How psychological and descriptive narratives modulate the perception of facial emotional expressions: an event-related potentials (ERPs) study

Abstract

The aim of the present study was to investigate whether stories with high and low narrative transport exert different effects on neural activation in response to facial emotional expressions. Thirty-one participants were randomly assigned to two groups based on the type of story they read: psychological narrative with high narrative transport (6 women and 10 men; age M = 34.38 ± 8.77); descriptive narrative with low narrative transport (9 women and 6 men; age M = 24.07 ± 7.38). The electroencephalographic activity of the participants in response to emotional facial expressions (joy, anger, fear, sadness) was recorded before (T0) and after (T1) the reading task. The findings indicated that the reading task modulated the early brain response (P1, N170) to emotional facial expressions, irrespective of the narrative type. However, only in the psychological narrative group was the amplitude of the P100 found to be positively associated with the extent to which an individual was transported into the narrative. In summary, the findings appear to indicate that an increased degree of transport into the narrative is associated with a greater internal simulation process of emotions and mental states. This, in turn, modulates the perception of the real social world after reading.

Introduction

ERPs and face processing

In the context of social interactions, faces serve as a rich source of information about others, including identity, mental states, and emotional states (e.g., Jack and Schyns 2015; Oruc et al. 2019a). Indeed, as highlighted by Oruc and colleagues (2019b), although the perception of a face is a remarkable computational achievement, it is recognized immediately and effortlessly by the human observer, this because an effective representation involves projecting a high-dimensional signal into a lower-dimensional space that retains the important characteristics of that signal for a particular task. In fact, a face representation can be conceptualized through semantic dimensions like facial attributes, or as a combination of fundamental functions that cover the so-called face-space.

From a neurophysiological perspective, event-related potentials (ERPs) studies have demonstrated a distinctive brain activation pattern associated with face processing. This is evidenced by a pronounced negative amplitude between 130 and 200 ms, commonly referred to as the N170 component, which is observed in response to facial compared to non-facial stimuli. This component is localized in occipital-temporal areas (e. (Bentin et al. 1996; Itier and Taylor 2004a; Roussel et al. 2004) and its source was identified in the fusiform gyrus and superior temporal sulcus (Itier and Taylor 2004b; Sadeh et al. 2010). Other ERP studies have also demonstrated that the N170 component is modulated by specific emotional facial expressions. Indeed, a meta-analysis conducted by Hinojosa and colleagues (2015) revealed that angry, fearful, and happy faces elicit greater N170 amplitudes compared to neutral facial expressions. Conversely, no significant results were observed for sad and disgusted faces. As posited by the authors, these results indicate that the processes underlying the N170 reflect a hierarchization of facial expressions likely not exclusively reliant on structural characteristics, but also on communicative and social factors. Indeed, unlike expressions of sadness or disgust, expressions of fear, anger or happiness require rapid socio-emotional reactions in the receiver, therefore, for their processing they would require rapid facial decoding mechanisms reflected by the N170 (Hinojosa et al. 2015). In addition to the N170, other early ERP components, such as the P100 and P200, and later components, such as the P300, appear to be modulated by emotion during face processing.

The P100 component is a pronounced positive amplitude that manifests between 80 and 120 milliseconds (ms) following the presentation of the stimulus. It is associated with the processing of perceptual information in visual brain regions, specifically the occipital lobes, and its amplitude is indicative of the recruitment of attentional control, selective attention, and the consumption of attentional resources (Allison et al. 1999; Dennis et al. 2007; Hillyard et al. 1995; 1998). The amplitude and latency of the P100 component are also influenced by emotional facial expressions. Indeed, several studies have found enhanced P100 amplitude in response to fearful, angry and happy faces in comparison to neutral faces (Batty and Taylor 2003; Kolassa et al. 2006; Moradi et al. 2017; Pourtois et al. 2004). Furthermore, a smaller amplitude was found in response to surprised faces compared to other emotional expressions (Meaux et al. 2014), and the latency was found to be faster for fear versus neutral faces (Williams et al. 2004). Additionally, several studies on patients diagnosed with schizophrenia (Caharel et al. 2007; Earls et al. 2016; Shah et al. 2018) have demonstrated a diminished amplitude and delayed latency in response to faces in patients when compared to healthy controls. Specifically, a reduced P100 amplitude was observed in response to sad, angry, and fearful facial expressions in patients compared to controls (Shah et al. 2018).

Nevertheless, a recent meta-analysis by Schindler and Bublatzky (2020) claims that the findings for the P100 cannot be interpreted in a univocal way as they are “highly diverse and either confirm (e.g., see Blechert et al. 2012; Foti et al. 2010; Muller-Bardorff et al. 2018) or reject emotional modulations (e.g., see Smith et al. 2013; Wieser et al. 2012)." (p. 364). In contrast, the data relating to two subsequent ERPs components, i.e. P200 and P300, appears to be considerably more consistent. P200 and P300 are typically present in the parietal and frontal brain regions and are associated with higher-order perceptual and attentional processing (Luck and Hillyard 1994; Moratti et al. 2011). In this context, Schindler and Bublatzky (2020) reported that greater amplitude of P200 and P300 is often found in response to angry and happy expressions (e.g., Carretié et al. 2013; Herbert et al. 2013; Karl et al. 2016; Liu et al. 2012; Santos et al. 2008; Smith et al. 2013; Syrjanen et al. 2018; Tortosa et al. 2013; Wu et al. 2019; Zhang et al. 2012), suggesting that these components are indeed modulated by emotional information.

Face processing, social cognition, and narrative

In support of the fact that the face is a rich source of information about others and is central to social interactions, several ERP and neuroimaging studies have demonstrated a positive correlation between face processing and social cognition abilities. Social cognition is a multidimensional construct that enables the perception, interpretation, and effective response to social stimuli (Patin and Hurlemann 2015; Pinkham et al. 2014; Happé and Bird 2017; De Jaegher et al. 2010; Frith and Frith 2008). It encompasses both low-order and high-order processes, including affective and cognitive empathy. The former refers to “an emotional reaction of the observer when perceiving that another is experiencing or is about to experience an emotion” (Dvash and Shamay-Tsoory 2014, p. 283). Cognitive empathy is the ability “to understand the feelings of others without necessarily implying that the empathizer is in an affective state himself” (Walter 2012, p. 10). Social cognition also includes theory of mind, which is defined as the ability to attribute mental states and beliefs to oneself and others in order to interpret and predict behaviours (Westra and Carruthers 2018). In a study conducted by Petroni and colleagues (2011), it was demonstrated that an individual's ability to comprehend the mental states of others, as measured by the Reading the Mind in the Eyes Test (Baron-Cohen et al. 2001), was positively correlated with the amplitude of the N170 and the activation of the fusiform gyrus, i.e., brain source of the N170. In alignment with these findings, several studies (Choi and Watanuki 2014; Meaux et al. 2014) have also indicated that the early ERPs (P100, N170) are influenced by the emotional expressions conveyed by a face, and that emotional abilities are associated with the N170, P200, P300 and late positive potential.

