Annual publication volume reflects the research output and developmental trajectory of a research field. Based on 2166 eligible articles published between January 2016 and May 2026, a publication trend graph was generated (see Figure 1Figure 1).

Figure 1. Publications on N400 (2016-May 2026).

As illustrated in the figure, publications on N400 research showed an overall fluctuating upward trend from 2016 to 2025, with output remaining at a high level. This period can be divided into three distinct stages. First, the phase of steady development spanned 2016 to 2018. Annual publications ranged from 170 to 182 with moderate growth, indicating stable research interest and continuous refinement of fundamental theories in this field. Second, the period from 2019 to 2020 witnessed rapid growth. The number of publications rose sharply from 182 in 2018 to 225 in 2019 and 230 in 2020, forming the first minor peak within the research timeframe. The average annual publication count exceeded 220 during this stage, accounting for 22.36% of the total publications. These figures reveal a research boom in N400 studies and a marked acceleration in research output. Third, the years 2021 to 2025 represented a period of high-level fluctuation and renewed growth. Although publication numbers declined slightly, they stayed above 200 per year and reached the highest point of 241 in 2025. It demonstrates that research enthusiasm did not wane; after a brief adjustment, the field entered another phase of rapid expansion with an enlarging research scale.

It should be noted that the publication data for 2026 only covered records up to May, with 110 publications in total, which cannot reflect the full-year output. Overall, the fluctuating increase in annual publications from 2016 to 2025 suggests growing global academic attention to N400 research. This field maintains strong academic vitality and sustained research interest, with promising prospects for future development.

Keywords are concise representations of research themes. Keyword co-occurrence analysis can intuitively reveal research hotspots, core topics, and thematic correlations within a field. In this study, standardized sample data were imported into CiteSpace, with the time span set from January 2016 to May 2026. Keywords were selected as the analysis item to generate the keyword co-occurrence map of N400 research.

To capture the micro-dynamic evolution of annual research hotspots in the N400 field, we set the CiteSpace time-slicing parameter to 1 year. For node threshold selection, we tested three schemes with Top N values of 50, 60, and 80, respectively. Comparative analysis revealed that the core cluster compositions and the distributions of the top 15 burst terms did not change substantially across the three schemes, indicating that the analytical results are not sensitive to the threshold choice. Further examination showed that at N = 50, high-centrality nodes were over-pruned, whereas at N = 80, increased noise nodes led to a looser modular structure. Based on a comprehensive balance between cluster silhouette clarity and network coherence, we ultimately adopted Top N = 60 as the threshold.

Using the optimized parameters, the resulting keyword co-occurrence network consists of 158 nodes and 180 links, with a network density of 0.0145. Node size is positively correlated with keyword frequency, and link thickness is positively correlated with keyword co-occurrence frequency. The overall network presents a clear structure with prominent core topics and a complete hierarchical layout (See Figure 2Figure 2).

Figure 2. Keyword co-occurrence knowledge map.

A total of 158 keywords were extracted in this study, with a cumulative total frequency of 10,141. Based on the literature from the Web of Science (WOS) Core Collection, the top 30 keywords are summarized in Table 1. Centrality indicates the importance of each keyword within the keyword co-occurrence network and reflects the prevailing research hotspots and themes in this field during the study period.

Table 1. Keyword frequency table.

The keyword co-occurrence network visually reveals a tiered distribution pattern of research themes within the collected literature, which can be summarized as core, intermediate and peripheral thematic clusters based on keyword frequency and centrality metrics. The core cluster gathers foundational technical keywords such as N400, ERP and brain potentials, reflecting the fundamental methodological framework adopted across most N400 publications. The intermediate cluster centers on terms related to linguistic cognition (language comprehension, semantics, sentence processing etc.), corresponding to the mainstream research directions that dominate existing empirical studies. The peripheral cluster covers keywords about individual variation and cross-language paradigms (age, older adults, Chinese etc.), which capture emerging expanding research themes in recent years. This hierarchical grouping merely describes the distribution of published research topics visualized via bibliometric mapping, rather than reflecting inherent hierarchical neural mechanisms of N400 semantic processing.

Keyword clustering classifies and summarizes research themes based on keyword co-occurrence maps. It can directly display core thematic clusters and the distribution of research hotspots, and serves as a core method to sort out the research system of a field.

Using the clustering function of CiteSpace, this study conducted a keyword clustering analysis on N400-related research from 2016 to May 2026 and generated a keyword clustering map. The modularity Q of the map is 0.8135, which is much higher than the critical value of 0.3, indicating a statistically significant cluster structure. The weighted mean silhouette S is 0.9419, exceeding the critical value of 0.7. This result demonstrates that the clustering outcomes are highly reliable and can fully and clearly reflect the distribution of research topics in this field (See Figure 3Figure 3).

Figure 3. Keyword clustering network.

It can be observed that ten major clusters are identified, namely cognitive control, mismatch negativity, sentence comprehension, language comprehension, networks, word recognition, lexical decision, familiarity, semantic processing, and comprehension. Accordingly, N400 research from 2016 to May 2026 is closely associated with cognitive control, language comprehension, semantic processing, word recognition and brain network mechanisms, forming a complete and logically coherent research framework.

