Numerical cognition is important in everyday life. It involves the capability to distinguish between numerical magnitudes and to contemplate this conscious awareness along with the skill of counting, which are important abilities for representing numbers mentally (Nieder, 2005). However, although number magnitude and counting stand out as prominent numerical properties (Shepard et al., 1975), the human understanding of numbers is not limited to such attributes. Various other characteristics of how numbers are mentally represented have been noted to influence one’s abilities in numerical tasks. Many studies that explored numerical representations assume the existence of a mental number line (MNL), wherein numbers and quantities are organized spatially. A well-established tradition suggests that individuals translate numerical representations into corresponding spatial positions (Dehaene et al., 1993; Fias, 1996; Gevers et al., 2006; Nourouzi Mehlabani et al., 2020). Given the importance of performance in MNL for mathematical achievements (Simms et al., 2016; Nazari et al., 2021), investigating the MNL associations with cognitive and emotional factors is important for understanding the neural underpinnings of numerical cognition. Recent research has highlighted the significance of distinct brain oscillations in understanding the neural dynamics of numerical processing using electroencephalography (EEG). These oscillations are also considered to have functional significance in cognitive and emotional processes (Ray and Cole, 1985; Bell and Diaz, 2012). The resting-state (rs) EEG involve characterizing specific EEG signals and reflect the intrinsic properties of neural systems (Yedukondalu and Sharma, 2023b). Understanding the links between rs-EEG and performance in MNL holds promise in elucidating the neural correlates underlying spatial-numerical associations. These insights can pave the way for establishing an empirical association between rs-EEG with respect to general emotional and cognitive traits, on the one hand, and aspects of numerical cognition, on the other. The aim of the current work was to offer insights into the correlation between rs- EEG cognitive and emotional trait measures in spatial-numerical associations.

Various approaches are employed to investigate the spatial-numerical associations, such as magnitude classification (Wood et al., 2008), the spatial-numerical associations of response codes (SNARC) effect (Dehaene et al., 1993), and spatial biases in mathematics (Pinhas and Fischer, 2008). MNL suggests that individuals mentally represent numbers on a line, with smaller numbers located to the left and larger numbers located to the right (Dehaene et al., 1993; Izard and Dehaene, 2008). The number-to-position (NP) task stands as a commonly utilized method for exploring how numbers are represented on the MNL (Siegler and Opfer, 2003; Berteletti et al., 2010; Nourouzi Mehlabani et al., 2020; Nazari et al., 2021). The NP task involves presenting a number line with marked endpoints (such as 0–100) alongside a designated target number. During each trial, participants are required to identify and mark the position of the target number on the number line. Research findings suggest that healthy adults demonstrate a tendency toward a leftward bias in spatial-numerical associations (Lourenco and Longo, 2009), which contrasts with observations influenced by various disorders (Cavézian et al., 2007; Förster et al., 2011; Barker et al., 2019).

Previous works suggested that individuals with right parietal lesions, combined with persistent left neglect, demonstrate a rightward bias in tasks involving the MNL. Specifically, when identifying the midpoint between two numbers, these patients tend to overestimate the midpoint within a given number pair (Zorzi et al., 2002; Rossetti et al., 2004). Also, a study involving a patient with right neglect has corroborated these findings, suggesting the potential parallelism between the spatial attention system utilized in MNL tasks (Pia et al., 2009). In contrast to individuals with unilateral right brain lesions, healthy adults exhibit a leftward bias in MNL (Loftus et al., 2009; Lourenco and Longo, 2009). This consistent leftward inclination is termed pseudoneglect (Bowers and Heilman, 1980a; Jewell and McCourt, 2000). It is suggested that hemispheric asymmetries in spatial attention occur in such direction biases. Moreover, following transcranial magnetic stimulation (TMS) of the right posterior parietal cortex, healthy adults showed a rightward bias (Fierro et al., 2001). These findings point to the special role of the parietal cortex in modulating attention that might lead to directional biases in MNL. However, our recent findings pointed to the importance of the emotional valence of the participants in MNL direction bias (Sabaghypour et al., 2023). Specifically, we uncovered evidence showing a distinct impact of emotional valence (using IAPS-based film clips inducing positive and negative emotions), on participants’ numerical estimations and spatial representations along the MNL. The investigation revealed that individuals experiencing positive emotional valence exhibited a notable rightward bias in their numerical estimations, whereas those in the negative valence group demonstrated a leftward bias. These insights expand upon previous studies suggesting a connection between emotions and numerical cognition (De Smedt et al., 2009; Holmes and Lourenco, 2011; Schukajlow et al., 2017; Fischer et al., 2021), highlighting the possible links between affective traits and spatial-numerical associations.