In the context of studies that analyse social cognitive processes, there is an interesting line of research that aims to investigate these processes through narrative (e.g., Black et al. 2021; Buck et al. 2014; Eekhof et al. 2022; Kidd and Castano 2013; Mar 2018a, b). Indeed, some fundamental characteristics of stories make them particularly suitable for the study of social cognition (e.g., Herman 2013; Oatley 1995). These characteristics are essentially related to characters: understanding a story means (among other things) understanding the emotions, motivations and goals that underlie the characters’ actions in a story (Adornetti et al. 2023; Altavilla et al. 2024; Ferretti 2022). In other words, the understanding of a story is an exercise in social cognition. In this regard, a substantial body of behavioural research has demonstrated that reading fiction can facilitate the development of social cognitive abilities (e.g., Djikic et al. 2013; Kidd and Castano 2013; Koopman 2015; Mar and Oatley 2008, Mar et al. 2011; Paluck 2009). As posited by Mar and colleagues (2006, 2009, 2010), the act of reading a narrative prompts the reader to experience a range of emotions, which leave an impression on the individual. Furthermore, as Oatley (1999) posits, the act of engaging with fiction entails the simulation of authentic challenges and, consequently, engenders tangible ramifications for the reader. Frequently, when an individual engages with a fictional narrative, the act of identification with the characters and emotional involvement in the story leads to a state of sympathetic resonance with the characters, and potentially even the experience of the events of the story as if the reader were directly experiencing them (see also Bal and Veltkamp 2013). Indeed, it can be argued that fictional narratives are more effective than non-fiction because they engage readers in a simulation process, forming new mental associations through the characters’ social experiences (Foroni and Mayr 2005; Hakemulder 2000; Hodson et al. 2009; Oatley 1999).

Neurobiological studies appear to corroborate the findings of behavioural studies (Ferstl and von Cramon 2002; Kuperberg et al. 2006; Mason et al. 2008; Virtue et al. 2006). For example, Mason and Just (2006, 2009) discovered that particular brain networks are implicated in both theory of mind tasks and narrative comprehension. These include the dorsomedial prefrontal cortex (dmPFC), the right temporal parietal junction (TPj), and the posterior superior temporal sulcus (pSTS), which can be conceptualized as a protagonist (or agent) perspective interpreter network. In particular, the dorsomedial prefrontal cortex is activated throughout the processing of a narrative, acting as an executive processor (protagonist monitor). In contrast, the temporal-parietal network appears to function as a simulator, generating expectations to comprehend the protagonist's intentions (Mason and Just 2009).

It is crucial to acknowledge the observations made by Walkington and colleagues (2020), which highlight the pivotal role of identification and transportation in eliciting social processes during the act of reading a story. As mentioned earlier, identification refers to the process by which the reader becomes one with the character of the story “and reacts to his or her experiences as if they were happening to the viewer” (Sestir and Green 2010, p. 275). Transportation is defined as "the state of feeling cognitively, emotionally and imaginatively immersed in a narrative world" (ibid., p. 275). In fact, several studies have shown that individuals who are highly immersed in a story show greater empathic responses after a week than their less immersed counterparts (Bal and Veltkamp 2013). They are also more likely to engage in prosocial behaviours (Johnson 2012) and show a reduction in critical thinking (Van Laer et al. 2013).

The present study

In light of the aforementioned background, the present study sought to investigate whether reading narratives that differ in their focus on the emotional and psychological dimensions of characters (and thus which might elicit varying levels of identification and transportation) modulates the neuronal response to facial expressions of diverse emotions. To the best of our knowledge, no event-related potentials (ERPs) studies have been conducted. The participants were divided into two groups with similar levels of empathy, and their electroencephalographic activity was recorded during a visual task involving images of facial emotional expressions of joy, fear, anger, and sadness. Each participant's neural response to facial expressions was recorded twice: before (T0) and after (T1) the reading of a narrative text, which was either psychological or descriptive in nature, according to their assigned group. Psychological narrative placed a strong emphasis on the characters' mental and emotional states, while descriptive narrative focused on describing the characters' actions and setting. The objective of this study was twofold: firstly, to examine the impact of psychological and descriptive narratives on the brain response to facial emotional stimuli; secondly, to determine the association between these brain responses and the extent to which individuals are transported into the narrative. The first hypothesis is that the psychological narrative group after the reading (T1) will exhibit a greater amplitude and an earlier latency of the ERPs components, particularly of the early components (P100 and N170), in response to the emotional facial expressions compared to T0 and compared to T1 of the descriptive narrative group. Moreover, it is predicted that this modulation will be found mainly in response to negative emotional facial expressions, given the presence of themes related to illness and mourning in the psychological narrative. The second hypothesis is that the transportation scores will be positively correlated with amplitude and negatively with latency of the earlier ERP components.

Methods

Participants

A total of forty-two participants were included in the study. The participants were divided into two groups: the first group, comprising twenty individuals, was assigned to the psychological narrative group (PG), while the second group, comprising twenty-two individuals, was assigned to the descriptive narrative group (DG). In order to be included in the study, participants had to be right-handed, free of any neurological and/or psychiatric diagnoses, and not be taking any medication or using any drugs. The present study was approved by the Ethics Committee of [Roma Tre University], and all participants provided informed consent prior to their involvement in the study.

Stimuli

The stimuli were composed of 120 black-and-white facial emotional expressions, 30 for each of the four emotional conditions (joy, fear, anger, and sadness), and 40 black-and-white neutral object images, which served as filler items (e.g., umbrella, watch, scissors).

The facial emotional expressions were selected from the Extended Cohn-Kanade Dataset (CK+, Kanade et al. 2000; Lucey et al. 2010), which is a comprehensive dataset for action unit and emotion-specific expression. The facial expressions were extracted from videos in which the subjects' faces exhibited a shift from a neutral expression to a targeted peak expression. The videos were recorded at a rate of 30 frames per second with a resolution of 640×480 pixels. In the present study, the frame of the facial expressions exhibiting a peak expression of ≥ 60% was selected. The images utilized as fillers were obtained from the International Affective Picture System (IAPS; Lang et al. 1997).

