1 Introduction

The autonomic nervous system (ANS), composed of the sympathetic and parasympathetic systems, regulates bodily functions, balancing activity to facilitate survival responses and post-threat recovery (McCorry 2007). Stress triggers sympathetic and parasympathetic responses, leading to changes in biomarkers and biosignals. On the basis of the literature review, heart rate variability (HRV) analysis and skin conductance response (SCR) are widely used as physiological indicators of stress (Battaglia et al. 2024; Jiménez-Mijangos et al. 2023). According to the neurovisceral integration model, the ANS interacts dynamically with the central nervous system (CNS) to regulate heart function during cognitive load and fear (McEwen and Morrison 2013; Thayer and Lane 2009). When mental demands rise, the prefrontal cortex exerts top-down control over the ANS, influencing heart rate via the vagus nerve (Di Gregorio et al. 2024). Simultaneously, the amygdala, critical in processing emotional stimuli, also impacts autonomic responses (Battaglia and Thayer 2022). This demonstrates how cognitive and emotional processes directly affect autonomic regulation, with HRV and SCR serving as key markers of cognitive load (Battaglia et al. 2024; Jiménez-Mijangos et al. 2023). Coordinated CNS–ANS interactions are essential for survival, as social and emotional factors can significantly impact physiological responses during performance monitoring (Battaglia et al. 2024; Di Gregorio et al. 2024). SCR, also known as electrodermal activity (EDA), is a peripheral indicator of the sympathetic nervous system (SNS) activity (Dolu, Acer, and Kara 2014; Visnovcova et al. 2016). Skin conductance assessment measures alterations in electrical conductivity attributed to changes in sweat production (Boucsein et al. 2012). It serves as a physiological indicator for wakefulness and attention (Martínez Vásquez, Posada-Quintero, and Rivera Pinzón 2023). Skin conductance measurements comprise two components: tonic (resting) and phasic (reactivity). Tonic skin conductance measures resting levels within seconds, whereas phasic responses involve brief oscillations triggered by stimuli such as fear or increased cognitive load, lasting a few seconds (Battaglia et al. 2024; Storm 2008). Particularly during mental tasks that elevate cognitive load, which is the mental effort used by working memory at any given time (Sazuka et al. 2024), combining skin conductance measurements with HRV metrics is recommended to comprehensively assess ANS responses (Nepal, Manandhar, and Jha 2019). HRV, a cardiac vagal tone indicator, is the fluctuation in the time intervals between successive heartbeats, known as inter-beat intervals (IBIs) (Wettstein, Kühne, and Tschacher 2020). A healthy heart displays dynamic oscillations, allowing the cardiovascular system to promptly adapt to physical and psychological challenges. Higher HRV, indicating greater cardiac vagal tone and a flexible ANS, enhances cognitive performance (Zeng et al. 2023). HRV parameters can be derived through time–domain, frequency–domain, and nonlinear analyses (Chan et al. 2015). Mostly used time–domain HRV parameters are the root mean square of successive differences between normal heartbeats (rMSSD), the standard deviation of the successive IBIs for normal sinus beats (SDNN), and the percentages of successive IBIs for normal sinus beats differing by more than 10 ms (PNN10), 30 ms (PNN30), and 50 ms (PNN50) (Malik and Camm 1993). Among these parameters, rMSSD measures the activity of the parasympathetic nervous system (PNS) (Shaffer and Ginsberg 2017) and is assessed as vagal tone activity (Zeng et al. 2023). Understanding the coordination between SNS and PNS is crucial for enhancing the ability to organize resources effectively to meet various demands. Stenfors et al. discovered that higher rMSSD and higher SDNN were linked to better performance in cognitive tests, including working memory (Stenfors et al. 2016). However, this association weakened when age was considered a covariate (Zeng et al. 2023). Working memory, essential for learning and reasoning, influences cognitive function and performance (Yin et al. 2020; Yoo and Collins 2022). Previously, it was found that young individuals with higher resting HRV performed better in working memory tasks (Landolt et al. 2017). Chronic stress-induced autonomic dysregulation can impair both decision-making speed and reaction times (RTs). Mental tasks that demand attention can trigger stress, which, in turn, affects ANS responses and consequently alters RTs (Landolt et al. 2017). Alternating attention allows individuals to shift focus between tasks with varying cognitive demands (Kim et al. 2020), whereas sustained and selective attention enable consistent focus during continuous activities and amidst distractions (Moscatelli et al. 2024). The prefrontal cortex plays a key role in these processes, as neurophysiological studies in nonhuman primates have shown its involvement in functions like working memory, decision-making, and action planning, highlighting its importance in managing complex cognitive tasks (Moscatelli, Monda, et al. 2023; Moscatelli, Toto, et al. 2023). Once responses to various levels of cognitive tasks are recorded along with electrophysiological signals such as SCR and electrocardiogram (ECG), ANS activity can be analyzed in relation to these tasks. To study stress levels and attention deficits in various conditions and diseases, it is crucial to use simpler, more affordable methods, especially when resources are limited (Nepal et al. 2018). Our study aimed to develop a reliable, cost-effective tool to investigate the relationship between ANS activities and cognitive load. Using a custom-built, computer-based test battery, we evaluated the impact of different cognitive load levels as measured through ANS function. Specifically, we utilized visual stimuli as cognitive tasks, synchronizing the recording of EDA and ECG to assess HRV parameters and SCLs. This methodology enabled us to investigate how varying cognitive load levels influence SCL, RTs, and HRV indicators, as well as how internal time estimation affects stress levels and RTs during these tasks.

