Content area
The attention hypothesis, which assumes that font emphasis captures readers' attention, is usually used to explain the mechanism by which such emphasis operates. This study further delineates the attention hypothesis by investigating the ways in which font emphasis captures attention and its effects on the integration of emphasized information into the previous context. We computed event-related potentials and frequency band-specific electroencephalographic power changes occurring while participants read sentences containing critical words that were either emphasized (i.e., displayed in a color different from the other words in the sentence) or not (i.e., shown in the same color as the rest of the sentence) and semantically congruent with prior words or not. The results showed that the emphasized words (as compared to control words) elicited a reduced N1 and increased P2, indicating that font emphasis reduced familiarity-based visuo-orthographic processing and instead increased controlled attentional processing. We also observed greater P300 and power decreases in the alpha and beta frequency range in response to critical words in the emphasized condition, suggesting that font emphasis enhances focal attention to promote a fuller integration of information into the sentence context. Furthermore, relative to the control condition, the emphasized condition induced delta and theta power increases for the incongruent words. These results suggest that font emphasis increases the efficiency of glyph processing, which facilitates lexical access.
Accepted: 16 August 2023 / Published online: 15 September 2023
©The Psychonomic Society, Inc. 2023
Abstract
The attention hypothesis, which assumes that font emphasis captures readers' attention, is usually used to explain the mechanism by which such emphasis operates. This study further delineates the attention hypothesis by investigating the ways in which font emphasis captures attention and its effects on the integration of emphasized information into the previous context. We computed event-related potentials and frequency band-specific electroencephalographic power changes occurring while participants read sentences containing critical words that were either emphasized (i.e., displayed in a color different from the other words in the sentence) or not (i.e., shown in the same color as the rest of the sentence) and semantically congruent with prior words or not. The results showed that the emphasized words (as compared to control words) elicited a reduced N1 and increased P2, indicating that font emphasis reduced familiarity-based visuo-orthographic processing and instead increased controlled attentional processing. We also observed greater P300 and power decreases in the alpha and beta frequency range in response to critical words in the emphasized condition, suggesting that font emphasis enhances focal attention to promote a fuller integration of information into the sentence context. Furthermore, relative to the control condition, the emphasized condition induced delta and theta power increases for the incongruent words. These results suggest that font emphasis increases the efficiency of glyph processing, which facilitates lexical access.
Keywords Font emphasis * Attention * Semantic integration * ERP * Brain oscillations
Introduction
Successful textual comprehension not only requires attention to the meaning of textual symbols (i.e., semantic information), but also to the hidden meanings behind such symbols (i.e., pragmatic information). Font emphasis is a paralinguistic cue that contains hidden meaning. It uses typography (e.g., boldface, italics, underlining, capitals, color variations, etc.) to convey the author's intention (Sanford et al., 2006), thus requiring additional processing by readers.
Can readers recognize the hidden meaning of font emphasis? Linguistic research has shown that readers do realize that emphasized words hold meaning beyond what they would if they appeared in plain case. For example, readers generally believe that capitalization strengthens the semantic strength of information (McAteer, 1992). Italics highlight the importance of information and enhance the contrast between texts (McAteer, 1992; Sanford et al., 2006). Converging evidence comes from psycholinguistic research, showing that emphasized information is processed differently from ordinary information during reading. For example, reading times of emphasized information increase when readers can only read one sentence/word at a time and not look back at previous information (Fraundorf et al., 2013; Lorch et al., 1995; McAteer, 1992). In normal reading, font emphasis attracts readers' attention, causing them to preview the word (Rayner & Schotter, 2014) and accelerate its lexical processing (Wu et al., 2021). By using a text-change paradigm, Sanford et al. (2006) found that font emphasis increased the probability of detecting words that changed across two consecutive presentations, indicating that information marked by font emphasis is more extensively processed.
Readers' differentiated treatment of emphasized information is also reflected in reading outcomes. A significant body of research has found that font emphasis improves readers' memory representation of emphasized information (e.g., Filik et al., 2006; Lorch, 1989; Lorch et al., 1995) and helps readers build more elaborate connections between emphasized information and associated content (Fraundorf et al., 2013; Wu et al., 2021). Taken together, both linguistic and psycholinguistic studies have shown that readers can recognize the hidden meaning emphasized by fonts and treat the emphasized information differently from the surrounding context.
The attention hypothesis is usually used to explain the mechanism of font emphasis. It assumes that font emphasis captures readers' attention such that the emphasized information receives different treatment, leading to deeper processing and stronger memory representation (Lorch, 1989; Lorch et al., 1995; Sanford et al., 2006). However, despite widespread support, the attention hypothesis does not explain in detail how font emphasis affects language processing during reading. It does not explain how font emphasis prompts the allocation of more attention resources to promote reading comprehension. Research of font implies some of the underlying mechanisms. Font is concerned only with the surface form of the writing system. However, electrophysiological experiments across various paradigms and writing systems have consistently shown that fonts affect the processing of words, including cognitive load and early perceptual/pre-lexical processes, both of which are reflected in the N1 component. N1 is sensitive to cognitive load when qualitative changes in font are manipulated (Bar-Kochva & Breznitz, 2012; Sulpizio & Job, 2018; Zhou et al., 2016). For example, Bar-Kochva and Breznitz (2012) found that words with diacritics elicited larger N1 amplitudes than did words lacking diacritics. Sulpizio and Job (2018) found that words starting with an upper-case letter elicited a larger N1 than did words starting with a lower-case letter. These authors suggested that upper-case letters would require more attentional demand than would lower-case letters, due to their low familiarity and frequency in the Italian writing system, making the process more difficult. In addition to cognitive load, N1 reflects the early perceptual/pre-lexical processes involved in the mapping of visual features onto locationspecific letter representations. This occurs when there is a slight change in font, for example, when a familiar/easy font has a larger N1 amplitude than a relatively unfamiliar/difficult one (Kuchinke et al., 2014; Vergara-Martinez et al., 2020; Xue et al., 2019).
Font emphasis changes the visual features of select words in a sentence/discourse. The change is subtle, as letter shapes are preserved. However, such a change alters the font familiarity of the emphasized words, as the visual form is rare and unexpected as compared to the other words. Thus, it is quite possible that during reading, font emphasis affects familiarity with emphasized words, making it more difficult to map emphasized words to higher word representations. At the same time, based on the fact that infrequent unfamiliar stimuli raise the level of arousal (Ulrich et al., 2006), the unfamiliar font used for the emphasized words may lead readers to perceive that font as signifying the importance of the words for which it is being used, and consequently deploy more attention resources in those areas.
