FIGURE 3.

Summary of co-keyword analysis completed across the Full Corpus, and both Decades. (A) Keyword co-occurrence matrices are visualized using graph theory methods, and modularity analytics applied to identify self-clustering modules. These modules are initially visualized within the context of the overarching network, with each highlighted by a block color (e.g., two modules are initially identified at the full decade, while four are identified for Decade 2). (B) To aid accessibility, these modules are visualized as colored concentric rings, each reflecting a module’s singular rank and theme (Psychological: Blue; Physiological: Red; Green: Interdisciplinary). Node size reflects normalized Eigenvector centrality – with larger nodes having higher levels of normalized Eigenvector centrality, and thus, empirically assessed as dominant across the module; such prominent internal nodes are labeled (see Supplementary Section 1 for full node listing used for thematic coding). (C) Modular internal and external connectivity metrics were extracted, normalized, and visualized relative to the normalized corpus median values via a strategic diagram. Providing a summary of internal connectivity along the Y-axis, this provides a measure of the cohesiveness of a thematic trend – with highly developed and interconnected themes displaying higher levels of internal connectivity. Summative metrics on the X-axis provide an overview of external connectivity, demonstrating the central dominance of a theme to the research domain. The range of corpus internal and external connections display significant growth between Decade 1 and Decade 2, in line with research proliferation (see Figure 2), coupled with simultaneous constriction in modularity, perhaps indicative of the development of cohesive research themes (Figure 2D)—see Supplementary Table 4. For yearly (1994, 2006, 2015) assessment see Supplementary Figure 3. Further, for larger high resolution figures please see https://figshare.com/s/9a0f7b0839fef2df1a6e.