2 Materials and Methods

2.4 Experimental Procedures

Participants were comfortably seated in the recording room and allowed to rest for 10 min before the test. They were informed about the experiment and instructed to remain still, avoid deep breaths, and maintain a relaxed posture throughout the test. They were instructed to remain still, avoid deep breaths, and maintain a relaxed posture throughout the test. The experimental procedure consisted of five steps, as detailed and outlined in Figure 1.

First, the skin conductance and ECG subunits were calibrated, followed by a 2-min baseline recording in a resting state without stimulus. Subsequently, participants completed four tasks, each lasting 2 min and progressively increasing in cognitive load. Visual stimuli were used in each task, as previously described and summarized in Table 1 (Karimi et al. 2023). For this study, the number of stimulus and task durations were adjusted to refine the experimental design and optimize data collection and analysis. The tasks were categorized into two main types: simple and cognitive, each incorporating a vigilance component, resulting in four subgroups: simple, simple vigilance (SV), cognitive, and cognitive vigilance (CV). The vigilance tasks required heightened attention due to the randomly varying stimulus intervals.

TABLE 1. Visual tasks design description.

Tasks	Visual Stimulus	Stimulus Exposure Time (s)	Number of Stimulus	Inter Stimulus Interval (s)	Descriptions
Simple (S)	Simple RED	0.5	30	2 Fixed	Participant was asked to press on Button-1 as quick as possible upon the onset of visual stimulus with index finger of the dominant hand
Simple vigilance (SV)	Simple RED	0.5	30	2–5 Random	Participant was asked to press on Button-1 as quick as possible upon the onset of visual stimulus with the index finger of dominant hand
Cognitive (C)	Cognitive 5 color (red, black, green, yellow, blue)	0.5	24 Non-target 6 Target	2 Fixed	Participant was asked to press on Button-1 as quick as possible if the stimulus is red colored (target), otherwise Button-2 with the index finger of dominant hand
Cognitive vigilance (CV)	Cognitive 5 color (red, black, green, yellow, blue)	0.5	24 Non-target 6 Target	2–5 Random	Participant was asked to press on Button-1 as quick as possible if the stimulus is red colored (target), otherwise Button-2 with the index finger of dominant hand

• Note: Non-target stimulus: black, green, yellow, and blue; target stimulus: red.

3 Results

3.2 Effect of Different Tasks on ECG-derived Parameters

3.2.1 Effect of Different Tasks on BPM, SDNN, and PNN

The percentage change in values for all parameters was compared to the baseline to evaluate task-induced alterations. PNN10, SDNN, and BPM percent changes did not differ significantly across tasks (Boucsein et al. 2012; Hansen, Johnsen, and Thayer 2009; Setz et al. 2010). However, there were notable differences in the percentage changes of PNN30 (p = 0.06) and PNN50 (p = 0.021) (Table 4).

TABLE 4. Comparing the effects of different tasks on the percentage changes in PNN30 and PNN50 values.