In addition, the attention hypothesis alone does not fully describe what processes font emphasis affects during reading. In the normal reading process, the reader must recognize the vocabulary, combine it with the previous context, and further connect it with the following content. Whether and how font emphasis influences word recognition and its integration with subsequent associated content are well documented. Previous studies have shown that font emphasis influences word recognition (Macaya & Perea, 2014; Wu et al., 2021). In a word-recognition task, Macaya and Perea (2014) found that low-frequency words with font emphasis were responded to more rapidly than were words with no emphasis, whereas there were no signs of a parallel effect for high-frequency words, indicating that font emphasis favors the process of lexical access. Using an eyetracking technique, Wu et al. (2021) found font emphasis to shorten first-fixation and gaze durations associated with lexical access during sentence reading. The effect of font emphasis on word recognition further affects the integration of the word with subsequent content. According to the landscape model, concepts fluctuate in their activation as readers progress through a text (van den Broek, 1995; van den Broek et al., 1996). Font emphasis drawing readers' attention increases the activation of the emphasized information (Gaddy et al., 2001), which may make that information more likely to remain in processing cycles when other words are being read. This is supported by the fact that font emphasis improves the memory representation of non-emphasized content adjacent to the emphasized information (Cashen & Leicht, 1970). Additional support for this conclusion can be found in that font emphasis increases participants' ability to reject false statements about alternatives to emphasized information, suggesting that font emphasis helps readers build elaborate connections between emphasized information and associated content in their subsequent integration (Fraundorf et al., 2013; Wu et al., 2021). Compared with word recognition and its connection to the following content, the impact of font emphasis on the integration of emphasized information with the previous context is still underexplored. Both the verbal efficiency theory (Perfetti, 1985) and the lexical quality hypothesis (Perfetti, 2007) claim that word identification is a limiting factor in higher-level processes of comprehension such as integration. Font emphasis has been shown to impact word identification. Therefore, it is also likely to affect the integration of emphasized information with its previous context, just as it affects the connection to subsequent content.
As described above, the attention hypothesis is the accepted explanation for font emphasis. However, little is known about how font emphasis specifically affects language processing during reading. The main goal of the current experiment was to address this question by exploring the way font emphasis captures attention and determining its effect on the integration of the emphasized information with the previous context. To this end, we conducted a sentence-reading experiment in which critical words were either emphasized (i.e., displayed in a color different from the other words in the sentence) or not (i.e., shown in the same color as the rest of the sentence). Here, we used the color red to manipulate the emphasis state of critical words. During reading, color variation denotes importance for the color varied information (de Koning et al., 2009; Lemarie et al., 2008). It automatically captures attention (Sanford et al., 2006) and has been shown to be a powerful means of inducing a steady emphasis effect (Schneider et al., 2018; Wu & Yuan, 2003). Among the various colors that could be used, red ink on a white background provides a desirable level of legibility (Tinker & Paterson, 1931). Moreover, the color red facilitates the determination of the importance of Chinese idioms of low importance (Sun et al., 2019), indicating that text written in red, at least in Chinese, benefits from an increase in subjective relevance attributed to the conveyed information. Therefore, we explored the effect of font emphasis on sentence reading by manipulating critical words through the use of the color red. We also manipulated the semantic congruence of the critical words with respect to words prior, in order to explore the impact of font emphasis on semantic integration. In this way, four experimental conditions were compared: Emphasized/Congruent, Emphasized/Incongruent, Control/Congruent, and Control/ Incongruent (see Table 1 for example stimuli).
The first goal of the present study was to explore whether font emphasis encountered during reading affects familiarity with emphasized words and attracts more attention due to unfamiliar visual features. However, behavioral measures such as reading time, including all orthographic processing processes, cannot reveal whether font emphasis affects the visuo-orthographic processing of words. In addition, reading time is only an indirect indicator of attention allocation (Gaddy et al., 2001). Moreover, in previous reading studies, for the same manipulation, both increased and decreased reading times have been attributed to attention allocation (Birch & Rayner, 1997, 2010; Chen & Yang, 2015; Lowder & Gordon, 2015). This makes reading time a somewhat invalid indicator. Instead, the electroencephalogram (EEG) has a sufficiently high time resolution to capture brain activities that reflect different stages of cognitive processes. Therefore, the effect of font emphasis was recorded by an EEG in the present study. Given that EEG signals contain both temporal and neutral oscillation information, event-related potentials (ERPs) and oscillatory patterns were combined to reveal the cognitive neural mechanism of font emphasis.
When there is a slight change in font, N1 reflects the early perceptual/pre-lexical processes involved in the mapping of visual features onto location-specific letter representations (Vergara-Martinez et al., 2020). Previous research has connected P2 to the allocation of attention (Hillyard & Miinte, 1984; Luck & Hillyard, 1994). A decrease in alpha power is functionally related to increased attention (Gerven & Jensen, 2009; Klimesch, 1999; Marshall et al., 2015). Therefore, per our hypothesis, if font emphasis encountered during reading reduces familiarity with emphasized words and attracts more attention due to unfamiliar visual features, then critical words in the emphasized condition (i.e., emphasized words) were predicted to be more difficult to map onto locationspecific letter representations and would receive more attentional resources than critical words in the control condition (i.e., control words). Correspondingly, the emphasized words were expected to elicit a smaller Nl, larger P2, and decrease in alpha power as compared to the control words.
Furthermore, based on the notion that font emphasis facilitates word recognition (Macaya & Perea, 2014; Wu et al., 2021), as well as verbal efficiency theory (Perfetti, 1985) and the lexical quality hypothesis (Perfetti, 2007) arguing that word identification is a factor influencing higher-level processes of comprehension (e.g., integration) such that more resources can be reserved to support comprehension processing if less resources are consumed by lexical-level processing, we expected that the processes for integrating the meaning of a particular word into a higher-order semantic interpretation would be easier in the emphasized condition and those in the control condition would be more effortful. N400 is a negative-going ERP that is sensitive to the reprocessing of semantically anomalous information; it occurs between 300 ms and 500 ms after stimulus onset and peaks at about 400 ms (Kutas & Hillyard, 1980b; van Berkum et al., 1999). Consequently, in contrast to the control condition, incongruent words in the emphasized condition were predicted to induce a larger N400. In addition, words in the emphasized condition were presented in a regular font, up to the key words. Thus, the visual form of the emphasized words was rare and unexpected as compared to other words. P300 is an ERP component highly sensitive to the occurrence of unexpected, surprising, or discrepant stimuli (Kutas & Hillyard, 1980a). It is similar in time course to N400, but different in polarity. Consequently, instead of N400, it was also quite possible to find a larger P300 for the critical words in the emphasized (rather than control) condition, indicating the surprise resulting from encountering emphasized words.
Method
Subjects
Twenty-eight right-handed, healthy undergraduate students (14 males; mean age 19.61 years, range 18-22 years) participated in the study. All were native Chinese speakers with normal or corrected-to-normal vision. They signed an informed consent form prior to the study and were paid after the study.
Materials
We constructed 160 sets of sentences (see Table 1 for examples). Critical words (all two-character nouns) appeared in the middle or latter part of the sentences and were followed by at least two additional words. The colors of the critical words were manipulated. In the emphasized condition, the critical words were displayed in red and the other words in the sentences shown in white. In contrast, the sentences were entirely in white in the control condition. In addition, the critical words were manipulated to be congruent or incongruent with the preceding verbs. The critical words in the two congruence conditions were the same in terms of word frequency (Congruent, M = 18605.79, SD = 70730.03; Incongruent, M = 23263.14, SD = 55186.10; t (159) = -0.66, p = .513 , d = -0.05) and stroke number (Congruent, M = 16.29, SD = 4.69; Incongruent, M = 15.94 , SD = 4.50; t (159) = 0.71, p = .480 , d = 0.06).
Three pretests were conducted for the materials. The first pretest assessed the plausibility of the sentences. Two lists of experimental sentences were created in such a way that participants could only see the congruent or the incongruent condition of the same item and there were an equal number of sentences from the two conditions. During the test, each sentence was presented in its entirety on a computer screen. Twenty subjects were asked to rate the congruence of the sentences on a scale of 1 (very implausible) to 7 (very plausible). The means (SDs) of the congruent and incongruent conditions were 5.80 (0.58) and 2.06 (0.52), respectively. A paired-sample t-test revealed a significant difference between the two conditions, t (159) = 59.42, p < .001, d = 4.70, suggesting that our manipulation of semantic congruence was successful.