When sorting the ten keyword clusters by research content, we can inductively divide these published thematic lines into three broad thematic categories for interpretive convenience: foundational neural mechanism studies, core linguistic application research, and extended frontier explorations. This categorical grouping is a subjective inductive summary of bibliometric clustering outputs, not an empirically validated structural framework of N400 cognitive function. The basic mechanism layer consists of Cluster #0 (cognitive control), Cluster #4 (networks) and Cluster #1 (mismatch negativity). It focuses on fundamental mechanisms of neural cognition, brain networks and ERP components, laying theoretical and technical foundations for the entire field. The core application layer includes Cluster #3 (language comprehension), Cluster #2 (sentence comprehension), Cluster #5 (word recognition), Cluster #6 (lexical decision), Cluster #8 (semantic processing) and Cluster #9 (comprehension). It centers on application scenarios covering the whole process of linguistic cognition and represents the core focus of N400 research. The in-depth extension layer is represented by Cluster #7 (familiarity), which explores influencing factors and regulatory mechanisms of the N400 effect. This research branch also expands continuously to cross-language studies, population differences and multimodal integration, reflecting the cutting-edge development trends of this field.

Keyword burst indicators identify terms with dramatic changes in occurrence frequency, which helps to explore the research frontiers of N400 studies. Standardized data were imported into CiteSpace, with the time range set from January 2016 to May 2026. The top 60 high-frequency items were retained for each time slice, and keywords were selected as the analysis unit. Burst detection was conducted based on the keyword co-occurrence map to generate the visualization of the top 25 burst keywords in N400 research from 2016 to 2026.

The figure shows that semantic integration had the highest burst strength of 8.96 during 2018 to 2020, followed by word frequency with a burst strength of 8.27 from 2023 to 2026. The term mismatch negativity (MMN) maintained a burst duration of six years from 2017 to 2022, indicating long-term and sustained academic attention to this research topic.

In terms of chronological evolution, violations emerged as the earliest research hotspot between 2016 and 2017. Studies in this period mainly focused on N400 responses to semantic violations, establishing fundamental experimental paradigms for subsequent research on semantic processing. From 2018 to 2020, semantic integration, upcoming words and speech perception became successive burst keywords. Researchers began to investigate the effects of contextual prediction and verbal input on the N400 effect, shifting the research focus from isolated semantic conflicts to dynamic semantic processing. Between 2021 and 2023, new burst keywords including older adults, syntax and concreteness appeared. Research extended to different populations and linguistic structural features, with particular attention paid to cognitive aging among the elderly and the interaction between syntactic and semantic processing. From 2023 to 2026, word frequency, age, Chinese and network became newly emerging hotspots. It demonstrates that N400 research has expanded from traditional Indo-European languages to logographic writing systems such as Chinese. Combined with brain network analysis and individual developmental differences, the field has achieved progress in both research breadth and depth.

According to the temporal distribution and research content of burst keywords, current frontiers of N400 research mainly include semantic processing mechanisms across languages (especially Chinese), the regulatory effects of word frequency and linguistic experience, dynamic characteristics of brain networks, and linguistic cognitive differences among older adults. These emerging hotspots provide clear directions for future research (See Figure 4Figure 4).

Figure 4. Keyword burst plot.

Timeline maps adopt horizontal bands ordered by time to present the intellectual trajectory of research evolution. They clearly illustrate the distribution, evolutionary paths and developmental trends of research hotspots, enabling researchers to accurately grasp the latest advances in a given field. In CiteSpace, keyword timeline maps arrange keywords chronologically from left to right. The size of circular nodes is positively correlated with keyword frequency; larger nodes represent higher research popularity and more abundant research outputs.

This study analyzed the evolution of research hotspots using CiteSpace. The time span was set to 2016-2026 (up to May) with a one-year time slice. The top 60 high-frequency keywords were retained for each time slice, and keywords were selected as the node type. The generated keyword cluster timeline map is presented below. The map contains 158 nodes and 180 links. The modularity value Q = 0.8135 indicates significant cluster structures and reasonable classification of research topics. The weighted mean silhouette value S = 0.9419 verifies the high reliability of clustering results, which can effectively reflect the evolutionary characteristics of N400 research (See Figure 5Figure 5).

Figure 5. Keyword timeline map.

From an overall perspective of the map, the evolution of keywords along the timeline shows a shift from a focus on fundamental topics to diverse research directions. The total number of nodes increases over time, indicating growing research interest and a continuously refined research system in the field of N400. Besides the core terms N400 and event-related potential (ERP), nodes for brain potentials, language comprehension and semantic integration are relatively large. These topics have remained stable research hotspots over the years, which also demonstrates the close links between N400 research and language comprehension, semantic processing, as well as basic electrophysiological studies. These themes form the core research framework and underpin the overall development of N400-related studies.

On the right side of the map are labels for ten keyword clusters: #0 cognitive control, #1 mismatch negativity, #2 sentence comprehension, #3 language comprehension, #4 networks, #5 word recognition, #6 lexical decision, #7 familiarity, #8 semantic processing, and #9 comprehension. These clusters follow a clear evolutionary pattern on the timeline. Early research centered on clusters related to basic mechanisms, such as #0 cognitive control and #1 mismatch negativity, which explored the neural regulatory mechanisms of N400 and fundamental properties of ERP components. In the middle stage, research gradually expanded to clusters concerning language processing, including #2 sentence comprehension and #5 word recognition, covering language comprehension at lexical and sentential levels. In the later period, studies extended to brain networks, cross-language research and individual differences. Clusters such as #4 networks and #7 familiarity remained active from 2023 to 2026, revealing a general trend that research has evolved from single-component analysis to multi-dimensional and interdisciplinary integration. Dense connections within each cluster reflect strong internal relevance and the gradual enrichment of relevant knowledge systems. Cross-temporal and cross-topic connections between different clusters also highlight the interdisciplinary integration of N400 research, which jointly drives the sustained progress of this field.