenomenon, which is the tendency to brand interventions as being based on neuroscience, when in fact they were developed on the basis of behavioral data alone. For instance, the testing effect is sometimes dubbed an example of education neuroscience (e.g., Mastin, 2018), yet the phenomenon, in which retrieval improves more with retrieval practice relative to restudying material, is purely behavioral.
Returning to the brain-training example, we note that prior researchers (including the first author of this article) have proposed that brain-plasticity training could be effective at improving core education competencies such as language comprehension and mathematical reasoning (Atkins, Bunting, Bolger, & Dougherty, 2012). The bridge implied by brain-training research is that because the brain can be modified through training, it should be possible to train the brain in a way that will directly improve education outcomes. For example, targeted training of neurological mechanisms underlying WM should in theory result in generalization (i.e., transfer) of training to nontrained tasks (i.e., educationally relevant tasks). In as much as WM is an important core competency for education success (Alloway & Alloway, 2010), improving WM through training should ultimately improve student outcomes.
The assumed link between the neuroscience of brain plasticity and education seemed obvious; however, there are a number of problems with the brain-training-to-education bridge. First and foremost, the available literature now strongly suggests that brain training does not reliably generalize beyond the trained tasks (Redick et al., 2013; Sprenger et al., 2013), certainly does not carry over to better education outcomes, and may in fact result in poorer performance on important indicator variables when students are removed from the classroom for the training intervention (e.g., Roberts et al., 2016). Before postulating that brain training is a viable intervention, it is necessary to first show that brain training leads to changes in basic cognitive processes by showing behavioral improvements on untrained measures of WM or intelligence (for similar arguments, see Harrison et al., 2013), an important intermediate bridge between the brain and the classroom. Unfortunately, it does not (Melby-Lervåg, Redick, & Hulme, 2016; Simons et al., 2016). Second, even if brain training did generalize, establishing the link between the neuroscience underlying WM training and education outcomes requires cognitive theory. Cognitive theory and, in particular, conceptual models of WM provide the landing spot required for establishing the plausibility of using brain training to improve education outcomes.
Indeed, understanding the fundamental nature of WM and how it relates to specific education outcomes (e.g., reading comprehension, mathematics) is both necessary and sufficient for motivating the use of brain-training methodologies for education. This critique applies to any other education intervention: Neuroscience may be useful for understanding brain mechanisms and establishing connections to cognitive theory, but it is largely irrelevant to considerations of education policy and classroom practices. Thus, the gap between neuroscience and education cannot be bridged without the intermediary stepping stone of cognitive theory. But more to the point, neuroscience is not even needed: Whether training (or any other intervention) results in changes in brain activation or structure is irrelevant to the question of whether brain training (or any other intervention) can improve education success. What is relevant is whether training leads to changes in core cognitive processes (e.g., WM or IQ) necessary for education success and whether improvements in those abilities generalize to authentic education settings. The theory of change requires specification at the cognitive level but not at the neurological level.
Figure 1 provides a schematic of the various proposed bridges between neuroscience and education. The arrows are shaded to indicate the strength of the link between two specific levels. The strength can also be thought of as how much one level is needed to understand the level to which it links. As discussed above, we believe that neuroscience by itself is not relevant for informing education practice or for the development of education interventions (light gray arrows). To date, there are few, if any, examples of successful education interventions that can be attributed to advances in neuroscience (cf. Bowers, 2016). Although we do not believe that neuroscience can inform education practice directly, we do believe that it can be useful for developing a deeper scientific understanding of cognitive and social phenomena (dark gray arrows). However, gaining a deeper understanding of the neuroscientific basis of cognitive and social phenomena is not the goal of intervention research. Neuroscience can and should also play a vital role in the development and refinement of formal models of cognition (dark gray arrows; see Forstmann, Wagenmakers, Eichele, Brown, & Serences, 2011). Formal models may well be useful for informing education research, although the strength of formal models as they relate to education interventions lies in their account of cognitive and social behavior (black arrows), not in their ability to accommodate neuroscience data. As suggested by Krakauer, Ghazanfar, Gomez-Marin, MacIver, and Poeppel (2017), it is reasonable to assert that neuroscience needs behavior more than behavior needs neuroscience, and we believe this assertion applies to behavior in the classroom. One can develop a high-level model of cognition without considering the neural implementation, but it is difficult to imagine how a neuroscientist might address questions of behavior without having a well-formed conceptual model of the behavioral task.
Graphical depiction of various proposed bridges between neuroscience and education. Arrows are shaded to represent the strength of the link between two specific levels. See the text for further details.