The second pretest was conducted to assess the predictability of the critical words. We asked another 20 subjects to complete a cloze probability task. During the task, the sentences were presented up to where the critical words would appear. The subjects were asked to continue the sentences with the first words that came to mind. The mean (SD) of the probability for writing the critical words of the congruent condition was 6.41% (14.14%). No participant completed the sentences with the critical words that were used to create the incongruent condition. A paired-sample t-test showed a significant difference between the two conditions, t (159) = 5.73, p < .001, d = 0.45. In addition, the sentence constraint was calculated. The result showed that the mean (SD) of the probability to write a specific word based on the sentence context was 27.88% (12.42%), suggesting that although the predictability of critical words under the congruent condition was higher than that under the incongruent condition, the predictability of both was low.
The final pretest assessed the importance of the critical words in sentences. Four lists of experimental sentences were created in such a way that participants could only see one condition of the same item and there were an equal number of sentences from each condition. During the test, each sentence was presented in its entirety on a computer screen. The subjects were asked to rate the importance of critical words in the sentences on a scale of 1 (not important at all) to 7 (very important). The means (SDs) of the importance rating were 4.92 (1.02), 4.60 (1.24), 3.38 (1.64), and 3.52 (1.58), respectively, for the Emphasized-Congruent, Control-Congruent, Emphasized-Incongruent, and ControlIncongruent conditions. Main effects for font emphasis, F (1, 39) = 6.21, p = .017, η2p = .137, and semantic congruence, F (1, 39) = 15.44, p < .001, η2p= .284, were found. Critical words in the emphasized or congruent conditions were considered to be more important than those in the control or incongruent conditions. More importantly, an interaction effect between font emphasis and semantic congruence was found, F (1, 39) = 4.55, p = .039, η2p = .105. Simple effect analysis revealed that in the congruent condition, emphasized words were rated more important than control words, F(l, 39) = 8.73, p = .005, η2p = .183, whereas in the incongruent condition, the importance rating of the emphasized and control words had no significant difference, F (1, 39) = 1.43, p = .239, η2p = .035. The results suggest that color is appropriate as a marker for font emphasis and our manipulation was successful.
The 160 sets of sentences were grouped into four lists via a Latin square design, such that each list contained one condition from each experimental set and an equal number of sentences from each condition. In addition, 80 fillers were added. Half of the filler sentences were semantically congruent. The remaining fillers were semantically incongruent so that one of the words could not be integrated with its preceding verb or context. Again, with the experimental sentences, half of the filler sentences (20 each for the congruent and incongruent sentences) were emphasized by font. However, the position of the font emphasis was randomly distributed to prevent readers from guessing the position of the emphasis. In total, there were 240 sentences during the study. They were equally divided into four blocks, with each block containing 60 sentences. The subjects read the sentences in a pseudorandom order and no more than three sentences of the same condition were presented in succession.
Procedure
The subjects were seated in a comfortable chair in front of a computer screen. They were asked to read the sentences. All sentences were presented in the middle of the computer screen, in 34 Chinese Songti font, on a black background. Each trial started with a fixation cross that remained for 500 ms. Then, the sentence began to be presented word by word. Each word was presented for 400 ms, followed by a black screen for 300 ms. To ensure that the subjects read for comprehension, we required them to judge a statement that was related to the presented sentence by pressing marked buttons as quickly and accurately as possible. One-third of the trials were followed by the statements. For the remaining trials, the word "blank" was presented following the sentences and the subjects had to press the space bar for a response. During the study, the subjects were instructed not to move or blink when the words were presented individually on the computer screen.
Each block lasted about 8 min. There was a short break between every two blocks. The whole study lasted about 1.5 h, in which the subjects' preparation, instructions, and a practice consisting of eight sentences were included.
EEG recording
EEG was recorded (0.05-400 Hz, sampling rate 1,000 Hz) throughout the study with AC amplifiers (SynAmps2, NeuroScan Inc.). An elastic cap that was equipped with 64 Ag/AgCl electrodes according to the International 10-20 system was used. An online reference electrode that was placed between Cz and CPz and an electrode that was placed between FCz and Fz served as the ground. Vertical eye movements and blinks were monitored with a supra-to suborbital bipolar montage. Horizontal eye movements were monitored with a right to left canthal bipolar montage. Most of the electrode impedances were kept below 5 kQ.
Data analysis
Event-related potential (ERP) analysis
The data were analyzed using EEGLAB (Delorme & Makeig, 2004) and ERPLAB (Lopez-Calderon & Luck, 2014), both MATLAB toolboxes. Raw data were first analyzed with EEGLAB. During the analysis, continuous EEG data were re-referenced to the average of the two mastoids, high-pass filtered with a 0.1-Hz half-amplitude cutoff, and low-pass filtered with a 30-Hz half-amplitude cutoff (6 dB/ octave). An independent component analysis (ICA) was used to remove the EEG components related to ocular and muscular artifacts. We then shifted to ERPLAB. We segmented the EEG data from 200 ms before to 800 ms after the onset of the critical words. The first 200 ms pre-stimulus was used as the baseline, and all epochs were corrected to it. For artifact detection, a moving window peak-to-peak threshold algorithm (threshold 150 pV, window size 200 ms, window step 100 ms) and simple voltage threshold algorithm (threshold 75 pV) were used. There were 36 remaining trials for each condition. For each participant, the trials were averaged for each of the four conditions.
Statistical analyses of the ERP data were implemented in SPSS 25.0. Based on previous studies (Chen et al., 2014; Mei et al., 2019; Wang et al., 2009; Wu et al., 2018) and visual inspection of the data, the following three time windows were selected for analysis: 100-170 ms (Nl), 180-270 ms (P2), and 300-500 ms (N400/P300). P300 is similar in time course to N400. However, P300 sometimes follows N2 or is followed by a long "slow wave" lasting several hundred milliseconds (Kutas & Hillyard, 1980a). Visual inspection showed a slow wave behind the 300- to 500-ms time window. Therefore, we also analysed the amplitudes in the 500- to 700-ms time window to identify ERP components. The mean amplitude values were first computed across all trials for each electrode, condition, subject, and selected time window. The selected electrodes were then grouped into nine regions of interests (shown in Fig. 1) and the mean amplitude values within each region of interest were computed. Finally, repeated-measures analyses of variance (ANOVAs) were performed with the within-subject factors of Font Emphasis (emphasized, control), Semantic Congruence (congruent, incongruent), Anteriority (anterior, central, posterior), and Laterality (left, midline, right). The Greenhouse-Geisser correction was applied when the sphericity test was significant. In these cases, the original degrees of freedom and corrected p values were reported. The present study mainly focused on the influence of font emphasis, so only the interplay effects (if any) related to Font emphasis were further analyzed.
