Introduction

Transcranial electrical stimulation (tES) is rapidly emerging as a promising neuroscientific method that has garnered increasing interest in the fields of neuroscience and clinical research. This non-invasive approach involves the application of controlled electrical currents to the scalp, offering a versatile means of modulating brain activity and treat various neurological and neuropsychiatric disorders1,2.

To date, the most widely used tES technique is transcranial direct current stimulation (tDCS). tDCS is a polarity-dependent technique consisting of a continuous electrical current (usually ranging from 0.5 to 2 mA) delivered via two or more electrodes, typically comprising an anode and a cathode. In general, anodal tDCS increases the excitability of brain areas by inducing membrane potential depolarization, while cathodal tDCS exerts opposite effects, thereby inhibiting cortical excitability3,4,5,6,7. Studies have shown that exposure to tDCS results in a reduction in Gamma-Aminobutyric Acid (GABA) levels6,7,8,9, alongside elevations in Brain-Derived Neurotrophic Factor (BDNF) concentrations4. Consistently, tDCS has the capability to induce neuroplasticity aftereffects, initiating processes akin to long-term-potentiation via mechanisms dependent on Ca2 + and N-Methyl-D-Aspartate (NMDA) receptor-driven plasticity6,8.

Another tES technique that is recently gaining popularity is transcranial random noise stimulation (tRNS). tRNS is a polarity-independent technique that involves applying a low-intensity current to the scalp through at least two electrodes, with intensities fluctuating randomly (e.g., ± 0.5 mA), over a broad range of frequencies (from 0.1 to 640 Hz)10. tRNS operates by introducing noise to a signal that, on its own, is too weak to surpass a certain threshold. This phenomenon is known as stochastic resonance11. However, the precise neural-level mechanisms of action for tRNS are still a subject of ongoing debate. Emerging evidence suggests that tRNS may potentially enhance cortical plasticity resembling long-term potentiation by promoting the repetitive opening of sodium channels, thereby shortening the hyperpolarization phase12,13,14.

To date, studies related to the application of tES predominantly involve adults with or without psychiatric/neurological disorders15. Even if promising results have been arising for the treatment of specific disorders in paediatrics16,17, the percentage of tES studies conducted on children and adolescents represents only approximately 8% of the total18.

Despite being “non-invasive”, these techniques can still lead to Adverse Events (AEs). Of importance, ensuring the tolerability and safety of any experimental treatment is paramount to the successful transferability of experimental treatments from controlled research settings to real-world clinical applications. In this regard, single studies conducted on both adults and paediatrics seem to demonstrate the tolerability of these techniques8,15,19. Crucially, evaluating the occurrence of AEs in developing brains requires special attention, especially when considering children and adolescents with neuropsychiatric or neurodevelopmental disorders, whose brain development may differ from typical patterns.

In this regard, a review by Krishnan and colleagues20 first resumed the safety and tolerability outcomes of tDCS from 16 studies in approximately 203 children and adolescents with various neurological conditions. Results indicated that the AEs of tDCS were few, mild and transient.

One year later, Bikson and colleagues15demonstrated the safety and tolerability of tDCS in around 2800 sessions from 500 children with neurodevelopmental or acquired disorders, reporting no serious AEs. Furthermore, Zewdie and colleagues21,22collected data from 440 tDCS sessions in 44 children and adolescents with perinatal stroke outcomes and schizophrenia. They demonstrated the tolerability and safety of tDCS regardless of the active or sham condition, and the absence of serious AEs. Recently, a systematic review by Buchanan and colleagues22 evaluated the tolerability and safety of 1010 tDCS sessions in 137 children and adolescents with neurological, neuropsychiatric, or neurodevelopmental disorders across 11 included studies. Overall, findings suggested that tDCS appears to be safe and well-tolerated in youth (aged 5–17) with a clinical condition when administered within the usual parameters (e.g., dose: 0.5–2 mA, duration: 10–20 min, number of sessions: up to 20).

Despite evidence suggests an overall tolerability and safety of tES techniques in paediatrics23, their widespread usability, especially in clinical conditions, requires further investigation. Indeed, literature still presents some gaps that need to be addressed:

  1. i.