Interventions Based on Cognitive Theory
In contrast to the above skeptical view regarding neuroscience-based education interventions, there is reason for optimism regarding cognitive-psychological-based interventions. There is an abundance of research based on basic cognitive psychological theory that has proven to be effective at improving learning in classroom settings. These interventions include general learning practices such as active learning (e.g., Freeman et al., 2014), retrieval practice (for a review, see Roediger & Butler, 2011), spaced versus massed learning (for a review, see Benjamin & Tullis, 2010), and blocked versus interleaved practice (for a review, see Schmidt & Bjork, 1992). These interventions have all been shown to be effective when incorporated into actual classroom settings at a variety of grade levels.
In addition to general practices, many interventions exist that target specific learning objectives. Siegler and colleagues have developed a number-based board game that improves a variety of math abilities in low-income preschoolers (Ramani, Siegler, & Hitti, 2012; Siegler & Ramani, 2009). Snow and colleagues created Word Generation, an intervention that improves academic vocabulary in middle school (Lawrence, Rolland, Branum-Martin, & Snow, 2014; Snow, Lawrence, & White, 2009). Additionally, the widely used cognitive-science-based intervention, Tools of the Mind, has been shown to improve a wide range of executive function skills and improve classroom performance (e.g., Barnett, Jung, Yarosz, Thomas, & Hornbeck, 2008; Diamond, Barnett, Thomas, & Munro, 2007). These are just a few of the many current practices and interventions based on behavioral science research from within cognitive and developmental psychology. For the vast majority of these interventions, the neural correlates are unknown. Although learning about underlying neural mechanisms may lead to a better understanding of the brain, it is unclear how that knowledge would directly lead to improvements in any of the interventions.
Summary and Recommendations
We offer two recommendations on the basis of the literature reviewed above. Our first recommendation concerns the funding climate. There is, in our view, a need to improve funding for basic and translational research in the areas of cognitive and social psychology to synergize evidence-based interventions in education science. Although basic neuroscience research is foundational to the understanding of brain mechanisms, the practical benefit for the development of education interventions remains quite limited. If one of the goals of federal funding agencies is to improve education outcomes and student success, then we believe that investment in the cognitive and social psychological sciences is more worthwhile in both the short term and the long term. Given that federal funding for research is shrinking and the pool of resources is limited, we strongly urge the government to not divert funding from the cognitive and social sciences to the neurosciences.
The second recommendation is a caveat venditor and pertains to the tendency that scientists sometimes have toward hyperbole when discussing the implications of their results. Neuroscience has found its way into the minds of educators, marketers, and consumers. Educators (and parents) are inundated with advertisements for neuroscience-based products that promise to improve everything from memory to attention to classroom performance, and marketers have begun using “neuroscience” as a tool for persuasion, as if neuroscience somehow brings legitimacy to an advertiser’s claims (Racine, Bar-Ilan, & Illes, 2005; Vaughan, 2015). Not to diminish the value of neuroscience, but we believe that it is important for scientists to regulate how they communicate the value of basic scientific findings as they relate to practical applications. This recommendation is not specific to neuroscience, although we believe that it is especially important for neuroscientists to exercise caution because the public is more likely to believe claims substantiated by neuroscience data (Weisberg, Taylor, & Hopkins, 2015; for a similar argument, see Schwartz, Lilienfeld, Meca, & Sauvigné, 2016) and because the gulf between neuroscience and application is so large. In an era in which researchers are increasingly asked to write press releases about their research, it is especially important that the public not be misled about the promise of basic research for addressing applied problems.
Although Bruer’s (1997) original critique is now more than two decades old, his arguments that the gap between neuroscience and education is “a bridge too far” still hold. But further, we believe that attempts to bridge neuroscience and education might well be a bridge astray. To draw an analogy with physics, it is one thing to find support for a theoretical proposition in the frictionless environment of the lab; it is quite another thing for the support to hold when friction is applied. Phenomena that are observed in the frictionless environment of the lab are much more complex when real-world friction is applied. Our worry is that a focus on neuroscience may well lead us astray in our quest for evidence-based education interventions in a friction-filled world.
Recommended Reading
Bruer, J. T. (2006). (See References). The foundational article for the argument that scientists should not try to leap directly from neuroscience findings to education interventions.
Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students’ learning with effective learning techniques: Promising directions from cognitive and educational psychology. Psychological Science in the Public Interest, 14, 4–58. An overview of effective learning techniques that have emerged from research in cognitive science and cognitive psychology.
Simons, D. J., Boot, W. R., Charness, N., Gathercole, S. E., Chabris, C. F., Hambrick, D. Z., & Stine-Morrow, E. A. L. (2016). (See References). A thorough review and critique of cognitive-training research.
Weisberg, D. S., Taylor, J. C., & Hopkins, E. J. (2015). (See References). An empirical study examining how the use of neuroscience terms affects people’s belief in scientific claims.
Footnotes
Action Editor
Randall W. Engle served as action editor for this article.
Declaration of Conflicting Interests
The author(s) declared that there were no conflicts of interest with respect to the authorship or the publication of this article.
Funding
This work was supported by National Science Foundation Grant BCS-1522699 (to M. R. Dougherty). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.