As per the valence hypothesis, the right hemisphere mediates negative emotions like sadness, while the left hemisphere deals with positive emotions like happiness (Borod et al., 1998; Martin and Altarriba, 2017; Roesmann et al., 2019; Sabourimoghadam et al., 2022). Consequently, people tend to associate positive emotions with the right side and negative emotions with the left side of the space (Casasanto, 2009; Freddi et al., 2016). Hence, based on the valence hypothesis, negative emotions might bias attention toward the left hemispace and positive emotions might bias attention more to the right (Foster et al., 2008). In addition, animal studies also demonstrate preferences in two crucial survival activities (Rogers, 2017). Feeding tendencies lean toward right-sided behavior, influenced by the left hemisphere (Andrew et al., 2000; Karenina et al., 2016). Conversely, avoiding predators is associated with behavior regulated by the right hemisphere (Robins and Rogers, 2004; Koboroff et al., 2008). Furthermore, a distinct association linking numbers and space in a left-to-right direction has been evidenced in non-human animals (Rugani et al., 2007, 2011; Adachi, 2014). Hence, it has been suggested that changes in numerosity toward larger or smaller magnitudes might be associated with increased activity in the left (positive valence) or right (negative valence) hemisphere (Vallortigara, 2018). Investigating the frontal asymmetry in different frequency bands [i.e., theta (3–7 Hz), alpha (8–12 Hz), and beta (13–25 Hz)] might advance this proposal by providing links to neural activities that may correlate with direction bias in MNL in humans. ‘Asymmetry’ is defined by measurements taken from both sides, with the difference between these measurements serving as an index of the difference between two hemispheres.

Moreover, frontal rs-EEG might be a good predicator of emotional processes (Jackson et al., 2000, 2003; Campagna et al., 2021). Neuroimaging research also suggests that individual differences in emotional traits may be reliably indexed by prefrontal activity at rest (Paradiso et al., 1995; Jackson et al., 2000; Dalton et al., 2002; Zhang et al., 2020). According to the research on frontal EEG asymmetry, increased left prefrontal activity—indexed by alpha and increased beta activity (e.g., at the F3 electrode)—is correlated with positive emotions. As well, increased left frontal theta activity is linked to negative emotional processing (Dharmadhikari et al., 2018; Koller-Schlaud et al., 2020). Conversely, elevated right prefrontal activity (e.g., at the F4 electrode) is associated with negative emotions (Jackson et al., 2003; Reznik and Allen, 2018). Therefore, exploring how this functional hemispheric imbalance correlates with direction bias in spatial-numerical associations could offer an intriguing path for understanding how the brain’s asymmetrical activity, particularly in the frontal regions, is linked to performance in MNL.