    Increasing evidence of tolerability and safety for tRNS, beyond tDCS. Systematically evaluated data mainly pertain to tDCS, despite the existence of emerging techniques like tRNS, which have been rarely assessed for tolerability and safety in paediatrics24,25,26,27,28;

  2. ii.

    The relation between demographic characteristics and the occurrence of AEs. The influence of age and gender in relation to the occurrence of AEs after tES sessions has been underexplored. Only two recent studies18,29compared the incidence of AEs among individuals with different ages, gender, and clinical conditions, finding no significant differences. Considering the influence of inter-individual factors on tES outcomes30, the exploration of this aspect in relation to tolerability and safety considerations results crucial;

  3. iii.

    The relation between emotional-behavioural symptoms and the occurrence of AEs. The influence of symptoms, such as baseline anxiety levels and the propensity to experience somatic complaints, have never been included in tES studies, despite being essential factors to consider from a preventive perspective;

  4. iv.

    Blinding. The effectiveness of blinding procedures and their relationship with the incidence of AEs has not been thoroughly explored.

With the aim of advancing knowledge regarding the tolerability, safety, and blinding efficacy of tES in paediatric clinical populations, the present study capitalized data from various experiments conducted in a single-centre, applying different tES techniques (i.e., tDCS and tRNS) in 92 children and adolescents with neuropsychiatric (i.e., anorexia nervosa) or neurodevelopmental disorders (i.e. attention deficit and hyperactivity disorder - ADHD, and developmental dyscalculia) for a total of 1032 sessions. The study will primarily focus on tolerability, defined as the incidence of unpleasant sensations (or AEs) related to tES sessions. Safety will be evaluated based on the occurrence of severe side effects that posed a risk to the participant’s physical and mental health.

Specifically, the present study will address the following Research Questions (RQs):

  1. i.

    What is the percentage of AEs reported by participants across all sessions they underwent (RQ1a)? Which types of AEs were reported (RQ1b)?

  2. ii.

    Are there differences in the frequency of AEs (considered as at least one AE in any of the sessions participants underwent) between tES conditions (active vs. sham; RQ2), type of tES (i.e., tDCS and tRNS; RQ3), demographic characteristics (i.e., gender, RQ4a; age, RQ4b), and emotional-behavioural symptoms (RQ5)?

  3. iii.

    To what extent was the blinding effective (RQ6)? Are there differences in the blinding efficacy based on tES techniques (i.e., tDCS and tRNS; RQ7)?

Materials and methods

Participants and procedures

Tolerability, safety, and blinding data from three different clinical trials (i.e., Experiment 1; Experiment 2; Experiment 3) applying tES in children and adolescents with neuropsychiatric and neurodevelopmental conditions were considered.

All the included studies were conducted at the I.Re.Ne Lab (Innovation and Rehabilitation in Neurodevelopment Lab) within the Child and Adolescent Neuropsychiatry Unit of Bambino Gesù Children’s Hospital in Rome. The studies were performed under the Declaration of Helsinki. Ethical approval for each study was granted by the Ethics Committee of Bambino Gesù Children’s Hospital, and all the studies were registered on ClinicalTrials.gov. All participants and their families were fully informed about the procedures and the purposes of the experiments, and signed a written informed consent prior to their participation.

Before administering any form of tES, all participants filled a tES screening designed to identify participants with noted contraindications (e.g., metal in the head, cardiac pacemaker, pregnancy) or those at increased risk of complications (e.g., unstable epilepsy, history of syncope, neurological disorders).

The Experiment 1 (ClinicalTrials.gov ID: NCT05674266) is a randomized, double blind, sham-controlled trial aimed to investigate the efficacy of tDCS treatment on clinical improvement in adolescents with anorexia nervosa. The study began on July 2019 and is still in the recruitment phase. Therefore, the results have not yet been published so far while the protocol study has been recently published31.

The Experiment 2 (ClinicalTrials.gov ID: NCT04964427) was a randomized sham-controlled trial aimed to assess the online effects of a single session of anodal tDCS, a single session of sham tDCS, and a single dose of Methylphenidate on inhibitory control and working memory in children and adolescents with ADHD. The study began in March 2021 and ended in May 2022. The results have been already published32.

The Experiment 3 (ClinicalTrials.gov ID: NCT04242680) is a randomized, double-blind, sham-controlled clinical trial aimed to explore the combined effects of tRNS and cognitive training on numerical cognition in participants with developmental dyscalculia. The study began on September 2019 and is still in the recruitment phase. The results have not yet been published so far while the protocol study has been already published33.

