Abstract
Educational neuroscience is a relatively new field. Where is it in relation to other research domains, such as education research, the psychology of learning, and the neuroscience of learning? Document co-citation analysis reveals that these research fields are tenuously connected. Currently, educational neuroscience sits between subfields of neuroscientific and psychological research but remains distant from issues and topics that are prominent in the education research literature. Part of the problem is that the educational neuroscience literature is a meta-scientific literature, more about the promise and pitfalls of applying neuroscience to education than it is about applications of neuroscience to education. The psychology literature retains a fundamental role in linking educational neuroscience, education research, and neuroscience into an integrated learning research enterprise.
Educational neuroscience is a relatively new and highly interdisciplinary research front. Its objective is to improve educational practice by applying findings from brain research. Other research fields in education, psychology, and neuroscience are also attempting to improve teaching and learning through applications of their work. Educational neuroscientists are well aware of this. However, it would be useful to see how different fields in the learning sciences might interrelate and complement one another. Where is educational neuroscience?
Using document co-citation analysis, this article explores how educational neuroscience relates to the fields of education research, the psychology of learning, and the neuroscience of learning. There are connections among the various research communities with these four fields, although some of the connections are tenuous. As of now, educational neuroscience sits between subfields of neuroscientific and psychological research. However, it remains distant from issues and topics that are prominent in the education research literature. Part of the problem is that the existing educational neuroscience literature is more of a meta-scientific literature than a scientific literature. It is a literature more about the promise and pitfalls of applying neuroscience to education than it is about applications of neuroscience to education. We will also see that the psychology literature plays a fundamental role in linking educational neuroscience, education research, and neuroscience into an integrated learning research enterprise. As educational neuroscience matures and moves down from the meta-level, it may become more highly integrated with education research, psychology, and neuroscience. Based on the co-citation analyses, likely points for future integration suggest themselves.
I will first identify sets of articles—literatures—representative of important work in educational neuroscience, education research, the psychology of learning, and the neuroscience of learning. A document co-citation analysis of each literature will allow us to characterize each field and to identify research subspecialties within them. I will call these subspecialties research communities. Finally, document co-citation analyses of the combined literatures will allow us to see how the communities within each research field relate to one another. We will be able to see, for example, how research communities within educational neuroscience relate to communities in education research, psychology, and neuroscience.
Data and Methods
The four research literatures considered here are educational neuroscience, education research, psychology of learning, and the neuroscience of learning. The Web of Science (WoS) assigns articles to a subject matter field (the “su =” field in its advanced search option). Among the WoS fields are education and education research, psychology, and neuroscience and neurology. A subject field search on Education and Education Research identified the education research literature. To identify the educational neuroscience literature, the topic words (words occurring in titles, abstracts, or key words) “brain OR neuro*” was conjoined to the Education and Educational Research subject field search. (WoS uses the asterisk as a truncation symbol.) For the psychology of learning, the topic words “learn* OR educ*” were conjoined to a search of the psychology subject field. For the neuroscience of learning, “learn* OR educ*” was conjoined to a subject fields search on neurosciences and neurology. Searches were limited to the time period 1997 through 2015. The starting year was chosen because Bruer (1997) was one of the first articles to discuss applications of then current neuroscientific findings to education. Subsequently, Byrnes and Fox (1998) argued for the centrality of cognitive neuroscience to educational research but not without evoking some skeptical response (Mayer, 1998; Stanovich, 1998). These articles, still highly cited in the educational neuroscience literature, could be considered the starting points for what has become a substantial research program in educational neuroscience. Finally, for each of the four searches, the top-cited 500 articles were retained and serve as the four article sets analyzed here. (The exact WoS search strategies and the four article sets are available from the author.) Global citation counts, that is, the number of times articles are cited by others within the WoS, are used only to identify the top-cited articles. Citation counts of individual articles play no further role in the subsequent document co-citation analyses.
