Heart rate dynamics and antioxidant capacity under cognitive–emotional load in obsessive-compulsive disorder Authors: Galina Portnova, Guzal Khayrullina, Elena Proskurnina, Rohit Verma, Olga Martynova Published: 2026 ABSTRACT AbstractThe relationship between autonomic nervous system function, indexed by heart rate (HR) and heart rate variability (HRV), and the oxidant-antioxidant defense system in obsessive-compulsive disorder (OCD) remains poorly understood. This study investigated the association between autonomic reactivity and salivary antioxidant capacity (AOC) in patients with OCD under cognitive-emotional load. Thirty-one patients with OCD and twenty-nine healthy controls (HC) underwent a cognitive-emotional antisaccade task. Saliva samples for AOC analysis and electrocardiogram (ECG) recordings were collected before and after the task. Participants were stratified into high- and low-HR subgroups based on baseline levels. The OCD group demonstrated significantly lower baseline AOC and a blunted AOC response to the task compared to HC, whose AOC increased significantly. Crucially, stratification by baseline HR revealed a physiologically distinct OCD subtype. OCD patients with low baseline HR exhibited a maladaptive profile, showing a decrease in AOC and a significant increase in normalized low-frequency power (LFnu). In contrast, OCD patients with high baseline HR showed an AOC increase comparable to that of healthy controls. Overall, the OCD group also demonstrated a more pronounced increase in very low-frequency power (VLF) and higher overall LFnu, indicating generalized sympathetic predominance. These findings point to a physiologically distinct OCD phenotype defined by low baseline HR, impaired antioxidant response, and sympathetic hyperreactivity under stress, and suggest that baseline HR may serve as a candidate biomarker of a more severe, maladaptive form of OCD with compromised compensatory mechanisms linking autonomic regulation and antioxidant defense.Clinical trial numberNot applicable.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12888-026-08244-1. Introduction The etiology of obsessive-compulsive disorder (OCD) remains incompletely understood, with prior work implicating multiple interacting vulnerability factors [1]. These include genetic predisposition, exposure to emotional, physical, or sexual abuse or neglect, and adverse social experiences such as isolation, teasing, and bullying [1]. Clinically, OCD is characterized by intrusive obsessions and repetitive compulsions that substantially impair quality of life. Patients typically remain aware of the irrational nature of their symptoms but often develop an overinflated sense of responsibility and magical thinking, intolerance of uncertainty, and rigid beliefs about the controllability of intrusive thoughts [2, 3]. Persistent exposure to negative psychological influences in everyday social functioning can result in chronic stress throughout the course of OCD. Compulsions may serve as a maladaptive coping strategy that temporarily attenuates heightened arousal of the nervous system, thereby reinforcing the compulsive cycle [4]. Once these conditioned behavioral patterns are established, they tend to be repeatedly activated in response to obsessional thoughts, creating a self-perpetuating vicious circle that further interferes with daily functioning and socialization and, in turn, exacerbates stress. Chronic stress may therefore accompany OCD across the lifespan and act as a trigger for symptom exacerbations, often associated with hypervigilance, increased negative affect, and sustained mental arousal, all of which can influence physiological responses in both the nervous and immune systems [5]. The inability to effectively regulate mounting chronic stress is thought to contribute to an imbalance between sympathetic and parasympathetic activity within the autonomic nervous system (ANS) in individuals with OCD [6]. Utilizing electrocardiography (ECG), altered heart rate variability (HRV) and heart rate (HR) are frequently observed across a range of mental disorders, and OCD is no exception [7]. However, relatively few studies have specifically examined HRV in OCD, and the available findings remain inconsistent or inconclusive. In particular, resting-state HRV in patients with OCD may be reduced in association with greater symptom severity and the use of psychotropic medication [8, 9]. At the same time, studies in patients with panic disorder, social anxiety disorder, and generalized anxiety disorder have generally not found robust effects of psychotropic medication on HRV [9]. By contrast, findings on HRV in OCD remain mixed, with some reports suggesting that medication status may influence HRV indices specifically in this population. There are also conflicting data on associations between OCD symptom severity and HRV that appear to depend, at least in part, on the body position in which HRV is assessed, with differing results reported for recordings obtained in sitting versus standing postures [10, 11]. A recent study further reported that patients with OCD exhibit higher HR and lower HRV compared with healthy participants [7]. Another study also found that patients with OCD are characterized by increased sympathetic and reduced parasympathetic activity compared with healthy controls [12]. The predictive value of HRV appeared to be age dependent: whereas higher HRV in young adulthood distinguished individuals with OCD from their healthy relatives, lower HRV was characteristic of relatives in older age groups [7]. Previous research in adolescents without diagnosed psychiatric disorders has suggested that low HRV, indexed by respiratory sinus arrhythmia, moderates the association between anxiety and depressive symptoms and psychosocial stress [13]. HR and HRV patterns may contribute to a more thorough understanding of the etiology of OCD [7] and could serve as a methodology to detect changes in parasympathetic and sympathetic activity of the ANS in OCD [14]. The vagus nerve centrally regulates HR at rest, leading to a reduction in HR [15, 16]. Low cardiac vagal tone may be a marker of psychopathological symptoms. HRV can be used to index specific aspects of parasympathetic regulation of the cardiac system [17]. This parameter has also been associated with higher levels of inhibitory control [18, 19] and voluntary emotion regulation [19, 20]. Since HRV tends to recover following effective anxiety treatment, it has been proposed as an objective indicator of treatment response in OCD [14]. An antioxidant system is a complex network of enzymatic and non-enzymatic components that neutralize reactive oxygen species. One way to characterize the activity of this system is antioxidant capacity (AOC), an integral parameter proportional to the amount of scavenged reactive oxygen species. Numerous studies, including meta-analyses, indicate that oxidative imbalance acts as a pathogenic factor in patients with OCD [21, 22]. Oxidative stress is reflected by elevated levels of lipid peroxidation markers and decreased levels of antioxidant vitamins C, A, and E [23, 24], as well as reduced activity of antioxidant enzymes and increased albumin oxidation [25]. The development of oxidative stress in OCD patients may be associated with polymorphisms in genes encoding mitochondrial superoxide dismetase, which result in mitochondrial dysfunction [26]. Oxidative stress in OCD appears to have not only a systemic but also a local character, as it has been observed in the caudate nucleus of OCD patients [27]. On the other hand, there are reports of normal oxidant and antioxidant status in OCD patients [28], although increased markers of DNA damage have also been described [29], which is most likely due to the complexity of oxidative homeostasis that cannot be fully captured by integral indices alone. DNA damage in OCD may be related to an imbalance in the glutathione pathway [30]. Given the complexity of local and systemic oxidative metabolism, antioxidant status in saliva represents an independent and useful parameter for elucidating the role of oxidative stress in patients with OCD. The associations between HR, emotional or cognitive arousal, and antioxidant capacity have previously been reported in several studies. Some researchers demonstrated that pleasant tactile stimulation in healthy participants increases salivary antioxidant potential, which correlates with HRV and subjective ratings of stimulus pleasantness [31]. The association between HRV and oxidative responses (malondialdehyde) has also been observed in asthmatic children, underscoring the inter-relationship between oxidative stress markers and ANS function [32]. Higher salivary antioxidant properties have furthermore been found to correlate negatively with HR in anthropometric studies of healthy children [33]. Oxidative stress may play a role in the clinical manifestations of several psychiatric disorders, including OCD and other anxiety and depressive disorders [22, 34, 35]. In OCD, oxidative stress has been linked, in part, to dysfunction in catecholamine metabolism [36, 37]. Studies investigating oxidative stress and antioxidant status in OCD have highlighted the importance of oxidative inflammation and an imbalance between the generation of reactive oxygen species and their neutralization by antioxidant defenses in OCD and comorbid conditions [38, 39], and have led to the hypothesis that deficiencies in antioxidant enzymes may be characteristic of “pure” OCD [40]. A meta-analysis further indicates that patients with OCD show a systemic oxidative imbalance that is not adequately buffered by the antioxidant system [22]. In this study, we aimed to investigate the association between HR and HRV parameters and salivary oxidant–antioxidant activity in patients with OCD during a cognitive–emotional task. We hypothesized that there would be no significant differences in baseline AOC levels between the control and OCD groups, whereas the changes in oxidant–antioxidant activity following cognitive–emotional load would differ between these groups. Moreover, we expected that task-related changes in oxidant–antioxidant activity would be associated with baseline HR levels in individuals with OCD. Methods Participants We used saliva samples and HR data from 31 volunteers with OCD and 29 healthy controls; the data for all 29 controls had been collected in a previous study [4]. The inclusion criteria for the OCD group were as follows: age 18–40 years, right-handedness, and a diagnosis of OCD based on the International Statistical Classification of Diseases, 10th Revision (ICD-10). All participants were additionally interviewed by clinical psychologists (G.P., G.K.) to confirm the diagnosis. OCD severity had to range from 16 to 40 on the Yale–Brown Obsessive Compulsive Scale (Y-BOCS), corresponding to moderate to extremely severe symptoms. Exclusion criteria for the OCD group comprised harmful or dependent substance use; neurological or psychiatric disorders other than OCD (including bipolar disorder, autism spectrum disorder (ASD), and schizophrenia); history of traumatic brain injury; paroxysmal or epileptiform activity on electroencephalography (EEG); and uncorrected visual impairment. The OCD sample included 9 unmedicated individuals, 11 patients receiving antidepressant monotherapy, and 11 patients receiving combination pharmacotherapy (antidepressants + antipsychotics, n = 4; antidepressants + anticonvulsants, n = 4; antipsychotics + anticonvulsants, n = 3). Patients with OCD were recruited from psychiatric clinics with ICD-10 diagnoses F42.0, F42.1, or F42.2 and were invited to participate voluntarily in the study at the Institute of Higher Nervous Activity and Neurophysiology, Russian Academy of Sciences (IHNA and NPh RAS). The control group of healthy participants was recruited via announcements on social media. The inclusion criteria for the control group were as follows: age 18–40 years, right-handedness, and no history of mental illness as assessed by a clinical interview conducted by a clinical psychologist. Y-BOCS scores had to range from 0 to 7, corresponding to subclinical OCD symptoms. The exclusion criteria for the control group were harmful or dependent substance use, neurological or psychiatric disorders, uncorrected visual impairment, EEG abnormalities, or a history of traumatic brain injury. Psychological assessment All participants also completed the following questionnaires: the Yale–Brown Obsessive Compulsive Scale (Y-BOCS; web-based form), the State–Trait Anxiety Inventory (STAI), and the Barratt Impulsiveness Scale (BIS-11). Demographic and clinical characteristics of the participants are summarized in Table 1. The OCD and control groups were matched for age and gender. Confirming the clinical profile of the groups, participants with OCD showed significantly higher scores on the Y-BOCS (both subscales) and the STAI compared with the control group (all p < 0.05). No significant differences on these clinical measures were observed between the high- and low-HR subgroups.Table 1Clinical and biochemical descriptive statistics: inter- and intragroup comparisonsPatients with OCD (Valid N = 31)Control group (Valid N = 29)Mean (SD)AllHigh HRLow HRp value*AllHigh HRLow HRp value*Age23.1 (5.1)24.3 (4.8)22.9 (5.5)0.8723.9 (6)23.5 (5.7)24.2 (5.3)0.99Y-BOCS obsessions16.89 (4.6)16.88 (4.3)16.91 (4.7)0.996.146.535.690.68Y-BOCS compulsions11.04 (3.8)11.53 (3.5)10.27 (3.2)0.43.53.333.690.8Y-BOCS Total20.45 (4.4)22.28 (5.1)17.45 (4.1)0.161.18 (0.2)1 (0.1)1.38 (0.2)0.64STAI57.86 (17.4)57 (20.2)59.27 (16.8)0.4145.36 (15.7)46.07 (15.1)44.54 (14.6)0.51BIS-1176.31 (21.1)77.44 (22.0)74.45 (20.4)0.3166.29 (17.8)66 (19.1)66.62 (18.4)0.781st saliva sample AOC29.07 (14.62)29.3 (11.2)26.8 (12.9)0.7636.94 (14.73)41.4 (10.5)30.9 (11.9)<0.012nd saliva sample AOC30.92 (21.94)37.7 (15.2)20.9 (19.8)<0.0152.77 (15.29)52.2 (12.5)54.1 (11.7)0.69HR83.85 (10.17)92.29 (7.56)72.7 (6.97)<0.0178.69 (10.31)86.95 (7.56)69.24 (6.97)<0.01AOC: antioxidative capacity; BIS-11: Barrat Impulsiveness Scale; HR: Heart Rate; SD: Standard Deviation; STAI: State-Trait Anxiety Inventory; Y-BOCS: Yale-Brown obsessive-compulsive Scale. The average values is indicated, standard deviation is given in parentheses: Mean (St.dev.)