If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Instituto D’Or de Pesquisa e Ensino (IDOR), Rua Diniz Cordeiro 30, Botafogo, 22281-100, Rio de Janeiro, BrazilPrograma de Pós-Graduação Em Ciências da Reabilitação, Centro Universitário Augusto Motta - UNISUAM, Rio de Janeiro, Brazil
Bihemispheric prefrontal tDCS is safe and feasible in hospitalized patients with COVID-19.
•
Bihemispheric prefrontal tDCS showed a large effect size on heart rate variability in COVID-19 inpatients.
•
Oxygen saturation may be improved through prefrontal tDCS.
•
Bihemispheric prefrontal tDCS seems promising to enhance clinical outcomes after severe infections.
Abstract
Background
maladaptive changes in the autonomic nervous system (ANS) have been observed in short and long-term phases of COVID-19 infection. Identifying effective treatments to modulate autonomic imbalance could be a strategy for preventing and reducing disease severity and induced complications.
Objective
to investigate the efficacy, safety, and feasibility of a single session of bihemispheric prefrontal tDCS on indicators of cardiac autonomic regulation and mood of COVID-19 inpatients.
Methods
patients were randomized to receive a single 30-minute session of bihemispheric active tDCS over the dorsolateral prefrontal cortex (2mA; n = 20) or sham (n = 20). Changes in time [post-pre intervention] in heart rate variability (HRV), mood, heart rate, respiratory rate, and oxygen saturation were compared between groups. Additionally, clinical worsening indicators and the occurrence of falls and skin injuries were evaluated. The Brunoni Adverse Effects Questionary was employed after the intervention.
Results
there was a large effect size (Hedges’ g = 0.7) of intervention on HRV frequency parameters, suggesting alterations in cardiac autonomic regulation. An increment in oxygen saturation was observed in the active group but not in the sham after the intervention (P = 0.045). There were no group differences regarding mood, incidence and intensity of adverse effects, no occurrence of skin lesions, falls, or clinical worsening.
Conclusions
a single prefrontal tDCS session is safe and feasible to modulate indicators of cardiac autonomic regulation in acute COVID-19 inpatients. Further research comprising a thorough assessment of autonomic function and inflammatory biomarkers is required to verify its potential to manage autonomic dysfunctions, mitigate inflammatory responses and enhance clinical outcomes.
The complications of coronavirus disease 2019 (COVID-19) have become a challenge for the global health and scientific community. Besides the respiratory system, the disease also affects the nervous system through different mechanisms, such as inflammatory, direct viral infiltration, cerebrovascular dysfunctions, and hypoxia [
]. Therefore, identifying effective interventions to restore autonomic function and minimize the inflammatory process may be a strategy to prevent and reduce clinical worsening following COVID-19 and other diseases.
Techniques targeting neuromodulation may have therapeutic potential for managing symptoms and complications associated with severe infections like COVID-19. A recent randomized control trial investigated the effects of high-definition transcranial direct current stimulation (HD-tDCS) applied over the diaphragmatic motor cortex on the recovery of critically ill COVID-19 patients requiring pulmonary rehabilitation. The authors reported that HD-tDCS was a safe and effective add-on intervention in enhancing the clinical condition, reducing hospitalization length and the number of days of mechanical ventilation in patients who received active stimulation compared to sham [
]. However, the amount of patients with COVID-19 requiring intensive care and mechanical ventilation is relatively low (about 5% at the initial phase of the pandemic) [
]. Thus, a distinct strategy could be identifying a tDCS configuration to prevent and treat clinical worsening instead of focusing on the diaphragm function.
In this sense, stimulating brain regions associated with the regulation of ANS may be a promising approach to treating patients with COVID-19. As this system plays an important role in the inflammatory process [
], its modulation could reduce the disease's inflammatory response besides restoring autonomic function. A recent study indeed showed improvements in the levels of inflammatory markers of hospitalized patients with COVID-19 after receiving a non-invasive vagus nerve stimulation therapy [
Efficacy and neurophysiological predictors of treatment response of adjunct bifrontal transcranial direct current stimulation (tDCS) in treating unipolar and bipolar depression.
]. Previous accounts indeed reported alterations in indicators of cardiac autonomic modulation, through heart rate variability (HRV) analysis, after a single session of prefrontal tDCS in healthy individuals [
]. These findings corroborate the potential benefit of neuromodulation in restoring autonomic dysfunctions observed in many diseases and COVID-19 survivors, besides preventing exacerbated inflammatory responses, as abovementioned. However, evidence about early tDCS efficacy and safety is still scarce in hospital settings, particularly in COVID-19.
The present is an exploratory, randomized, sham-controlled, double-blinded study aiming to provide insights into the safety and feasibility of a single session of bihemispheric prefrontal tDCS in COVID-19 inpatients and its acute effects on indicators of cardiac autonomic modulation assessed through the HRV. Additionally, since the relationship between HRV and mood alterations has been well established in the literature [
], the patient's mood was also an outcome investigated in this study.