Document co-citation analysis is often used to delineate the specialty structure of research fields (Small, 1980). If Document A cites Documents B and C, then B and C are co-cited. Co-citations are counted by comparing the cited references that appear in the bibliographies of documents. The articles a document cites situate that document in the space consisting of all published documents. The higher the co-citation count between two articles, the more likely those articles are to be conceptually or thematically related. Note that although Document A is in the original article set, Documents B and C need not be, although Document A can also appear among the list of co-cited documents. Thus, co-citation networks situate Document A, directly or indirectly, within a wider network of relevant documents. For example, the 500-article neuroscience article set cites nearly 40,000 unique articles.
The co-citation data are represented in co-citation networks, where the nodes are cited references, and weighted edges represent co-citation counts. Co-citation networks are weighted, undirected networks and often consist of numerous, unconnected subnetworks. These subnetworks are called weak components. Each of the four literature networks contains dozens of weak components, ranging in size from hundreds to pairs of cited references. Here, the weak components of each literature will be interpreted as representing research communities within that literature.
To facilitate interpretation and visualization of these networks, some simplifications are helpful. First, in all the document co-citation networks, a threshold of five co-citations is required for two articles to be included in the network. This is a relatively low threshold. Second, for each literature, the focus will be on the largest three to four research communities, with one exception. Third, to facilitate visualization of the networks, pathfinder network analysis is used to remove redundant edges from the network. In the visualizations, only the strongest link between two cited references, the edge with the highest citation weight, appears. In graph–theoretic terms, these pathfinder networks are minimum spanning trees of the entire network. Thus, the visualizations show the strongest co-citation link between two cited references, not necessarily the only link. Where edge strengths in the pathfinder networks are high, one would also expect other weaker links to connect the two nodes. Where edge strengths are low or at threshold, one can infer that additional paths are less likely and that if they do exist, they are at or below the threshold. Finally, the pathfinder networks are further trimmed by eliminating all nodes that are linked to only one other node. This has the effect of trimming ends off branches in the networks. There are no instances where this trimming affects the interpretations. The Sci2 tool (Sci2Team, 2009) was used to generate the networks and Gephi to provide visualizations of the networks.
Overview of the Four Article Sets
Before turning to the co-citation networks, an overview of the four article sets yields insights about the possible relations among the literatures. Table 1 shows how many articles (bold) and cited references (italic) appear in the intersections of the article sets. Once would expect the educational neuroscience literature to be a subset of the education research literature because the educational neuroscience literature was identified by conjoining topic words to a search of the Education and Education Research subject domain. This does not occur because the global citation counts for the top 500 education research articles (range, 113–790) are considerably higher than those for the top 500 in educational neuroscience (range, 1–300). The two article sets have only 3 items in common: Drake, McBride, Lachman, and Pawlina (2009); Klingberg (2010); and Micheloyannis et al. (2006). One should always keep in mind that educational neuroscience is a relatively small research fields compared to the other three. Furthermore, the educational neuroscience article set has no articles in common with either the psychology or the neuroscience article sets. Psychology has a substantial intersection with education research (41 articles) and neuroscience (90 articles).
Data Summary.
Note. In each nondiagonal cell, the bold number is the number of articles common to the two article sets. The italic number is the number of cited references common to two article sets. The column at the right is the total number of unique cited references found in each article set.
However, what matters for document co-citation is not the number of articles common to two article sets but the number of shared cited references. The cited reference counts for these literatures range from around 17,000 to nearly 40,000. The number of commonly cited references between the article sets is the italicized number shown in the cells of Table 1. Although the educational neuroscience article set contains no articles found in the neuroscience of learning article set, there are still 954 unique cited references common to these two literatures. This is a relatively small number compared to the nearly 39,000 cited references in the neuroscience of learning literature, but it is not zero. Notice that the largest cited reference intersections occur between psychology and the three other literatures. This suggests that psychology might play an integrative role in connecting the research communities within each of these four literatures.