*Mann–Whitney U test between High HR and Low HR groups Table 1Clinical and biochemical descriptive statistics: inter- and intragroup comparisonsPatients with OCD (Valid N = 31)Control group (Valid N = 29)Mean (SD)AllHigh HRLow HRp value*AllHigh HRLow HRp value*Age23.1 (5.1)24.3 (4.8)22.9 (5.5)0.8723.9 (6)23.5 (5.7)24.2 (5.3)0.99Y-BOCS obsessions16.89 (4.6)16.88 (4.3)16.91 (4.7)0.996.146.535.690.68Y-BOCS compulsions11.04 (3.8)11.53 (3.5)10.27 (3.2)0.43.53.333.690.8Y-BOCS Total20.45 (4.4)22.28 (5.1)17.45 (4.1)0.161.18 (0.2)1 (0.1)1.38 (0.2)0.64STAI57.86 (17.4)57 (20.2)59.27 (16.8)0.4145.36 (15.7)46.07 (15.1)44.54 (14.6)0.51BIS-1176.31 (21.1)77.44 (22.0)74.45 (20.4)0.3166.29 (17.8)66 (19.1)66.62 (18.4)0.781st saliva sample AOC29.07 (14.62)29.3 (11.2)26.8 (12.9)0.7636.94 (14.73)41.4 (10.5)30.9 (11.9)<0.012nd saliva sample AOC30.92 (21.94)37.7 (15.2)20.9 (19.8)<0.0152.77 (15.29)52.2 (12.5)54.1 (11.7)0.69HR83.85 (10.17)92.29 (7.56)72.7 (6.97)<0.0178.69 (10.31)86.95 (7.56)69.24 (6.97)<0.01AOC: antioxidative capacity; BIS-11: Barrat Impulsiveness Scale; HR: Heart Rate; SD: Standard Deviation; STAI: State-Trait Anxiety Inventory; Y-BOCS: Yale-Brown obsessive-compulsive Scale. The average values is indicated, standard deviation is given in parentheses: Mean (St.dev.)*Mann–Whitney U test between High HR and Low HR groups Clinical and biochemical descriptive statistics: inter- and intragroup comparisons AOC: antioxidative capacity; BIS-11: Barrat Impulsiveness Scale; HR: Heart Rate; SD: Standard Deviation; STAI: State-Trait Anxiety Inventory; Y-BOCS: Yale-Brown obsessive-compulsive Scale. The average values is indicated, standard deviation is given in parentheses: Mean (St.dev.) *Mann–Whitney U test between High HR and Low HR groups Experiment procedure All participants were asked to get a good night’s sleep before testing. Detailed instructions regarding the procedure were provided to all participants. The behavioral task was designed to induce both cognitive and emotional load. The cognitive load stems from the executive function demands of the antisaccade task, specifically response inhibition, while the emotional load is introduced by the presentation of affective stimuli. The cognitive–emotional load was implemented using a modified antisaccade task with an overlap design, in which affective pictures served as fixation stimuli [4]. The task consisted of three identical blocks lasting 30–35 minutes each, and the entire experimental procedure, including the preceding clinical interviews, lasted approximately 1.5–2 hours (Fig. 1). The study was carried out using the equipment of the Research Resource Center for Functional Brain Mapping at IHNA and NPh RAS. The paradigm comprised 300 antisaccade trials (100 trials per block). The stimuli and paradigm have been described in detail previously (see Figure S1 in the Supplementary Material) [4]. The task was designed to induce a combined cognitive and emotional load. The cognitive load stems from the executive function demands of the antisaccade task, specifically response inhibition, whereas the emotional load is concurrently introduced by the presentation of affective stimuli. This task has been previously validated both in patients with OCD and in healthy individuals [6].Fig. 1Experimental design Experimental design The volunteers were seated comfortably in a chair, with their head stabilized using a chin and head rest, and were instructed to avoid large movements. Calibration and validation of the eye-tracking camera were then performed. The experiment commenced only if the participant successfully passed the calibration and validation procedures; calibration was accepted if the average error after validation was < 0.30° of visual angle and the maximum error was < 0.50°. The task instructions were presented on the screen before each block. Participants were allowed to rest for approximately 3–5 minutes between blocks. Before the next block, the calibration and validation procedures were repeated. The cognitive‑emotional load while performing the antisaccade task was validated in our previous studies [4, 6]. Saliva collection The first saliva sample was collected 10–15 minutes before the start of the experimental procedure, and the second sample was collected immediately after the end of the procedure (see Fig. 1). The duration of the antisaccade task, including 1–2 minute breaks between blocks, ranged from approximately 15.5 to 19.5 minutes. Participants were instructed to follow specific rules to ensure valid trace element analysis in saliva: to sleep at least 8 hours the night before the experiment, to adhere to their usual daily routine, and to abstain from stimulants (caffeine immediately before the experiment, alcohol, and recreational drugs the night before) and other sedatives, unless these were part of their regular OCD medication regimen. Saliva was collected in a fasting state by passive drooling into a collection tube. Before saliva collection, participants rinsed their mouths with room‑temperature water. The volume of saliva collected for each sample was at least 1 ml. Saliva samples were frozen at −20 °C and stored for no longer than one month before trace element analysis. Analysis of salivary antioxidant capacity The salivary AOC was quantified using an enhanced chemiluminescence protocol as described previously [41]. The chemiluminescent system consisted of the free radical generator 2,2’-azobis(2-amidinopropane) dihydrochloride (ABAP, Sigma) and the chemiluminescent probe luminol (Sigma). A 1 mmol/L luminol solution (Sigma) and a 50 mmol/L ABAP solution were prepared by dissolving the weighed reagents in phosphate buffer (100 mM KH2PO4, pH 7.4, Sigma). Chemiluminescence was recorded at 37 °C until a stable baseline was reached, after which a 200 μL aliquot of saliva was added. Recording continued until a new steady-state level had been achieved. Antioxidant potential was calculated as the area of signal depression following the