Time-frequency analyses
Event-related spectral perturbation was employed to characterize the oscillatory brain activity in different time-frequency (TF) bands, averaged across single trials. The raw data were first analyzed with EEGLAB and ERPLAB. The procedures were almost the same as with the ERP analysis, except that we applied a low-pass filter at 100 Hz and segmented the EEG data from 500 ms before to 1,500 ms after the onset of the critical words. The remaining analyses were performed using MATLAB's Fieldtrip toolboxes (https://www.fieldtriptoolbox. org/). First, the low- and high-frequency ranges were computed separately within the 2,000-ms epoch in 50-ms steps. For the lower-frequency range, TF representations were obtained by applying a Hanning taper with a length of 500 ms, in a frequency range from 2 Hz to 30 Hz in 2-Hz steps. For the higherfrequency range, we used multi-tapers (Mitra & Pesaran, 1999) with a 400-ms time smoothing and 5-Hz frequency smoothing, in a frequency range from 25 Hz to 100 Hz in 2.5-Hz steps. Then the power changes in the post-stimulus interval were expressed as a relative change from the baseline interval (from -0.35 to -0.05 s). After acquiring the TF representations of single trials, we averaged the power estimates over trials separately for the four conditions, for each subject.
For statistical analysis, the cluster-based random permutation test implemented in the Fieldtrip software package was used. This non-parametric statistical procedure optimally handles the multiple-comparisons problem. The permutation test was performed within 0-1,500 ms after the onset of the critical words (in 50-ms steps) over 62 electrodes (CB1 and CB2 were not included). Simple dependent t-tests were performed on each data point (electrode by time by frequency) comparing two conditions. For the low-frequency range, 29 frequencies from 2 Hz to 30 Hz were calculated; for the high-frequency range, 76 frequencies from 25 Hz to 100 Hz were calculated. All adjacent data points exceeding a preset significance level (0.05%) were grouped into clusters. Cluster-level statistics was calculated by taking the sum of the t-values for every cluster. The probability of significance of the clusters was calculated using the Monte Carlo method with 1,000 random draws.
The main effect of font emphasis and its interaction with semantic congruence were both analyzed. For the main effect, the emphasized condition (i.e., the average of the power of "Emphasized/Congruent" and "Emphasized/Incongruent") and control condition (i.e., the average of the power of "Control/Congruent" and "Control/Incongruent") were calculated. For the interaction, the two subtracted power values (i.e., (Emphasized/Congruent minus Control/Congruent) vs. (Emphasized/Incongruent minus Control/Incongruent)) were compared. Follow-up tests were then done if this interaction was significant.
Results
Behavior results
The averaged accuracy of the statement judgments performance was 95.57%, indicating that the subjects did read the sentences for comprehension.
ERP results
Figure 2 shows the grand average waveforms elicited by the critical words at nine representative electrodes (F3, Fz, F4, C3, Cz, C4, P3, Pz, and P4) in different conditions.
Table 2 shows the results of the repeated measures ANOVAs for the amplitudes in the time windows of 100-170 ms, 180-270 ms, 300-500 ms, and 500-700 ms. In the time window of 100-170 ms, a main effect for Font Emphasis was found. Mean amplitude was less negative going for the emphasized words than the control words. In addition, there was a marginally significant interaction between Font Emphasis and Anteriority. Further analysis revealed that the emphasized words elicited a smaller negativity than the control words over all the regions (the anterior region, F (1, 27) = 15.99, p < .001, η2p = .372; the central region, F(l, 27) = 10.92, p = .003, η2p; = .288; the posterior region, F (1, 27) = 8.75, p = .006, η2p = .245). More importantly, a marginally significant interaction effect for Font Emphasis, Semantic Congruence, Anteriority and Laterality was found. We then performed three three-way ANOVAs (Font Emphasis X Semantic Congruence X Laterality) for the anterior, central, and posterior regions. Only a triple marginally significant interaction was found in the central region (the anterior region, F < 1; the central region, F (2, 54) = 2.76, p = .072, η2p = .093; the posterior region, F (2, 54) = 1.97, p = .150, η2p = .068). However, further analysis revealed that there was no interaction between Font Emphasis and Semantic Congruence over all the central regions, Fs < 1.
The overall analysis of amplitude at 180-270 ms showed a main effect of Font Emphasis with larger positivity for the emphasized than the control condition. The effect of Font Emphasis was modulated by Anteriority. Simple effect tests revealed that the effect was present at the anterior region, F (1, 27) = 22.68, p < .001, η2p= .456, and the central region, F (1, 27) = 7.61, p = .010, η2pp = .220, but not the posterior region, F (1, 27) = 3.04, p = .092, η2p = .101. The effect of Font Emphasis was also modulated by Laterality. Further analysis found that the effect was only present at the midline region (the left region, F(1, 27) = 3.07, p = .091, η2p = .102; the midline region, F (1, 27) = 6.91, p = .014, η2p= .204; the right region, F (1, 27) = 3.75, p = .063, η2p = .122).
In the 300- to 500-ms time window, both the main effects of Font Emphasis and Semantic Congruence were found. The critical words elicited a smaller negativity in the emphasized or congruent condition than in the control or incongruent condition. The effect of Font Emphasis was modulated by Anteriority and Laterality, respectively. More importantly, there was a significant three-way interaction between Font Emphasis, Anteriority, and Laterality. We then performed three two-way ANOVAs (Font Emphasis X Laterality) for the anterior, central and posterior regions. The interaction effects were found for all the three regions (the anterior region, F (2, 54) = 9.26, p < .001, η2p = .255; the central region, F (2, 54) = 10.96, p < .001, η2p = .289; the posterior region, F (2, 54) = 19.03, p < .001, η2p = .413). Further analysis revealed that the effect was present at all regions except the left-posterior region (the left-anterior region, F (1, 27) = 32.30 , p < .001, η2p = .545; the midlineanterior region, F (1, 27) = 39.65 , p < .001, η2p = .595; the right-anterior region, F (1, 27) = 26.21 , p < .001, η2p = -493; the left-central region, F (1, 27) = 15.99 , p < .001, η2p = .372; the midline-central region, F (1, 27) = 28.14 , p < .001, η2p = .510; the right-central region, F(l, 27) = 30.05 , p < .001, η2p = .527; the left-posterior region, F (1, 27) = 2.27, p = .143, η2p = .078; the midline-posterior region, F (1, 27) = 19.91, p < .001, η2p = .424; the right-posterior region, F(l, 27) = 12.63,p = .001, η2p = .319). Similarly, there was a significant three-way interaction between Congruence, Anteriority, and Laterality. We then performed three twoway ANOVAs (Congruence × Laterality) for the anterior, central and posterior regions. The interaction effects were found for all the three regions (the anterior region, F (2, 54) = 7.58, p = .001, η2p = .219; the central region, F (2, 54) = 10.96, p < .001, η2p = .289; the posterior region, F (2, 54) = 6.66, p = .003, η2p = .198). Further analysis revealed that the effect was present at all regions except the left-central region (the left-anterior region, F (1, 27) = 4.29, p = .048, η2p = .137; the midline-anterior region, F (1, 27) = 15.32 , p = .001, η2p = .362; the right-anterior region, F (1, 27) = 9.28 , p = .005, η2p = .256; the left-central region, F (1, 27) = 2.08 , p = . 161, η2p = .071; the midline-central region, F (1, 27) = 13.03 , p = .001, η2p = .326; the right-central region, F (1, 27) = 9.42 , p = .005, η2p = .259; the left-posterior region, F (1, 27) = 4.35, p = .046, η2p = .139; the midline-posterior region, F (1, 27) = 9.17, p = .005, ηp = .254; the rightposterior region, F (1, 27) = 8.58, p = .007, p = .241).
