INTRODUCTION
The estimated global annual prevalence of major depressive disorder (MDD) is approximately 4%-6% in adults, with lifetime prevalence estimates reaching up to 20% [1]. MDD is a multifactorial condition arising from a combination of genetic and environmental factors (such as poverty, recent adverse life events, and childhood maltreatment), psychological factors (cognitive patterns), and biological factors (inflammatory responses and monoamine pathways) [2,3]. Most research indicates that MDD significantly diminishes quality of life, impairing multiple aspects of social functioning, including financial stability, relationships, and educational performance [4].
Executive dysfunction is a common symptom of major depression [5], often manifesting as difficulties in planning, initiating, and completing goal-directed activities. This domain comprises essential components such as inhibitory control (IC), decision-making, working memory, and set shifting [6]. Executive function is typically assessed using tools such as the Stroop, Wisconsin Card Sorting Test, and the Continuous Performance Test (CPT). Notably, CPT is recognized for its ability to evaluate IC and reaction time, a key component of executive function, as well as measures of attention and impulsivity [6-8]. It is widely used in various clinical settings to evaluate impairments in executive function [9].
Numerous studies have highlighted the impact of depressive symptoms on executive function, which may arise from a combination of factors, including impairments in reaction time, attention, memory, IC, fatigue/energy loss, and the influence of immunological and dopaminergic pathways [10-13]. Emerging research also suggests the potential role of autonomic nervous system (ANS) dysregulation in these impairments [14,15]. This interaction is particularly notable in terms of the parasympathetic nervous system’s (PNS) role, often conceptualized through the lens of the “polyvagal theory.” [16] This theory posits that humans adapt to the environment through the vagal system, and that dysregulation of the vagal nerve leads to maladaptation. Indeed, our article in press also revealed that the vagal nerve responds to IC in healthy young adults [17].
Heart rate variability (HRV) is noninvasive, highly applicable, and predominantly reflects the PNS modulation of the heartbeat [18]. Recently, HRV has garnered attention as a valuable physiological marker of the ANS in mental health research [19,20]. However, while these data may reveal correlations with individuals at risk of depression, there is considerable variation in how studies record and report HRV, making it challenging to draw clinically meaningful conclusions owing to bias in the existing research [21,22].
The relationship between MDD and HRV, or MDD and executive function has been confirmed in numerous studies. However, no studies have combined both factors to examine their associations. To address this issue, we recruited a group of healthy individuals and individuals with depression. We recorded HRV parameters and analyzed the related reactivities during the CPT. Our aim was to investigate the relationship between HRV and executive function as well as how HRV parameters change during the task.
METHODS
Participants
The health group was approved by the Institutional Review Board of the National Yang Ming Chiao Tung University (IRB: YM108115EF), whereas the depressive group was approved by the Institutional Review Board of the Tri-Service General Hospital (IRB: 1-105-05-119). We recruited 29 healthy males and 25 males diagnosed with depression by two psychiatric physicians, based on the criteria of the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (Text Revision). Patients initially diagnosed with depression had not yet been treated with antidepressants. We conducted an agematched study consistent with our previous research, and the experimental design similarly followed the earlier protocol, utilizing continuous HRV to measure ANS activity and the Conners Continuous Performance Test-II (CCPT-II) to assess behavioral responses [17]. All participants had a body mass index between 18.5 and 28 kg/m2 and had never previously taken the CCPT-II. The study protocols were explained in detail to the participants, and their basic information was collected after providing written informed consent. Through participant interviews, we excluded individuals receiving treatment for conditions such as cancer, renal or cardiovascular diseases, diabetes, psychotic disorders, or any condition related to the substance withdrawal or use disorders. The Beck Depression Inventory-II (BDI-II) was used to quantify depression severity, whereas the Beck Anxiety Inventory (BAI) was used to measure anxiety levels. In addition, to avoid HRV changes caused by hormonal fluctuations related to the menstrual cycle [23], this study included only young male participants.
CCPT-II
The CCPT-II is characterized by its simple instructions, making it easily applicable across various age groups and in individuals with cognitive impairments. It is a widely used assessment tool for measuring sustained attention, reaction time, and executive function and is known for its internal consistency and test-retest reliability [9,24]. The CCPT-II software (Multiple-Health Systems, Inc.) conducts a 14-minute test, during which characters appear for 250 ms. Participants were seated in front of a computer in a quiet environment and must react to quickly changing stimuli, pressing the space bar for “non-X” signals and not pressing it for “X” signals. At the end of the test, the software assessed the commission, detectability (d’), and reaction time. Of these metrics, d’ is a psychophysical measure of attentive executive performance calculated based on the signal detection theory [25,26]. We used both commission and d’ as indices of IC. We also used the reaction time as the indicator of frontal lobe-related psychophysiological states [27].