2. Material and methods
2.1 Study design, randomization, and participants
The present study consisted of a randomized, sham-controlled, double-blinded design conducted at the semi-intensive care unit of the Hospital Quinta D’Or (Rio de Janeiro, Brazil). Given the exploratory nature of the study and the limited resources for data collection [
], a sample of 40 participants was enrolled. This study is part of a protocol previously registered with clinicaltrials.gov (NCT04808284) and executed with no major deviations. It was approved by the institutional ethics committee (No. 4.156.081) and reported according to CONSORT guidelines [
]. All patients provided informed written consent before enrolment according to the Declaration of Helsinki.
Patients were admitted consecutively and randomly assigned to receive either sham or active-tDCS. Randomization was performed through a web-based system, and group assignments were concealed from patients and investigators who conducted the evaluation protocol.
Recruitments were based on medical records or a physician referral according to the following inclusion criteria: 1) individuals of both sex, aged between 18-80 years; 2) confirmed infection by SARS-CoV-2; 3) ability to understand and execute the proposed protocol; 4) stable vital signs (body temperature <38 °C; respiratory rate between 12 and 30rpm; and blood pressure between 90/60mmHg and 140/90mmHg; heart rate between 50 and 120 bpm; SpO2 ≥ 90%). Exclusion criteria were: 1) dyspnoea or signs of respiratory effort; 2) hemodynamic instability; 3) deep vein thrombosis, active bleeding, use of cardiac pacemaker; 4) injury, pain, or metallic implants in the skull or scalp; 5) seizure history; 6) suspected or confirmed pregnancy; 7) concomitant or previous rheumatic or neurological diseases; 8) severe psychiatric diseases (schizophrenia, bipolar disorder, intellectual disability); 9) severe musculoskeletal and integumentary disorders; 10) severe liver or kidney disease. COVID-19 severity was classified using the Ordinal Scale for Clinical Improvement proposed by a special WHO committee [
To investigate the safety, feasibility, and acute effects of bihemispheric prefrontal tDCS on indicators of autonomic modulation of COVID-19 inpatients, the following evaluations were performed: clinical data collection, assessment of adverse effects of tDCS, mood assessment, and HRV recording.
The entire protocol was conducted in the patient's semi-intensive care room. One investigator was aware of group randomization for the patients and responsible for setting up the stimulator according to the protocol for sham or active-tDCS. This investigator was not involved in any of the evaluations. Two other investigators were responsible for applying the evaluations and, as for the patients, they were unaware of the group assignment.
2.3 Stimulation protocol
Patients were asked to rest in a sitting position on the bed or chair to feel comfortable during the entire stimulation session. A single tDCS session was applied for both groups by a DC stimulator (NeuroConn, Germany), with a pair of saline-soaked sponge electrodes (surface 35cm2) positioned and fixed to the scalp using elastic bands. For active-tDCS, electrodes were placed bilaterally over the dorsolateral prefrontal cortex (DLPFC), specifically on F3 (anode) and F4 (cathode), according to the 10–20 EEG system [
]. For sham-tDCS, the current intensity was ramped up for 30s until it reached 2mA, and after 30s, the intensity was ramped down for 15s. To ensure the placebo effect during sham-tDCS, patients remained with the electrodes attached to the scalp for 30min. Investigators and patients were unaware of when stimulation was turned off [
Each patient used disposable electrodes, sponges, and elastic bands to reduce the risk of infections, and all instruments were disinfected with alcohol 70° after each experimental session. In addition, strict adherence to the appropriate collection, biosafety, and personal protective equipment protocols were adopted [
]. One of the investigators monitored patients' vital signs throughout the evaluation and intervention process. The protocol would be immediately interrupted if any exclusion criteria were identified.
2.4 Clinical data
Clinical data were extracted from patients' medical records, including comorbidities, clinical signs and symptoms at hospital admission, and drugs used during hospitalization. The following parameters were collected before and after tDCS intervention during the rest period that heart rate variability (HRV) was recorded: heart rate, respiratory rate, and oxygen saturation measured by pulse oximetry. The presence of severe skin lesions, fall events, and acute clinical worsening, including transfer to the Intensive Care Unit (ICU), orotracheal intubation, and death, were evaluated within one hour after the intervention.
2.5 Adverse effects of tDCS
The adverse effects of tDCS were assessed through the Brunoni Adverse Effects Questionnaire [
]. This questionnaire contains 11 items in which patients must score the intensity of specific symptoms (headache, neck ache, scalp pain, itching, tingling, burning sensation, redness on the skin, sleepiness, difficulty concentrating, sudden change of mood) during stimulation, on a rating scale of 1–4 (1, absent; 2, mild; 3, moderate; 4, severe). In addition, patients quantified how their experienced adverse effects were related to tDCS on a Likert-type items (1, none; 2, remote; 3, possible; 4, probable; 5, definite). They were also asked about the presence of unexpected adverse effects. The parameters considered for analysis were the occurrence, intensity, and perceived relationship of adverse effects with tDCS. Patients were also questioned whether they believed the stimulation was real or fictitious.
