Comparative efficacy of targeted structural patterns of electroencephalography neurofeedback in children with inattentive or combined attention deficit hyperactivity disorder

Abstract Objective To evaluate and compare the effects of three courses of different structural patterns of electroencephalography neurofeedback on predominantly inattentive attention deficit hyperactivity disorder (ADHD‐PI) and combined ADHD (ADHD‐CT). Methods Thirty‐eight ADHD‐PI and ADHD‐CT children were selected and completed three courses of different structural patterns of electroencephalography neurofeedback according to their ADHD type. Before and after each course, relative power value of electroencephalography, including θ, β, α, SMR and their ratios (θ/β, θ/α), and eighteen integrated visual and auditory continuous performance test (IVA/CPT) quotients were obtained and compared. Data were analyzed by SPSS software, and p < .05 was considered statistically significant. Results After one course, θ, three IVA/CPT quotients in both types and two comprehensive quotients in ADHD‐CT changed significantly (all p < .05). After two courses, θ/α, θ/β and five IVA/CPT quotients in both types, θ and α in ADHD‐PI, four comprehensive quotients, and four respond control quotients in ADHD‐CT varied significantly compared to before treatment and after one course (all p < .05). After three courses, α, β, θ, θ/α, θ/β and ten IVA/CPT quotients in both types changed significantly compared to before treatment and after one course (all p < .05). In addition, six IVA/CPT quotients in both types after three courses were significantly higher than those after two courses (all p < .05). Conclusion Different structural patterns of electroencephalography neurofeedback targeted for ADHD‐CT and ADHD‐PI were both effective and feasible. Three courses of EEG neurofeedback were most effective.


INTRODUCTION
Attention deficit hyperactivity disorder (ADHD) is one of the most common neurological and behavioral disorders during childhood (Wolraich et al., 2019). A recent study showed pooled worldwide ADHD prevalence of 7.2% among children (Thomas et al., 2015), with estimates from some community-based samples being somewhat higher at 8.7% to 15.5% (Rowland et al., 2015;Wolraich et al., 2014). Children with ADHD often show signs of distraction, excessive activity, poor impulse control, and poor overall control, which do not correspond to their age of learning or socialization (Wolraich et al., 2019). The personal learning and interpersonal contact abilities of ADHD patients are sometimes seriously affected, and patients can even develop cognitive, conduct, and emotional disorders (Baijot et al., 2017). In the clinic, ADHD is divided into the predominantly inattentive type (ADHD-PI, 45% of ADHD patients), the predominantly hyperactive/impulsive type (ADHD-HI, 21% of ADHD patients), and the combined type (ADHD-CT, 34% of ADHD patients) (Cueli et al., 2019;Valmiki et al., 2021).
Over the past few decades, different treatments and interventions aimed at inattention, hyperactivity, and impulsivity have been used.
Medications, including methylphenidate, amphetamine, atomoxetine, and guanfacine, have been widely used in the acute treatment of ADHD (Van Doren et al., 2019) and reported to be effective for some features in short-term (Lin et al., 2021); however, medications failed to improve mood, self-esteem, and social relationships (Harpin et al., 2016). Furthermore, the compliance of ADHD patients generally decreases after one year's medications (Frank et al., 2015). In addition, some psychosocial treatments for ADHD children and adolescents, including behavioral therapy (parent training, classroom management, peer interventions), and cognitive training have been demonstrated to be effective (Chan et al., 2016;Che et al., 2021;Evans et al., 2018;Sonuga-Barke et al., 2013).
Besides medications and psychosocial treatments, neurofeedback is also considered as a valid option (Arns et al., 2020;Enriquez-Geppert et al., 2019). This treatment method can avoid side effects of drugs and has already gained popularity in recent years (Arns et al., 2014). As a non-invasive and relatively new approach for treating brain-related conditions, electroencephalography (EEG) neurofeedback can normalize disrupted brain waves that associate with ADHD by means of repeated training based on largely operant conditioning (Razoki, 2018).

