Elsevier

Neuroscience

Volume 478, 1 December 2021, Pages 24-38
Neuroscience

Research Article
Neurofeedback Training of the Control Network in Children Improves Brain Computer Interface Performance

https://doi.org/10.1016/j.neuroscience.2021.08.010Get rights and content

Highlights

  • Training for 2 months and the effect last for at least ten days.

  • Up-regulating parietal alpha-amplitude NFT significantly improves the attention of subjects.

  • Up-regulating parietal alpha-amplitude NFT significantly improves BCI accuracy, ITR, and SNR.

  • Compared with adults, neurofeedback training may performed better on children.

  • The brain can be divided into the control network and the functional network. They could be regulated independently.

Abstract

In the past 20 years, neural engineering has made unprecedented progress in the interpretation of brain information (e.g., brain-computer interfaces) and in neuromodulation (e.g., electromagnetic stimulation and neurofeedback). However, there has been little research aiming to improve the performance of brain-computer interfaces (BCIs) using neuromodulation. The present study presents a novel design for a neurofeedback training (NFT) method to improve the operation of a steady-state visual evoked potential (SSVEP)-based BCI and further explores its underlying mechanisms. The use of NFT to upregulate alpha-band power in the user’s parietal lobe is presented in this study as a new neuromodulation method to improve SSVEP-based BCI in this study. After users completed this NFT intervention, the signal-to-noise ratio (SNR), accuracy, and information transfer rate (ITR) of the SSVEP-based BCI were increased by 5.8%, 4.7%, and 15.6%, respectively. However, no improvement was observed in the control group in which the subjects did not participate in NFT. Moreover, a general reinforcement of the information flow from the parietal lobe to the occipital lobe was observed. Evidence from a network analysis and an attention test further indicates that NFT improves attention by developing the control capacity of the parietal lobe and then enhances the above SSVEP indicators. Upregulating the amplitude of parietal alpha oscillations using NFT significantly improves the SSVEP-based BCI performance by modulating the control network. The study validates an effective neuromodulation method and possibly contributes to explaining the function of the parietal lobe in the control network.

Introduction

The study of cortical oscillations through electroencephalography (EEG) has made significant progress. The effects and influences of various rhythms on people have been gradually explored since neural oscillations were first discovered by Berger (1929). However, the physiological mechanisms by which cortical oscillations affect human behavior are still not fully understood. Some studies have indicated that the roles of neural oscillations include feature binding, information transfer mechanisms, and the generation of rhythmic motion output (Klimesch, 1999, Buzsaki, 2004, Canolty et al., 2006, Buzsáki and Moser, 2013, Iaccarino, 2016). In recent decades, a few breakthroughs have been made in this field through electroencephalography and neuroimaging. It has been found that neural oscillations play an essential role in neural information processing. However, there is little experimental evidence regarding the function of neural oscillations thus far. It is unlikely that the role of neural oscillation can be explained based on the existing evidence alone. Further insights into neural oscillations and their relationship to cognitive processes are necessary. Furthermore, attempts should be made to regulate and control the specific characteristics of these oscillations to achieve purposeful functional changes.