Procedure

The participant was situated in front of a screen and outfitted with sensors on the head for the purpose of acquiring electroencephalographic data during a visual task. The visual experimental task was constructed using E-Prime software (version 2.0; Psychology Software Tools, Inc.). The visual task started with the follow instructions: “Please pay attention and watch the images that appear on the screen. When you are ready, press the space bar to begin.” As illustrated in Fig. 1, the trial began with the presentation of a fixation cross for a duration of 800 ms, after which the stimulus (joy, fear, anger, sadness, or filler) was displayed on the screen for a duration of 1500 ms. The trial was concluded with an inter-trial interval, which was randomly assigned to be either 350 or 500 ms. A total of 160 trials were presented in a randomized order.

Fig. 1.

Experimental visual task

Upon the conclusion of the initial presentation (T0), the instructions were displayed on the screen. "Please read the text to your right in a safe and internal manner. You don't have a time limit. When you finish reading, press any key to restart the presentation." The narrative text (psychological or descriptive) was placed on the desk at the participant's side. Right after reading the text, participants then repositioned themselves in front of the screen and pressed a button to restart the visual presentation of the stimuli (T1). The reading time of the participants of the psychological narrative group was M = 33’36” SD = 05’46’’, and of the descriptive narrative group was M = 35’41’’ SD = 07’28’’.

Reading text

The aim of this study was to investigate the effect of narrative type on the neural response to facial emotional expressions. Behavioural studies (e.g., Bal and Veltkamp 2013) have shown that people are more transported into the story when it focuses on the mental and emotional states of the characters. As van Krieken et al. (2017, p. 4) observed, the significance of "privileged access to the perceptions, evaluations, and goals of a character" serves as a crucial mechanism through which readers are transported into the protagonist's world and able to identify with the characters. Based on this assumption, two different narratives were chosen: a psychological narrative, which emphasized the emotional and psychological dimension of the characters, and a descriptive narrative, which focused more on the narration of actions and the environmental context, or setting, of the story. Specifically, participants assigned to PG read the short story A Small, Good Thing from Raymond Carver's collection of short stories, Cathedral (1983, pp. 59–89, Italian translation 2014) (see “Appendix A”). This novel is about an eight-year-old boy named Scott who dies after being hit by a car on his way to school on his birthday. The selection of this text was driven by the fact that the entire narrative revolves around the emotions and mental states of the protagonists. This is exemplified by the boy's parents, who are portrayed as struggling with pain and feelings of uncertainty as they wait for their child to be hospitalized for three days. Given the prevalence of themes related to illness and mourning, as well as the emotions associated with them, such as fear, it was hypothesized that reading this type of narrative would facilitate transposition into the story and identification with the characters.

For the DG, participants read the short story “Big two-hearted river”, extracted from “In our time”, the inaugural American volume of short stories by Ernest Hemingway (1925; the Italian translation was extracted from the book “Pietre, piume e insetti. L'arte di raccontare la natura” edited by Sturani 2013, pp. 320–338). The story narrates the journey of a protagonist named Nick who returns to his father's town after the war and takes a path along the railroad tracks until he reaches a river (see “Appendix B”). In contrast to the psychological narrative, the focus of this novel is on the character's actions, such as fishing or eating trout, and his connection to nature. The environmental context is described in detail and with remarkable realism, while emotional or psychological dimensions are largely absent.

The two texts were converted to a uniform format. This involved transcribing them into a computer and reproducing the original organization in paragraphs, using the same font (Arial), size (15), and line spacing (1.5). The psychological and descriptive texts were 30 pages (9625 words) and 25 pages (7630 words), respectively. Participants read the printed version on white A4 paper, which did not indicate the title of the story. Therefore, they did not know which novel they were reading.

To ensure that the reading task was completed with the necessary degree of accuracy, a five-question comprehension test of the narrative was administered at the end of the experimental procedure. All participants achieved at least three correct answers and were included in the EEG data processing.

Transportation and empathy assessment

To assess the extent to which participants were transported in the experimental reading narratives, Green and Brock's (2000, 2013) Transport Narrative Questionnaire was administered. The questionnaire consists of a 12-item scale that encompasses three dimensions of transport: cognition, emotion, and imagination. These three dimensions could be treated as subscales: cognitive (items 1, 3 and 4 from “Appendix C”), affective (items 5, 7, 11) and imagery (item 12). To the present study, only the total score of the transport scale was considered. Participants are asked to rate their experience of being immersed in a narrative on a 7-point Likert scale (1 = "not at all," 7 = "very much"), (see “Appendix C” for the full list of items).

To ensure that the results were not influenced by the different levels of empathy exhibited by the groups, empathy was assessed using the Interpersonal Reactivity Index (IRI) (Davis 1980), a 28-item self-report measure.

Electroencephalographic data processing

Electroencephalographic data were recorded continuously at a sampling rate of 1000 Hz using Net Station software (version 5.3.0.1, Electrical Geodesic, Inc., Eugene, OR, USA) and a 64-hydrocel geodesic sensor net, with impedances kept below 50 kΩ and reference to the vertex (Cz).

EEG data were processed using Net Station software (ibidem). The digital 30Hz low-pass filter was applied offline. The EEG data of each participant were segmented into epochs from -100 ms to 600 ms after stimulus onset. A baseline correction was applied 100 ms before stimulus onset. Artifact detection was set to 200 μV for bad channels, 150 μV for eye blinks, and 100 μV for eye movements (Electrical Geodesic, Inc., Eugene, OR, USA; Altavilla et al. 2022; 2021; Cecchini et al. 2013; Chiera et al. 2022; McPartland et al. 2010; Picton et al. 2000). The segments with an eye blink, an eye movement or more than 30% bad channels were excluded.

The ERPs components extracted were: the peak amplitude and latency of the P100 (80–150 ms) on the occipital [left electrode: 35(O1); right electrode: 39(O2)] and temporal electrodes [left: 24(T7); 23(T9); right: 52(T8); 55(T10)], the peak amplitude and latency of the N170 (150–210 ms) on occipital, temporal and temporo-parietal electrodes [left: 25(TP7); 29(TP9); 27(P5); 30(P7); 32(P9); right: 48(TP8); 47(TP10); 45(P6); 44(P8); 43(P10)], and the mean amplitude of the P200 (210-350 ms) and P300 (350-450 ms) on temporo-parietal and fronto-central electrodes [left: 16(C1); 20(C3); 7(FC1); 15(FC3); 14(FC5); 9(F1); 12(F3); 13(F5); right: 51(C2); 50(C4); 54(FC2); 53(FC4); 57(FC6); 3(F2); 60(F4); 59(F6)].