Tasks	PNN30 percentage change and (interquartile range values)	PNN50 percentage change and (interquartile range values)
Simple (S)	7% (91)	62% (218)
Simple vigilance (SV)	−7% (55)	−2% (86)
Cognitive (C)	18 % (126)	41% (341)
Cognitive vigilance (CV)	−3% (77)	−13% (179)
	p = 0.006	p  = 0.021

• Note: PNN30: Percentages of successive IBIs for normal sinus beats differing by more than 30 ms; PNN50: percentages of successive IBIs for normal sinus beats differing by more than 50 ms. Values represent the median of percent changes relative to the baseline value. Positive values indicate an increase relative to the baseline as a percentage change. Negative values indicate a decrease relative to the baseline as a percentage change.

Specifically, we observed statistical variances between SV and C tasks, as well as between C and CV tasks in PNN30. Moreover, differences were evident between SV and C tasks in PNN50.

3.2.2 Effects of Different Tasks on rMSSD

The percentage change of rMSSD values compared to the baseline was then analyzed to assess task-induced changes. Friedman analysis indicated significant differences in the percentage change of rMSSD values during the tasks (p = 0.042) (Table 5). In pairwise comparisons, it was found that the main difference lies between S and CV tasks.

TABLE 5. Comparing the effects of different tasks on the percentage changes in RMSSD values.

Tasks	rMSSD percentage change and (interquartile range values)
Simple (S)	12 % (36)
Simple vigilance (SV)	−6 % (25)
Cognitive (C)	10 % (40)
Cognitive vigilance (CV)	−4 % (45)
p = 0.042

• Note: rMSSD: root mean square of successive differences between normal heartbeats. Values represent the median of percent changes relative to the baseline value. Positive values indicate an increase relative to the baseline as a percentage change. Negative values indicate a decrease relative to the baseline as a percentage change.

3.2.3 Comparisons Between High and Low Groups Based on Baseline rMSSD Values

Comparisons were conducted between high and low groups on the basis of baseline rMSSD values. This grouping was performed in two distinct manners, using both the mean and median rMSSD values. The mean baseline RMSSD value for participants was 30.39, whereas the median baseline rMSSD value was 24.

3.2.4 Comparison of SCL Values and RTs between Low and High rMSSD Groups

Comparing groups formed based on the median baseline rMSSD value, the SCL percentage changes differed significantly across various tasks. Although the task significantly affected the percentage change in SCL (p_task = 0.002, interaction = 0.530, p_RMSSD_group = 0.635), the RMSSD groups had no statistically significant effect, nor did the interaction (Figure 2). Similarly, when comparing SCL values between low and high RMSSD groups based on the mean baseline RMSSD value, the task significantly affected the percentage change in SCL (p_task = 0.002, interaction = 0.837, p_RMSSD_group = 0.376), but the rMSSD groups and the interaction effect were not statistically significant (Figure 3).

3.2.5 Comparison of RTs Values between Low and High Groups

Comparing groups based on the median baseline rMSSD value, RTs percentage changes differed significantly across tasks. The task significantly affected the percentage change in SCL (p_task < 0.001), but the rMSSD groups and the interaction did not (interaction = 0.356, p_rMSSD_group = 0.247; Figure 4). Similarly, when comparing RTs values between low and high RMSSD groups based on the mean baseline RMSSD value, the task significantly affected the percentage change in SCL (p_task < 0.001), but the RMSSD groups and the interaction were not significant (interaction = 0.814, p_rMSSD_group = 0.847; Figure 5).

4 Discussion

Our analysis showed that participants’ SCL increased during the tasks, indicating that these tasks effectively triggered SNS activation, resulting in heightened arousal and stress. This outcome is in consistent with the earlier studies. Nepal et al. demonstrated that skin conductance is a reliable indicator of sympathetic activation during cognitive stressors such as mental subtraction tasks. Similarly, Remier et al. observed elevated SCR during cognitive tasks, further supporting this relationship. Additionally, Kindermann et al. reported increased SCR during stress periods, and Scavone et al. found a strong correlation between SCL and emotional stress, reinforcing the role of SCR as a marker for sympathetic activity across different types of stressors (Kindermann and Werner 2014; Nepal et al. 2018; Reimer and Mehler 2011; Scavone, Kadziolka, and Miller 2020).