EEG neurofeedback
The BioNeuro biofeedback instrument from Canadian Thought Technology was used for ADHD treatment, and it measured changes in the relative power of EEG. According to our previous study (Shi et al., 2012) and the internationally accepted 10-20 Electrode Placement System, Cz (positive electrode), right lobe A2 (negative electrode), and left lobe A1 (reference electrode) were selected as the most stable EEG detection areas. Participants were required to wash their hair before testing; Cz, A1, and A2 areas needed to be wiped with alcohol; the head sebum needed to be removed with scrubbing, and the conductive cream needed to be applied to the Cz area before the electrodes were placed.
The test room was indoors, quiet, and with soft lighting. The use of psychostimulant drugs and stimulant drinks were banned 48 h before treatment, respectively. Firstly, subjects were accompanied by a pretest instructor and remained in the test room for 10 min to familiarize the surroundings and reduce their tension and anxiety. Thereafter, subjects were instructed to close their eyes and relax for 1 min.
Researchers measured the EEG baseline for 3 min, and the treatment threshold was set. According to the protocol, the treatment commenced and researchers guided children to strive for and image the target, and children would be rewarded when the set target was achieved. After treatment, the EEG baseline was measured for 3 min again.

Integrated visual and auditory continuous performance test (IVA/CPT)
According to the literature (Corbett & Constantine, 2006;Kim et al., 2015), IVA/CPT quotients (Braintrain, USA) were adopted to reflect changes in cognitive ability, including attention and respond control, attention and respond control ability quotients were basic, and comprehensive quotients were calculated from basic quotients. The basic and comprehensive quotients had different emphases, that is, the former quotients were single correlation quotients that reflected a single skill, whereas the latter quotients comprehensively evaluated the control ability of subjects.

Quality control
Our study was carried out by a well-trained research staff along with doctors, and they were trained before the study. The training content included purpose and procedures of the study and how to implement the tests. The instruments used in the trial were standardized by the local technical supervision bureau. Patients were diagnosed and classified by two physicians with the rank of associate chief or above.
Before treatment, subjects were given training guidance to master the principles of brain waves and the target task of neurofeedback. A health education program was also administered to parents of subjects.
The purpose and significance of treatment were clarified three times so that they could supervise and urge subjects to cooperate during the treatment.

Statistical analysis
All data were collected and analyzed by use of SPSS software (version 22.0 SPSS Inc., Chicago, IL, USA). Numerical variables were expressed as mean ± standard deviation. Age, IQ, and the sex ratio of children between ADHD-PI and ADHD-CT groups were compared by the independent sample t test and Chi-square test, respectively. The changes or differences in IVA/CPT quotients and brain waves among each treatment course or between two groups were analyzed by repeated measures analysis (general linear model). The Greenhouse-Geisser corrected F test (calibration degree of freedom ranged from 1 to 3) was used when the data did not conform to Mauchly's test of sphericity.
The Bonferroni test was used for comparisons of the confidence interval of main effect, and the F test was used for interactions between two groups and indexes among each course of treatment. All tests were two-tailed, and p < .05 was considered statistically significant.

Age, gender, and IQ of subjects
Before treatment, age, gender and IQ of children between two groups were balanced and comparable. All subjects were aged from 6.02 to 11.78 years, with a mean of 8.29 ± 1.46 years, and there was no significant difference between ADHD-PI (8.06 ± 1.28) and ADHD-CT (8.36 groups.