In the 500- to 700-ms time window, a main effect for Font Emphasis was found. The critical words elicited a larger positivity in the emphasized than in the control conditions. In addition, there was a marginally significant interaction between Font Emphasis and Laterality and a significant three-way interaction between Font Emphasis, Anteriority, and Laterality. We then performed three two-way ANOVAs (Font Emphasis X Laterality) for the anterior, central, and posterior regions. The interaction effects were found in the posterior region (the anterior region, F < 1; the central region, F (2, 54) = 2.25, p = .115, η2p = .077; the posterior region, F (2, 54) = 6.95, p = .002, η2p = .205). Further analysis revealed that the effect was present at the midline-posterior and the right-posterior regions (the left-posterior region, F (1, 27) = 1.79, p = .192, η2p = .062; the midline-posterior region, F (1, 27) = 11.95, p = .002, η2p = .307; the posteriorregion, F (1, = 6.30, p = .018, η2p = .189).
To summarize, compared with the control condition, the critical words in the emphasized condition elicited a smaller negativity in the 100- to 170-ms and 300- to 500-ms time windows with a broad scalp distribution, and a greater positivity in the 180- to 270-ms time window with an anteriorcentral and midline scalp distributions. Figure 2 has shown the scalp topographies of the effects. In addition, we found a greater positivity for the critical words in the emphasized condition in the 500- to 700-ms time window with a midlineand right- posterior scalp distributions compared to the control condition. Finally, the critical words elicited a greater negativity for the incongruent than the congruent condition with a latency of 300- to 500-ms post-stimulus onset. However, none of the interactions between Font Emphasis and Semantic Congruence and/or any other topographical factor was significant.
Time frequency (TF) results
The TF representations of each condition, as well as their contrast, are shown in Fig. 3 (Emphasized and Control conditions) and Fig. 4 (Congruent and Incongruent conditions). For the low frequencies (2-30 Hz), the contrast of the emphasized and control conditions revealed two significant clusters. Relative to the control condition, the critical words in the emphasized condition induced power decrease with right-hemisphere scalp distribution in the alpha-frequency range (8-12 Hz) in the time window of 0.35-0.95 s (p = .011) and the beta-frequency range (12-18 Hz) in the time interval of 0.3-0.8 s (p = .006). The contrast between the congruent and incongruent conditions showed one significant cluster (p = .003) in the beta frequency range (14-22 Hz) between 0 and 1 s. Beta power exhibited a decrease in the incongruent condition with central to posterior scalp distribution.
In addition, the permutation test revealed significant differences between the difference powers ((Emphasized/ Congruent minus Control/Congruent) vs. (Emphasized/ Incongruent minus Control/Incongruent)) in the deltafrequency range (2-4 Hz; within 0-0.2 s, p = .023) and the theta-frequency range (4-8 Hz; within 0.05-0.3 s, p = .009), suggesting a two-way interaction of Font Emphasis and Semantic Congruence in the delta and theta band. Follow-up analyses of the delta band revealed that, for the congruent words, no significant differences between the emphasized and the control conditions were observed (p = .214); in contrast, for the incongruent words, the emphasized condition induced power increase (post-right distribution) as compared to the control condition (p = .046). Similarly, follow-up analyses of the theta band showed no significant differences between the emphasized and the control conditions for the congruent words (p > 1); however, the emphasized condition induced power increase (over right scalp) as compared to the control condition (p = .032).
For the high frequencies (30-100 Hz), the contrasts showed no significant difference.
In short, compared to the control condition, critical words in the emphasized condition induced power decrease in the alpha- (8-12 Hz) and beta-frequency range (12-18 Hz). In addition, compared to the control condition, the emphasized condition induced delta (21 Hz) and theta (4-8 Hz) power increases for incongruent words. Additionally, a beta (14-22 Hz) decrease in response to semantic anomalies was found.
Discussion
As described in the introduction, the goal of this experiment was to further delineate how font emphasis affects semantic integration of words in sentence reading by investigating the way it captures attention and its effect on the integration of emphasized information with the previous context. To examine these questions, we computed ERPs and frequency band-specific EEG power changes occurring while participants read sentences that contained a critical word that was either emphasized or not and semantically congruent with prior words or not. In the ERP data, we observed a smaller negativity in the 100- to 170-ms time window, larger positivity in the 180- to 270-ms time window, smaller negativity in the 300- to 500-ms time window, and lager positivity in the 500- to 700-ms time window for emphasized words. We also found a larger negativity in the 300- to 500-ms time window for incongruent words. However, the effects of font emphasis and semantic congruence did not interact. In terms of the frequency band-specific EEG power, we observed power decreases in the alpha- and beta-frequency ranges in response to critical words in the emphasized condition. Moreover, relative to the control condition, the emphasized condition elicited delta and theta power increases for incongruent words.
The way font emphasis garners increased attention
Consistent with our expectations, we found that compared to the control condition, the critical words in the emphasized condition evoked a smaller negativity in the 100- to 170-ms time window (i.e., Nl). When there is a slight change in font, Nl reflects the early perceptual/pre-lexical processes involved in mapping visual features onto location-specific letter representations (Vergara-Martinez et al., 2020). Font emphasis subtly changes the visual features of words, as the letter shapes are preserved. Thus, our results indicate that the words printed in the emphasized font were more difficult to encode and map onto higher word representations. Previous research has consistently reported larger amplitudes for Nl as a function of expertise in orthographic processing (Vergara-Martinez et al., 2020; Xue et al., 2019). Participants in the present study may have been more familiar with words presented in regular font (in the control condition), as most materials were presented in this font. In contrast, the visual form of the emphasized words was rare and unfamiliar. Therefore, compared with the control condition, the critical words were more difficult to map onto higher word representations, evoking a reduced N1 in the emphasized condition.
As was our expectation, the emphasized words attracted more attention. This was reflected in an increase in P2 and decrease in alpha power. We observed a larger P2 (i.e., positivity in the 180- to 270-ms time window) and a decrease in alpha activity when critical words were marked by font emphasis. P2 has previously been connected to the allocation of attention (Hillyard & Miinte, 1984; Luck & Hillyard, 1994). A decrease in alpha power is functionally related to increased attention (Gerven & Jensen, 2009; Klimesch, 1999; Marshall et al., 2015). Therefore, these results suggest that font emphasis does capture attention during reading. Results from the previous study show that infrequent unfamiliar stimuli raise the level of arousal (Ulrich et al., 2006). Therefore, the unfamiliar visual feature of emphasized words leads readers to perceive them as signifying the importance of the word and to deploy more attention resources. Although supporting the conclusion put forth by the attention hypothesis, these results also offer a further, more novel finding: direct evidence of the assumption that font emphasis draws one's attention during reading. In previous studies, it was reasonable to suppose that the memory advantage (Lorch et al., 1995), longer reading times for one-time reading (Lorch et al., 1995), and a higher probability of detecting words that changed across two consecutive presentations (Sanford et al., 2006) were the result of increased attention. It was also reasonable to conclude that in the gaze-contingent boundary paradigm, emphasizing previewed words leads to shorter fixations on target words, and thus font emphasis draws more attention to a word even before the reader fixates (Rayner & Schotter, 2014). However, the above measures were indirect indicators leading to the inference that emphasized information receives more attention during reading; no direct evidence was provided. Our research provides this evidence and strengthens the existing explanation.