HRV
HRV parameters were analyzed through time- and frequency-domain methods utilizing 5-minute windows of electrocardiogram (ECG) signals. Recent studies have extensively evaluated the importance of HRV parameters [18,28]. For time-domain HRV parameters, we calculated the mean RR interval and standard deviation of normal-to-normal heartbeats (SDNN) for each time window. It has been suggested that SDNN more effectively reflects PNS activity when measured during short-term resting conditions [18]. In our frequency-domain analysis of HRV, we focused on the high-frequency power (HF, 0.15-0.40 Hz) linked to PNS modulation and respiratory effects, often termed the respiratory band. We also evaluated the low-frequency power (LF, 0.04-0.15 Hz), previously thought to represent both sympathetic nervous system (SNS) and PNS activities. However, recent findings have indicated that the LF primarily arises from baroreflex mechanisms under resting conditions [29-31]. We used logarithmic transformation of the HF and LF power data to adjust for possible skewness in their distributions [32].
HRV measurement
Prior to the test, participants were asked to relax for 10 minutes in a quiet room, where the temperature was controlled at 25°C with an air conditioner. An ECG patch was affixed to gather continuous ECG data immediately after consent was obtained. Following a 5-minute baseline recording of the ECG data, participants undertook the 14-minute CCPT-II test. After completing the test, the participants were asked to remain seated for an additional minute. ECG data were recorded throughout the test period. Participants were instructed to sit still, minimize movement, and breathe naturally without hyperventilating or engaging in paced breathing [33]. An HRV analyzer (model WG-103A; wegene) was used to acquire and process the ECG signals. This device employs an 8-bit analog-to-digital converter with a sampling frequency of 256 Hz, enabling the real-time analysis of ECG signals. Subsequently, a fast Fourier transformation was performed to transform the stationary RR interval data into a power spectrum, following the method described earlier [32]. To monitor real-time HRV changes throughout consecutive CCPT-II sessions, we implemented a continuous 5-minute time window method, with each window overlapping the next by 4 minutes, resulting in a 1-minute interval between successive time windows [34]. The HRV value from the first time-window was set as the baseline (t0), and the average HRV data (tmean) were used to represent HRV throughout the CCPT-II process.
Statistical analysis
Statistical analyses were performed using SPSS Statistics (version 20.0; IBM Corp.). The Shapiro-Wilk test did not show a significant deviation from normality in the distribution of the investigated values. We used multivariate analysis of variance (MANOVA) to determine whether there were any significant differences in the HRV parameters between each group. To detail the phasic changes in HRV during the CCPT-II in both groups, we used a repeated-measures ANOVA controlled by BAI with Bonferroni correction. To explore the association between HRV and executive function, we calculated the correlation coefficients between the CCPT-II and HRV parameters at baseline (t0) and t14 and the mean HRV data (tmean) of the health and depression groups. For high comorbidity of anxiety and depression, all values were corrected using participants’ BAI [35,36]. Significance was tested at a 95% confidence interval. p<0.05 indicated statistical significance.
RESULTS
Characteristics and demographic data of participants
There was no significant difference in age between the groups. Regarding depression and anxiety levels as measured by the BDI and BAI, the health group had significantly lower scores than the depression group. Specifically, the health group showed lower BDI scores than the depression group (F=21.4, p<0.001) and the BAI scores (F=8.24, p=0.001).
In terms of cognitive performance, measured using the CCPT-II, there was a significant difference in d’. The health group had a higher d’ value than the depression group (p=0.034), indicating higher IC in the health group.
The HRV at different time windows also revealed significant differences between the healthy and depression groups. At t0, t14, and tmean, the high-frequency power with logarithm (lnHF) was significantly higher in the health group than in the depression group. Additionally, the SDNN was significantly higher at the t14 and tmean time windows in the health group than in the depression group.
The other parameters of HRV, such as low-frequency power with logarithm (lnLF) and RR, revealed no significant differences between the health and depression groups. The results are summarized in Table 1.
Comparing HRV parameters with baseline separately in health and depression group
The HRV parameters during the CCPT-II in both the health and depression groups are shown in Figure 1. There was a significant difference only in lnHF between the groups (F=6.31, p=0.017). The significant lowered values compared with t0 were observed in SDNN of health (t9, p=0.047) and that of the depressive group (t5, p<0.001; t6, p<0.001; t7, p<0.001; t8, p<0.001; t9, p<0.001; t10, p<0.001; t11, p<0.001; t12, p<0.001; t13, p<0.001; t14, p<0.001) and the lnLF of health (t8, p=0.005; t9, p=0.003) and depressive group (t4, p=0.007; t5, p=0.002; t6, p=0.001; t7, p=0.001; t8, p=0.001; t9, p<0.001; t10, p=0.001; t11, p=0.020; t12, p=0.010).