2.6 Mood assessment
Depressive and anxious symptoms were assessed before and after the intervention through the Brazilian-validated versions of the Beck Depression Inventory-II (BDI-II) [
] respectively. Both BDI-II and BAI were composed of 21 questions; each item was scored from 0 to 3.
2.7 Heart rate variability (HRV)
In the present study, indicators of cardiac autonomic modulation were defined through parameters of HRV. A heart rate monitor (Polar® RS800, Polar Electro Oy Inc., Kempele, Finland) was secured with a strap around the patient's chest and positioned at the height of the xiphoid process. A conductive gel (Mercur, Rio Grande do Sul, Brasil) was placed in the electrode areas. RR intervals were recorded at a sampling frequency of 1000Hz for six minutes before and after the tDCS intervention. The patients remained at rest during and at least 15 min before data acquisition. The patient's room was also controlled regarding any external disturbances, like noise or light changes [
]. One investigator remained outside the room to prevent anyone from entering during the acquisition. In case of external disturbances, the HRV record was restarted.
Signal processing was undertaken by the Kubious HRV software (Kubios Oy, Kuopio, Finland) [
] after excluding the first minute of data collected in both resting periods (before and after intervention). After visual inspection of the time series of RR interval (tachogram), HRV data from four patients (three sham and one active-tDCS) were excluded due to low-quality signals. The HRV parameters computed followed the recommendations in HRV literature for short-term recordings [
Rate variability, of measurement, physiological interpretation and clinical use, American TF of TES of C and TN, Electrophysiology S of P and. Guidelines Heart rate variability.
]. Temporal parameters included the standard deviation of the NN intervals (SDNN), and the root mean square of successive differences between RR intervals (RMSSD). Power spectral analysis of RR interval variability was performed using fast Fourier transformation and quantified into the following standard frequency-domain measurements: very-low-frequency (VLF; frequency range: ≤0.04Hz), low-frequency power (LF; frequency range: 0.04–0.15Hz) and high-frequency power (HF; frequency range: 0.15–0.40Hz), recommended by “Task force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology” [
Rate variability, of measurement, physiological interpretation and clinical use, American TF of TES of C and TN, Electrophysiology S of P and. Guidelines Heart rate variability.
]. LF and HF powers were also computed in normalized units (LFnu and HFnu), which corresponds to the relative value of each power component (LF and HF) in proportion to the total power minus the VLF component [
]. The normalization process allows a more reliable comparison of frequency components across studies, making normalized parameters less varied by methodological differences and less sensitive to variations in total power [
Rate variability, of measurement, physiological interpretation and clinical use, American TF of TES of C and TN, Electrophysiology S of P and. Guidelines Heart rate variability.
]. The LF/HF ratio was also computed since it is usually considered to assess the sympathetic and parasympathetic modulation of the ANS and sympathovagal balance [
Rate variability, of measurement, physiological interpretation and clinical use, American TF of TES of C and TN, Electrophysiology S of P and. Guidelines Heart rate variability.
Parametric, inferential statistics were used when data followed a normal distribution (Kolmogorov-Smirnov test; p > 0.05), and the homogeneity of variance was confirmed (Levene's test; p > 0.05). Otherwise, non-parametric tests were applied. Differences between groups (sham and active-tDCS) related to adverse effects were analyzed using the Fishers exact test. The independent sample t-test was applied to test for group differences in age, symptoms' duration, and BDI-II score. Mann-Whitney U test was used to test for between-group differences related to disease severity, hospitalization length, the intensity of each adverse effect of tDCS, and BAI score. Missing HRV data (4 out of 40 patients) were replaced through the mean imputation method. Differences between assessments [change = post-pre intervention] between groups for HRV parameters, heart rate, respiratory rate, and oxygen saturation were assessed by one-way analysis of covariance (ANCOVA) [
]. Although the randomization process aims at balancing the distribution of patients' characteristics between treatment groups, considering the variability usually observed for HRV results during short-term recordings [
], we also performed an ANCOVA including HRV baseline values and age as covariates. The effect size was computed using the Hedges' g, and statistical power was performed through a post hoc power analysis [
]. Statistical analyses were done using MATLAB (Mathworks, USA) and the software STATISTICA 7.0 (StatSoft, USA). The significance level was set at p < 0.05.
3. Results
We screened 254 patients, of which 144 did not meet the eligibility criteria, 18 refused to participate, 20 were included in another trial, and 32 were either discharged from the hospital or transferred to other units. The study was conducted in 2021, and no SARS-CoV-2 vaccine was administered to the recruited patients. Forty patients were randomized into sham or active-tDCS groups (Fig. 1). There were no dropouts during the experimental protocol.
Fig. 1Flow diagram of the study giving an overview of the hospitalized patients included in the study and analyzed. The main inclusion criteria patients excluded did not meet (N = 144) were: age greater than 80 years; unstable vital signs; and negative COVID-19 diagnosis. Abbreviation: tDCS, transcranial direct current stimulation; HRV, heart rate variability.