addition of the antioxidant sample to the system. The mean of three parallel measurements was used for analysis. Measurements were performed using a 12‑channel Lum‑1200 chemiluminometer (DISoft, Russia), and signal processing was carried out with PowerGraph 3.3 Professional software. ECG acquisition and preprocessing For all participants, a 15‑minute resting‑state ECG was recorded either between 9:00 a.m. and 12:00 p.m. (morning) or between 1:00 p.m. and 3:00 p.m. (afternoon), before the onset of the cognitive–emotional task, and ECG recording was continued throughout the three blocks of task performance. The recording cabin was temperature controlled (22–24 °C), dimly lit, and participants were comfortably seated. ECG data were recorded with a Brain Products amplifier (Brain Products GmbH, Gilching, Germany) using a blood pulse sensor attached to the middle finger of the right hand. The raw signal was band‑pass filtered at 0.05–40 Hz and preprocessed with automatic R‑peak detection implemented in a Matlab script, followed by visual inspection and correction of the detected R‑peaks. Following automated detection, the R‑peak annotations were manually inspected and corrected to address errors such as missed beats and noise‑related false positives. Only datasets without substantial uncorrectable artifacts (e.g. prolonged signal loss) after this procedure were retained for analysis. Heart rate patterns Using the R‑peak intervals, time‑domain and frequency‑domain HRV indices were calculated from the ECG data. The average HR was computed across the three experimental blocks. HR was then used to divide each group of participants into two subgroups according to their individual mean HR (low HR and high HR), as shown in Table 1. We applied this procedure separately within the control and OCD groups (Fig. 3a). In the OCD group, the overall mean HR was 85.1 beats per minute; in the low‑HR subgroup it was 74.6, and in the high‑HR subgroup 93.2 beats per minute. In the control group, the overall mean HR was 77.3 beats per minute; in the low‑HR subgroup it was 70.0, and in the high‑HR subgroup 87.7 beats per minute. For frequency-domain HRV markers, a Fast Fourier Transform (FFT) was performed to estimate the following variables [12]:High-frequency power (HF): spectral power in the 0.15–0.40 Hz band, reflecting parasympathetic (vagal) modulation of cardiac activity.Low-frequency power (LF): spectral power in the 0.04–0.15 Hz band, which reflects a complex mixture of sympathetic and parasympathetic modulation, with the relative contribution of each varying across physiological contexts [42, 43]. Other studies have suggested that LF primarily reflects sympathetic activity, with a smaller parasympathetic contribution [44].Very low-frequency power (VLF): spectral power in the 0.003–0.04 Hz band, thought to reflect slow-acting regulatory processes involving both sympathetic and parasympathetic influences, as well as non-neural mechanisms. Although VLF is less stable than LF and HF components in short-term recordings, it can still provide useful information about autonomic dysregulation and underlying regulatory mechanisms. It has been used to assess autonomic dysfunction in a range of clinical conditions, including heart failure, depression, and post–myocardial infarction states [45, 46]. Furthermore, several studies suggest that VLF can yield meaningful insights even in short-term segments, particularly when advanced signal processing methods are applied [47].Normalized low-frequency power (LFnu): calculated as LF expressed in normalized units, i.e. as the percentage of LF relative to the sum of LF and HF of the R‑peak interval time series. LFnu is widely used as an index of sympathovagal balance, reflecting the relative weighting of low‑frequency versus high‑frequency modulation [43]. At the same time, some authors interpret LFnu primarily as reflecting the proportion of sympathetic activity within overall autonomic modulation and view it as complementary to normalized high-frequency power (HFnu) [12]. High-frequency power (HF): spectral power in the 0.15–0.40 Hz band, reflecting parasympathetic (vagal) modulation of cardiac activity. High-frequency power (HF): spectral power in the 0.15–0.40 Hz band, reflecting parasympathetic (vagal) modulation of cardiac activity. Low-frequency power (LF): spectral power in the 0.04–0.15 Hz band, which reflects a complex mixture of sympathetic and parasympathetic modulation, with the relative contribution of each varying across physiological contexts [42, 43]. Other studies have suggested that LF primarily reflects sympathetic activity, with a smaller parasympathetic contribution [44]. Low-frequency power (LF): spectral power in the 0.04–0.15 Hz band, which reflects a complex mixture of sympathetic and parasympathetic modulation, with the relative contribution of each varying across physiological contexts [42, 43]. Other studies have suggested that LF primarily reflects sympathetic activity, with a smaller parasympathetic contribution [44]. Very low-frequency power (VLF): spectral power in the 0.003–0.04 Hz band, thought to reflect slow-acting regulatory processes involving both sympathetic and parasympathetic influences, as well as non-neural mechanisms. Although VLF is less stable than LF and HF components in short-term recordings, it can still provide useful information about autonomic dysregulation and underlying regulatory mechanisms. It has been used to assess autonomic dysfunction in a range of clinical conditions, including heart failure, depression, and post–myocardial infarction states [45, 46]. Furthermore, several studies suggest that VLF can yield meaningful insights even in short-term segments, particularly when advanced signal processing methods are applied [47]. Very low-frequency power (VLF): spectral power in the 0.003–0.04 Hz band, thought to reflect slow-acting regulatory processes involving both sympathetic and parasympathetic influences, as well as non-neural mechanisms. Although VLF is less stable than LF and HF components in short-term recordings, it can still provide useful information about autonomic dysregulation and underlying regulatory mechanisms. It has been used to assess autonomic dysfunction in a range of clinical conditions, including heart failure, depression, and post–myocardial infarction states [45, 46]. Furthermore, several studies suggest that VLF can yield meaningful insights even in short-term segments, particularly when advanced signal processing methods are applied [47]. Normalized low-frequency power (LFnu): calculated as LF expressed in normalized units, i.e. as the percentage of LF relative to the sum of LF and HF of the R‑peak interval time series. LFnu is widely used as an index of