HRV is a well-established marker of parasympathetic activity (Shaffer and Ginsberg 2017; Zeng et al. 2023). Previous studies have shown that individuals with anxiety disorders tend to have lower resting-state HRV (Chalmers et al. 2014; Cheng et al. 2022). In our study, significant changes in vagally mediated HRV during tasks, compared to resting state, highlight the impact of cognitive load on the parasympathetic branch of the ANS. Although previous studies focused on tonic HRV, our research examined phasic HRV, which measures the change in HRV from baseline to task periods, offering deeper insight into how HRV fluctuations during cognitive tasks relate to performance.

The examination of RTs further supports the influence of cognitive load on physiological responses. We found significant differences in RTs between simple and cognitive tasks, with RTs increasing in parallel with the rise in mental load. This finding aligns with our previous study and other research, which suggests that more complex cognitive tasks require greater cognitive resources, leading to longer processing times and delayed responses (Fan et al. 2020; Karimi et al. 2023; Parrington, MacMahon, and Ball 2015).

The detailed aspects of our results include as follows:

SCL changes:

The percentage change in SCL from baseline recordings was found to be higher in vigilance tasks compared to their non-vigilance counterparts. The vigilance tasks were designed to heighten participants’ alertness, leading to an increased stress response. However, as participants initially engaged in non-vigilance tasks before transitioning to vigilance tasks, it was anticipated that sympathetic stress would diminish over subsequent task sessions. These findings align with Steptoe and Greer's 1980 study, which observed lower levels of autonomic arousal as training progressed (Steptoe and Greer 1980). Additionally, the internal timing processes used to estimate event occurrences in non-vigilance tasks may create a greater cognitive load, impacting the sympathetic response more than simply awaiting random stimuli, as occurs in vigilance tasks (Dolu et al. 2004; Golob and Starr 2000). Moreover, to determine the cause of the observed decrease in SCL values during vigilance tasks, the order of the tasks can be changed to rule out the influence of learning effects.

RTs changes:

As anticipated, RTs increased with task difficulty. Although there was a trend toward longer RTs for target stimuli, this difference did not reach statistical significance. The lower frequency of target stimuli, which occurred at least four times less often than non-target stimuli, likely contributed to slower RTs by reducing the activation of rapid decision-making processes. This observation is consistent with findings by Lucci et al., who reported that a higher frequency of target stimuli is associated with faster RTs (Lucci et al. 2016). However, it is important to consider potential confounding variables that could influence these results. For instance, individual differences in baseline ANS activity may have affected the relationship between cognitive load and RTs. For example, in our study, the age of participants, which ranged from 20 to 40 years, could also influence ANS activity and, consequently, the response patterns under cognitive load. This suggests that future studies could benefit from categorizing participants by age to better understand these effects.

5 Conclusion

Our findings indicate that cognitive tasks increase sympathetic outflow, causing significant changes in SCL and RTs, and significantly affect the brain–heart connection, leading to notable changes in HRV. Our system's flexibility in modifying task parameters, such as type, duration, and stimulus intervals, makes it ideal for analyzing ANS responses under cognitive load. Future research studies could utilize this system to explore the impact of cognitive load on physiological responses, including changes in ANS function during aging, as previously mentioned. It is also well-suited for investigating ANS activity across a range of conditions, including psychiatric disorders like attention deficit hyperactivity disorder (ADHD) and non-psychiatric conditions such as cardiologic diseases, where ANS dysregulation is prevalent.

Author Contributions

Nazli Karimi Ahmadi: conceptualization, investigation, methodology, validation, writing–review and editing, formal analysis, project administration, data curation, supervision, writing–original draft, visualization, software. Sezgi Firat Ozgur: conceptualization, methodology, writing–review and editing, software, formal analysis, validation. Erhan Kiziltan: conceptualization, investigation, methodology, writing–review and editing, software, formal analysis, validation.

Ethics Statement

This study has received approval from the Hacettepe University Institutional Review Board and Ethics Committee under the project number GO 23/234 and decision number 2023/07-49. The study conforms with The Code of Ethics of the World Medical Association (Declaration of Helsinki).

Consent

Written informed consent was obtained from all participants prior to the initiatiion of the study.

Conflicts of Interest

The authors declare no conflicts of interest.