The role of font emphasis in integration
The N400 component has usually been associated with reprocessing of semantically anomalous information. Compared to semantically appropriate words, semantically inappropriate words elicit a larger N400 (Kutas & Hillyard, 1980b; van Berkum et al., 1999). Our result of incongruent words evoking a larger negativity in the 300-500ms time window (i.e., N400) than congruent words replicates this canonical pattern.
We also observed an effect in the 300- to 500-ms time window from emphasis manipulation. Compared with words printed in regular font, the words printed in emphasized font elicited a smaller negativity. One might also consider this negativity as N400. However, careful inspection of the waveforms indicated that the font emphasis effect within the 300to 500-ms time window was more like the P300 rather than the N400 effect. P300 is similar in time course to N400, but different in polarity. P300 is one of the ERP components most sensitive to the occurrence of unexpected, surprising, or discrepant stimuli, sometimes after N2 or before a long "slow wave" lasting several hundred milliseconds (Kutas & Hillyard, 1980a). Visual inspection showed a slow wave (500- to 700-ms time window) behind the 300- to 500-ms time window. Our statistical analysis further confirms that the slow wave only followed the font emphasis rather than the semantic congruence effect, suggesting that the two effects are different, and that the font emphasis effect is more like P300 than N400 in the 300- to 500-ms interval.
Another piece of evidence comes from the results of the neural oscillatory activities. P300 has been proposed as a result of neuro-inhibition that is engaged when incoming stimuli garner attentional processes to facilitate memory encoding (Polich, 2012). We found evidence of neuroinhibition in that the alpha and beta band powers decreased in the emphasized condition, as compared to the control. Alpha power suppression is functionally related to increased attention (Gerven & Jensen, 2009; Klimesch, 1999; Marshall et al., 2015). Beta power suppression reflects changes in the current cognitive set (Prystauka & Lewis, 2019) and engagement of task-relevant brain regions (Wang et al., 2012). In this study, words in the emphasized condition were presented in a regular font, up to the key word. When the emphasized word was encountered, the reader perceived the striking, infrequent, low-probability font as signaling the importance of the word, as evidenced by the pretest reflecting importance. A neural inhibitory mechanism then became active to enhance focal attention (as measured by decreased alpha power), with the respondent putting more effort into integrating the emphasized word into the current sentence representation (as measured by beta power suppression); this resulted in a greater P300 amplitude. Taken together, a subsequent "slow wave" and the neuro-inhibition mechanism initiated suggest that the font emphasis effect is more like a P300 effect within the 300- to 500-ms time window.
The finding of a P300 rather than an N400 effect suggests that the integration mechanism for font emphasis is different from general semantic integration. This is compatible with previous research (Kutas & Hillyard, 1980a) revealing different electrophysiological responses for semantic and physical deviations. This separation is also reflected in the beta activity. We found a beta power decrease in response to semantic anomalies, a typical feature of semantic anomaly processing (Prystauka & Lewis, 2019) that indicates the semantic unification of words into the preceding sentence context (Wang et al., 2012). We also found beta power suppresion in the font emphasized condition. However, no interaction between Font Emphasis and Semantic Coherence was found. We concluded that unlike general semantic integration, font emphasis attracts attention to promote a deeper and more integrated processing of the information highlighted by the reader. As for the reason why font emphasis attracts attention, previous research has argued that visual highlighting markers like font emphasis have both low-level visual properties and high-level linguistic signalling functions with which skilled readers are familiar and both of which attract attention (Y. Wu et al., 2021). More specifically, the current findings suggest that words highlighted by font amount to infrequent, low-probability stimuli that readers perceive as important, suppressing irrelevant activity and enhancing focal attention.
In addition to the two questions we planned to address, the present study reveals the mechanism underlying the lexical access advantage of emphasized words that was found in the previous studies (e.g., Wu et al., 2021). This advantage benefits from more efficient glyph matching. Our oscillatory results showed that compared to the control condition, incongruent words in the emphasized condition elicited a stronger power increase in delta (0-0.2 s) and theta (0.05-0.3 s) bands. These results agree with the findings of research manipulating semantic congruence and sentential constraint, in which the delta and theta activity were linked to the working memory load associated with the detection of anomalies or lexico-semantic retrieval processes (Hald et al., 2006; Kielar et al., 2015; Rommers et al., 2017). Thus, our results suggest that when incongruent words were emphasized with fonts, participants did more work to evaluate whether the words could be successfully integrated into the sentence context or to retrieve unexpected words. Each of these functional roles for the delta and theta activities indicates that participants responded more strongly to the interruption when incongruent words were emphasized. This violation detection occurred almost at the time the incongruent words appeared and before semantic access (as reflected in the N400). Based on findings indicating that the brain generates form estimates based on lexical-semantic predictions (Dikker & Pylkkanen, 2011; Laszlo & Federmeier, 2009) and the fact that in this study congruent words were more predictable than incongruent words, we concluded that the delta and theta results were linked to the violation detection of the visuo-orthographic information of the words. The increasing power of the emphasized condition suggests that font emphasis strengthens the process of glyph matching, leading to stronger violation detection. This efficient glyph matching facilitates lexical semantic processing.
Throughout the results, the ERP analyses showed very early font emphasis effects, whereas the effects of font emphasis in the TF analyses were only found in later time windows. Also, while the ERP effects from font emphasis were not modulated by semantic congruence, several interaction effects were seen in the TF analysis. The different patterns of results stem from the different calculations of ERP and oscillation activities. Time-locked event-related responses can be either phase-locked or non-phase-locked. The former manifested as ERP that can be easily observed in the time domain after averaging across several trials. The latter manifested as event-related oscillations that become invisible in the time domain by directly averaging across several trials, as the voltage at a given post-stimulus time point will be positive in some trials and negative in others (Lu & Hu, 2019). However, it is possible to perform a timefrequency analysis that extracts the amplitude at a given frequency independent of its phase prior to averaging. This makes it possible to see the time course of stimulus-elicited oscillations (Luck, 2014). The different calculations of ERP and oscillation activities dictate that a particular effect does not always manifest itself consistently in ERPs and neural oscillations. This is the reason why we found different patterns in this study. Beyond these inconsistent results patterns, we found many consistent results. For example, in terms of the font emphasis effect, P300 and alpha/beta together reflected that the emphasized information required more attention to be more fully integrated. In terms of the effect of semantic coherence, both classical N400 and beta effects were found. Furthermore, the results overlapped a great deal in time. These consistent results illustrate the reliability of our data. Conversely, the inconsistent results complement one another and more fully reveal the cognitive neural mechanisms by which font emphasis affects sentence reading.
In brief, our results reveal that font emphasis facilitates early processing and reduces visual familiarity of words. These findings refine attention theory. In previous studies, researchers only speculated in general terms that font emphasis attracts more attention during reading than does the surrounding context (Rayner & Schotter, 2014; Sanford et al., 2006). In this study, we demonstrated that font emphasis reduces the familiarity with the emphasized words and attracted more attention due to their unfamiliar visual features. In addition, combined with previous studies (Fraundorf et al., 2013; Wu et al., 2021), the finding that font emphasis allows information to be more fully integrated with the preceding context indicates that it facilitates the entirety of reading processing, including lexical accessing and integration processes.
Our study also establishes font emphasis as an important linguistic signal. We found that font emphasis allowed information to be more fully integrated regardless of whether it was semantically congruent with the previous context. Font emphasis attracts attention with their fewer and more eye-catching advantages such that readers indiscriminately integrate the emphasized information. It is such an important linguistic signal that readers cautiously treat so as not to miss important information.