Alpha-band oscillations are the dominant oscillations in the human brain. Although the physiological basis of these oscillations is still unclear, their origin is thought to be related to the thalamo-cortical circuitry (Lopes da Silva et al., 1973, Nunez et al., 2001, Nicolelis and Fanselow, 2002, Lőrincz et al., 2009, Hughes, et al., 2011, Schmid et al., 2012, Howells et al., 2018). An increasing body of recent research on alpha-band function has shown that the alpha band performs a significant inhibitory function in information processing by the brain. Previous work has suggested that information is routed by functionally blocking off task-irrelevant pathways (gating by inhibition). Importantly, this inhibition is reflected by oscillatory activity in the alpha band. The task dependent activation and inhibition of different regions form part of the mechanism of attention (Jensen and Mazaheri, 2010). Moreover, evidence has shown that alpha-band oscillations reflect cognitive and memory performance and state. Alpha-band oscillations enable voluntary orientation of attention and controlled knowledge access, since they are closely linked to the fundamental processes of attention regulation (suppression and selection) (Klimesch, 2012). The traditional argument regarding the role of alpha-band activity in attention is that alpha oscillations desynchronize in response to anticipatory attention and in the absence of stimuli. This type of event-related desynchronization (ERD) is termed a “prestimulus, anticipatory ERD”. A large prestimulus ERD is relevant to high detection performance (Ergenoglu et al., 2004, Hanslmayr et al., 2005a, Hanslmayr et al., 2005b) and reflects the maintenance of target information in an attentional buffer. Furthermore, an increase in the amplitude of event-related synchronization (ERS) reflects inhibition, whereas a decrease in the amplitude of ERS reflects release from inhibition. Examples have been found in tasks that vary the side on which stimulation is applied (e.g., stimulation of the right vs. left visual hemifield), the stimulus modality, or the stimulus processing domain (e.g., related to the ventral vs. dorsal processing stream). Findings show that when attention is focused on the auditory element of a compound auditory-visual stimulus, alpha oscillation is increased in the visual cortices (Foxe et al., 1998). When a task engages the ventral stream, alpha power is increased in parietal regions (Jokisch, 2007). Moreover, the ability of humans to respond to visual stimuli based on their spatial locations requires the parietal cortex (Corbetta, 2000) to direct visual attention. One theory maintains that the parietal cortex controls the autonomous orientation of attention toward a location of interest (Mesulam, 1981, Heilman et al., 1985, Halligan and Marshall, 1994). Another emphasizes its role in reorienting attention toward visual targets appearing at unattended locations (Posner et al., 1984, Morrow and Ratcliff, 1988, Di Pellegrino, 1995, Friedrich et al., 1998). All of the abovementioned studies suggest that the parietal lobe is closely related to attention to visual stimuli and plays a crucial role in control and monitoring.

Neurofeedback training (NFT) is an operant conditioning method that helps subjects to control or change their brain activity (Schafer and Moore, 2011). The training enables the voluntary modulation of neural oscillations through real-time feedback on the brain's electrical parameters. Neurofeedback training is regarded as a safer and more stable method than electromagnetic stimulation for controlling or changing neural oscillations. Since its inception, NFT has been applied in several scenarios (Zoefel et al., 2011), such as efforts to cure attention-deficit hyperactivity disorder (Arns et al., 2009, Sonuga-Barke et al., 2013), epilepsy (Zhao et al., 2009), and Huntington's disease (Papoutsi et al., 2017); to improve cognitive performance (Vernon et al., 2003, Hanslmayr et al., 2005a, Hanslmayr et al., 2005b); and to enhance self-regulation of emotional and behavioral control (Mayeli et al., 2019).

At present, NFT is also applied to improve the operation of various types of BCIs. Steady-state visual evoked potential (SSVEP)-based interfaces are a classical BCI paradigm that forms an essential part of many BCIs for various applications. SSVEP can be understood as phase-locked human brain activity that spontaneously responds to stimulation; the signal-to-noise ratio (SNR) is the measure most commonly used to evaluate the signal quality. Despite decades of research, the performance of SSVEP-based BCIs is still heavily relies on the specific subjects and algorithms (Neuper and Pfurtscheller, 2010). SSVEP signals and performance can be influenced by NFT to such an extent that Mayra, in a recent study on an SSVEP-based BCI algorithm, excluded subjects who had undergone any form of NFT before data collection (Mayra and Maurits, 2018). A study by Ordikhani-Seyedlar suggested that the power of the SSVEP signal was regulated by various degrees of attention, whether covert or overt (Ordikhani-Seyedlar et al., 2014). Additionally, Collura and Thomas designed experiments to demonstrate the relationship between periodic SSVEPs and short-term attention (Collura and Thomas, 2002). Moreover, their work pointed out that attention-sensitive SSVEPs in response to stimulation could serve as the basis for a type of neurofeedback and could likely be used in training to overcome attention deficits. Regarding the application of neurofeedback derived from SSVEPs, a study by Bruno has claimed that visual and auditory cognition can be improved by binaural audio and SSVEP neurofeedback (Bruno et al., 2017).