All patients were hospitalized in a semi-intensive care unit due to COVID-19 diagnosis, determined by reverse-transcriptase-polymerase-chain-reaction (RT-PCR) positive for SARS-CoV-2 infection from nasopharyngeal or nasal sampling. They were evaluated within 4–23 days from the onset of symptoms, and in most cases (58 out of 60 patients), COVID-19 severity was classified as mild [
]. Statistical analysis showed no differences between groups for sex, age, disease severity, symptoms duration, and hospitalization length (n = 40; p > 0.49 for all cases; Table 1). Clinical data for each group are shown in Supplementary Table 1 and include comorbidities, clinical signs, and symptoms at hospital admission. Additionally, details concerning the reasons for hospitalization, the stage of the infection (defined as from the onset of symptoms), and medications (including names and dosages) are described in Supplementary Table 2.
Table 1Comparison of clinical profile and mood between sham and active-tDCS group.
Data are expressed as mean (standard deviation). Abbreviation: tDCS, transcranial direct current stimulation; BDI-II, Beck Depression Inventory-II; BAI, Beck Anxiety Inventory.
Table 1 reports results for mood assessments, heart rate, respiratory rate, and oxygen saturation. Positive changes were observed for oxygen saturation after the neuromodulation in the active-tDCS group with respect to negative changes in the sham group (Fig. 2A). No differences were found between groups for changes in BDI-II and BAI scores, heart rate, and respiratory rate (p > 0.05 for all cases).
Fig. 2Changes in oxygen saturation (A) and HRV normalized parameters, HFnu (B) and LFnu (C) from baseline to after intervention within sham and active-tDCS groups. Data are expressed as mean and standard error of the mean. Light and dark gray bars represent results from sham and active-tDCS groups, respectively. The asterisk indicates a statistically significant (p < 0.05) difference between groups. Abbreviations: tDCS, transcranial direct current stimulation; n.u., normalized units.
Table 2 shows the results for both ANCOVA analyses performed for HRV parameters. When only age was considered as a covariate, HRV normalized parameters were significantly different between assessments [post-pre intervention] among groups. HFnu, a parameter suggested to be related to the activation of the parasympathetic branch of the ANS, increased after bihemispheric prefrontal stimulation, whereas it decreased in the sham group after intervention. Likewise, LFnu, a parameter suggested to be associated with both ANS branches’ activity, decreased in the active-tDCS group and increased in the sham group. When the baseline was included as an additional covariate, there was a significant baseline effect and no group and age effects for changes in time [post-pre intervention] in the parameters LFnu and HFnu. However, there was a large effect size of tDCS intervention on these HRV frequency parameters (Table 2; Fig. 2B and C). No significant effects were found for changes in SDNN, RMSSD, absolute values of LF and HF, and LH/HF ratio (p > 0.05 for all cases).
Table 2Comparison of heart rate variability between sham and active-tDCS group.
Outcomes, Post-Pre tDCS
sham-tDCS
active-tDCS
Hedges' g
F value (p-value)
F value (p-value)
Group factor
Age factor
Group factor
Age factor
Baseline factor
SDNN [ms]
1.93(8.01)
1.08(4.60)
0.130
0.11(0.741)
1.66(0.206)
0.23(0.632)
2.52(0.122)
1.61(0.213)
RMSSD [ms]
1.61(10.93)
1.87(6.35)
0.029
0.02(0.883)
1.22(0.276)
0.001(0.974)
1.40(0.245)
0.28(0.598)
LF [ms2]
45.38(146.73)
−2.94(79.82)
0.407
1.43(0.240)
0.56(0.458)
1.06(0.311)
1.46(0.235)
3.03(0.091)
HF [ms2]
55.92(328.95)
21.12(54.84)
0.149
0.18(0.675)
0.24(0.627)
0.20(0.656)
0.26(0.612)
0.03(0.857)
LFnu
5.54(9.53)
−1.57(9.93)
0.714
4.85(0.034)*
0.74(0.395)
1.47(0.233)
2.48(0.125)
18.89(0.000)*
HFnu
−5.41(9.34)
1.58(9.95)
0.707
4.76(0.036)*
0.72(0.401)
1.43(0.241)
2.40(0.131)
18.39(0.000)*
Ratio LF/HF
2.31(9.09)
−1.43(6.43)
0.470
2.00(0.166)
0.00(0.973)
2.24(0.145)
0.00(0.980)
0.51(0.478)
Data are expressed as mean (standard deviation). Abbreviations: tDCS, transcranial direct current stimulation; SDNN, standard deviation of normal to normal NN intervals; RMSSD, root mean square of successive differences between RR intervals; LF, low-frequeny component; HF, high-frequency component; nu, normalized unit;
Based on the effect size estimated from our data, a single tDCS session ensured 70% statistical power. Supplementary Table 3 shows group means for each outcome evaluated before (pre-tDCS) and after (pos-tDCS) the intervention, with respective coefficients of variation and confidence intervals.