A detailed description of the procedures, eligibility criteria, and interventions of each experiment can be found in Supplementary Materials (S1. Methods).

A total of 92 children and adolescents were enrolled (gender: 54 females; age in years: M = 12.69, SD = 2.39, age range: 8–17 years; non-verbal Intelligence Quotient, nvIQ: M = 111.67, SD = 10.33).

On average, each participant completed approximately 10 sessions, with a range from 2 to 18, as following. Specifically, participants with anorexia nervosa underwent 342 active and 378 sham tDCS sessions (for a total of 720 sessions); participants with developmental dyscalculia underwent 160 active and 100 sham tRNS sessions (for a total of 260 sessions); participants with ADHD underwent 26 active and 26 sham tDCS sessions (for a total of 52 sessions).

Out of 1032 sessions, 528 sessions were conducted in a blinded active mode and 504 were in a blinded sham mode. See Table 1 for a comprehensive overview of the studies’ characteristics.

Table 1 Characteristics of tES studies.
Full size table

Interventions

Across all conditions and techniques, in each experiment, stimulation was delivered by a battery-operated stimulator (Brain-Stim stimulation by E.M.S. S.R.L-Bologna, Italy) using rectangular (i.e., tDCS experiments) or circular (i.e., tRNS experiment) 25 cm2 identical, saline-soaked, sponge electrodes. To minimize irritation, a small amount (1/8 of an inch) of gel cream was applied to the sponge surfaces. Moreover, electrode impedance was checked before starting each stimulation session, ensuring it remained below 10 kΩ. Consistent with previous paediatric tES studies34,35,36,37, for each experiment, the stimulation duration did not exceed 20 min, and the current intensity did not exceed 1 mA (i.e., 1 mA for tDCS experiments; ±0.75 mA for tRNS experiment). Both participants and their families were kept blinded to the condition (i.e., active or sham) of tES sessions.

In both sham tDCS and tRNS sessions, participants underwent the same procedures as in the active conditions, including identical electrode placement and an initial 30-second period of stimulation. The current was then ramped down to 0 mA without the participant’s awareness, ensuring blinding to the treatment condition. Both participants and their families were kept unaware of whether the sessions were sham or active, effectively controlling for placebo effects and maintaining the integrity of the study.

For further methodological details regarding the included studies, see Supplementary Materials (S1. Methods).

Concomitant activities

In relation to Experiment 1, participants did not engage in any activities during the tDCS sessions. Conversely, in Experiment 2, participants underwent neuropsychological tasks from the 10th minute of stimulation. Lastly, in Experiment 3, participants were involved in cognitive training, concurrently with the initiation of stimulation. For more details see Supplementary Materials (S1. Methods).

Measures

Tolerability and safety questionnaires

In all the experiments (i.e., Experiment 1, Experiment 2, Experiment 3), tolerability and safety questionnaires38,39 were administrated immediately after the end of each stimulation session. In particular, in Experiment 1, the safety and tolerability has been evaluated for 18 sessions for each participant, in Experiment 2 for 2 sessions for each participant32 and in Experiment 3 for 10 sessions for each participant.

Participants were asked to report on any AEs experienced during the session. They were asked to report on severity of the AEs (4-point Likert scale: 0 = absent; 1 = mild; 2 = moderate; 3 = severe). Across each experiment, they were asked to report on pain (including headache, neck pain, scalp pain), itching, burning sensation, and AEs classified as unspecified symptoms (i.e., “other symptoms”, including tingling, metallic taste, feeling of warmth, sleepiness, concentration problems, severe mood changes, and fatigue). Symptoms such as metallic taste, severe mood changes, and fatigue have never been reported across sessions and were therefore excluded from the statistical analysis/results. Measures on skin redness/irritation were collected only in the Experiment 2 and Experiment 3, therefore, were not considered for the main statistical analysis/results. However, for transparency, measures reported on skin redness/irritation were provided in the Supplementary Materials (see S2. Statistical analysis and results).

Blinding assessment

For the two studies employing a between-subjects design (i.e., Experiment 1 and Experiment 3), the assessment of blinding efficacy occurred at the conclusion of the final follow-up meeting (i.e., participants were asked whether they belonged to the active tES group or to the sham tES group based on their current perception). For the within-subject design study, at the end of the second tES session, participants were asked in which of the two tES sessions they thought they had received active or sham stimulation.