In addition to its role in emotional processes, EEG findings also shed light on the correlation between specific frequency bands and cognitive processing, emphasizing the importance of distinct EEG measures (Yedukondalu and Sharma, 2023a; Jain et al., 2024; Yedukondalu et al., 2024). For instance, the theta/beta ratio (TBR) at central areas (like the Cz electrode) (Clarke et al., 2019), and the individual frontal alpha peak frequency (APF) as trait-like parameters of EEG (Sambrook and Goslin, 2015), could serve as indicative measures of cognitive processing capacity and performance (Finley et al., 2022). Alpha frequency is also divided into 8–10 Hz (low alpha) and 10–12 Hz (high alpha), allowing for comparison between these sub-bands (Pfurtscheller, 1989; Zoefel et al., 2011). This classification relies on the assumption that alpha rhythms stem from at least two separate neural groups, one oscillating below 10 Hz and the other above 10 Hz. Regarding the amplitude of sub-alpha bands, results revealed that high alpha interacts with visual cognitive tasks and semantic memory demands, whereas low alpha signifies a broad attentional demand not specifically tied to the task at hand (Klimesch, 1999). As well, the increased beta amplitude at occipital and frontal regions observed in individuals with trait-anxiety might contribute to decreased cognitive performance (Gordeev, 2007; Price and Budzynski, 2009).

In the present work, EEGs were recorded from participants at rest who were subsequently directed to complete a NP task. In keeping with the hypothesis of a prominent role of frontal EEG asymmetry in emotional processing, the main purpose (see Figure 1) of the present work was that higher baseline activation in the left hemisphere (identified through decreased alpha and/or increased beta) would correlate with a tendency toward rightward bias or, at minimum, a reduced leftward bias commonly observed in the spatial-numerical associations (Göbel et al., 2006; Loftus et al., 2008; Rotondaro et al., 2015). Moreover, building upon the role of EEG measures in cognitive processing, the investigation also aimed to evaluate their relation to absolute errors and reaction times (RTs) in MNL performance. Overall, the exploration of distinct oscillations intends to uncover their implications with respect to the accuracy and speed of MNL performance.

The present work investigated the rs-EEG regional characteristics and their correlation with mean direction error, mean absolute error, and RT in the CNP task. In our experiment, absolute errors indicated the extent of deviation of numerical magnitudes estimations, whereas direction bias indicated the deviation based solely on whether it is higher or lower than the expected value. Frontal asymmetry during rest is considered a trait variable with links to emotional processing (Paradiso et al., 1995; Jackson et al., 2000; Dalton et al., 2002; Zhang et al., 2020). Considering the recent insights highlighting the impact of emotions on MNL bias (Vallortigara, 2018; Sabaghypour et al., 2023), the idea of a potential correlation between asymmetry in frontal activity and direction bias was considered. Moreover, as a transfer of results from prior rs-EEG studies linking TBR, TSR, lowAlpha/highAlpha ratio, APF in centro-frontal regions to cognitive function (Klimesch, 1999; Sambrook and Goslin, 2015; Clarke et al., 2019; Finley et al., 2022), and the link between beta activity and trait-anxiety (Gordeev, 2007; Price and Budzynski, 2009), these rs-EEG measures were also examined.

First of all, since almost all our participants exhibited leftward direction errors, we replicated previous reports that neurologically healthy participants exhibit a general leftward bias in MNL tasks (Göbel et al., 2006; Longo and Lourenco, 2007; Loftus et al., 2009). Bowers and Heilman (1980b) observed that in healthy individuals, there was an inclination toward the left while performing line bisections (Bowers and Heilman, 1980b). They termed this tendency the “pseudoneglect” effect, which has also been referred to as “left-side underestimation” (Bradshaw et al., 1987) or “right hemispatial inattention” (Weintraub and Mesulam, 1988). A meta-analysis on pseudoneglect confirmed a moderate effect size for spatial pseudoneglect, establishing it as a reliable phenomenon (Jewell and McCourt, 2000). The prevailing hypothesis explaining this leftward bias involves differences in attention and activation between the right and left parietal hemispheres. Specifically, the leftward bias observed in healthy individuals is attributed to the allocation of attention toward the left hemispace, stemming from increased activation of the right hemisphere during spatial tasks (Jewell and McCourt, 2000). Prior research has proposed that performance in spatial-numerical association tasks might be influenced by various biases, including emotional valence (Sabaghypour et al., 2023). Regarding emotions, one perspective posits that the evidence supports a prefrontal neural index of valence (Jansari et al., 2000; Drake and Myers, 2006), implying a difference between positive and negative emotional traits.