The protocol was safe. No skin lesions, fall events, transfers to the ICU, orotracheal intubations, or deaths were found for any included patients up to one hour after the intervention. There were no statistical differences between groups for both occurrence and intensity of all the symptoms evaluated using the Brunoni tDCS Adverse Effects Questionnaire (p > 0.34 for all cases; Fig. 3). Furthermore, no patient reported severe symptoms. For the sham-tDCS group, headache and sleepiness were the most reported symptoms (20% for each symptom), while for the active-tDCS group, tingling was the most frequent symptom (30%). Other symptoms were presented in a maximum of 15% (n = 3) of patients per group. In addition, even when comparing the occurrence of adverse effects among groups separately for each intensity level (i.e., absence, mild, moderate), we still did not find statistical differences between groups for any reported symptoms (p > 0.34 for all cases).
Fig. 3Summary of patients' reported adverse effects as frequency and intensity. Abbreviation: tDCS, Direct Current Transcranial Stimulation.
Regarding the perceived relationship of adverse effects with tDCS (Table 3), the symptoms itching and burning sensation were among those with higher median scores for all groups. Finally, most patients in the sham-tDCS group (17 out of 20) considered stimulation was real. Regarding patients in active tDCS groups, 18 responded correctly that stimulation was real, while one patient said it was fictitious. Four patients (N sham-tDCS = 3; N active-tDCS = 1) answered they did not know.
Table 3Perceived relationship of adverse effects reported by patients with tDCS.
Adverse Effect
sham-tDCS
active-tDCS
Headache
3.5(2–5)
4(4–4)
Neck pain
4(1–4)
1(1–1)
Scalp pain
5(5–5)
–
Itching
5(5–5)
4(3–5)
Tingling
4(2–4)
3(2–5)
Burning sensation
5(5–5)
4(3–5)
Sleepiness
2(1–4)
1(1–2)
Trouble concentrating
2(2–2)
1(1–1)
Spatial disorientation
–
2(2–2)
Abbreviation: tDCS, transcranial direct current stimulation. Data are expressed as median (minimum-maximum).
In the present study, we explored the efficacy, safety, and feasibility of a single session of bihemispheric tDCS over the DLPFC on indicators of cardiac autonomic regulation assessed through HRV and mood of COVID-19 inpatients. Patients were randomized into two groups, sham-tDCS and active-tDCS. Our findings showed the intervention was safe and feasible since all patients completed the proposed protocol, and there were no mood changes, clinical worsening, presence of skin lesions, or fall events associated with the stimulation session. Furthermore, there were no differences between groups regarding the occurrence and intensity of symptoms investigated with the Brunoni tDCS Adverse Effects Questionnaire. An improvement of oxygen saturation was observed in COVID-19 patients who received active stimulation. There was a large effect size of intervention on normalized spectral HRV indexes.
4.1 Safe and feasible bihemispheric prefrontal tDCS in hospitalized COVID-19 patients
Despite the absence of large-scale studies with tDCS in hospitalized patients and the novelty of treating patients with COVID-19, methodological issues were carefully considered to ensure a stimulation protocol as safest and tolerable as possible. The current intensity (2mA) and electrodes’ size (35cm2) were chosen in accordance with previous studies that applied tDCS protocols over the prefrontal cortex and did not report serious adverse effects [
Cognitive control dysfunction in emotion dysregulation and psychopathology of major depression (MD): evidence from transcranial brain stimulation of the dorsolateral prefrontal cortex (DLPFC).
Repeated stimulation of the dorsolateral-prefrontal cortex improves executive dysfunctions and craving in drug addiction: a randomized, double-blind, parallel-group study.
]. In addition, we used sponge electrodes dampened with saline (NaCl) solution as the method of conductivity with the scalp, frequently used in tDCS studies [
This study had no dropouts or interruptions throughout the intervention protocol and no occurrence of serious adverse effects, as discussed as follows. The intensity and frequency of adverse effects reported by the patients after the tDCS protocol were not group-dependent. Overall, the symptoms presented were classified as mild intensity and occurred at a similar rate between active and sham groups. No patient reported severe symptoms. Headache was mainly reported by patients in the sham group and may have been a symptom of the disease itself [
]. The symptom of tingling, reported primarily in the active stimulation group, is expected during transcutaneous electrical stimulations, likely due to the activation of specific cutaneous receptors [
]. Our findings agree with evidence reporting headache and tingling as common mild symptoms during tDCS protocols applied in healthy and pathological individuals [
]. Other adverse effects were observed in the abovementioned studies, e.g., pain, fatigue, nervousness, and discomfort. These symptoms, however, were not presented in our sample.
In addition to the adverse effects, aspects that could contraindicate the tDCS application, such as worsening health conditions, were considered. Those outcome measures are not commonly reported in tDCS safety studies, likely because most of the available protocols were performed on outpatients. Even if the present study included critically ill patients from a semi-intensive care unit, there were no episodes of falls, and no patient was transferred to ICU, intubated, or died within one hour after the intervention applied in our research.