Child behaviour checklist

Behavioural and emotional symptoms were evaluated using the parent-report questionnaire Child Behaviour Checklist – CBCL40. For all considered studies, the CBCL was administered only at baseline for clinical eligibility, before recruiting participants, to detect potential psychopathological comorbidities.

The CBCL’s hierarchical structure comprises 113 items and diverse scales, organized as follows:

Syndrome Scales (Anxious/Depressed, Withdrawn/Depressed, Somatic Complaints, Social Problems, Thought Problems, Attention Problems, Rule-Breaking Behaviour, and Aggressive Behaviour), Broad Band Scales (Internalizing Problems, Externalizing Problems, Total Problems);

DSM-Oriented Scales (Affective Problems, Anxiety Problems, Somatic Problems, ADHD Problems, Oppositional Defiant Problems, Conduct Problems), 2007-Other Scales (Sluggish Cognitive Tempo, Obsessive-Compulsive Problems, Post-traumatic Stress Problem). Raw scores were converted in T scores in accordance with the cut-off thresholds established by Achenbach and Rescorla40. T scores > 69 were categorized as clinically relevant, T scores between 65 and 69 were considered borderline, and T scores < 65 indicated nonclinical symptoms.

In the manuscript text, we provide a detailed examination of findings related to Anxiety Problems and Somatic Problems from the DSM-Oriented Scales, as these dimensions are likely to be more closely linked to complaints and perceived AEs during stimulation when compared to the other dimensions examined across the CBCL. However, for a comprehensive overview of CBCL results, we have also included the statistical analysis/results pertaining to the remaining scales.

Due to issues in questionnaire completion, the analyses included only CBCL data obtained from 88 out of 92 participants.

Statistical analyses

Concerning AEs reports, an AE occurrence was coded (0 = no; 1 = yes) if reported at any intensity or duration29. Data were presented as frequency and percentage.

For RQ1, a comprehensive overview of the occurrence of AEs across all sessions (i.e., 1032) was provided (RQ1a) as well as the type of AEs (RQ1b) reported.

For RQ2 and RQ3, the occurrence of AEs was considered as the percentage of tES sessions in which at least one AE was reported by each participant. This allows us to merge data from all the three examined experiments regardless of their methodological differences (i.e., different study designs). The type of AEs reported was considered as the percentage of tES sessions in which any type of AE was reported by each participant. Chi-square (χ2) contingency tables were run and Odds Ratios (OR) with confidence intervals at 95% (CI 95%) were reported to compare differences in the frequency of tES sessions in which at least one AE was experienced by participants, in relation to tES conditions (active vs. sham; RQ2) and type of tES (tDCS and tRNS; RQ3). Bonferroni’s correction (p-value = 0.05/4 χ2 = 0.0125) was applied for multiple comparisons.

For RQ4 and RQ5, the percentage of participants who did not report AEs (coded as 0) and the percentage of participants who reported at least one AE (coded as 1) across all sessions were considered spitting participants according to demographic characteristics [i.e., gender: males vs females, RQ4a; age: children (≤ 10 years old) vs. adolescents (≥ 11 years old according to Word Health Organization, 2018); RQ4b], and emotional-behavioural symptoms [participants with borderline/clinically relevant T-scores (≥ 65) vs.participants with nonclinical T-scores (< 65) in the CBCL scales, RQ5]. Chi-square (χ2) contingency tables were run and Odds Ratios (OR) with confidence intervals at 95% (CI 95%) were reported. Bonferroni’s correction (p-value = 0.05/19 χ2 = 0.0026) was applied for multiple comparisons.

For RQ6 and RQ7, referring to blinding efficacy, the percentage of participants who did not identify the experimental condition to which they were exposed (coded as 0), those who correctly identified the experimental condition to which they were exposed (coded as 1), and those who were unable to provide a response were considered (RQ6). Moreover, χ2 contingency tables were run and OR with confidence intervals at 95% (CI 95%) were reported to assess any differences between tES techniques (i.e., tDCS and tRNS; RQ7) in the frequency of participants who correctly identified or failed to identify the experimental condition to which they were exposed. A p-value ≤ 0.05 was considered statistically significant.