sympathovagal balance, reflecting the relative weighting of low‑frequency versus high‑frequency modulation [43]. At the same time, some authors interpret LFnu primarily as reflecting the proportion of sympathetic activity within overall autonomic modulation and view it as complementary to normalized high-frequency power (HFnu) [12]. Normalized low-frequency power (LFnu): calculated as LF expressed in normalized units, i.e. as the percentage of LF relative to the sum of LF and HF of the R‑peak interval time series. LFnu is widely used as an index of sympathovagal balance, reflecting the relative weighting of low‑frequency versus high‑frequency modulation [43]. At the same time, some authors interpret LFnu primarily as reflecting the proportion of sympathetic activity within overall autonomic modulation and view it as complementary to normalized high-frequency power (HFnu) [12]. The following processing steps were used to calculate these frequency-domain parameters. The sequence of R‑R intervals was converted into a beats‑per‑minute (BPM) time series, and the spectrum function from MATLAB’s Signal Processing Toolbox was then used to transform this time series into the frequency domain. The indices listed in items 1–4 were calculated as the area under the resulting power spectrum within the specified frequency bands. Statistics All statistical analyses were performed using Statistica 13. A repeated-measures mixed analysis of variance (rmANOVA) was used to evaluate the effects of the between-subject factor Group (2 levels: OCD, control) and the within-subject factors AOC (2 levels) and HRV dynamics (3 levels). Separate rmANOVAs were conducted for each HRV metric. Post hoc comparisons were corrected using the Bonferroni method. Partial eta-squared (η2 p) was calculated as a measure of effect size, with values of 0.01, 0.06, and 0.14 interpreted as small, medium, and large effects, respectively. Correction for multiple rmANOVA tests across different HRV metrics was not applied, as each metric addresses a distinct conceptual question [12]. Additionally, correlation analyses were performed using Spearman’s rank correlation coefficient to examine relationships between HR, HRV, AOC metrics, and clinical scores (BDI, Y‑BOCS). Spearman correlations are appropriate here given the non-normal distribution and ordinal nature of some variables. An analysis of covariance (ANCOVA) was conducted to statistically control for medication effects. Two coding strategies were employed: an ordinal scale indexing overall medication load and a set of binary variables for each drug class. The following covariates were included: number of different medication types (0 = no medication, n = 9; 1 = single medication, n = 11; 2 = combination of medications, n = 11); antidepressant use (0 = no, n = 12; 1 = yes, n = 19); anticonvulsant use (0 = no, n = 24; 1 = yes, n = 7); and antipsychotic use (0 = no, n = 24; 1 = yes, n = 7). No significant medication effects were detected (see Supplementary Table S1). Results Our analysis revealed three key sets of findings regarding antioxidant capacity (AOC), heart rate (HR), and heart rate variability (HRV) in patients with OCD compared with healthy controls. Salivary antioxidant capacity (AOC) Salivary antioxidant capacity (AOC) was significantly lower in the OCD group than in the healthy control (HC) group both before and after the cognitive–emotional task. A repeated-measures ANOVA revealed a significant main effect of Group (F(1, 58) = 15.997, p = 0.00018, η2 p = 0.22), confirming higher overall AOC in healthy participants. There was also a significant main effect of Time (F(1, 58) = 14.330, p = 0.00037, η2 p = 0.20), indicating that AOC increased across the session. However, this increase was driven exclusively by the HC group, as evidenced by a significant simple effect (F(1, 58) = 8.9527, p = 0.00406, η2 p = 0.13). In contrast, the OCD group showed no significant change in AOC following the task (see Fig. 2).Fig. 2Antioxidant capacity (AOC) dynamics in patients with OCD and healthy controls. Thick horizontal lines represent group means. Shaded bar contours indicate values within one standard deviation (SD). Error bars extending from the mean represent ±1.96 SD. Red and blue circles denote individual data points for each participant. The asterisk (*) indicates a statistically significant difference (p < 0.05) Antioxidant capacity (AOC) dynamics in patients with OCD and healthy controls. Thick horizontal lines represent group means. Shaded bar contours indicate values within one standard deviation (SD). Error bars extending from the mean represent ±1.96 SD. Red and blue circles denote individual data points for each participant. The asterisk (*) indicates a statistically significant difference (p < 0.05) Heart rate (HR) The analysis of heart rate (HR) data was guided by a significant negative correlation between baseline HR and change in antioxidant capacity (AOC) within the OCD cohort. To further investigate this relationship and test for potential physiological subtypes, we stratified both the OCD and healthy control (HC) groups into “high‑HR” and “low‑HR” subgroups using their respective median baseline HR values (Fig. 3a). At the same time, in the full sample, mean HR was significantly higher in the OCD group than in the HC group (group effect: F(1, 56) = 4.8248, p = 0.03221, η2 p = 0.08).Fig. 3(a) Individual mean HR values and the scheme for subgroup division into high‑HR and low‑HR groups; (b) changes in AOC in patients with OCD and healthy controls with high and low HR (a) Individual mean HR values and the scheme for subgroup division into high‑HR and low‑HR groups; (b) changes in AOC in patients with OCD and healthy controls with high and low HR Antioxidant capacity (AOC) in sub-groups with low and high baseline heart rate (HR) A significant interaction was observed between change in antioxidant capacity (AOC) and mean heart rate (HR) (Wilks’ λ = 0.85222, F(2, 55) = 4.7688, p = 0.012, η2 p = 0.15), as illustrated in Fig. 3b. Post hoc analyses revealed a divergent pattern within the OCD group. The high‑HR OCD subgroup showed a significant increase in AOC from the first to the second saliva sample (p = 0.0083). In contrast, the low‑HR OCD subgroup exhibited a non‑significant trend toward a decrease in AOC (p = 0.091). Among healthy controls (HC), both subgroups showed a significant increase in AOC; however, this increase was more pronounced in the low‑HR control subgroup than in the high‑HR control subgroup (p < 0.0001 and p = 0.018, respectively). Furthermore, average AOC values differed significantly between the high‑HR and low‑HR OCD subgroups (mixed‑effect: F(2, 116) = 3.556, p = 0.033, η2 p = 0.08; post hoc p = 0.0079), but not between the corresponding