4.2 Prefrontal tDCS induced HRV and oxygen saturation changes in COVID-19 patients
HRV results indicated a single bihemispheric prefrontal tDCS session may be a potential approach to modulate the cardiac autonomic responses of hospitalized patients with COVID-19. The literature well established that autonomic regulation is a key regulator of the HRV power spectrum. The high-frequency (HF) component of HRV is strongly associated with cardiovagal activity [
]. When we performed the ANCOVA analysis considering age as a covariate, our results showed an increase in the HFnu parameter and a decrease in the LFnu parameter after the intervention in the active-tDCS group, compared to the sham group. Nevertheless, the inclusion of baseline as a covariate in addition to age showed it had a significant effect on the changes in the parameters LFnu and HFnu. These findings indicate that large baseline differences for these parameters contributed to the group differences found when ANCOVA analysis considered only age as a covariate. Despite the lack of a statistically significant group effect, we would like to point out the large effect size of tDCS intervention on both HRV frequency parameters (Hedges’ g: LFnu = 0.714 and HFnu = 0.707 [
]). The other HRV parameters evaluated did not show statistical changes between groups or large effect sizes. By observing their coefficient of variation (CV) values in Supplementary Table 3, these data show a larger variability when compared to LFnu and HFnu, decreasing the probability of finding different responses between conditions. Moreover, the normalization process is less sensitive to variations in total power, allowing a more reliable comparison of frequency components [
Rate variability, of measurement, physiological interpretation and clinical use, American TF of TES of C and TN, Electrophysiology S of P and. Guidelines Heart rate variability.
]. Therefore, as observed in our findings, the consideration of baseline in the statistical model was fundamental to show the need for planning studies with larger samples to explore the effects of bihemispheric prefrontal tDCS on indicators of cardiac autonomic responses. The observed effects here can be useful for such planning.
Overall, the large effect size found for the normalized spectral parameters indicated that the bihemispheric dorsolateral prefrontal cortex (DLPFC) may be a promising tool to modulate the cardiac autonomic response of patients. The magnitude of the effect and the validity of the results may be more important than the statistical significance since they emphasize the clinical relevance of the treatment [
]. These results corroborate previous accounts that observed alterations in HRV parameters following a single session of tDCS over the DLPFC of healthy individuals [
]. There is evidence that the HFnu and LFnu changes found in the present study suggest an increase in patients’ vagal activity and attenuation of sympathetic activity [
Electrodiagnostic assessment of the autonomic nervous system: a consensus statement endorsed by the American autonomic society, American academy of neurology, and the international federation of clinical neurophysiology.
] would be suitable to further confirm such modulations in SNA. Lastly, BDI-II and BAI scores did not change after the intervention for both groups, suggesting HRV indexes results were not associated with mood changes [
The electrical field generated by the active-tDCS may have caused a direct stimulation of cortical neurons, modulating the excitability of brain regions associated with autonomic cardiovascular control. These findings corroborate a previous meta-analysis [
] that identified the DLPFC as an effective brain region in regulating HRV, likely due to functional connectivity between the DLPFC and deeper brain regions associated with autonomic cardiovascular control, such as the subgenual anterior cingulate cortex (sgACC) [
], may also have been a mechanism underpinning our findings. A tingling sensation suggests peripheral nerves, such as the trigeminal and occipital nerves, were likely stimulated in the patients evaluated [
]. The brain circuits associated with sensory information processing are still unknown in the tDCS field. The tDCS protocol we applied in our study does not allow us to discriminate the contribution of transcranial and/or transcutaneous mechanisms. Boekholdt et al. (2020) [
] suggest including specific control groups in future tDCS protocols, in which transcranial and transcutaneous mechanisms can be evaluated independently. The approaches proposed by the authors to investigate each mechanism are, respectively, blocking the peripherical stimulation through topical anesthetic creams and targeting peripheral sites more distant from the brain. However, more studies are required in this field.
Studies have also reported symptoms of autonomic dysfunction in the early and long stages of COVID-19 infection, such as orthostatic intolerance [
Sympathetic neural overdrive, aortic stiffening, endothelial dysfunction, and impaired exercise capacity in severe COVID-19 survivors: a mid-term study of cardiovascular sequelae.
]. The possible mechanisms underlying the sympathetic overactivity in COVID-19 may be the overproduction of pro-inflammatory cytokines, alterations in blood gases (e.g., hypoxia), emotional distress, and dysregulation of the renin–angiotensin–aldosterone system [
] reported reduced HRV in patients with diagnosis of chronic atrial fibrillation during COVID-19 acute phase, with respect to the period before the pandemic, suggesting a reduced cardiac parasympathetic tone. Therefore, the HRV results observed in our study prompt further investigation to verify whether bihemispheric prefrontal tDCS could be a potential therapeutic tool for restoring sympathovagal imbalance caused by this disease, preventing the development of dysautonomia and subsequent involvement of several organs [
]. They are among the most common symptoms reported in mid-term and long-term phases of COVID-19 (the second known as long COVID), reaching up to 40% of survivors and associated with risk of mortality [
]. Consequently, researchers have highlighted the necessity of developing intervention strategies to reduce the burden of such disorders due to COVID-19. Therefore, bihemispheric prefrontal tDCS can be a promising adjunctive therapy to modulate the SNA of COVID-19 survivors and help to treat neuropsychiatric manifestations in this population.