F I G U R E 1
Relative power and ratio of brain waves in both types before and after each course

The relative power and ratio of brain waves in the two groups changed significantly after three treatment courses
The interactions between different ADHD types and treatment courses had no significant effects on θ, α, β, SMR, θ/α, and θ/β (all p > .05, Figure 1). Therefore, it was necessary to interpret main effects of treatment course.
There were significant differences in main effects of treatment course in the relative power and ratio of brain waves in both types (all p < .05, Table 1).
After one treatment course of EEG neurofeedback, θ in both types decreased significantly (all p < .05).
Compared to before treatment and after one treatment course, θ/α and θ/β in both types, θ in ADHD-CT decreased significantly, whereas α in ADHD-PI increased after two treatment courses of EEG neurofeedback (all p < .05). Similarly, after two treatment courses of EEG neurofeedback, θ in ADHD-PI decreased significantly and α in ADHD-CT increased significantly compared to before treatment (all p < .01). The data are presented in Table 1.
After three treatment courses of EEG neurofeedback, compared to before treatment and after one treatment course, θ/α, θ/β, and θ in both types decreased significantly, whereas α and β in both types increased (all p < .05). In addition, SMR in ADHD-PI was significantly higher than that before treatment (all p < .05, Table 1). However, there were no significant differences in the relative power and ratio of brain waves in both types between two and three treatment courses.

Most IVA/CPT quotients in the two types increased significantly after three treatment courses
The interactions between different ADHD types and treatment courses had no significant effect on all 18 IVA/CPT quotients (all p > .05, Figure 2). Therefore, it was necessary to interpret main effects of treatment course.
There were significant differences in the main effects of the treatment course in IVA/CPT quotients in both types (all p < .05, Table 2).
After one treatment course of EEG neurofeedback, VAQ, VIGA, FOCA, and FOCV in both types, as well as FAQ and AAQ in ADHD-CT increased significantly (all p < .05). The data are presented in Table 2. VAQ, VIGA, VIGV, FOCA, FOCV, SPDA, and SPDV in both types, as well as ARCQ, VRCQ, FAQ, AAQ, CONA, CONV, STAA, and STAV in ADHD-CT after two treatment courses of EEG neurofeedback were significantly higher than those before treatment and after one treatment course (all p < .05). PRUA and PRUV in ADHD-CT, as well as ARCQ, VRCQ, FAQ, AAQ, STAA, and STAV in ADHD-PI after two treatment courses of EEG neurofeedback were significantly higher than those before treatment (all p < .05). The data are presented in Table 2.
As shown in Table 2, FAQ, VAQ, VIGA, VIGV, FOCA, and FOCV in both types, as well as AAQ in ADHD-CT after three treatment courses of EEG neurofeedback were significantly higher than those after two treatment courses (all p < .05). FAQ, AAQ, VAQ, STAV, VIGA, VIGV, FOCA, FOCV, SPDA, and SPDV in both types, as well as ARCQ, VRCQ, PRUV, CONA, CONV, and STAA in ADHD-CT after three treatment courses of EEG neurofeedback were significantly higher than those TA B L E 1 Comparison of relative powers and ratios of brain waves before and after every treatment course in each type (x ± s)