control subgroups. In summary, the physiological response of OCD patients in the high‑HR subgroup was comparable to that of healthy controls, with both groups showing a significant increase in AOC. Conversely, the low‑HR control subgroup exhibited the most robust AOC increase. Most notably, OCD patients in the low‑HR subgroup showed a distinct and maladaptive pattern, characterized by a decrease in AOC, which clearly distinguished them from all other subgroups. Heart rate variability (HRV) As shown in Table 2, VLF power increased from the first to the second and third blocks of the task (main effect of Block: F(2, 116) = 73.522, p < 0.0001, η2 p = 0.55). This effect was significantly more pronounced in the OCD group (Block × Group interaction: F(2, 116) = 3.892, p = 0.0242, η2 p = 0.07), with VLF values in the second and third experimental blocks being significantly higher in the OCD group (LSD post hoc tests, p < 0.0094).Table 2Heart rate variability parameters calculated separately for each of the three task blocks in the OCD and control groupsgroupValid NHRV parametersBlock1Block2Block3MeanSDMeanSDMeanSDOCD31LF0.030.020.040.030.050.05HF0.020.020.030.030.030.03VLF23.929.9634.9412.2336.9327.82LFnu0.580.130.630.190.630.16Control29LF0.030.030.050.050.060.09HF0.020.020.030.020.030.03VLF23.176.7331.899.7131.8610.34LFnu0.530.160.540.160.550.15HF: High-Frequency Power; HRV: Heart rate variability; LF: Low-Frequency Power; LFnu: normalized Low-Frequency Power; SD: Standard Deviation; VLF: Very Low Frequency Power Table 2Heart rate variability parameters calculated separately for each of the three task blocks in the OCD and control groupsgroupValid NHRV parametersBlock1Block2Block3MeanSDMeanSDMeanSDOCD31LF0.030.020.040.030.050.05HF0.020.020.030.030.030.03VLF23.929.9634.9412.2336.9327.82LFnu0.580.130.630.190.630.16Control29LF0.030.030.050.050.060.09HF0.020.020.030.020.030.03VLF23.176.7331.899.7131.8610.34LFnu0.530.160.540.160.550.15HF: High-Frequency Power; HRV: Heart rate variability; LF: Low-Frequency Power; LFnu: normalized Low-Frequency Power; SD: Standard Deviation; VLF: Very Low Frequency Power Heart rate variability parameters calculated separately for each of the three task blocks in the OCD and control groups HF: High-Frequency Power; HRV: Heart rate variability; LF: Low-Frequency Power; LFnu: normalized Low-Frequency Power; SD: Standard Deviation; VLF: Very Low Frequency Power To corroborate these findings, non-parametric analyses were also performed. A Friedman test indicated a significant effect of Block on VLF across all participants (Friedman ANOVA: χ2 (N = 83, df = 2) = 82.14, p < 0.0001). Wilcoxon matched-pairs tests showed that VLF in the second and third blocks was significantly higher than in the first block (z > 7.18, p < 0.0001), whereas the second and third blocks did not differ from each other (z = 1.06, p = 0.29). Mann–Whitney U tests revealed that VLF values in the second (z = 2.00, p = 0.045) and third (z = 2.14, p = 0.03) blocks were significantly higher in the OCD group than in the HC group. Thus, power in the very low-frequency (VLF) band increased in both groups during the task, but this increase was significantly more pronounced in the OCD group, suggesting heightened autonomic reactivity to cognitive–emotional stress (see Fig. 4a).Fig. 4(a) Dynamics of VLF power across the three task blocks in patients with OCD and healthy controls; (b) normalized low-frequency power (LFnu) in patients with OCD and healthy controls (a) Dynamics of VLF power across the three task blocks in patients with OCD and healthy controls; (b) normalized low-frequency power (LFnu) in patients with OCD and healthy controls The OCD group had significantly higher normalized low-frequency power (LFnu) throughout the task (group effect: F(1, 58) = 8.2772, p = 0.00561, η2 p = 0.12; Figure 4b), consistent with a pattern of sympathetic predominance in cardiac autonomic regulation. Heart rate variability (HRV) in sub-groups with low and high baseline heart rate (HR) The dynamics of normalized low-frequency power (LFnu) further distinguished the heart rate (HR) subgroups (Fig. 5). A significant Group × Time interaction (F(2, 112) = 3.24, p = 0.043, η2 p = 0.06) indicated divergent responses. The high‑HR OCD subgroup exhibited consistently elevated LFnu across all task blocks compared with the high‑HR control subgroup, consistent with a sustained pattern of sympathetic predominance.Fig. 5Lfnu dynamics in patients with OCD and healthy controls stratified by high and low heart rate (HR) Lfnu dynamics in patients with OCD and healthy controls stratified by high and low heart rate (HR) In contrast, the low‑HR OCD subgroup demonstrated a reactive and dysregulated autonomic profile. These patients showed a significant increase in LFnu from the first to the second and third blocks (post hoc p = 0.0067 and p = 0.0051, respectively), whereas LFnu remained stable in the high‑HR OCD subgroup and in both control subgroups. This marked reactivity points to a particularly maladaptive and hyper‑responsive pattern of sympathetic regulation in OCD patients with low baseline HR. Correlation analysis In the full sample, Spearman rank-order correlations revealed significant associations between questionnaire scores and AOC. AOC in the second saliva sample correlated negatively with STAI scores (r = −0.57, p < 0.0001), BDI scores (r = −0.57, p < 0.0001), and BIS‑11 scores (r = −0.50, p < 0.0001). The increase in AOC over the course of the study (ΔAOC; see Figure S2) also showed significant negative correlations with STAI (r = −0.42, p = 0.001), BDI (r = −0.42, p = 0.0012), and BIS‑11 scores (r = −0.29, p = 0.0288). Separately conducted correlation analyses in the OCD and control groups showed significant associations between HR and AOC. In the OCD group, mean HR correlated positively with AOC in the second saliva sample (r = 0.44, p = 0.0029) and with the increase in AOC over the course of the study (r = 0.38, p = 0.016). In the control group, mean HR correlated positively with AOC in the first saliva sample (r = 0.46, p = 0.0016) but negatively with the increase in AOC during the study (r = −0.58, p < 0.0001). We did not observe significant correlations between participant age and AOC, HR measures, or questionnaire scores. Discussion The AOC was significantly higher in healthy participants than in the OCD group and increased significantly after the cognitive–emotional task only in healthy volunteers. Previous work has shown that an imbalance between oxidant and antioxidant activity and impaired antioxidant defenses is associated with OCD [48]. Consistent with this, several studies have reported elevated levels of lipid peroxidation products such as malondialdehyde in patients with major