ANS modulation could also be a strategy to regulate exacerbated inflammatory responses to severe infections, such as COVID-19. The ANS is responsible for regulating inflammatory and immune responses throughout the modulation of vagus nerve activity, the main component of the parasympathetic nervous system [
]. The activation of vagal efferent fibers induces two distinct mechanisms of the immune response: the cholinergic anti-inflammatory pathways (ChAP) and the hypothalamic pituitary adrenal axis (HPAA). The ChAP suppresses pro-inflammatory cytokines through acetylcholine release induced by activation of T-cells in the spleen. The HPAA releases anti-inflammatory glucocorticoids via modulation of cells in adrenal glands [
]. Both mechanisms can be activated through vagus nerve stimulation, as previous studies reported reduced inflammation in distinct pathological conditions [
Transcutaneous vagal nerve stimulation blocks stress-induced activation of Interleukin-6 and interferon-γ in posttraumatic stress disorder: a double-blind, randomized, sham-controlled trial.
]. In COVID-19, encouraging results have been reported with decreases in levels of inflammatory markers following therapies with non-invasive vagus nerve stimulation [
Treatment of stage 3 COVID-19 with transcutaneous auricular vagus nerve stimulation drastically reduces interleukin-6 blood levels: a report on two cases.
]. If confirmed the hypothesis that active bihemispheric prefrontal tDCS can heighten vagal activation of COVID-19 patients, it would be a potential intervention to mitigate hyperinflammatory syndrome associated with COVID-19 and other diseases.
Our findings also showed an increase in oxygen saturation of patients in the active tDCS with respect to the sham group after the intervention, with a medium estimated effect size (Hedges’ g = 0.61 [
An emergency system for monitoring pulse oximetry, peak expiratory flow, and body temperature of patients with COVID-19 at home: development and preliminary application.
]. Although we do not have data to support this assumption, oxygen saturation and autonomic changes may be related. The sympathetic discharge has an important role in the development of pulmonary dysfunction. It may affect the pulmonary vascular bed through adrenergic receptors, leading to local inflammatory response due to activation of macrophages by catecholamines and subsequent release of pro-inflammatory cytokines [
], reducing the distribution of blood to peripheral alveoli and hindering a proper gas exchange. On the other hand, parasympathetic activity elicits pulmonary vasodilator responses under vasoconstriction tone [
], besides induces anti-inflammatory effects as abovementioned. Therefore, future studies should investigate whether bihemispheric prefrontal tDCS can indeed attenuate sympathetic activation and heightened parasympathetic activity, and whether these changes can contribute to reducing pulmonary vascular resistance and local inflammation, enhancing oxygen saturation. If this assumption is confirmed, bihemispheric prefrontal tDCS may be a helpful add-on therapy to re-establish sufficient oxygen supply, usually required in about 70% of COVID-19 inpatients [
Clinical course and outcomes of critically ill patients with SARS-CoV-2 pneumonia in Wuhan, China: a single-centered, retrospective, observational study.
The hypothesis discussed above corroborates studies that observed acute improvements (i.e., until 90 min after intervention) in respiratory symptoms (e.g., dyspnea) and levels of inflammatory markers after applying non-invasive vagus nerve stimulation therapy in patients with acute bronchoconstriction and primary Sjögren's syndrome, respectively [
]. The possibility of suppressing pro-inflammatory cytokines, restoring autonomic balance, and enhancing oxygen saturation through bihemispheric prefrontal tDCS therapy, may hinder clinical worsening during severe infections like COVID-19, which can result in acute respiratory distress syndrome (ARDS), multiple organ failure, and death [
]. Further research investigating the safety and efficacy of more tDCS sessions on modulation and inflammatory markers would be helpful to support the clinical benefits of such intervention.
4.3 Study limitations
The present study was a first approach to investigate the safety and feasibility of a single bihemispheric prefrontal tDCS session in hospitalized patients with COVID-19 and its acute effects on indicators of cardiac autonomic modulation and mood. A priori power analysis (i.e., sample size calculation) was not performed because neuromodulation studies in a hospital setting are rare, and to our knowledge, no previous work investigated bilateral prefrontal stimulation. However, based on the effect size estimated from our data, a single tDCS session ensured a relatively high (70%) statistical power for the outcomes evaluated.
Since this study was conducted during the pandemic period, with a high risk of SARS-CoV-2 virus transmission, a non-COVID-19 control group was not evaluated. Thus, it is not possible to infer whether the acute findings observed after the intervention were specific to COVID-19 disease. However, bihemispheric prefrontal tDCS showed to be safe and feasible in this population.
In addition to HRV, other tests should be used to thoroughly assess the efficacy of bihemispheric prefrontal tDCS on the autonomic function of COVID-19 patients, including evaluation of cardiovascular sympathetic adrenergic, cardiovagal, and sudomotor (sweating) function [
Electrodiagnostic assessment of the autonomic nervous system: a consensus statement endorsed by the American autonomic society, American academy of neurology, and the international federation of clinical neurophysiology.