Morgan provided further direct evidence regarding the relationship between attention and SSVEPs. Specifically, he asked subjects to gaze at the center of a screen and showed blocks of flicker stimuli of different frequencies on both the left and right sides (Morgan et al., 1996). This attention-bias experiment found that subjects’ EEG signals reflected the frequency of stimulation on the side to which attention was being focused, thus demonstrating the correlation between attention and SSVEPs.

Based on the above descriptions, since alpha oscillation exerts an inhibitory physiological function and SSVEP-based BCI performance appears to require considerable attention, we hypothesized that improving the alpha amplitude in the parietal lobe can increase attention and the SNR of the SSVEP signal to enhance SSVEP performance.

Section snippets

Design

The flow chart in Fig. 1 shows the experimental design in detail. For each subject in the NFT group, the experiment consisted of twenty-four training sessions within two months. Participants were assigned to complete three sessions per week, with each session lasting an hour. In each session, subjects were asked to play an EEG feedback game. Each week, they were free to perform their three training sessions at any time as long as the sessions were separated by at least one day. To capture the

NFT results

We recruited twenty-four participants; each one completed 24 training sessions in two months (for a total training duration of 24 h). Among these participants, 91.7% were successfully trained (22 of 24). Successful training meant that a subject’s alpha-band power was higher after training than before. As shown in Fig. 7, the average alpha-band power ratios of the 24 subjects before and after training were 32.93 and 38.58, respectively. Thus, the training improved the alpha-band power ratio by

Discussion

As hypothesized, NFT succeeded in strengthening the regulation of parietal lobe activity. The results demonstrated that the training method in this study was reinforced the executive control network and thus enhanced the ability to complete the task. At the same time, the designed test scale also verified the effectiveness of the training in improving attention.

The alpha band provides active inhibition (Mathewson et al., 2009, Jensen and Mazaheri, 2010, Klimesch, 2012); conversely,

Ethics approval and consent to participate

A written consent form was completed by all participants’ guardians. This project was reviewed by the Ethics Committee of the Tsinghua University School of Medicine (Protocol number 2019006 approved on 3/4/2019).

Consent for publication

Consent to publish the results was obtained from the parent or legal guardian of each subject.

Availability of supporting data

The raw/processed data required to reproduce these findings cannot be shared at this time, as the data are also part of an ongoing study. We will consider publishing some of the essential data once the ongoing study is published.

Competing interests

The authors declare that they have no competing interests.

Funding

This work was supported by the National Key R&D Program of China under grant 2017YFB1002505 and by the National Natural Science Foundation of China under grant 61431007.

Authors’ contributions

J. Sun contributed to the entire research process, including the experimental design, data collection, result analysis and article writing. J. He participated in the data collection and helped to improve and polish the article. As the corresponding author, Prof. X. Gao guided the research and the writing of the article. He also critically revised the manuscript content and verified the quality of the research. All authors read and approved the final manuscript.

Acknowledgements

The authors would like to thank all of the participants who took part in this study, as well as the staff at the Beijing Jinbo Zhihui Education Technology Co., Ltd., and Beijing CUSoft Co., Ltd., for their support throughout this project. We would also like to thank laboratory colleagues Xiaoyang Li et al. for their help in data collection.

References (76)

  • T. Ros et al.

    Mind over chatter: plastic up-regulation of the fMRI salience network directly after EEG neurofeedback

    Neuroimage

    (2013)
  • M.C. Schmid

    Thalamic coordination of cortical communication

    Neuron

    (2012)
  • D. Vernon et al.