χ2 analyses and OR were computed using R Studio (version 2024.04.0, RStudio, Boston, MA, USA), with particular reliance on the Epitools package.

Results

Occurrence (RQ1a) and type of AEs (RQ1b)

No participant reported serious AEs in any experiment (i.e., Experiment 1, Experiment 2, Experiment 3) and conditions (i.e., active and sham).

No participant chose to discontinue the trial or was excluded due to tolerability and safety issues. Overall, only AEs classified as “mild” were reported.

Out of 1032 tES sessions, at least one mild AE occurred in 241 sessions (see Fig. 1, panel A). The most frequently AE was itching (169 sessions), followed by pain (56 sessions), and “other symptoms” (47 sessions), which were further divided into concentration problems (18 sessions), sleepiness (17 sessions), tingling (11 sessions), and a feeling of warmth (1 session). The least common AE was a burning sensation (24 sessions) (Fig. 1, panel B).

Fig. 1
figure 1

Occurrence and Type of AEs. (A) Percentage of tES sessions in which at least one AE or none was reported. (B) Type of symptoms experienced, expressed as a percentage of the sessions in which they were observed.

Full size image

Comparison of AEs between conditions: active vs. sham (RQ2)

Out of 528 active tES sessions, at least one AE occurred in 151 sessions (28.60%). Whereas out of 504 sham tES sessions, at least one AE occurred in 90 sessions (17.86%). A significant difference emerged between active vs. sham in the frequency of sessions in which at least one mild AE occurred, indicating that the active condition had a higher probability of experiencing at least one mild AE in any of the sessions participants underwent (χ²(1) = 16.62, OR = 1.84, CI 95%: 1.37–2.48, p = 0.000046). See Fig. 2, panel A.

For further details regarding the type of AEs reported, please refer to the Supplementary Materials (see S2. Statistical analysis and results).

Comparison of AEs between types of tES: tDCS vs. tRNS (RQ3)

Overall, out of 772 tDCS sessions, at least one AE occurred in 129 sessions (16.71%), whereas out of 260 tRNS sessions, at least one mild AE occurred in 112 sessions (43.08%). A significant difference emerged between tDCS sessions vs. tRNS sessions in the frequency of sessions in which at least one mild AE occurred, indicating that tDCS had a lower probability of inducing at least one AE across sessions (χ²(1) = 75.54, OR = 0.26, CI 95%: 0.19–0.36, p < 0.00001). See Fig. 2, panel B.

With regard to tDCS, out of 368 active sessions, at least one AE occurred in 79 sessions (21.47%). Whereas out of 404 sham sessions, at least one mild AE occurred in 50 sessions (12.38%). A significant difference emerged between active tDCS vs. sham tDCS in the frequency of sessions in which at least one mild AE occurred, indicating that the active tDCS condition had a higher probability of experiencing at least one AE across sessions (χ²(1) = 11.44, OR = 1.93, CI 95%: 1.32–2.85, p = 0.00072). See Fig. 2, panel C.

With regard to tRNS, out of 160 active sessions, at least one AE occurred in 72 sessions (45%). Whereas out of 100 sham sessions, at least one AE occurred in 40 sessions (40%). No significant difference emerged between active tRNS vs. sham tRNS in the frequency of sessions in which at least one AE occurred (χ²(1) = 0.63, OR = 1.23, CI 95%: 0.74–2.04, p = 0.43). See Fig. 2, panel D.

For further details regarding the type of AEs reported, please refer to the Supplementary Materials (see S2. Statistical analysis and results).

Fig. 2
figure 2

Comparison of AEs Across tES Conditions and Techniques. (A) Comparison of AEs occurrence between active and sham conditions. (B) Comparison of AEs occurrence between tDCS and tRNS. (C) Comparison of AEs occurrence between active tDCS and sham tDCS. (D) Comparison of AEs occurrence between active tRNS and sham tRNS.

Full size image

Comparison of AEs based on demographic characteristics (RQ4a and RQ4b)

Gender (RQ4a). Out of 38 males, 25 participants (65.79%) reported at least one mild AE in any of the sessions they underwent (regardless of the type of tES and conditions).

Whereas out of 54 females, 30 participants (55.55%) reported at least one mild AE in any of the sessions they underwent (regardless of the type of tES and conditions).