Table 2 presents the five most prominent key words, authors (by times cited within each article set), cited references, and source journals for the four literatures. Among the keywords, there are no surprises. It is apparent, however, that the neuroscience of learning is unique in its emphasis on cellular mechanisms involved in learning and memory. The only author that appears among the top-cited five in more than one literature is Joseph Ledoux in psychology and neuroscience. The highly occurring cited references point to primary research foci are in each of the fields. The educational neuroscience literature cites articles that provide rationales for and against a research program linking neuroscience to education. This provides the first hint about the meta-status of the educational neuroscience literature. Educational research starting with the heavy citation of the 1996 National Research Council’s report on National Science Education Standards (Standards, 1996) and including Brown (1992), Driver (1996), and Lederman (1992) appears to have a primary focus on science education. The psychology of learning features research on motivation with Ames (1992), Dweck and Leggett (1988), and Elliot and Church (1997) and on working memory with Baddeley (1992) and Baddeley (1986). Judging from the cited references in the neuroscience of learning literature, the primary focus is on the neural underpinnings of memory and learning as studied using animal models. The leading source journals are also unique to each field. Around 17% of the 500 neuroscience articles were published in the Journal of Neuroscience. Mind Brain and Education published 13% of the educational neuroscience articles. This contrasts with the Journal of Educational Psychology publishing around 5% of the psychology articles and the Review of Educational Research around 8% of the education research articles. The highly cited neuroscience and educational neuroscience articles are less dispersed among source journals than are the education research and psychology articles.
Top Five Key words, Authors, Cited References, and Source Journals for the Four Article Sets.
Note. Educ. Neuro. = educational neuroscience; Educ. Res. = educational research; Psych. = psychology; Neuro. = neuroscience; Refs.= references.
The Co-Citation Networks and Research Community Structure of the Four Literatures
Document co-citation analysis allows us to examine the structure of each literature, to identify research communities within them and eventually to investigate the relationships that obtain among these research communities across the literatures.
Education Research
Let us begin with the education research literature, the literature that might be closest to issues of concern in classroom teaching and learning. The simplified and trimmed co-citation network generated from the over 27,000 cited references found in the 500 top-cited education research articles contains 189 cited references that were co-cited 5 or more times (Figure 1). These articles are distributed among 17 weak components, or research communities, containing from 3 to around 55 articles. The three largest research communities are science education/cognition, motivation, and cognitive load theory as described subsequently.
The education research co-citation network (trimmed). Node size is scaled to the number of other articles to which the reference is linked at threshold five (range 2–16). Edges are scaled to co-citation weights (range 5–13).
Among the cited references appearing in the science education/cognition community is the highly cited and highly influential 1996 U.S. National Research Council’s National Science Education Standards (Standards, 1996). Developing, implementing, and improving these educational standards have been a major focus within educational research over the past 2 decades. One can also discern some structure within this community. Linked to the Standards report are Blumenfeld et al. (1991); Driver, Newton, and Osborne (2000); and Lederman (1992). The Blumenfeld article is on project-based learning and how it can affect student’s motivation and thought. The Lederman article is a review of student and teachers’ conceptions of science and how those conceptions influence science learning. The Driver et al. (2000) article is on the use of scientific argumentation in the classroom. These are articles written by science educators. The work of cognitive psychologists and situated learning theorists also appear here. Kuhn (1993) also argues that attention to scientific argumentation improve classroom teaching and learning. White and Frederiksen (1998) and Schauble, Glaser, Duschl, Schulze, and John (1995) are on applications of cognitive psychology to classroom science instruction. Vygotsky (1978) and Lave (1991) present research in the tradition of situated cognition. The articles in this community are published primarily in science education journals and educational psychology journals. The focus within this community is on improving classroom science instruction by applying insights from the study of human cognition.