Correlations between HRV and IC in health group
Table 2 presents the relationship between CCPT-II and HRV parameters at different time points in the health group: baseline (t0), t14, and mean (tmean) time window.
At baseline (t0), no significant correlations were found between HRV and CCPT-II parameters. At t14 time window, there was a significant correlation between HRV and CCPT-II variables. The commission negatively correlated with SDNN (r=-0.445, p=0.017), lnLF (r=-0.353, p=0.049), and lnHF (r=-0.397, p=0.030); d’ conversely correlated with RR (r=0.411, p=0.026), SDNN (r=0.560, p=0.003), lnLF (r=0.473, p=0.011), and lnHF (r=0.448, p=0.016); hit reaction time positively correlated with SDNN (r=0.469, p=0.012), lnLF (r=0.464, p=0.013), and lnHF (r=0.359, p=0.046). During the CCPT-II (tmean), commission significantly negative correlation with SDNN (r=-0.347, p=0.045) and lnHF (r=-0.360, p=0.038); d’ positively correlated with RR (r=0.352, p=0.042), SDNN (r=0.474, p=0.008), lnLF (r=0.346, p=0.045), and lnHF (r=0.432, p=0.016). No significant correlation was observed between reaction time and HRV variables during the CCPT-II in this study.
Correlations between HRV and IC in depression group
Table 3 presents the correlation between the CCPT-II and HRV parameters at different time windows, including baseline (t0), t14 and during CCPT-II (tmean) for individuals with depression. A key finding is the significant positive correlation observed between d’ and RR at t0 (r=0.412, p=0.032) and tmean (r=0.375, p=0.047). However, no other significant correlations between the CCPT-II and HRV variables were found in the depression group.
DISCUSSION
This study extensively examined the intricate dynamics between executive function and ANS regulation in a depression group, as manifested by HRV measurements during CCPT-II. Our findings emphasize the impact of depression on autonomic function, reaction time, and related IC. In this article, we detail how depression not only reduces vagal regulation but also leads to a decline in IC and reaction time modulation. This decrease in IC appears to stem from a broader impact on neural networks. This highlights the extensive neurophysiological effects of depression, impacting the ANS and critically influencing cognitive function.
Initially, comparisons of the HRV parameters between the health and depression groups, including the RR interval, SDNN, and lnLF, showed no significant differences. However, a significant disparity was observed in the lnHF, which is considered to be indicative of vagal modulation [18]. This suggests that in young individuals with depression who perform the CCPT-II, ANS regulation differs from that in healthy individuals, particularly in terms of parasympathetic activity. Previous studies indicated that depression is associated with reduced vagal modulation [37]. The significant difference in the baseline lnHF between the two groups in this study corroborated these findings, highlighting diminished vagal activity as a characteristic of depression in this demographic group. Additionally, PNS regulation within the depression group during CCPT-II, as evidenced by the significant difference at tmean and between groups in lnHF, indicates that PNS dysregulation in the depression group extensively affects the entire executive process. Consistent with prior research, the lower lnHF values observed in the depression group than in the healthy cohort further underscore the prevalence of reduced parasympathetic activity associated with depression, highlighting its pervasive effect on executive functions [37,38]. However, during the process, both groups exhibited declines in SDNN and lnLF compared to baseline, yet no significant differences between the groups were observed. Although the depression group showed signs of vagal dysregulation, the overall ANS retained regulatory capabilities during the CCPT-II task. This suggests that, despite the challenges posed by depression, a functional level of autonomic control remains that can potentially be leveraged in therapeutic interventions [39,40].
Second, within the CCPT-II, detectability not only represents the accuracy of recognition but also serves as a crucial indicator of IC [41]. Higher values indicate better IC, while lower values suggest poor impulse control. Previous research on detectability has primarily focused on populations with attention-deficit/hyperactivity disorder and traumatic brain injuries [42,43]. Our study revealed that, compared to the health group, the depression group exhibited lower detectability scores, indicating diminished signal recognition accuracy under the influence of depressive symptoms. This aligns with prior studies that reported a higher error rate in the depression group, supporting our findings [44,45]. Additionally, despite a significant decrease in detectability in the depression group compared with the health group, the reaction time remained unchanged. This suggests that the decline in IC among those with depression may not stem from a speed-accuracy tradeoff. Instead, other neuropsychological aspects should be considered to fully understand this phenomenon [46]. This observation also complies with the lack of a significant correlation between HRV and reaction time in the depression group.