No difference emerged between males vs. females in the frequency of participants who reported at least one mild AE in any of the sessions they underwent (χ²(1) = 0.79, OR = 1.47, CI 95%: 0.62–3.60, p = 0.37), regardless of the type of tES and conditions. See Fig. 3, panel A.

Age (RQ4b). Out of 24 children (≤ 10 years old), at least one mild AE was experienced by 14 participants (58.33%) in any of the sessions participants underwent, regardless of the type of tES and conditions. Whereas out of 68 adolescents (≥ 11 years old), at least one mild AE was experienced by 41 participants (60.29%) in any of the sessions participants underwent, regardless of the type of tES and conditions. See Fig. 3, panel B.

No difference emerged between children vs. adolescents in the frequency of participants who reported at least one mild AE in any of the sessions participants underwent (χ²(1) = 0.03, OR = 0.92, CI 95%: 0.36–2.44, p = 0.87), regardless of the type of tES and conditions.

Fig. 3
figure 3

Comparison of AEs by Gender and Age. (A) Number and percentage of participants reporting at least one AE during a tES session, categorized by gender (male vs. female). (B) Number and percentage of participants reporting at least one AE during a tES session, categorized by age (≤ 10 years vs. ≥11 years).

Full size image

Comparison of AEs based on emotional/behavioural symptoms (RQ5)

Anxiety Problems. Out of 43 participants presenting borderline or clinically relevant T-scores in the Anxiety Problems scale, at least one mild AE was experienced by 25 participants (58.14%) across sessions, regardless of the applied tES and conditions. Whereas out of 45 participants presenting nonclinical T-scores in the Anxiety Problems scale, at least one mild AE was experienced by 26 participants (57.78%) in any of the sessions participants underwent, regardless of the type of tES and conditions. See Fig. 4, panel A.

No difference emerged between participants presenting or not borderline/clinically relevant T-scores in the Anxiety Problems scale emerged in the frequency of participants who reported at least one mild AE in any of the sessions participants underwent (χ²(1) = 0.001, OR = 1.01, CI 95%: 0.43–2.39, p = 0.97), regardless of the type of tES and conditions.

Somatic Problems. Out of 19 participants presenting borderline or clinically relevant T-scores in the Somatic Problems scale, at least one mild AE was experienced by 9 participants (47.37%), regardless of the type of tES and conditions. Whereas out of 69 participants presenting nonclinical T-scores in the Somatic Problems scale, at least one AE was experienced by 42 participants (60.87%), regardless of the type of tES and conditions. See Fig. 4, panel B.

No difference between participants presenting or not borderline/clinically relevant T-scores in the Somatic Problems scale emerged in the frequency of participants who reported at least one mild AE (χ²(1) = 1.11, OR = 0.58, CI 95%: 0.20–1.65, p = 0.29), regardless of the type of tES and conditions.

Fig. 4
figure 4

Comparison of AEs by Anxiety and Somatic Problems. (A) Number and percentage of participants reporting at least one AE during a tES session, categorized by Anxiety Problems scale score (borderline/clinically relevant vs. nonclinical T-scores). (B) Number and percentage of participants reporting at least one AE during a tES session, categorized by Somatic Problems scale score (borderline/clinically relevant vs. nonclinical T-scores).

Full size image

Similarly, analyses conducted on all other CBCL scales did not yield significant results (see Table 2).

Table 2 Incidence of AEs based on the remaining emotional-behavioural measures (CBCL scales).
Full size table

Efficacy of blinding procedures (RQ6)

Out of the 92 participants included in the study, 3 are still awaiting the final follow-up and have not yet provided a response regarding blinding.

Among the remaining 89 participants, 39 correctly identified their treatment condition (active or sham), 42 did not correctly identify their treatment condition, and 8 were unable to provide an answer. See Fig. 5, panel A.

Comparison of blinding procedures efficacy between tES techniques (RQ7)

Out of the 63 participants in the tDCS experiments who responded to the blinding question, 35 correctly identified their treatment condition (55.55%). Whereas out of the 26 participants in the tRNS group who responded to the blinding question, 7 correctly identified their treatment condition (26.92%). A significant difference emerged between participants in the tDCS group vs. tRNS group, indicating that the tDCS group had a higher probability of identifying their treatment condition (χ²(1) = 4.96, OR = 3.39, CI 95%: 1.25–9.21, p = 0.026). See Fig. 5, panel B.