Statistical analysis

The repeated-measures ANOVAs 2 Group (psychological narrative vs. descriptive narrative) × 2 Time (pre- vs. post- reading)) × 4 Emotion (joy vs. fear vs. anger vs. sadness) × 2 Hemisphere (left vs. right) with Group as between factor and Time, Emotion, and Hemisphere as within factors were performed on the amplitude and latency of the P100 and N170, and on the amplitude of the P200 and P300 on each group of electrodes.

For each emotion, the repeated-measures ANOVAs 2 Group (psychological narrative vs. descriptive narrative) × 2 Time (pre- vs. post-reading) × 2 Hemisphere (left vs. right) with Group as between factor and Time and Hemisphere as within factors were performed on the amplitude and latency of the P100 and N170 and on the amplitude of the P200 and P300 on each group of electrodes. Fisher LSD post-hoc was applied.

Correlational analyses (Pearson's r) were performed between transport score and ERPs data at T1 for each emotion (joy, fear, anger, and sadness) in each group (psychological narrative and descriptive narrative). All statistical analyses were performed using Statistica v.8, StatSoft, Inc. 2007.

Results

After cleaning the EEG data, only participants with at least 20% artifact-free trials in each condition at both recordings (T0 and T1) were included in the in statistical analyses: sixteen in the psychological narrative group (6 women and 10 men; age M = 34.38; SD = 8.77) and fifteen in the descriptive narrative group (9 women and 6 men; age M = 24.07; SD = 7.38). Means and standard deviation of the trials included in PG at T0 joy (24.63 ± 4.47), fear (24.94 ± 4.82), anger (23.88 ± 4.90), sadness (24.3 ± 5.08) and at T1 joy (22.88 ± 6.84), fear (22.3 ± 7.87), anger (22.75 ± 6.50), sadness (23.00 ± 5.85); in DG at T0 joy (22.47 ± 5,32), fear (23.47 ± 4.45), anger (23.67 ± 5.01), sadness (22.00 ± 5.36) and T1 joy (19.67 ± 7.28), fear (20.80 ± 7.34), anger (21.33 ± 5.73), sadness (20.80 ± 7.44).

The results of the independent t-test analysis indicated that there was no significant difference between the two groups on the total IRI empathy score [t₍₂₉₎= − 1.25; p = 0.222]. However, a significant difference was observed between the two groups on the transportation score, with the participants of the psychological narrative group demonstrating a greater ability to transport themselves into the narrative compared to the participants of the descriptive narrative group [t₍₂₉₎= 3.10; p = 0.004].

ERPs amplitude

As shown in Table 1, the 2 (Group) × 2 (Time) × 4 (Emotion) × 2 (Hemisphere) repeated-measures ANOVAs showed a main effect of Time in P100 on occipital [F(1,29) = 11.58; p = 0.002] and temporo-parietal [F(1,29) = 6.90; p= 0.014] electrodes, in N170 on temporal [F(1,29) = 13.08; p= 0.001] and temporo-parietal [F(1,29) = 12.43; p = 0.001] electrodes, with a lower amplitude at T0 compared to T1 (Fig. 2). The main effect of Hemisphere was found at P100, P200, and P300 in all electrode groups, with a lower amplitude in the left hemisphere compared to the right.

Table 1.

ANOVAs Group [psychological narrative (PG) vs. descriptive narrative (DG)] × Time [pre-(T0) vs. post-(T1) reading] × Emotion [joy (J) vs. fear (F) vs. anger (A) vs. sadness (S)] × Hemisphere [left (L) vs. right (R)] on amplitude and latency in the P100 and N170, and on amplitude in the P200 and P300 on each left (L) and right (R) electrodes on which the components are extracted

Component	Electrodes Effects	Post-hoc
P100	Occipital Time F(1,29) = 11.58; p = 0.002 ηp2 = 0.29; G–G ε = 1	T0 < T1
	Hemisphere F(1,29) = 6.61; p = 0.016 ηp2 = 0.19; G–G ε = 1	L < R
	Temporo-parietal Time F(1,29) = 6.90; p = 0.014 ηp2 = 0.19; G–G ε = 1	T0 < T1
	Hemisphere F(1,29) = 5.97; p = 0.021 ηp2 = 0.17; G–G ε = 1	L < R
	Emotion × Hemisphere F(3,87) = 3.68; p = 0.015 ηp2 = 0.11; G–G ε = 0.77; G–G(Adj) p = 0.025	J (L)< J (R) p < 0.001; J (L)> A (L) p = 0.003 J (L)< S (R) p < 0.001; J (R)> F (L) p< 0.001 J (R)> F (R) p < 0.001; J (R)> A (L) p < 0.001 J (R)> A (R) p < 0.001; J (R)> S (L) p < 0.001 F (L)< S (R) p< 0.001; F (R)< A (L) p= 0.002 F (R)< S (R) p < 0.001; A (L)< A (R) p =0.021 A (L)< S (L) p = 0.008; A (L)< S (R) p < 0.001 A (R)< S (R) p < 0.001; S (L)< S (R) p < 0.001
N170	Temporal Time F(1,29) = 13.08; p <0.001; ηp2 = 0.35; G–G ε = 1	T0 > T1
	Temporal Time × Emotion × Hemisphere F(3,87) = 3.99; p = 0.010 ηp2 = 0.12;G–G ε = 0.89;G–G(Adj) p = 0.013	T0 A(L) > all conditions p <.01 T0 F(L) > T0 S(R) p = 0.035 T0 F(L) > T1 F(R) p = 0.035 T0 F(R) > T0 S(R) p = 0.041 T0 F(R) > T1 F(R) p = 0.042 T0 S(L) > T1 S(R) p = 0.045 T0 S(L) > T1 F(R) p = 0.045
	Temporo-parietal Time F(1,29) = 12.43; p = 0.001 ηp2 = 0.30; G–G ε = 1	T0 > T1
	Time × Hemisphere F(1,29) = 5.72; p = 0.023 ηp2 = 0.16; G–G ε = 1	T0(L) > T0(R) p = 0.035 T0(L) > T1(L) p < 0.001 T0(L) > T1(R) p < 0.001 T0(R) > T1(L) p < 0.001 T0(R) > T1(R) p < 0.001
P200	Temporo-parietal Hemisphere F(1,29) = 8.56; p = 0.007 ηp2 = 0.23; G–G ε = 1	L < R
	Time × Hemisphere F(1,29) = 4.56; p = 0.041 ηp2 = 0.14; G–G ε = 1	T0(L) < T0 (R) p = 0.002 T0(L) < T1(L) p < 0.001 T0(L) < T1(R) p < 0.001 T0(R) < T1(R) p < 0.001 T1(L) < T1(R) p < 0.001
	Fronto-central Hemisphere F(1,29) = 10.87; p = 0.003 ηp2 = 0.27; G–G ε = 1	L < R
P300	Temporo-parietal Hemisphere F(1,29) = 10.41; p = 0.003 ηp2 = 0.26; G–G ε = 1	L < R
	Fronto-central Hemisphere F(1,29) = 13.78; p< 0.001 ηp2 = 0.32; G–G ε = 1	L < R

In bold the main post-hoc interaction effects discussed

Fig. 2.