The second prominent education research community centers on Ames (1992) and Pintrich and Degroot (1990). This community addresses issues of student motivation, achievement goals, and naive theories of intelligence. Research in this tradition has established that the students’ learning goals as well as their beliefs about the nature of intelligence influences student achievement. Bandura (1977, 1997) on social learning theory and self-efficacy also form part of this research cluster. I will call this community the education motivation community.
Sweller (1991) and Sweller, van Merrienboer, and Paas (1998) are among the most prominent papers in the third cluster. This community does research on cognitive load theory. Cognitive load theory, a theory developed by Sweller, addresses how well-designed classroom instruction can reduce demands on student working memory capacity and thereby enhance learning. Two papers by R. E. Mayer (Mayer, 2001; Mayer & Anderson, 1992) present research in the tradition of cognitive load theory, on how well-designed multimedia instruction can improve learning. Miller (1956) suggests an eventual link between cognitive load theory and psychological research on working memory. This will be called the cognitive load community.
Based on papers the educational research community has co-cited over the last two decades, one can infer that improving science instruction in schools has been a major research focus, with an emphasis on how to impart science content or subject matter knowledge to students. In addition, research on motivation and cognitive load theory are prominent. Cognitive psychology contributes to this research, although we might characterize it as applied cognitive psychology in contrast to what we will encounter within the psychology of learning. There is no hint of neuroscience or educational neuroscience anywhere within the education research co-citation network.
Psychology of Learning
Unlike education research, neuroscience has a major presence in the psychology of learning. Figure 2 shows the three largest research communities within this literature. It also shows a smaller community of two (four in the untrimmed network) cited references because this community will eventually link two research communities found in the neuroscience literature. Two of the psychology communities have authors in common with communities in education research.
The psychology co-citation network (trimmed). Node size is scaled to the number of other articles to which the reference is linked at threshold five (range 2–22). Edges are scaled to co-citation weights (range 5–14).
The largest psychology research community contains, almost exclusively, publications by neuroscientists and cognitive neuroscientists. Miserendino, Sananes, and Melia (1990) is central to this network and is co-cited with numerous articles in this literature. This article on cellular mechanisms in the amygdala underlying fear-potentiated startle captures the essence of this community with its focus on synaptic mechanisms. Many of the cited references in this community, and in particular those along the long spine of the community network, are authored or coauthored by Ledoux and his collaborators, Quirk and Rogan (LeDoux, 2000; Ledoux, Iwata, Cicchetti, & Reis, 1988; Quirk, Armony, & LeDoux, 1997; Quirk, Repa, & Ledoux, 1995; Rogan & Ledoux, 1995; Rogan, Staubli, & LeDoux, 1997). This basic neuroscience research investigates long-term potentiation (LTP) in the hippocampus, making the case that LPT is the synaptic mechanism underlying learning and memory. Ledoux and colleagues employ fear conditioning in rodents to investigate this mechanism as well as to study emotion circuits in the brain. Toward the top of the community structure, beginning with LaBar, Gatenby, Gore, LeDoux, and Phelps (1998) and including Bechara et al. (1995), studies on the human amygdala and hippocampus appear. On the right branch, the Bliss citations (Bliss, 1973; Bliss & Collingridge, 1993) report work on LTP in animals (McClelland, McNaughton, & Oreilly, 1995), on connectionist models of learning and memory, and (Scoville & Milner, 1957) on amnesia also appear in this area of the network. Thus, it appears that animal work predominates in this community, but human studies appear on the edges of the community network. Research at the level of synapses and circuits conducted on animal models would appear to be at a substantial remove from educational practice. For brevity, I will call this the synaptic mechanisms community.
A second research community again contains articles on motivation, with Ames (1992), Dweck and Leggett (1988), and Elliot and Church (1997) being most prominent. This is the motivation community in psychology. Here one sees motivation research embedded in a psychological context, rather than limited to applications in education. The motivation community in the psychology literature (12 cited references) is much smaller than the motivation community within education research (54 cited references). All authors of cited references in the psychology motivation community appear in the education research community as do 10 of the 12 cited references.