Third, the neurovisceral integrated model elucidated the link between prefrontal lobe-related IC and ANS modulated HRV [47]. In this study, the significant correlations between all HRV parameters controlled by the ANS and CCPT-II parameters in the health group demonstrate the involvement of the ANS in IC. In addition, enhanced IC contributes to the management of negative thinking and emotional regulation [48]. However, in contrast to the significant correlations observed in the health group, the depression group only showed a significant positive correlation between RR intervals and d’, indicating that the ANS’s role in modulating IC is not prominent in the depression population. In fact, depression primarily affects the wide frontal lobe related neural network [49,50], which, according to the neurovisceral integrated model and recent studies on the frontoparietal network’s role in task performance [47,51-53], suggests that the decreased correlation between d’ and the HRV may originate from the debilitating of the frontal cortex related neural network in the depression group.
Fourth, this study found that in the health group at the t14 time-window, HRV parameters were significantly positively correlated with reaction time, yet no similar correlation was observed at t0 or tmean. Hansen et al. [54] proposed that higher HRV, indicative of vagally mediated HRV, enhances IC through its connection to the frontal lobe, thereby facilitating more efficient attentional regulation and reducing reaction time. However, subsequent research revealed that the SNS also influences reaction time [55,56]. Our study discovered that, in addition to the significant positive correlations between lnHF and SDNN with reaction time at t14, lnLF, representing both SNS and PNS, also showed a significant positive correlation with reaction time. This supports the conclusion that reaction time was jointly affected by both SNS and PNS. However, we did not observe similar results for the t0 and tmean. This suggests that the dynamics of HRV and reaction time may exhibit more pronounced associations at specific times during cognitive task. Indeed, the reaction time, as a behavioral output, involves the frontal cortex and the visuomotor circuit [46,57]. Recent research highlights the role of the parietal lobe, pre-supplementary motor area, and anterior cingulate cortex in the visual-to-motor transformation, elucidating its underlying complex connections [46,58,59]. Evidence further suggests that both learning and past experiences influence reaction time, indicating its association with adaptation [60-62]. These increasingly understood neural networks demonstrate that macroscopic cortical activity, through repeated practice, becomes more efficient compared to unfamiliar scenarios [61]. Utilizing the neurovisceral integration model, we posit that such connections, including frontal cortex and cingulate cortex, could further link to the ANS [63]. Moreover, the correlation between HRV and the adaptation related brain regions, such as frontal and cingulate cortex, and thalamus has been studied [64-66]. This may explain our finding that the correlation between reaction time and HRV may not be consistent across different time measurements. We hypothesize that the modulation between the cortex related to the visuomotor circuit and the ANS becomes more efficient through repeated signal processing. This leads to an efficient correlation between the ANS and reaction time, particularly in the later stages of the CCPT-II in health group. However, these observations warrant further investigation.
Limitation
Although this study offers valuable insights into the relationship between autonomic function and cognitive performance in both health and depression groups, its limitations must be acknowledged. The relatively small sample size and inclusion of only male participants may limit the generalizability of our findings. The male design provided us with a result that was not affected by menstrual hormones. Additionally, the cross-sectional design constrained our ability to draw causal inferences between autonomic function and cognitive performance. Cognitive dysfunction and intellectual impairment should have been considered at the beginning of this study. Our research primarily focused on analyzing IC and reaction time, without evaluating baseline cognitive function or intelligence. Future studies with larger sample sizes should include assessments of general cognitive abilities and intellectual function to allow for more comprehensive subgroup classification and comparison. The severity of depression, as measured by the BDI, was predominantly mild to moderate, which necessitates further exploration of whether these findings extend to individuals with severe depression. However, recruiting participants with severe depression can be challenging as they often exhibit lower motivation to engage in research. Future research should use a longitudinal design to establish the causal relationships between HRV and cognitive performance over time. Moreover, intervention studies focusing on autonomic regulation, such as biofeedback or mindfulness training, can provide deeper insight into how enhancing autonomic function may improve cognitive performance in individuals with depression.
Conclusion
The significant associations between HRV and reaction times indicate that HRV may serve as a practical indicator of inhibitory function in healthy controls and a valuable biomarker for individuals with depression. In our study, autonomic regulation appeared to play a crucial role in the cognitive performance and reaction time in the health group. The diminished correlation between HRV and CCPT-II parameters in the depression group signified the generalized effect of depression not only in the ANS, but also in its relationship with IC and reaction time. This underscores the need for clinical strategies that not only address ANS fluctuations and functional impairments in depression but also aim to restore autonomic function as a component of comprehensive psychiatric care.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