ERPs grand average of the emotions (joy, fear, anger and sadness) of both groups (psychological and descriptive) at T0 compared to T1 on left and right occipital, temporal, temporo-parietal and fronto-central electrodes.

The interaction effect of Time per Hemisphere was found in N170 [F(1,29) = 5.72; p =0.023] and P200 [F(1,29) = 4.56; p =0.041] on temporo-parietal electrodes where the right hemisphere at T1 presented a greater amplitude compared to left at T0 and at T1. An interaction effect Emotion per Hemisphere was found in P100 at temporo-parietal electrodes [F(3,87) = 3.68; p = 0.015], with joy eliciting a greater amplitude in the right hemisphere compared to anger and sadness, and sadness eliciting a greater amplitude in the right hemisphere compared to fear and anger (p < 0.001).

The 2 (Group) × 2 (Time) × 2 (Hemisphere) repeated-measures ANOVAs on the amplitude in response to each emotion showed the main effect of Time in P100 in response to joy [F(1,29) = 8.81; p = 0.006], fear [F(1,29) = 8.44; p= 0.007], anger [F(1,29) = 10.36; p= 0.003], and sadness [F(1,29) = 7.65; p = 0. 010] on occipital electrodes with lower amplitude at T0 compared to T1; in N170 in response to fear [F(1,29) = 6.67; p = 0.015] on temporo-parietal electrodes, in response to anger on temporal [F(1,29) = 6.84; p = 0.014] and temporo-parietal [F(1,29) = 9.94; p = 0.004] electrodes, and in response to sadness [F(1,29) = 4.94; p = 0.034] on temporal electrodes with less negative amplitude at T0 compared to T1. The main effect of hemisphere was found for the amplitude of almost all components and for almost all groups of electrodes, with the left hemisphere showing lower amplitude compared to the right hemisphere in response to all emotions. The interaction effect of Group per Time was found in P100 on occipital electrodes [F(1,29) = 4.59; p = 0.041] and in P300 on temporo-parietal electrodes [F(1,29) = 5.26; p = 0.031] in response to anger, with the descriptive narrative group presenting a greater amplitude at T1 compared to T0 and compared to the psychological narrative group at T1. The interaction effect of Group per Time per Hemisphere was found in N170 on temporal electrodes [F(1,29) = 4.28; p = 0.047] in response to joy, post hoc analyses showed that the descriptive narrative group presented a greater negative amplitude in the left hemisphere at T1 compared to T0. In Tables 1 and 2, all post-hoc results are shown.

Table 2.

Significant effects of the ANOVAs Group [psychological narrative (PG) vs. descriptive narrative (DG)] × Time [pre-(T0) vs. post-(T1) reading] × Hemisphere [left (L) vs. right (R)] on each emotion (joy, fear, anger and sadness) on amplitude and latency in the P100 and N170, and on amplitude in the P200 and P300 on each left (L) and right (R) electrodes on which the components are extracted

Emotion	Component	Electrodes Effects	Post-hoc
Joy	P100	Occipital Time F(1,29) = 8.81; p = 0.006 ηp2 = 0.23; G–G ε = 1	T0 < T1
		Hemisphere F(1,29) = 6.38; p = 0.017 ηp2 = .18; G-G ε= 1	L < R
	N170	Temporal Group × Time × Hemisphere F(1,29) = 4.28; p = 0.047 ηp2 = 0.13; G–G ε = 1	DG T0 L > DG T1 L p <0.001 DG T0 R > DG T1 L p = 0.004 DG T0 L > DG T1 R p = 0.013
	P200	Temporo-parietal Hemisphere F(1,29) = 7.08; p = 0.013 ηp2 = 0.20; G–G ε = 1	L< R
		Fronto-central Hemisphere F(1,29) = 6.25; p = 0.015 ηp2= 0.19; G–G ε = 1	L < R
	P300	Temporo-parietal Hemisphere F(1,29) = 7.99; p = 0.008 ηp2 = 0.22; G–G ε = 1	L < R
		Fronto-central Hemisphere F(1,29) = 7.21; p = 0.012 ηp2 = 0.20; G–G ε = 1	L < R
Fear	P100	Occipital Time F(1,29) = 8.44; p = 0.007 ηp2 = 0.23; G–G ε = 1	T0 < T1
		Hemisphere F(1,29) = 6.78; p = 0.014 ηp2 = 0.19; G–G ε = 1	L < R
		Group × Hemisphere F(1,29)= 5.04; p = 0.033 ηp2= .15; G-G ε= 1	PG L < PG R p =0.041
		Group × Time × Hemisphere F(1,29) = 4.83; p = 0.043 ηp2 = 0.13; G–G ε = 1	PG T0 L < PG T0 R p = 0.023
	N170	Temporo-parietal Time F(1,29) = 6.67; p = 0.015 ηp2 = 0.19; G–G ε = 1	T0 > T1
		Tempo-parietal Group × Time × Hemisphere F(1,29) = 7.16; p = 0.012 ηp2 = 0.20; G–G ε = 1	PG T0 L > PG T0 R p = 0.014 PG T0 R < PG T1 L p = 0.008 PG TO R < PG T1 R p = 0.013 DG T0 L < DG T0 R p = 0.020 DG T0 L < DG T1 L p = 0.018 DG T0 L < DG T1 R p = 0.049
	P200	Temporo-parietal Hemisphere F(1,29) = 5.18; p = 0.030 ηp2 = 0.15; G–G ε = 1	L < R
		Fronto-central Hemisphere F(1,29) = 8.33; p = 0.007 ηp2 = 0.22; G–G ε = 1	L < R
	P300	Temporo-parietal Hemisphere F(1,29) = 5.22; p = 0.030 ηp2 = 0.15; G–G ε = 1	L < R
		Fronto-central Hemisphere F(1,29) = 13.46; p = 0.001 ηp2 = 0.32; G–G ε = 1	L < R
Anger	P100	Occipital Time F(1,29) = 10.36; p = 0.003 ηp2 = 0.26; G–G ε = 1	T0 < T1
		Group × Time F(1,29) = 4.59; p = 0.041 ηp2 = 0.14; G–G ε = 1	PG T0 < DG T1 p = 0.007 DG T0 < DG T1 p < 0.001
		Temporal Time F(1,29) = 4.83; p = 0.036 ηp2 = 0.14; G–G ε = 1	T0 > T1
	N170	Temporal Time F(1,29) = 6.84; p = 0.014 ηp2 = 0.19; G–G ε = 1	T0 > T1
		Temporo-parietal Time F(1,29) = 9.94; p = 0.004 ηp2 = 0.26; G–G ε = 1	T0 > T1
		Temporal Time × Hemisphere F(1,29) = 8.39; p = 0.007 ηp2 = 0.22; G–G ε = 1	T0 L > T0 R p < 0.001 T0 L > T1 L p < 0.001 T0 L > T1 R p < 0.001
	P200	Temporo-parietal Hemisphere F(1,29) = 7.47; p = 0.011 ηp2 = 0.20; G–G ε = 1	L < R
		Fronto-central Hemisphere F(1,29) = 11.56; p = 0.002 ηp2 = 0.29; G–G ε = 1	L < R
	P300	Temporo-parietal Hemisphere F(1,29) = 7.41; p = 0.011 ηp2 = 0.15; G–G ε = 1	L < R
		Gruppo × Time F(1,29) = 5.16; p = 0.031 ηp2 = 0.20; G–G ε= 1	PG T1 < DG T1 p = 0.040
		Fronto-central Hemisphere F(1,29) = 14.54; p < 0.001 ηp2 = 0.33; G–G ε = 1	L < R
Sadness	P100	Occipital Time F(1,29) = 7.65; p = 0.010 ηp2 = 0.21; G–G ε= 1	T0 < T1
		Hemisphere F(1,29) = 6.14; p = 0.019 ηp2 = 0.17; G–G ε = 1	L < R
	N170	Temporal Time F(1,29) = 4.94; p = 0.034 ηp2 = 0.15; G–G ε = 1	T0 > T1
	P200	Temporo-parietal Hemisphere F(1,29) = 7.65; p = 0.010 ηp2 = 0.21; G–G ε = 1	L < R
		Fronto-central Hemisphere F(1,29) = 10.357; p = 0.003 ηp2 = 0.27; G–G ε = 1	L < R
	P300	Temporo-parietal Hemisphere F(1,29) = 6.72; p = 0.015 ηp2 = 0.19; G–G ε= 1	L < R
		Fronto-central Hemisphere F(1,29) = 12.62; p = 0.001 ηp2 = 0.30; G–G ε = 1	L < R