In the third component, Baddeley (1986) and Miller (1956) are highly cocited. The upper portion of the community network contains cognitive psychological research on numerical cognition (Geary, 1993) and reading (Daneman & Carpenter, 1980; Siegel & Ryan, 1989; Wagner & Torgesen, 1987). The lower portion of the network, connected to Miller (1956), contains cited references familiar from the cognitive load community in education research. I will call this the working memory community. When we look at the combined analysis, it is this psychology community that links the cognitive load theory community within the education literature to the other learning science research communities. Also, the motivation community within psychology coalesces with the motivation community in education.
The fourth, small research community in psychology contains two cited references in the trimmed network but four references in the untrimmed version: Gallagher, Graham, and Holland (1990); Baxter, Parker, Lindner, Izquierdo, and Murray (2000); Schoenbaum, Chiba, and Gallagher (1999); and Schoenbaum, Chiba, and Gallagher (1998). The research in this community is on the role of orbitofrontal cortex in learning.
Document co-citation analysis indicates that there are three major research communities within the psychology of learning (synaptic mechanisms of learning and memory, motivation, and working memory) and smaller one of eventual interest (orbitofrontal cortex in learning). Motivation and working memory share cited references with comparable communities in education research. Neuroscience is prominent with psychology, but the neuroscience tends to favor basic research on cellular mechanisms using animal models.
Neuroscience of Learning
The two largest research communities in the neuroscience of learning include nearly 90% of all the cited references appearing in the co-citation network (Figure 3). Schultz, Dayan, and Montague, (1997) is the central node among a cluster of articles that report work on reward conditioning. Most of the 41 papers in the cluster are unique to the neuroscience literature but four also appear as cited references in the psychology community on synaptic mechanisms: Bechara, Damasio, Damasio, and Anderson (1994); Schultz et al. (1997); Knowlton, Mangels, and Squire (1996); and Rolls and Baylis (1994).
The neuroscience co-citation network (trimmed). Node size is scaled to the number of other articles to which the reference is linked at threshold five (range 2–32). Edges are scaled to co-citation weights (range 5–18).
The trimmed version of largest neuroscience community contains 82 cited references, 27 of which also appear in the synaptic mechanisms community in psychology. The similarities are sufficient to consider this the neuroscience network on synaptic mechanisms. If one looks at the upper right branch of this network starting from Bourtchuladze et al. (1994), one first sees cited work by Ledoux and colleagues that appeared in the psychology literature and toward the end of the branch (Bechara et al., 1995; Labar, Ledoux, Spencer, & Phelps, 1995) on studies of human amygdala and hippocampus that also appeared in psychology literature. A paper by Bliss also appear within this community (Bliss & Collingridge, 1993) on LTP in the hippocampus as a synaptic model of memory. Starting form Bliss, the lower right branch includes O’Keefe (1978) and Scoville and Milner (1957), also present in the psychology co-citation network.
The majority of work found in the neuroscience of learning co-citation network consists of animal studies on the cellular and synaptic mechanisms underlying learning and memory. Again as in psychology, human work appears at the periphery of the cluster. This community on mechanisms is very similar to the comparable synaptic mechanisms community within psychology. The other major research focus is on reward conditioning.
Educational Neuroscience
The educational neuroscience co-citation network literature contains 54 cited references. Of these, 46 appear in one major research community: the educational neuroscience community. The trimmed pathfinder network as shown in Figure 4 contains 12 cited references.
The educational neuroscience co-citation network (trimmed). Node size is scaled to the number of other articles to which the reference is linked at threshold five (range 2–20). Edges are scaled to co-citation weights (range 5–16).
This network is centered on Bruer (1997). The cited references appearing in this network have an odd feature, which differentiates the educational neuroscience research community from other communities we have discussed. In the other literatures, the communities could be readily identified as research subfields within the larger research area. In education research, there was a community on science instruction; in the psychology literature, a community on working memory; and in the neuroscience literature, a community on reward conditioning. The educational neuroscience community is different. Rather than a research community, it is a meta-community about educational neuroscience.