depressive disorder comorbid with OCD [40] and in patients with pure OCD [49]. Meta-analytic work indicates that patients with OCD show statistically significant increases in oxidant markers, whereas findings for antioxidant markers are less consistent [22, 50]. In contrast to our results, several studies have described an oxidative imbalance shifted toward the antioxidant side and alterations in antioxidant vitamin levels in OCD [34, 49, 51]. At the same time, the use of antioxidant agents targeting oxidative stress has emerged as a novel treatment strategy and has been shown to produce significant improvements in OCD symptoms [38]. Apparently, the relationship between antioxidant status and OCD symptoms is complex and likely nonlinear, and depends on the compensatory capacities of the nervous system that are recruited when coping with stress. One of the key systems involved in restoring and normalizing stress levels is the autonomic nervous system (ANS). Our results revealed opposite associations between HR and changes in AOC after the cognitive–emotional task in the OCD and control groups: in patients with OCD, higher mean HR was associated with greater increases in AOC, whereas in healthy volunteers higher HR was associated with smaller AOC increases or even decreases. In the OCD group, higher HR was thus linked to elevated post‑task AOC levels. When we divided participants into subgroups with high and low baseline mean HR, we further clarified that the direction of the HR–AOC association differed in OCD compared with controls. Both high‑HR and low‑HR subgroups in the control group showed increases in AOC after task performance; however, this increase was significantly more pronounced in participants with initially lower HR. In contrast, OCD subgroups with low and high HR differed markedly in both AOC changes and average AOC levels: patients with low HR had substantially reduced AOC and a blunted or absent AOC increase, whereas patients with higher HR showed higher AOC levels and a significant task‑related AOC increase similar to that observed in healthy controls. These findings suggest that OCD patients with elevated HR may exhibit more adaptive ANS and oxidant–antioxidant functioning under cognitive–emotional load, with physiological responses that more closely resemble those of healthy individuals than the responses observed in OCD patients with low baseline HR. According to previous findings, HR in OCD appears to vary as a function of factors that are not yet fully understood; consequently, some studies have reported elevated HR in OCD [7, 12], whereas others have not found significant HR alterations in comparison with healthy controls [9, 11]. Consistent with this heterogeneity, significant correlations between elevated HR and OCD symptom severity have not been observed consistently across studies. It is therefore possible that increased HR in patients with OCD is not necessarily a purely negative indicator; rather, in at least a subset of patients, it may reflect a compensatory mechanism involving coordinated adjustments in both ANS activity and oxidant–antioxidant balance. Significant between‑group differences in HRV were observed for LFnu and VLF dynamics. The physiological interpretation of the VLF component remains complex and is less well established than that of the HF and LF bands. Earlier work, such as [52], linked VLF power to reduced parasympathetic activity, whereas more recent evidence suggests that VLF is shaped by multiple slow-acting mechanisms, including thermoregulatory processes, the renin–angiotensin–aldosterone system, vascular tone, and other long‑period sympathetic and humoral influences [53, 54]). In our study, VLF power increased in both groups, with a more pronounced elevation in the OCD group. This heightened VLF reactivity may reflect stronger engagement of slow‑acting regulatory systems in response to cognitive–emotional stress in OCD. Although contemporary research tends to view LFnu as an index of overall autonomic modulation, reflecting the combined influence of both sympathetic and parasympathetic branches of the ANS [55], the majority of earlier studies interpreted this parameter more narrowly, often treating it as a direct marker of sympathetic tone or sympathovagal balance [42, 43]. Therefore, our conclusion regarding an altered autonomic state in OCD is not based on LFnu in isolation. Instead, it rests on a convergent pattern of multiple indicators: elevated VLF reactivity, higher LFnu, and their correlation with HR. We interpret these combined patterns as evidence of an altered autonomic state under cognitive load, in a way that is consistent with a broader stress response profile in the OCD group while avoiding over‑attribution of the findings specifically to sympathetic predominance. Our data are also consistent with previous findings indicating that OCD is characterized by decreased parasympathetic tone and abnormal sympathetic reactivity compared with healthy controls [12, 56]. However, most prior work has focused primarily on dysfunction of the parasympathetic branch of the ANS [10, 57]. In particular, several studies have reported negative associations between HRV parameters indexing parasympathetic activity (such as HF power and RMSSD) and OCD severity [10, 57]. The observed links between reduced parasympathetic activity and greater OCD symptom burden may be partly explained by previously reported reductions in serotonin transporter availability in patients with OCD [57–59]. Although this line of research is still evolving, it is becoming increasingly clear that the effectiveness of specific pharmacological treatments in OCD must be evaluated in the broader context of homeostatic parameters, including autonomic function, oxidative–antioxidant balance, trace elements, and endocrine factors. The functioning of the sympathetic and parasympathetic components of the ANS is highly interconnected and should be considered jointly when examining the pathogenesis of OCD [58]. This idea is supported by our HRV findings in OCD subgroups stratified by baseline HR. In the OCD group, individuals with lower HR showed a significant increase in LFnu after the first task block, indicating heightened reactivity of this parameter that sharply contrasted with the relatively stable LFnu response in the low‑HR control subgroup. Consequently, the pattern of ANS responses to cognitive–emotional load appeared less balanced in OCD patients with lower HR, which may point to reduced compensatory capacity of the ANS in this subgroup. In the context of OCD, chronic anxiety and repetitive intrusive thoughts act as persistent internal stressors that can drive tonic sympathetic hyperactivity [60]. The resulting elevation in heart rate (HR) may