In bold the main post-hoc interaction effects discussed

ERPs latency

As shown in Table 1, the 2 (Group) × 2 (Time) × 4 (Emotion) × 2 (Hemisphere) repeated-measures ANOVAs showed an interaction effect of Time per Emotion per Hemisphere in N170 on temporal electrodes [F(3,87) = 3.99; p = 0.010], with anger showing a later latency at T0 compared to all conditions (p < 0.001), and fear showing a later latency at T0 compared to T1 in the right hemisphere.

The 2 (Group) × 2 (Time) × 2 (Hemisphere) repeated-measures ANOVAs on the latency in response to each emotion showed a main effect of time in P100 at temporal electrodes [F(1,29)= 4.83; p = 0.036] in response to anger, with shorter latency at T0 compared to T1.

An interaction effect of Group per Hemisphere was found in P100 at occipital electrodes [F(1,29)= 5.04; p = 0.033] in response to fear: the psychological narrative group showed a shorter latency in the left hemisphere compared to the right hemisphere. An interaction effect of Time per Hemisphere was found in N170 on temporal electrodes [F(1,29) = 8.39; p = 0.007] in response to anger: at T0, the left hemisphere showed a later latency compared to the right hemisphere, and at T1, it showed a later latency compared to T0. The interaction effect of Group per Time per Hemisphere was found at P100 on occipital electrodes [F(1,29) = 4.83; p = 0.043] and at N170 on temporo-parietal electrodes [F(1,29)= 7.16; p =0. 012] in response to fear: the results indicated that the psychological narrative group showed a later latency at T1 compared to T0 in the right hemisphere, while the descriptive narrative group showed a later latency at T1 compared to T0 in the left hemisphere. Tables 1 and 2 present the complete post hoc results.

Correlation analyses between transportation score and ERPs data

As shown in Table 3, in the psychological narrative group, correlational analyses (Pearson's r) indicated a positive correlation between the transportation score and the amplitude of the P100 on the right occipital electrode in response to joy and on the right temporal electrodes in response to joy (r = 0.64; p = 0.008), fear (r = 0.51; p = 0.043), and sadness (r = 0.50; p = 0.048).

Table 3.

Significant correlations (Pearson’s r) between transportation score and amplitude and latency of ERP components (P100, N170, P200, P300) at T1 on each left (L) and right (R) electrodes on which the components are extracted

	Group	Emotion	Component
			P100	N170	P200	P300
Transportation score	Psychological narrative	Joy	R Temporal r = 0.64; p = 0.008	ns	ns	ns
		Fear	R Occipital r = 0.51; p = 0.043 R Temporal r= .59; p = 0.016	ns	ns	ns
		Anger	ns	ns	ns	ns
		Sadness	R Temporal r = 0.50; p = 0.048	ns	ns	ns
	Descriptive narrative	Joy	ns	R Temporal r =0.54; p = 0.037	ns	ns
		Fear	ns	ns	ns	ns
		Anger	ns	ns	ns	ns
		Sadness	ns	ns	ns	ns

In the descriptive narrative group, there was a positive correlation between transportation score and N170 latency in response to joy on the right temporal electrodes (r = 0.54; p = 0.037).

Discussion

The face is one of the most informative and complex vehicles of social interaction. Several ERPs and neuroimaging studies have shown that the processing of faces and facial emotional expressions is modulated by social cognitive abilities. In recent years, a significant body of research has demonstrated the efficacy of narrative as a tool for investigating social cognition and its impact on social processes (Djikic et al. 2013; Kidd and Castano 2013; Koopman 2015; Mar and Oatley 2008, Mar et al. 2011; Paluck 2009). These studies have consistently shown that reading narrative influences social cognition, particularly when individuals report high levels of engagement with the story. To date, however, no neurophysiological studies appear to have examined whether reading a narrative that focuses more on the mental and emotional states of the character modulates neural activation in response to facial emotional expressions differently from reading a descriptive and more neutral narrative. Furthermore, it is unclear whether this activation correlates with the degree to which an individual is transported into the story. To address this question, the present study recorded the electrical brain activity of participants in two groups, the psychological narrative group and the descriptive narrative group, in response to facial emotional expressions (joy, anger, fear, and sadness) before and after a reading task.