If one examines the cited references in Figure 4, one finds that they fall into two categories. Eleven of the articles discuss the merits and demerits, promise, and pitfalls of attempting to apply neuroscientific research to education. Most of the articles make a positive case for the importance and future development of educational neuroscience: Byrnes and Fox (1998); Fischer et al. (2007); Goswami (2004, 2006); Hinton and Fischer (2008); and Willingham and Lloyd (2007). Many of these articles provide arguments and counterexamples to Bruer (1997). Stern (2005) is an editorial critical of aspects of the educational neuroscience effort. Weisberg, Keil, and Goodstein (2008) present data on how neuroscientific evidence affects appraisal of psychological explanations.
Fischer and Bidell (2006) is a review of work on the dynamic development of thought and action for which the authors have suggested there is neuroscientific support.
Dehaene, Piazza, Pinel, and Cohen (2003); Temple et al. (2003); and B. A. Shaywitz et al. (2002) are instances of the second type of article one finds in this community. These articles present cognitive neuroscientific research in the areas of numerical cognition (Dehaene) and more commonly on reading and dyslexia (Temple and Shaywitz). Within the educational neuroscience literature, articles like these serve as “exemplar” articles, articles that are cited because they provide examples of the kind of cognitive neuroscience that does or might have educational relevance. In the untrimmed version of this network, B. A. Shaywitz et al. (2002) is linked to 11 other exemplar articles, not including Temple, all of which are studies of word recognition or dyslexia and 10 of which are brain imaging studies.
Thus, the educational neuroscience literature, as identified here, tends to cite articles that are not about research in educational neuroscience but about the field of educational neuroscience itself. Currently, at the foundation of the educational neuroscience literature is a meta-literature, not a research literature. (Alas, the article you are now reading is another contribution to the meta-literature.) It is not clear how this meta-literature will connect to education research and the psychology and neuroscience of learning.
Integrating Education, Psychology, Neuroscience, and Educational Neuroscience
Examining the four literatures individually has revealed 10 research communities within the literatures. For educational neuroscience, there is a single educational neuroscience community. Within educational research, there are three major communities: science education/cognition, motivation, and cognitive load theory. There are three large research communities within psychology: synaptic mechanisms, motivation, and working memory as well as the small community on the orbitofrontal cortex and learning. Two major research communities emerged from the neuroscience of learning literature, synaptic mechanisms, and reward conditioning.
A document co-citation analysis of the four combined literatures, using a co-citation threshold of five, shows the connections and relations among the 10 identified research communities. From hereon, our primary interest will be the co-citation linkages that obtain among the research communities. For this reason, the nodes in the following co-citation networks, rather than being the individual cited references, will be the 10 research communities. The internal structure of the communities changes very little as the various literatures are combined. The edges connecting the communities will be made explicit, showing the pairs of cited references across communities responsible for the linkage. The most interesting finding that emerges from these combined analyses is that the psychology literature and its research communities account for all the connections among research communities that arise in the combined analysis. No combinations of the other three literatures generate any co-citation links between their six original research communities.
Figure 5 shows the results of the combined co-citation analysis. First, note that although the motivation communities within education research and psychology merge into a single community in the combined analysis, motivation still remains a separate research community. None of the co-citation linkages between the communities shown in Figure 5 exceed weight six. The synaptic mechanism communities in neuroscience and psychology are essentially the same community in both literatures. They coalesce in the combined network. However, in the combined network reward conditioning, a separate community within neuroscience becomes linked to the coalesced synaptic mechanism community. The orbitofrontal cortex community identified within the psychology community is responsible for this connection. Gallagher et al. (1990) from the psychology orbitofrontal cortex community becomes linked with Killcross, Robbins, and Everitt (1997) in the neuroscience synaptic mechanisms community, and Schoenbaum et al. (1999) from the orbitofrontal community becomes linked with Watanabe (1996) from the neuroscience reward-conditioning community.