For further comparisons related to blinding procedures, see Supplementary Materials (Fig. S1).

Fig. 5
figure 5

Blinding Efficacy and Comparison by tES Technique. (A) Percentage of participants who correctly guessed, incorrectly guessed, or were unsure of their assigned condition (active vs. sham). (B) Number and percentage of participants categorized by tES technique (tDCS vs. tRNS) who correctly guessed, incorrectly guessed, or were unsure of their assigned condition.

Full size image

Discussion

The current study represents one of the most extensive reports in paediatrics addressing the tolerability and safety of tDCS and tRNS.

We particularly examined the tolerability of these techniques, defined as the frequency and intensity of unpleasant sensations (or AEs) associated with the application of two different tES methods. Investigating these aspects, especially in clinical populations, is essential for evaluating the practical applicability of these techniques in paediatrics and for providing an indirect measure of their safety. Our findings endorse the safety of these methods, given that “moderate” or “severe” AEs were not reported across more than 1000 tES sessions in children and adolescents with different clinical conditions. Even considering the reported “mild” AEs, our results provide support for the use of tES in paediatric clinical populations, as approximately 77% of sessions had no recorded AEs. Regarding the type of symptoms experienced, the majority of AEs were related to itching (approximately 17% of sessions). Our findings align with prior studies where itching consistently emerged as one of the most frequently reported symptoms in paediatrics15,20,21,22,26and adults19,38.

When comparing the experimental conditions (i.e., active vs. sham), active sessions were found to be more likely to induce AEs compared to sham sessions (28.60% vs. 17.86%, respectively). Some previous studies24,26,28,41,42,43,44,45,46have statistically compared the incidence of AEs in active and sham sessions in paediatrics. Among these, only two studies have reported a higher incidence of AEs during active stimulation compared to sham stimulation24,43 while the others did not detect significant differences. Overall, previous studies have examined differences between active and sham stimulation in the incidence of individual symptoms (e.g., itching, burning sensation, pain, etc.), while our results considered the presence of AEs from a more macroscopic perspective, examining the number of sessions in which at least one AE was reported by the participant. Additionally, it is worth noting the difference in sample sizes between our study (i.e., 92 participants) and those mentioned above (i.e., ranging from a minimum of 11 to a maximum of 30 participants). These two aspects could explain, at least in part, the difference between our results and those of most previous studies.

When considering the different tES techniques included, it seems that tRNS is more likely to induce AEs regardless of the experimental condition (i.e., active or sham) compared to tDCS (43.08% vs. 16.71%, respectively). These results would be in contrast to previous neurophysiological47and behavioural39,48,49studies, which have indicated that tRNS presents higher skin perception thresholds and lower incidence of AEs than tDCS. However, when comparing the active and sham conditions within the context of tRNS, no significant differences in the incidence of AEs emerged (45% vs. 40%, respectively). Conversely, in tDCS, the active condition was significantly associated with a higher incidence of AEs compared to the sham (21.47% vs. 12.38%, respectively). Overall, in the present study, the comparison between tDCS and tRNS in terms of safety appears to be of limited utility in determining the actual higher or lower incidence of AEs in the two techniques, especially considering the differences in the experiments [i.e., (i) setup (respectively, elastic band vs. headset, square electrodes vs. round electrodes), (ii) treatment intensity (respectively, 2–3 times per week vs. 5 times per week), (iii) concurrent activities (respectively, none or partial vs. throughout the session), (iv) different clinical populations (respectively, anorexia nervosa and ADHD vs. developmental dyscalculia), (v) partially different stimulated brain regions (respectively, only prefrontal vs. prefrontal and parietal cortex), (vi) partially different experimental designs (respectively, between-subject and within-subject vs. only between-subject), and (vii) significantly different total number of sessions between the two techniques (respectively, 772 vs. 260 sessions)]. Given all the differences, we deemed it more beneficial and insightful to utilize the variances between active and sham conditions within the techniques, rather than comparing the two conditions across the different techniques. Indeed, previous studies have shown that differences in experimental methods can influence perceptibility and aspects related to the safety and tolerability of tES38,50,51,52,53,54. Furthermore, the lack of difference in the incidence of AEs between active and sham conditions in tRNS, which apparently has more AEs, could suggest that the higher incidence is due to those different experimental factors rather than specific functional characteristics of the technique (e.g., frequency spectrum, current distribution, etc.). The present study may serve as an initial step in comparing the two techniques over a considerable number of sessions. However, future studies should compare tDCS and tRNS using identical experimental paradigms to mitigate confounding factors.