It is worth noting that in the present study, the information to be emphasized was randomly determined. However, in daily life, emphasis is used selectively and purposively to highlight a given body of content. This kind of emphasis is based on top-down background knowledge and sentence/ discourse context. The difference between and interaction of top-down and bottom-up emphasis has yet to be determined.
In conclusion, the present study offers additional details regarding the attention hypothesis and effect of font emphasis. Font emphasis reduces visual familiarity with emphasized words. The resulting unfamiliarity leads readers to perceive those words as signifying importance and thus deploy more attention resources, leading to full integration.
Funding This work was supported by the Fujian Social Science Planning Project (Grant numbers [FJ202IB 105]), Natural Science Foundation of Fujian Province (Grant numbers [2023JO 1287]), National Social Science Foundation for Education (Grant numbers [BBA200038]), and China Postdoctoral Science Foundation (Grant numbers [2019M652239]).
Data availability The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.
Code availability Not applicable.
Declarations
Competing interests The authors have no relevant financial or nonfinancial interests to disclose.
Ethics approval The study was approved by the Ethics Committees at the authors' affiliations.
Consent to participate and publication Informed consent was obtained from all individual participants included in the study.
References
Bar-Kochva, I., & Breznitz, Z. (2012). Does the reading of different orthographies produce distinct brain activity patterns? An ERP study. PloS One, 7(5), e36030. https://doi.org/10.1371/journal. pone.0036030
Birch, S., & Rayner, K. (1997). Linguistic focus affects eye movements during reading. Memory & Cognition, 25(5), 653-660. https://doi. org/10.3758/BF03211306
Birch, S., & Rayner, K. (2010). Effects of syntactic prominence on eye movements during reading. Memory & Cognition, 38(6), 740-752. https://doi.Org/10.3758/MC.38.6.740
Cashen, V. M., & Leicht, K. L. (1970). Role of the isolation effect in a formal educational setting. Journal cf Educational Psychology, 61(6 Pt.l), 484-486. https://doi.org/10.1037/h0030286
Chen, L., & Yang, Y. (2015). Emphasizing the only character: EMPHASIS, attention and contrast. Cognition, 136, 222-227. https://doi. org/10.1016/j.cognition.2014.11.015
Chen, L., Wang, L., & Yang, Y. (2014). Distinguish between focus and newness: An ERP study. Journal cf Neurolinguistics, 31, 28-41. https://doi.Org/10.1016/j.jneuroling.2014.06.002
de Koning, B. B., Tabbers, H. K, Rikers, R. M. J. P., & Paas, F. (2009). Towards a Framework for Attention Cueing in Instructional Animations: Guidelines for Research and Design. Educational Psychology Review, 21(2), 113-140. https://doi.org/10.1007/ S10648-009-9098-7
Delorme, A., & Makeig, S. (2004). EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. Journal cf Neuroscience Methods, 134(1), 9-21. https://doi.Org/10.1016/j.jneumeth.2003.10.009
Dikker, S., & Pylkkanen, L. (2011). Before the N400: Effects of lexical - semantic violations in visual cortex. Brain and Language, 118(1-2), 23-28. https://doi.Org/10.1016/j.bandl.2011.02.006
Filik, R., Purdy, K., Gale, A., & Gerrett, D. (2006). Labeling of medicines and patient safety: evaluating methods of reducing drug name confusion. Human Factors, 48(1), 39-41. https://doi.org/ 10.1518/001872006776412199
Fraundorf, S. H., Benjamin, A. S., & Watson, D. G. (2013). What happened (and what did not): Discourse constraints on encoding of plausible alternatives. Journal cf Memory and Language, 69(3), 196-227. https://doi.org/10.1016/j.jml.2013.06.003
Gaddy, M. L., van den Broek, P., & Sung, Y.-C. (2001). The influence of text cues on the allocation of attention during. In T. Sanders, J. Schilperoord, & W. Spooren (Eds.), Text Representation: Linguistic and psycholinguistic aspects (pp. 89-110). John Benjamins Publishing Company.
Hald, L. A., Bastiaansen, M. C. M., & Hagoort, P. (2006). EEG theta and gamma responses to semantic violations in online sentence processing. Brain and Language, 96(1), 90-105. https://doi.org/ 10.1016/j?bandl.2005.06.007
Hillyard, S. A., & Miinte, T. F. (1984). Selective attention to color and location: An analysis with event-related brain potentials. Perception & Psychophysics, 36(2), 185-198. https://doi.org/10.3758/ BF03202679
Kielar, A., Panamsky, L., Links, K. A., & Meltzer, J. A. (2015). Localization of electrophysiological responses to semantic and syntactic anomalies in language comprehension with MEG. NeuroImage, 105, 507-524. https://doi.Org/10.1016/j.neuroimage.2014.ll.016
Klimesch, W. (1999). EEG alpha and theta oscillations reflect cognitive and memory performance: A review and analysis. Brain Research Reviews, 29(2-3), 169-195. https://doi.org/10.1016/ S0165-0173(98)00056-3
Kuchinke, L., Krause, B., Fritsch, N., & Briesemeister, B. B. (2014). A familiar font drives early emotional effects in word recognition. Brain and Language, 137, 142-147. https://doi.org/10.1016/jbandl.2014.08.007
Kutas, M., & Hillyard, S. A. (1980). Event-related brain potentials to semantically inappropriate and surprisingly large words. Biological Psychology, 11(2), 99-116. https://doi.org/10.1016/03010511(80)90046-0
Kutas, M., & Hillyard, S. A. (1980). Reading senseless sentences: brain potentials reflect semantic incongruity. Science, 207(4421), 203205. https://doi.org/10.1126/science.7350657
Laszlo, S., & Federmeier, K. D. (2009). A beautiful day in the neighborhood: An event-related potential study of lexical relationships and prediction in context. Journal cf Memory and Language, 61(3), 326-338. https://doi.Org/10.1016/j.jml.2009.06.004
Lemarie, J., Lorch, R. F. J., Eyrolle, H., & Virbel, J. (2008). SARA: A text-based and reader-based theory of signaling. Educational Psychologist, 43(1), 27-48. https://doi.org/10.1080/0046152070 1756321
Lopez-Calderon, J., & Luck, S. J. (2014). ERPLAB: An open-source toolbox for the analysis of event-related potentials. Frontiers in Human Neuroscience, S(1 APR), 1-14. https://doi.org/10.3389/ fnhum.2014.00213
Lorch, R. F. (1989). Text-signaling devices and their effects on reading and memory processes. Educational Psychology Review, 1(3), 209-234. https://doi.org/10.1007/BF01320135
Lorch, R. F., Pugzles, E., & Klusewitz, M. A. (1995). Effects of Typographical Cues on Reading and Recall of Text. Contemporary Educational Psychology, 20(1), 51-64. 0361-476X/95.
Lowder, M. W, & Gordon, P. C. (2015). Focus takes time: structural effects on reading. Psychonomic Bulletin & Review, 22, 1733- 1738. https://doi.org/10.3758/sl3423-015-0843-2
Lu, X., & Hu, L. (2019). Electroencephalography, Evoked Potentials, and Event-Related Potentials. In L. Hu & Z. Zhang (Eds.), EEG Signal Processing and Feature Extraction (pp. 23-42). Springer.
Luck, S. J., & Hillyard, S. A. (1994). Electrophysiological correlates of feature analysis during visual search. Psychophysiology, 31, 291-308. https://doi.Org/10.llll/j.1469-8986.1994.tb02218.x
Luck, S. J. (2014). An introduction to the Event-Related potential technique (2nd ed.). The MIT Press.