DISCUSSION
In the connective process of brain function, neural resource allocation is different when attention is used for task control (Hale et al., 2014). Task execution would be interfered when the connectivity of brain function decreases, and then cognitive function is impaired (Silberstein et al., 2016 treatment time. In addition, θ/β and θ/α in both types showed a downward trend after two and three treatment courses. By contrast, α and β in both types increased significantly after two and three treatment courses. These results indicated that both structural patterns of EEG neurofeedback for ADHD-PI and ADHD-CT patients were effective. In addition, changes in EEG were observed after two treatment courses, and the curative effect showed an obviously increasing trend with the prolongation of treatment course, suggesting that at least two consecutive treatment courses are required so as to achieve relatively satisfactory therapeutic effects. CPT is applied to evaluate the continuous attention and reaction control of cognitive function, and it is basically free from subjective factors. Researchers have already demonstrated the feasibility of CPT as an assessment and diagnostic tool of ADHD (Gilbert et al., 2016).
IVA/CPT quotients quantify and standardize the core symptoms of ADHD. IVA/CPT quotients measure the sustained attention of visual and auditory response through random appearance of audio-visual mixed signals to identify the random response, impulse conflict, and fatigue status (Simões et al., 2017). These factors can be judged by quantitative scores (quotients). The integration of IVA/CPT quotients improved ADHD diagnosis and better reflected complexity and heterogeneity of ADHD (Berger et al., 2017). In this study, after two treatment courses, besides PRUA, PRUV, CONA, CONV in ADHD-PI and FRCQ in both types, other IVA/CPT quotients in both types increased significantly. In response to this, most symptoms of ADHD patients were relieved. After three treatment courses, in addition to CONA in ADHD-PI, FRCQ in both types also increased significantly. FAQ, VAQ, VIGA, VIGV, FOCA, FOCV in both types, and AAQ in ADHD-CT were significantly higher than those after two treatment courses. And along with this, the symptoms of ADHD patients improved further during the third treatment course. This change trend of IVA/CPT quotients with the increase of treatment course was consistent with the change rule of EEG neurofeedback, that is, the potential function of each brain wave needs to be stimulated and used rationally in mental applications and subsequently stabilized through long-term and multi-course training.
Therefore, compared with the results of previous studies, which mostly applied two treatment courses, our study further indicated the necessity for consistent three-course EEG neurofeedback in the treatment of ADHD.
Six attention quotients, namely vigilance, focus, and speed quotients of visual and auditory signals, were used to measure the errors of omission, sensitivity to change, and reaction speed during the test, respectively. Taken collectively, they reflect the attention problems associated with slow mental activity or the attention deficit of subjects. In this study, from the second treatment course of EEG neurofeedback, vigilance, focus, and speed quotients in both types increased unanimously.
Meanwhile, the three comprehensive quotients reflecting attention ability (FAQ, AAQ, and VAQ) in both types increased gradually. Namely, subjects' attention improved significantly after two treatment courses, and further improved and stabilized after three treatment courses.
The results also indicated that different structural patterns of the EEG neurofeedback targeted for different ADHD types could improve the attention retention ability and concentration of patients with different ADHD types, and the therapeutic effects are basically same.
Six control quotients, namely prudence, consistency and stamina quotients of visual and auditory signals, were used to measure the ability to stop, think, recognize, respond correctly to disturbances, maintain the response consistency over time, and remain stable throughout the test, respectively. They jointly reflect the overall coordination and volitional control abilities of the subjects. In ADHD-CT, prudence, consistency, and the stamina quotients of children increased significantly after two treatment courses, revealing that the ability to deal with disturbance and attention control of these children improved dramatically. And further improvement and stabilization of these two abilities were achieved after three treatment courses. In ADHD-PI, only the stamina quotients of children increased after two treatment courses. This was reasonable because the ability to deal with disturbance of external stimuli or information and attention control of ADHD-PI patients was still within normal range. And after each treatment course, the corresponding quotients reflecting these two abilities did not change significantly. Above results together suggested that different structural patterns of EEG neurofeedback targeted for characteristics of brain function of different ADHD types was necessary and feasible. In the other hand, three comprehensive quotients reflecting the respond control ability (FRCQ, ARCQ, and VRCQ) in both types improved, except these changes were more obvious in ADHD-CT. Taken collectively, these results indicate that different structural patterns of EEG neurofeedback targeted for different ADHD types adopted in this study can effectively improve the cognitive and executive functions of different types of ADHD patients. And the more treatment courses, the better treatment effect.
This study had some limitations. The sample size of this study was comparatively small and only met the basic statistical requirements.
The statistical efficiency would have improved if blind grouping was used and/or the sample size was increased appropriately. Furthermore, all subjects were from Jiaxing City. As a medium-sized city, although Jiaxing City has a good representative, there might have regional limitations. Therefore, further studies are needed to verify the results of this study.

CONCLUSION
In summary, the results of this study indicated that different structural patterns of EEG neurofeedback targeted for ADHD-CT and ADHD-PI were both effective and feasible, especially in cognitive and executive functions, including attention and respond control abilities. Two treatment courses of EEG neurofeedback had a good therapeutic effect in ADHD patients, but three treatment courses were most effective. In the future, this targeted and different structural pattern multi-course treatment of EEG neurofeedback deserves to be applied in the treatment of ADHD to achieve a better therapeutic effect.