In relation to the potential influence of demographic factors, our results did not indicate any significant differences between males and females (65.79% vs. 55.55%, respectively) or between children and adolescents (58.33% vs. 60.29%, respectively) in terms of AEs’ outcomes. Our findings align with the only two studies that compared these factors, finding no differences in the incidence of AEs18,29. The influence of stable factors such as gender and age on tES effects has been documented15,50,51,52,53,54,55,56, and contributed to the inter-subject variability that is often observed in tES studies30. Of note, especially regarding the age variable, the disparities in the number of children and adolescents may have prevented the emergence of significant differences. To identify possible predictors to tES response and develop tailored intervention protocols, further studies are needed to elucidate the influence of participants’ baseline characteristics, not only with regard to tES effects, but also concerning the safety and tolerability profile of these techniques.

One of the objectives of the present study was also to examine the potential influence of specific personal and emotional-behavioural variables as predisposing factors in the perception of discomfort associated with tES. Within this context, we focused on symptoms related to generalized anxiety and the inclination to express somatic complaints. Our findings did not yield significant differences in AEs between individuals falling within a borderline/clinical range and those within a nonclinical range for these factors (58.14% vs. 57.78% and 47.37% vs. 60.87%, respectively). As for age variable, the discrepancies in the number of participants under “clinical/borderline” and “nonclinical” in the somatic complaints scale may have hindered the emergence of significant differences. To the best of our knowledge, no study has ever included this kind of variables as predictive factors for a higher or lower incidence of AEs. However, from a preventive standpoint and with the aim of optimizing the target clinical populations, further studies considering the influence of additional personal characteristics are needed.

The present study aimed also to assess the efficacy of blinding procedures, documenting that approximately half of the respondents identified the experimental condition to which they belonged (i.e., 43.82%). In particular, a significant difference was observed between tDCS and tRNS in the ability to correctly guess the assigned condition (55.55% vs. 26.92%, respectively). This finding seems to align with previous studies showing that tRNS better maintains blinding integrity47. Although the effectiveness of blinding procedures in the context of tES in paediatrics has not been systematically examined, recent studies have cast doubts on the actual efficacy of blinding procedures using sham stimulation as a placebo57,58. Even in light of our results, it would be prudent for future studies to explore different and potentially more effective blinding methods.

The current study presents some limitations.

Concerning safety measures, one limit is the absence of objective measures such as neurophysiological data, as well as medical evaluations (e.g., heart rate and blood pressure) to integrate a safety assessment that, in our study, substantially relies on self-reported responses. Given that most studies, including the present one, have primarily focused on self-reported tolerability of tES, future research should prioritize the evaluation of safety through more objective measures. To date, beyond the absence of serious AEs, it remains unclear what objective safety indicators are applicable to these techniques in paediatric populations. Future research in tES will need to identify standardized parameters for assessing safety in addition to tolerability. Importantly, evaluating safety parameters is crucial for considering the potential future integration of these techniques into treatment practices. With regard to the assessment of blinding procedures, a limitation could be related to the different number of sessions among the 3 experiments (i.e., 18 vs. 10 vs. 2) and to the different experimental designs employed (i.e., 2 between-subjects vs. 1 within-subjects). These factors could have influenced the ability to discern whether the received stimulation was active or sham. In addition, the impossibility of controlling randomly responses must be taken into account, especially when dealing with paediatric participants.

In conclusion, our study provides a comprehensive examination of the tolerability and safety of tES in paediatric populations. The high level of tolerability and the absence of “moderate” or “severe” AEs in over 1000 tES sessions underscore the overall safety of these techniques in children and adolescents across diverse clinical conditions. The comparison of active and sham sessions, as well as tRNS and tDCS, reveals nuanced insights, challenging some previous findings and emphasizing the importance of considering experimental features and setup specifics. Notably, the study highlights potential factors influencing both AEs and blinding procedures, calling for further exploration and refinement in future research.