The primary results showed a main effect of Time on the P100 at occipital and temporo-parietal electrodes, and the N170 at temporal and temporo-parietal electrodes, with greater amplitude observed at T1 compared to T0. The ANOVAs on each emotion also showed a main effect of Time in the P100 in response to joy, anger, and sadness on occipital electrodes, with greater amplitude at T1 compared to T0, and in the N170 in response to fear, anger, and sadness on temporo-parietal electrodes, and anger on temporal electrodes, with greater negative amplitude at T1 compared to T0.

Previous studies have shown a correlation between greater electrophysiological activation, particularly in the temporal N170, and improved social cognition (Choi and Watanuki 2014; Meaux et al. 2014; Petroni et al. 2011). Although emotional recognition was not assessed in the present study, the results seem to indicate that reading narratives, whether psychological or descriptive, increases attention to socioemotional stimuli.

It is crucial to highlight the early onset of this process, as evidenced by the observed effects on the amplitude of the early ERP components (P100 and N170), which are associated with low-order perceptual processes. This perceptual modulation can be explained by the fact that, as claimed by several authors (Mar and Oatley 2008; Mar et al. 2009; 2010; 2018; Mason and Just 2009; Oatley 1999), narration seems to provide an opportunity to simulate social scenarios, assume different perspectives, understand and anticipate the mental states and goals of the characters. This internal simulation "training" elicits greater socio-emotional perception of the real world. Consistent with these assumptions, the finding is also present in the group that reads the descriptive narrative probably because even in the case in which only the character's actions are described, and not his emotions and mental states, simulation and transport processes are activated to understand his intentions and goals.

Our results also revealed a significant Time per Emotion per Hemisphere interaction effect in latency. A post-hoc analysis indicated a shorter latency of the N170 in response to fear on the right temporal electrodes and in response to anger on the left ones after reading (T1) both psychological and descriptive narratives compared to T0. This result seems to confirm the previous interpretation and also suggests a faster perception of fear and anger facial expressions after reading the narrative, with a hemispheric lateralization related to the type of emotion. In line with the evolutionary approach to affective neuroscience research (e.g., Montag and Panksepp 2016; 2017; Panksepp 1998), fear and anger are considered negative basic emotions that are beneficial for individual survival, as they trigger the activation of a "fight-flight-freeze" system in dangerous situations. Greater and faster perceptual sensitivity to these emotions in a social context could facilitate more effective emotional regulation by enhancing emotional responsiveness and synchronization with others during social interaction.

Consistent with the latter point, the interaction effect of Group per Time per Hemisphere showed a longer latency of the N170 at temporo-parietal electrodes to fear following reading. This was observed specifically in the right hemisphere in the psychological narrative group and in the left hemisphere in the descriptive narrative group. This prolonged latency may indicate that after a rapid initial processing of emotion, more extended temporal processing in the temporo-parietal regions is required for the implementation of top-down emotional regulation.

By returning to the hemispheric lateralization, the greater involvement of the left hemisphere in response to anger and the right hemisphere in response to fear can be explained by reference to the approach/withdrawal motivational dimension of emotions, in line with several previous studies that have used visual and auditory emotional stimuli (Carver and Harmon-Jones 2009; Demaree et al. 2005; Gadea et al. 2011; Palomero-Gallagher and Amunts 2022). Consistent with the results of the present study, these investigations have shown that emotions that direct the individual toward the stimulus, such as anger and joy, are associated with the left hemisphere, whereas emotions that motivate withdrawal from the stimulus, such as fear and sadness, are associated with the right hemisphere.

Furthermore, it is important to highlight that, unlike several studies in which participants explicitly pay attention to or label the presented emotions into categories as well as some studies that have directly incorporated different tasks to assess to what extent task demands can modulate ERPs based on emotional faces (e.g., Wronka and Walentowska 2011; Rellecke et al. 2012; Itier and Neath-Tavares 2017; see Schindler et al. review 2023), in the present study the different neural response to different emotions was detected during an implicit task in which no explicit action of the participant was required suggesting bottom-up processes of emotional discrimination. In future research it would be interesting to include an explicit recognition task to also evaluate high-order processes.

Another important result was the interaction effect of Group per Time found in P100 on occipital electrodes and in P300 on temporo-parietal electrodes in response to anger with the descriptive narrative group presenting a greater amplitude at T1 compared to T0 and compared to psychological narrative group at T1, and the interaction effect of Group per Time per Hemisphere found in N170 on temporal electrode in response to joy with the descriptive narrative group presenting a greater negative amplitude in the left hemisphere at T1 compared to T0. The greater post-reading amplitudes observed in the descriptive narrative group can be attributed to the type of emotion evoked by the narrative. This narrative is characterized by an extremely long, emphatic, and meticulous description of details, including both the actions of the character and the description of the setting. The passage about the trout in the river (see “Appendix B”) is a prime example. Such a narrative may have elicited feelings of nervousness and boredom, which in turn may have more strongly modulated the response to emotions such as anger.

As predicted by the experimental manipulation, the psychological narrative group was found to have a higher transport score than the descriptive group. The psychological narrative may have evoked strong feelings of fear and distress, given the theme of illness and grief. Correlation analysis showed a positive correlation between transport score and P100 amplitude in response to joy, fear, and sadness only in this group after reading. The strong correlation also with joy shows that the more one immerses oneself in a stressful text, the greater the neural response to a positive emotion seems to be, suggesting the detection of an emotional incongruence with respect to the themes of illness and mourning present in the psychological narrative.

These results are of great importance for the purposes of the present study, as it supports the hypothesis that transportation into the story mediates the early brain perceptive response to socio-emotional stimuli, as shown by several behavioural studies (Bal and Veltkamp 2013; Johnson 2012; Sestir and Green 2010; Walkington et al. 2020). The greater the transport into the narrative, the greater the internal simulation process of emotions and mental states, which will modulate the perception of the real social world after reading. On the contrary, in the descriptive group, the transportation score correlates only with the latency of the N170 in response to joy, so immersion in an emotionless narrative seems to delay the processing of a positive emotion.