With respect to hemispheric lateralization, the valence hypothesis of frontal asymmetry suggests that emotions are specialized and distributed between the left and right frontal hemispheres (Davidson, 1992). According to this hypothesis, the left frontal regions predominantly deal with positive motivations like hope and happiness (Harmon-Jones and Winkielman, 2007; Berkman and Lieberman, 2010; Cao et al., 2022). However, the right frontal regions are associated with vigilant attention that often occurs during certain negative affective states (Shackman et al., 2009; Grimshaw and Carmel, 2014). In contrast to the focus on parietal contributions in numerical representations (Hubbard et al., 2005; Knops et al., 2009), the significance of frontal cortical involvement in numerical cognition has been somewhat underappreciated. However, a bunch of recent research indicates that frontal contributions might hold particular importance in this research domain (Doricchi et al., 2005; Bachmann et al., 2010; Sabaghypour et al., 2023). Nevertheless, limited knowledge exists regarding the electrophysiological measures linking the individual variations with asymmetric orientation along the MNL.

Our findings on frontal asymmetry (F4-F3 difference) revealed significant positive correlations with MNL direction bias considering theta and significant negative associations concerning beta activity. Although there was an observed positive correlation tendency, the results did not reach statistical significance for alpha activity. Increased left frontal theta activity (negative difference percent in our measures) and increased right frontal beta activity (i.e., positive difference percent) have been reported to be linked to negative emotional traits (Coan and Allen, 2003; Dharmadhikari et al., 2018; Zhao et al., 2018; Koller-Schlaud et al., 2020). The correlation analysis of theta and beta in our results suggests that individuals exhibiting a larger positive theta (and larger negative beta) difference, which is correlated with positive emotional traits, displayed a greater inclination toward rightward attention, thereby moderating the general directional bias in MNL. These results are consistent with research that demonstrated shifts in emotions can affect attention with positive emotions inducing rightward bias, and negative emotions inducing leftward bias (Dimond et al., 1976; Levick et al., 1993). Our findings also align with the suggestion that the opposite side of the frontal cortex activity is linked to directing attention to the left and right side (Wright et al., 1995).

Furthermore, a significant negative and positive correlation between theta and beta asymmetries and absolute errors was observed, respectively. The reverse pattern regarding absolute errors might be attributed to a favorability toward positivity: Positive emotional traits have been linked to cognitive advantages, including increased cognitive flexibility, problem-solving abilities, and better decision-making (Isen, 2002). These results indicate that positive emotional traits could improve cognitive functions and promote effective cognitive performance (Fredrickson and Cohn, 2008).

Cognitive-related measures derived from rs-EEG data concerning mean absolute error and RT were explored to elucidate the possible correlation between brain activity patterns and performance in MNL. The findings revealed notable negative and positive correlations between APF and the low-alpha/high-alpha ratio with regard to absolute errors, respectively. APF detected through rs-EEG has been correlated to cognitive performance among healthy participants. APF that represents the frequency of the highest magnitude (or the center of gravity) within the range of 8–12 Hz and reflects a stable neurophysiological trait marker of cognitive preparedness (Angelakis et al., 2004; Grandy et al., 2013). A faster APF indicates increased magnitude (linked with improved cognitive performance) within the upper range of the alpha spectrum and vice versa. Our findings regarding APF and lowAlpha/highAlpha ratio, represent the first evidence linking this EEG index to numerical cognition, consistent with prior studies (Klimesch et al., 1990; Vogt et al., 1998; Klimesch, 1999; Rathee et al., 2020) that highlighted significant correlations between individual differences in APF and various cognitive measures. Theoretical explanations of these correlations suggest that alpha oscillations serve as a general mechanism for regulating neural inhibition (Klimesch et al., 2007) as well as controlling the flow of information within the brain (Mazaheri and Jensen, 2010), both of which contribute to improved cognitive performance.