]. Considering the clinical conditions of COVID-19 inpatients and the limitations of performing an experimental protocol in a hospital setting, a short-term HRV recording was suitable to investigate the safety and acute effects of bihemispheric prefrontal tDCS on indicators of autonomic modulation. We computed and interpreted the reliable HRV parameters for short-term recordings, as recommended in HRV literature [
Rate variability, of measurement, physiological interpretation and clinical use, American TF of TES of C and TN, Electrophysiology S of P and. Guidelines Heart rate variability.
]. However, for HRV analysis, long-term recordings are considered the reference standard for clinical evaluation. It has a higher predictive power and is expected to provide a more accurate reflection of the overall responsiveness of the cardiovascular system than short-term recordings [
Finally, the use of concomitant medications by the patients evaluated in this study can be a confounding factor. Although questionable, some medications like beta blockers and ACE inhibitors may affect HRV [
To understand the clinical benefits of bihemispheric prefrontal tDCS in COVID-19 survivors, further research with greater samples is needed to investigate the efficacy and long-term effects of a therapy with more sessions, including a control group, a complete assessment of autonomic function, and analyses of inflammatory markers.
5. Conclusions
Our findings are the first insights into the safety and feasibility of a single bihemispheric prefrontal tDCS session in COVID-19 inpatients and its acute effects on cardiac autonomic modulation and mood. Bihemispheric prefrontal tDCS was safe and feasible since no mood impairments, clinical worsening or fall events were observed in the patients after the intervention, and adverse effects were neither severe nor tDCS-dependent. An improvement in oxygen saturation was also observed after the stimulation. A large effect size was observed for HRV frequency indexes suggesting the intervention may have attenuated sympathetic and increased vagal activities.
Our findings also reinforced the importance of considering baseline measures when assessing intervention effects on studies outcomes, even in exploratory research with relatively small sample sizes. Further randomized clinical trials with greater sample sizes and more tDCS sessions, applying a thorough assessment of autonomic function and inflammatory biomarkers, can foster the investigation of bihemispheric prefrontal tDCS as a potential adjuvant therapeutic tool to help manage autonomic dysfunctions, mitigate inflammatory responses, and enhance clinical outcomes.
Funding
This work was supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq); the Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ); the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior(CAPES); and the Instituto D’Or de Pesquisa e Ensino (IDOR). TPP received stipends from CNPq and FAPERJ. EAF received stipends from CAPES (Finance Code 001). The funding sources had no involvement in study design, in the collection, analysis and interpretation of data, in the writing of the report, and in the decision to submit the article for publication.
CRediT authorship contribution statement
Talita P. Pinto: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Data curation, Writing – original draft. Jacqueline Cunha Inácio: Methodology, Investigation, Writing – original draft. Erivelton de Aguiar Ferreira: Methodology, Investigation, Writing – original draft. Arthur de Sá Ferreira: Formal analysis, Writing – review & editing. Felipe Kenji Sudo: Methodology, Writing – review & editing, Funding acquisition. Fernanda Tovar-Moll: Conceptualization, Resources, Writing – review & editing, Supervision, Funding acquisition. Erika Rodrigues: Conceptualization, Methodology, Resources, Writing – original draft, Supervision, Project administration, Funding acquisition.
Acknowledgments
We thank Dr. Thiago Lemos and Dr. Myriam Monteiro for their support during the research project development. We also thank Dr. Patricia Vigário and MSc Roberto Costa for their support in methodological procedures.
Appendix A. Supplementary data
The following is/are the supplementary data to this article.
Efficacy and neurophysiological predictors of treatment response of adjunct bifrontal transcranial direct current stimulation (tDCS) in treating unipolar and bipolar depression.
Rate variability, of measurement, physiological interpretation and clinical use, American TF of TES of C and TN, Electrophysiology S of P and. Guidelines Heart rate variability.
Cognitive control dysfunction in emotion dysregulation and psychopathology of major depression (MD): evidence from transcranial brain stimulation of the dorsolateral prefrontal cortex (DLPFC).
Repeated stimulation of the dorsolateral-prefrontal cortex improves executive dysfunctions and craving in drug addiction: a randomized, double-blind, parallel-group study.
Electrodiagnostic assessment of the autonomic nervous system: a consensus statement endorsed by the American autonomic society, American academy of neurology, and the international federation of clinical neurophysiology.
Sympathetic neural overdrive, aortic stiffening, endothelial dysfunction, and impaired exercise capacity in severe COVID-19 survivors: a mid-term study of cardiovascular sequelae.
Transcutaneous vagal nerve stimulation blocks stress-induced activation of Interleukin-6 and interferon-γ in posttraumatic stress disorder: a double-blind, randomized, sham-controlled trial.
Treatment of stage 3 COVID-19 with transcutaneous auricular vagus nerve stimulation drastically reduces interleukin-6 blood levels: a report on two cases.
An emergency system for monitoring pulse oximetry, peak expiratory flow, and body temperature of patients with COVID-19 at home: development and preliminary application.
Clinical course and outcomes of critically ill patients with SARS-CoV-2 pneumonia in Wuhan, China: a single-centered, retrospective, observational study.
01701-1&id=gr1.jpg){kind=link}
01701-1&id=gr2.jpg){kind=link}
01701-1&id=gr3.jpg){kind=link}