Finally, the interaction effect of Time per Hemisphere found at N170 and P200 on temporo-parietal electrodes showed a greater amplitude of the right hemisphere at T1 compared to the left at T0 and at T1. The greater involvement of the right hemisphere after reading could support the interpretation of greater attention to socioemotional stimuli. Indeed, several electrophysiological and neuroimaging studies show that the brain correlate of emotional processes and social skills, such as empathy and theory of mind, predominantly involves the right hemisphere (Adolphs 2002; Adolphs et al. 2000; Damasio 1994; Schore 2005). In this regard, the results of a recent ERP study (Altavilla et al. 2024) investigating how first- and third-person narrative modulates social cognitive ability also report a greater involvement of the right hemisphere. In this study, neural responses to eye expressions were recorded during an experimental explicit recognition task before and after reading the narrative. The results showed a smaller N220-400 on right fronto-central electrodes in response to eye expressions in both groups, and a larger N100 on left fronto-central electrodes and a larger P220-400 on right temporo-parietal electrodes in response to eye expressions only in the group reading the third-person story. No significant differences were found in the behavioural accuracy data after the reading task, probably due to the small sample size.

Overall, given the findings, it is plausible to hypothesize that the results of the present study may also be consistent with previous behavioural studies that have demonstrated improvements in social cognition following narrative reading (Djikic et al. 2013; Kidd and Castano 2013; Koopman 2015; Mar and Oatley 2008; Mar et al. 2011; Paluck 2009). However, further research using a behavioural emotion recognition task and a larger sample size, as also suggested by Altavilla and colleagues (2024), is needed to substantiate this claim.

In fact, these last points can be seen as two of the limitations of the present study: missing behavioural data and the small sample size. The small sample size is due to the double electroencephalographic recording, which led to the exclusion of eleven participants due to the presence of excessive artefacts. Although other ERP studies have reported similar sample sizes (e.g., Altavilla et al. 2022; 2024; Massaro et al. 2018; Ernst et al. 2013), it would be important to increase the sample size in future research to reduce the presence of artefacts due to double recording. One possibility would be to project the text onto the screen instead of providing it in paper form. Although this would reduce the naturalness and spontaneity of the reading, it might increase the likelihood of having a lower number of artefacts. It is also important to emphasise that, given the exploratory nature of the study, the data were not corrected for multiple comparisons, so further studies with even larger samples of homogeneous age are needed to generalise the data of the present study. Finally, in future research, a larger sample size with men and women homogeneously distributed across groups would also allow for gender-based analyses.

In conclusion, the present study suggests that reading narratives, whether psychological or descriptive, modulates the early brain response to socioemotional stimuli, i.e., facial emotional expressions. Moreover, this modulation is positively related to the degree to which an individual allows himself or herself to be transported into the story. Regarding this last point, it would be beneficial to conduct further research in the future to investigate the individual differences that might influence the ability to be transported into a story, including psychological variables, emotional regulation abilities, and attachment styles.

Appendix A: An excerpt of the psychological narrative text

[…] "Why won't he wake up?" Ann said. "Howard? I want some answers from these people." Howard didn't say anything. He sat down again in the chair and crossed one leg over the other. He rubbed his face. He looked at his son and then he settled back in the chair, closed his eyes, and went to sleep. Ann walked to the window and looked out at the parking lot. It was night, and cars were driving into and out of the parking lot with their lights on. She stood at the window with her hands gripping the sill and knew in her heart that they were into something now, something hard. She was afraid, and her teeth began to chatter until she tightened her jaws. She saw a big car stop in front of the hospital and someone, a woman in a long coat, get into the car. She wished she were that woman and somebody, anybody, was driving her away from here to somewhere else, a place where she would find Scotty waiting for her when she stepped out of the car, ready to say Mom and let her gather him in her arms.

In a little while, Howard woke up. He looked at the boy again. Then he got up from the chair, stretched, and went over to stand beside her at the window. They both stared out at the parking lot. They didn't say anything. But they seemed to feel each other's insides now, as though the worry had made them transparent in a perfectly natural way.

Appendix B: An excerpt of the descriptive narrative text

[…] The river was there. It swirled against the log spiles of the bridge. Nick looked down into the clear, brown water, coloured from the pebbly bottom, and he watched the trout keeping themselves steady in the current with fin wavering As he watched them they changed their positions by quick angles, only to hold steady in the fast water again. Nick watched them a long time! He watched them holding themselves with their noses into the current, many trout in deep, fast-moving water, slightly distorted as he watched far down through the glassy convex surface of the pool, its surface pushing and swelling smooth against the resistance of the log-driven piles of the bridge. At the bottom of the pool were the big trout. Nick didn't see them at first. Then he saw them at the bottom of the pool, big trout looking to hold themselves on the gravel bottom in a varying mist of gravel and sand, raised in spurts by the current! […]

He saw the trout in the water jerking with his head and body against the shifting tangent of the line in the stream. Nick took the line in his left hand and pulled the trout, thumping tiredly against the current, to the surface. flis back was mottled the clear, water-over-gravel colour, his side flashing in the sun. The rod under his right arm, Nick stooped, dipping his right hand into the current. He held the trout, never still, with his moist right hand, while he unhooked the barb from his mouth, then dropped him back into the stream.

Appendix C: Questionnaire on narrative transportation (translated in Italian from Green and Brock 2000, 2013)

Circle the number (from1-Not at all to 7-Very much) under each question that best represents your opinion about the narrative you just read.

1. While I was reading the narrative, I could easily picture the events in it taking place.

2. While I was reading the narrative, activity going on in the room around me was on my mind.

3. I could picture myself in the scene of the events described in the narrative.

4. I was mentally involved in the narrative while reading it.

5. After the narrative ended, I found it easy to put it out of my mind.

6. I wanted to learn how the narrative ended.

7. The narrative affected me emotionally.

8. I found myself thinking of ways the narrative could have turned out differently.

9. I found my mind wandering while reading the narrative.

10. The events in the narrative are relevant to my everyday life.

11. The events in the narrative have changed my life.

12. I had a vivid mental image of [character name].

Notes: Items 2, 5, and 9 are reverse-scored.

Item 12 can be repeated for the number of main characters in the story, substituting a different

character name for each item.

Author contribution

D.A.: wrote original draft, methodology, results, conceptualization; V.D.: writing, methodology, conceptualization; A.C.: review and editing, conceptualization; S.C.: methodology; I.A.: writing, conceptualization; F.F.: review and editing, conceptualization, supervision. All authors reviewed the manuscript.

Funding

Open access funding provided by Università degli Studi Roma Tre within the CRUI-CARE Agreement.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interest

The authors declare that there is no conflict of interest. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.