Shifting focus to the beta frequency range, the results revealed a noteworthy negative correlation between beta activity at the occipital region (O1) and mean absolute errors in MNL. This negative relation might be interpreted in light of the fact that increased beta (especially in occipital areas with closed eyes) is linked with stress and anxiety traits (Gordeev, 2007; Díaz et al., 2019). Such results are also consistent with findings that linked anxiety traits to cognitive performance (Eysenck et al., 2007) including visuospatial (Eysenck et al., 2005) and numerical processing (Lindskog et al., 2017; Grežo and Sarmány-Schuller, 2018).

In terms of relative amplitude indices, significant positive correlations were observed between TSR and TBR at Cz with regard to the mean absolute error of participants. The observed significant positive correlations between TSR and TBR and absolute errors in participants’ performance align with findings suggesting a potential relationship between these ratios and cognitive performance in healthy adults (Putman et al., 2014; Clarke et al., 2019; Marlats et al., 2020). Few published studies explored the behavioral correlates of the TBR and TSR. In line with our findings, it has been evidenced that a high TBR/TSR is linked to a higher frequency of behavioral errors (van Dongen-Boomsma et al., 2010; Guleken et al., 2020). Furthermore, the TSR was the only significant measure that predicted the mean RTs. This finding aligns with previous research that reported an increase in central-midline TSR ratio could be linked to increased RT in behavioral tasks by recurrent lapses in attention (Ogrim et al., 2012).

To the best of our knowledge, this is the only study investigating the differential roles of rs-EEG measures on MNL performance. Set against the importance of our findings, it is essential to acknowledge the limitations of our research. Our study primarily focused on trait-related measures of frontal asymmetry, assuming relatively stable individual differences over time (Wheeler et al., 1993; Davidson, 1998). We speculated based on these trait asymmetries but did not include specific behavioral measures directly related to emotions. Future research might benefit from incorporating state-oriented inquiries to investigate changes in frontal EEG asymmetry in response to specific emotional states. Including behavioral measures such as the Positive and Negative Affect Schedule (PANAS) (Crawford and Henry, 2004) could offer deeper insights and more conclusive results into the observed relationships. Furthermore, it is essential to note that our study was constrained by limited resources, which impacted the extent of EEG data collection. The limited electrode coverage might have restricted our ability to capture more localized neural activity. Future research should consider employing a broader array of electrodes across various brain regions to obtain a more comprehensive understanding of relevant findings.

In light of these findings, the implications for future neuromodulation research are noteworthy. Since, successful performance in MNL tasks has been linked to mathematical achievements (Schneider et al., 2009; Simms et al., 2016), understanding the correlations between various EEG measures and performance in this task sheds light on potential avenues for utilizing neuromodulation approaches. Considering these findings, we shall also propose that employing techniques like transcranial alternating current stimulation (tACS) or TMS with a focus on theta or beta stimulation protocol might offer evidence supporting the causal role of frontal asymmetry in the direction bias.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: doi: 10.17605/OSF.IO/FM8CJ.

The studies involving humans were approved by the local Ethical Committee approved the study, adhering to the ethical principles outlined in the Declaration of Helsinki concerning human subjects. Prior to participation, all individuals provided informed consent and received academic credit as compensation for their involvement. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

SS: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing. FN: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing. NB: Formal analysis, Investigation, Project administration, Visualization, Writing – review and editing. FS: Data curation, Investigation, Project administration, Writing – review and editing. NZ: Data curation, Formal analysis, Investigation, Writing – review and editing.