not represent a purely pathological endpoint but rather a sustained preparatory state that enhances vigilance, attention, and the speed of cognitive processing in order to monitor and respond to perceived threats [61]. This state can, in turn, support the high cognitive load associated with obsessive rumination as well as the increased motor activity linked to compulsions. From an evolutionary perspective, such a profile resembles a conserved adaptive stress response, which may become maladaptive when chronically engaged in the absence of external danger. Conversely, a lower HR in individuals with OCD—a group typically characterized by elevated sympathetic activation—may reflect an exhaustion phase of the stress response. From the perspective of Selye’s General Adaptation Syndrome, prolonged, unremitting stress can ultimately deplete the capacity for a normal sympathetic response, leading to maladaptation [62]. However, the interpretation of low baseline HR as reflecting an exhaustion phase of the stress response, although grounded in Selye’s General Adaptation Syndrome framework, remains speculative in the context of OCD and should be examined in future prospective studies. This is manifested as autonomic imbalance, particularly in heart rate regulation [63]. Such a pattern, which has been described in various forms of long-term stress, is thought to indicate a depletion of the body’s reserve capacity and its ability to mount effective stress resistance [64]. In our sample, biochemical data from OCD patients showed a positive correlation between increased antioxidant activity and higher HR, suggesting a compensatory role of sympathetic nervous system activation in patients with elevated HR. By contrast, patients with a lower baseline HR may exhibit a depletion of this compensatory mechanism and a shift toward exhaustion-related states. Previous work supports a close link between oxidative stress and autonomic dysfunction. For example, in patients with vertigo, elevated levels of reactive oxygen species (ROS) have been associated with reduced parasympathetic activity [65], and in arterial hypertension, increased sympathetic drive has been repeatedly linked to oxidative stress [66, 67]. Elevated oxidative stress accompanied by reduced HRV has been observed not only in specific clinical populations, such as patients with chronic kidney disease, but also in community samples [68, 69]. Interestingly, recent data suggest that the impact of environmental exposures such as metals and phenols on HRV may be at least partly mediated by oxidative stress [70]. In contrast to the VLF and LFnu bands, no significant group differences were observed in the HF component, which is widely regarded as a specific marker of parasympathetic (vagal) activity. This pattern suggests that autonomic dysregulation in our OCD cohort under cognitive load may be driven primarily by alterations in sympathetic-related modulation, rather than by a pronounced shift in cardiac vagal tone. This interpretation is further supported by the HRV profile of the OCD subgroup with low heart rate, whose state may reflect an exhaustion phase of the stress response and a maladaptive configuration of both autonomic branches, rather than a simple increase in parasympathetic activation. Future studies employing different stress paradigms, longer recordings, and larger samples will be needed to clarify the role of vagal functioning in OCD. Conclusions We found that patients with OCD exhibit a marked impairment in salivary antioxidant capacity and in its responsiveness to cognitive–emotional load. This impairment is not uniform but appears to be strongly shaped by the individual’s baseline autonomic state. Our exploratory data point to a potentially distinct subgroup of OCD patients with lower baseline heart rate who show the most maladaptive profile: a decrease in antioxidant capacity under stress combined with a hyperreactive sympathetic nervous system. In contrast, OCD patients with higher baseline heart rate display a more adaptive physiological response, more closely resembling that of healthy controls. Taken together, these findings suggest that baseline heart rate may represent a promising biomarker for identifying a physiologically distinct and potentially more severe OCD subtype, characterized by a failure of compensatory autonomic and antioxidant mechanisms. Limitations This study has several limitations that should be taken into account when interpreting the results. First, the sample size, while adequate to detect our primary group effects, constrained our ability to perform more fine‑grained analyses. The marked heterogeneity of OCD, which includes distinct symptom dimensions (e.g., contamination, checking, hoarding), means that our results may not generalize equally across all clinical subtypes. A larger cohort would allow us to test whether the physiological subgroups identified here (high‑HR vs. low‑HR) map onto specific symptom profiles or trajectories. Second, although we statistically adjusted for psychotropic medication, residual confounding due to medication effects on autonomic and redox markers cannot be fully excluded, given known influences of psychotropic drugs on HRV and related physiological parameters. The use of a median split for HR stratification reduces statistical power and should be considered an exploratory, data‑driven approach to illustrating heterogeneity within the OCD group, warranting replication in larger samples. While our ANCOVA models indicated that medication status did not significantly affect the primary outcomes, the diversity of medications and combinations in the patient group introduces additional heterogeneity; future studies with larger, medication‑naïve cohorts will be important to confirm these findings. Third, we did not stratify participants by age into distinct developmental groups (e.g., younger vs. older adults). Because both autonomic function and antioxidant activity change across the lifespan, age‑related influences may have affected our results, and examining these relationships in a broader age range would help clarify the developmental trajectory of autonomic–antioxidant interactions in OCD. Finally, the cross‑sectional design precludes causal inferences: although we identified associations between autonomic tone and antioxidant capacity, the direction and temporal dynamics of these relationships remain unclear. Longitudinal studies following patients across treatment, and investigations in first‑degree relatives of individuals with OCD, will be crucial for assessing genetic contributions and the stability of these physiological subtypes over time [71]. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary material 1 Supplementary material 1