    The effect of training distinct neurofeedback protocols on aspects of cognitive performance

    Int J Psychophysiol

    (2003)
  • Z. Yan et al.

    Functional connectivity analysis of steady-state visual evoked potentials

    Neurosci Lett

    (2011)
  • L. Zhao et al.

    Changes in EEG measurements in intractable epilepsy patients with neurofeedback training

    Prog Natl Sci

    (2009)
  • B. Zoefel et al.

    Neurofeedback training of the upper alpha frequency band in EEG improves cognitive performance

    Neuroimage

    (2011)
  • V. Zotev et al.

    Self-regulation of human brain activity using simultaneous real-time fMRI and EEG neurofeedback

    NeuroImage

    (2014)
  • E. Angelakis et al.

    Eeg neurofeedback: a brief overview and an example of peak alpha frequency training for cognitive enhancement in the elderly

    Clin Neuropsychol

    (2007)
  • M. Arns et al.

    Efficacy of neurofeedback treatment in ADHD: the effects on inattention, impulsivity and hyperactivity: a meta-analysis

    Clin EEG Neurosci

    (2009)
  • M. Bauer

    Tactile spatial attention enhances gamma-band activity in somatosensory cortex and reduces low-frequency activity in parieto-occipital areas

    J Neurosci

    (2006)
  • H. Berger

    Uber das Elektrenkephalogramm des Menschen (On the EEG in human)

    Archiv fur Psychiatrie und Nervenkrankheiten

    (1929)
  • B.-V. Bruno et al.

    Neurofeedback training system with audiovisual stimuli for the attention state induction during cognitive processes

  • G. Buzsaki

    Neuronal oscillations in cortical networks

    Science

    (2004)
  • György Buzsáki et al.

    Memory, navigation and theta rhythm in the hippocampal-entorhinal system

    Nat Neurosci

    (2013)
  • R.T. Canolty et al.

    High gamma power is phase-locked to theta oscillations in human neocortex

    Science

    (2006)
  • X. Chen et al.

    A high-itr ssvep-based bci speller

    Brain-Comput Interfaces

    (2014)
  • X. Chen et al.

    High-speed spelling with a noninvasive brain-computer interface

    Proc Natl Acad Sci

    (2015)
  • X. Chen et al.

    Filter bank canonical correlation analysis for implementing a high-speed ssvep-based brain-computer interface

    J Neural Eng

    (2015)
  • Collura et al.

    Application of repetitive visual stimulation to EEG neurofeedback protocols

    J Neurother

    (2002)
  • M. Corbetta

    Voluntary orienting is dissociated from target detection in human posterior parietal cortex

    Nat Neurosci

    (2000)
  • A.K. Engel et al.

    Dynamic predictions: oscillations and synchrony in top-down processing

    Nat Rev Neurosci

    (2001)
  • E. Eqlimi et al.

    EEG correlates of learning from speech presented in environmental noise

    Front Psychol

    (2020)
  • J.J. Foxe et al.

    Parieto-occipital similar to 10 hz activity reflects anticipatory state of visual attention mechanisms

    NeuroReport

    (1998)
  • F.J. Friedrich et al.

    Spatial attention deficits in humans: a comparison of superior parietal and temporo-parietal junction lesions

    Neuropsychology

    (1998)
  • J.H. Gruzelier

    Eeg-neurofeedback for optimising performance. i: a review of cognitive and affective outcome in healthy participants

    Neurosci Biobehav Rev

    (2013)
  • P.W. Halligan et al.

    Toward a principled explanation of unilateral neglect. Special Issue: The cognitive neuropsychology of attention

    Cognit Neuropsychol

    (1994)
  • S. Hanslmayr et al.

    Increasing individual upper alpha power by neurofeedback improves cognitive performance in human subjects

    Appl Psychophysiol Biofeedback

    (2005)
  • Heilman, K. M., Watson, R. T., Valenstein, E. in Clinical Neuropsychology (eds. Heilman, K. M. & Valenstein, E.)...
  • View full text