Macaya, M., & Perea, M. (2014). Does bold emphasis facilitate the process of visual-word recognition? Spanish Journal cf Psychology, 17(e2), 1-5. https://doLorg/10.1017/sjp.2014.2
Marshall, T. R., O'Shea, J., Jensen, X. O., & Bergmann, T. O. (2015). Frontal Eye Fields Control Attentional Modulation of Alpha and Gamma Oscillations in Contralateral Occipitoparietal Cortex. The Journal cf Neuroscienc, 35(4), 1638-1647. https://doi.org/ 10.1523/JNEUROSCI.3116-14.2015
McAteer, E. (1992). Typeface emphasis and information focus in written language. Applied Cognitive Psychology, 6(4), 345-359. https://doi.org/lo. 1002/acp.2350060406
Mei, Y., Dai, Y., & Lei, Y. (2019). The influence of hierarchical masks on masked repetition priming: Evidence from event-related potential investigation. Frontiers in Human Neuroscience, 13, 70. https://doi.org/10.3389/fnhum.2019.00070
Mitra, P. P., & Pesaran, B. (1999). Analysis of Dynamic Brain Imaging Data. Biophysical Journal, 76(2), 691-708. https://doi.org/10. 1016/S0006-3495(99)77236-X
Perfetti, C. A. (1985). Reading ability. Oxford University Press.
Perfetti, C. A. (2007). Reading ability: Lexical quality to comprehension. Scientific Studies cf Reading, 11(4), 357-383. https://doi. org/10.1080/10888430701530730
Polich, J. (2012). Neuropsychology of P300. In S. J. Luck & E. S. Kappenman (Eds.), The Oxford Handbook cf Event-Related Potential Components (pp. 159-188). Oxford University Press. https://doi. org/10.1093/oxfordhb/9780195374148.013.0089
Prystauka, Y., & Lewis, A. G. (2019). The power of neural oscillations to inform sentence comprehension: A linguistic perspective. Language and Linguistics Compass, 13(9), el2347. https://doi. org/10.1111/lnc3.12347
Rayner, K., & Schotter, E. R. (2014). Semantic preview benefit in reading english: The effect of initial letter capitalization. Journal cf Experimental Psychology: Human Perception and Performance, 40(4), 1617-1628. https://doi.org/10.1037/a0036763
Rommers, J., Dickson, D. S., Norton, J. J. S., Wlotko, E. W, & Federmeier, K. D. (2017). Alpha and theta band dynamics related to sentential constraint and word expectancy. Language, Cognition and Neuroscience, 32(5), 576-589. https://doi.org/10.1080/23273 798.2016.1183799
Sanford, A. J. S., Sanford, A. J., Molle, J., & Emmott, C. (2006). Shallow processing and attention capture in written and spoken discourse. Discourse Processes, 42(2), 109-130. https://doi.org/10. 1207/sl5326950dp4202_2
Schneider, S., Beege, M., Nebel, S., & Rey, G. D. (2018). A metaanalysis of how signaling affects learning with media. Educational Research Review, 23, 1-24. https://doi.Org/10.1016/j.edurev.2017. 11.001
Sulpizio, S., & Job, R. (2018). Early and multiple-loci divergency of proper and common names: An event-related potential investigation. Neuropsychologic, 119, 107-117. https://doi.org/10.1016/jneuropsychologia.2018.07.034
Sun, E., Zhang, Q., & Zhang, J. (2019). Impact of Calligraphy Style and Color on Comprehension of Importance in Chinese Idiom. Studies cf Psychology and Behavior, 17(2), 160-169.
Tinker, M. A., & Paterson, D. G. (1931). Studies of typographical factors influencing speed of reading. VII. Variations in color of print and background. Journal cfApplied Psychology, 15(5), 471-479. https://doi.org/10.1037/h0076001
Ulrich, R., Nitschke, J., & Rammsayer, T. (2006). Perceived duration of expected and unexpected stimuli. Psychological Research, 70(2), 77-87. https://doi.org/10.1007/s00426-004-0195-4
van Gerven, M., & Jensen, O. (2009). Attention modulations of posterior alpha as a control signal for two-dimensional brain-computer interfaces. Journal cf Neuroscience Methods, 179(1), 78-84. https://doi.Org/10.1016/j.jneumeth.2009.01.016
van Berkum, J. J. A., Hagoort, P., & Brown, C. M. (1999). Semantic integration in sentences and discourse: evidence from the N400. Journal cf Cognitive Neuroscience, 11(6), 657-671. https://doi. org/10.1162/089892999563724
van den Broek, P. (1995). A "Landscape" Model of Reading Comprehension: Inferential Processes and the Construction of a Stable Memory Representation. Canadian Psychology/Psychologie Canadienne, 36(1), 53-54. https://doi.org/10.1037/h0084723
van den Broek, P., Risden, K., Fletcher, C. R., & Thurlow, R. (1996). A "landscape" view of reading: Fluctuating patterns of activation and the construction of a stable memory representation. In B. K. Britton & A. C. Graesser (Eds.), Model cf understanding text (pp. 165-187). Lawrence Erlbaum Associates.
Vergara-Martinez, M., Perea, M., & Leone-Fernandez, B. (2020). The time course of the lowercase advantage in visual word recognition: An ERP investigation. Neuropsychologia, 146, 107556. https://doi.Org/10.1016/j.neuropsychologia.2020.107556
Wang, L., Hagoort, P., & Yang, Y. (2009). Semantic illusion depends on information structure: ERP evidence. Brain Research, 1282, 50-56. https://doi.Org/10.1016/j.brainres.2009.05.069
Wang, L., Jensen, O., Van Den Brink, D., Weder, N., Schoffelen, J., Magyari, L., Hagoort, P., & Bastiaansen, M. (2012). Beta Oscillations Relate to the N400m During Language Comprehension. Human Brain Mapping, 33(12), 2898-2912. https://doi.org/10. 1002/hbm.21410
Wu, J.-H., & Yuan, Y. (2003). Improving searching and reading performance: the effect of highlighting and text color coding. Information & Management, 40(7), 617-637. https://doi.org/10.1016/ S0378-7206(02)00091-5
Wu, Y., Yang, X., & Yang, Y. (2018). Importance conveyed in different ways: Effects of hierarchy and focus. Journal cf Neurolinguistics, 47, 37-49. https://doi.Org/10.1016/j.jneuroling.2018.02.005
Wu, Y., Wang, Z., Lin, W., Ye, Z., & Lian, R. (2021). Visual salience accelerates lexical processing and subsequent integration: an eye-movement study. Journal cf Cognitive Psychology, 33(2), 146-156. https://doi.org/10.1080/20445911.2021.1879817
Xue, L., Maurer, U., Weng, X., & Zhao, J. (2019). Familiarity with visual forms contributes to a left-lateralized and increased N170 response for Chinese characters. Neuropsychologia, 134,107194. https://doi.Org/10.1016/j.neuropsychologia.2019.107194
Zhou, A., Yin, Y., Zhang, J., & Zhang, R. (2016). Does font type influence the N200 enhancement effect in Chinese word recognition? Journal cf Neurolinguistics, 39, 57-68. https://doi.Org/10.1016/j. jneuroling.2016.03.001
Copyright Springer Nature B.V. Jan 2024