Co-citation linkages between research communities. The co-citation linkages between research communities never exceed weight six.
Next, educational neuroscience becomes connected to the neuroscience and psychology synaptic mechanisms community. A connecting edge of weight five links (Immordino-Yang, 2007) the last node in the original educational neuroscience community with (Damasio, 1994) the last node in the synaptic mechanisms community. This link appears because the total co-citation of Immordino-yang (2007) and Damasio (1994) reach threshold in the combined literatures. Immordino-yang (2007) discusses the implications of emotion and social cognition for education. Damasio (1994) is a book written for a general audience that presents his research, much of which is relevant to the study of emotion. The five articles that co-cite Immordino-yang (2007) and Damasio (1994) are generally about the benefits of the mind, brain, and education approach to education and education research. Three of the five co-citing articles were published in the journal Mind Brain and Education, and three of the five articles were written by Immordino-yang. Immordino-yang and Damasio have also coauthored an article on emotion and learning (Immordino-Yang & Damasio, 2007).
Bradley and Bryant (1983) occurs as a cited reference in both the educational neuroscience and the psychology literatures. It is the last node in educational neuroscience community, connected to B. A. Shaywitz et al. (2002) at weight five. Within psychology’s working memory and cognition community, Bradley and Bryant (1983) is connected to Wagner and Torgesen (1987) at weight six. The articles wherein these co-citations appear present research on reading abilities, phonological awareness, and dyslexia (see e.g., Ehri, Nunes, Stahl, & Willows, 2001; Ehri, Nunes, Willows, et al., 2001). These co-citing articles tend to be published in education–psychology journals, such as Reading Research Quarterly.
The working memory and cognition community are linked to the cognitive load community via the co-citation of two classic, cognitive psychology publications (Baddeley, 1986 and Miller, 1956) on working memory that occurs within the psychology literature. Toward the Baddeley end of the merged communities, one finds cited references from the psychology community on working memory that includes work on numerical cognition (Geary, 1993) and reading (Daneman & Carpenter, 1980). Toward the Miller end, one finds cited references to work in cognitive load theory (e.g., Kalyuga, Ayres, Chandler, & Sweller, 2003). Articles co-citing Baddeley and Miller also include work by psychologists on more general aspects of working memory such as Burgess and Hitch (1999). The co-citing articles appear in a mix of educational psychology and psychology journals.
Sweller et al. (1998) and Renkl (1997) provide the co-citation link between cognitive load theory and the science education/cognition community. Renkl is coauthor on three of the six co-citing articles. All of the co-citing articles address how classroom learning is more efficient, when instructional design is sensitive to students’ working memory limitations. Atkinson, Derry, Renkl, and Wortham (2000) and Kalyuga, Ayres, Chandler, and Sweller (2003). The co-citing articles are published in educational psychology and education research journals.
Table 1 showed that the psychology literature shared the largest number of cited references with the other three fields. It is for this reason that all of the merging and linking of the original 10, distinct research communities into a single giant component, plus motivation, arises from co-citation linkages attributable to the psychology literature. The psychology literature plays a significant role in connecting neuroscientific research and educational neuroscience with education research.
Conclusion
Where is educational neuroscience? Based on the literatures used here and document co-citation analysis, educational neuroscience is near the middle of the 10 research communities identified in the original co-citation analyses of the educational neuroscience, education, psychology, and neuroscience literatures. It is connected to neuroscientific–psychological research on synaptic mechanisms by a common interest among neuroscientists and educational neuroscientists in the role of emotion in effective learning. This is a new and fragile connection but possibly one that will strengthen over time. It is not now clear how this line of research will influence what we have seen to be the major foci of education research, science education, cognitive load, and motivation in classroom and instructional settings. An eventual link with the, still isolated, motivation research community would seem most likely. At its other end, the educational neuroscience community, is linked to research on working memory and cognition. This link arises via shared interests in these two communities in reading and dyslexia. This connection of educational neuroscience with cognitive psychology arises through educational neuroscience’s “exemplar” literature, typified by B. A. Shaywitz et al. (2002). We can be encouraged that educational neuroscience is connected to other research communities within the learning sciences, but we still must be concerned about its apparent distance from education research and educational practice.
It is true that these co-citation linkages are at, or near, threshold. Based on pathfinder network analysis, the minimal spanning trees identify the strongest links between these literatures and their respective research communities. On the other hand, given the generality of the search strategies used to identify the relevant literatures, one might be surprised that a co-citation network connecting most of the research communities in these four research fields exists. Maybe its like the dancing bear. The communities may be only weakly connected, but it is remarkable that they connected at all.
What about the “meta” problem within the educational neuroscience literature? One might dismiss the tendency to reflect on one’s own novel research program and to defend it from detractors as part of the emergence and maturation of any new field. Of course, one would like more evidence that the meta-tendency is characteristic of new research fields. The “meta” problem will dissipate if and when educational neuroscience delivers usable results, maybe when we begin to see the presence of neuroscientific ideas within the education research literature with greater frequency. Part of the problem may be that educational neuroscience was born under a bad star. Educational neuroscience emerged in the mid to late 1990s. Neuroscience was in the news as the result of claims made by early childhood advocates about the implications of neuroscience (primarily developmental neurophysiology) for learning and parenting. These claims were based on overgeneralization and oversimplification of long-standing findings about developmental synaptogenesis, critical periods, and enriched environments. Bruer (1997, 1999) argued that these claims were overblown. These claims became characterized in the literature as “neuro-myths” (Goswami, 2004; Organization for Economic Cooperation and Development [OECD], 2007). Educational neuroscientists recognized the need to differentiate their research program from the kind of pop-neuroscience found within the policy community and the mass media. Educational neuroscientists and its critics devoted energy and publications to arguing about what kind of bridges, if any, could be made between neuroscience and educational practice. This original debate still casts a long shadow over the structure of the educational neuroscience co-citation network. If bridges are built, there will be no need to argue about whether they can be built.
Bruer (1997) also suggested that as a practical research strategy, it seemed reasonable to attempt to bridge from education to cognitive neuroscience via cognitive psychology. The analyses earlier give some support to that suggestion. The closet connection of neuroscience or educational neuroscience to education research is via cognitive neuroscientific studies of dyslexia and cognitive psychological research on working memory. All of the connections between the research communities shown in Figure 5 are due to co-citations that arise from within the psychology literature. It would seem psychology remains an important bridging discipline.
Where might educational neuroscience go in the future? One place I would look is at the cited references that appear on the edges of the synaptic mechanisms community. What I have glibly dubbed synaptic mechanisms research is at a finer level of analysis a network of varied, interrelated research programs. At the ends of that community structure, cited references to articles that report work on human cognition, rather than on animal models, begin to appear. This seems a likely place for future merging of research interests within the learning sciences. Already, the existing link between educational neuroscience and neuroscience is via the path from Immordino-Yang (2007) to Damasio (1994), which is an end node in the untrimmed synaptic mechanism network. At this end of the network, the cited references also include LaBar et al. (1998) and Bechara et al. (1995), which are studies on the human amygdala and hippocampus. Bechara’s work is also cited in Immordino-Yang and Damasio (2007). On another end of this network, on a branch emanating from Hebb (1949), one finds articles by McClelland, McNaughton, and Oreilly (1995) on connectionist models of learning and memory and Knowlton et al. (1996) on human habit learning. These seem likely areas where basic research and computational modeling might eventually be linked more closely with educational psychology and education research.
One virtue of co-citations analysis is that is it forward looking. Citation and co-citation patterns change as research moves forward. This initial attempt to locate educational neuroscience within the learning sciences is exactly that, an initial attempt. Let’s see where educational neuroscience is 10 years from now.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.