EEG Biofeedback of low beta band components: frequency-specific effects on variables of attention and event-related brain potentials
Introduction
Various parameters of the human electroencephalogram (EEG) can be brought under operant control by means of a training process involving the real-time display of ongoing changes in the EEG via an EEG biofeedback loop. The principal feasibility of learned self-regulation has been demonstrated for evoked potentials (EPs) (Rosenfeld et al., 1969), event-related potentials (ERPs) (Birbaumer et al., 1981), slow cortical potentials (SCPs) (Birbaumer, 1984, Hardman et al., 1997), and EEG frequency components (Kamiya, 1968), with the latter two being of particular interest due to their reported intrinsic clinical benefits. For instance, the operant modulation of positive and negative SCP shifts has been found to facilitate control over epileptic seizures (Birbaumer et al., 1991, Rockstroh et al., 1993), and has been employed to spectacular effect as a brain-computer communication device for totally paralysed patients (Birbaumer et al., 1999).
Arguably the best established clinical application of AC EEG frequency component training consists of the treatment of epilepsy through learned self-regulation of the 12–15 Hz sensorimotor rhythm (SMR) recorded from central scalp regions over sensorimotor cortex (for review see Sterman, 2000). While the term SMR originally referred exclusively to the occurrence of phasic EEG spindles, the term will here be used to cover both phasic and tonic 12–15 Hz activity over the sensorimotor strip. Trained enhancement of the SMR has been demonstrated to result in increased seizure thresholds in response to exposure to eliptogenic agents in monkeys (Sterman et al., 1978), and to lead to reduced seizure incidence in human epileptics (Sterman and Friar, 1972, Sterman et al., 1974, Sterman and MacDonald, 1978, Sterman and Shouse, 1980). SMR activity over sensorimotor cortex is probably generated through thalamocortical interactions during burst firing activity in ventrobasal thalamic relay nuclei (Harper and Sterman, 1972), associated with the suppression of somatosensory afferent gating (Howe and Sterman, 1972). In consideration of its effects on cortical excitability in epilepsy, it has been concluded that SMR neurofeedback training appears to facilitate thalamic inhibitory mechanisms (Sterman, 1996).
From the apparent impact of SMR training on sensorimotor excitation, Lubar and colleagues have extrapolated the application of SMR training to the treatment of hyperactivity disorder (HD) (Lubar and Shouse, 1976, Shouse and Lubar, 1979). Subsequently, the operant enhancement of SMR, trained concurrently with suppression of slower theta (4–8 Hz) components, has often been complemented with or supplanted by training of higher beta band components, such as beta1 (15–18 Hz) in the treatment of attention deficit disorder (ADD) and attention deficit hyperactivity disorder (ADHD) (Lubar and Lubar, 1984, Lubar et al., 1995, Linden et al., 1996). A stated assumption in the use of the SMR and beta1 protocols is that the former addresses problems of hyperactivity and impulse control, while the latter is held to alleviate symptoms of inattentiveness (Lubar and Lubar, 1984, Lubar, 1991, Othmer et al., 1999). However, while recent controlled studies of beta band neurofeedback have produced promising results in the treatment of ADHD (Rossiter and LaVaque, 1995, Fuchs et al., 2003, Monastra et al., 2002), no protocol-specific differential effects between SMR and beta1 training have been demonstrated.
The controlled assessment of specific cognitive and electrocortical effects from training self-regulation of these frequency bands would be of great value in order to provide an empirical rationale for their clinical application to specific symptoms, and for furthering an understanding of their etiology. In a recent study, Egner and Gruzelier (2001) have supplied the first systematic evidence for protocol-specific effects of beta band protocols. They showed that, in a group of healthy volunteers who were trained on both the SMR and beta1 protocols, significant changes in attention performance and event-related potentials (ERPs) could be predicted on the basis of the subjects’ individual neurofeedback learning rates, and that learning rates on the two protocols showed differential relations with dependent measure changes. Specifically, it was found that SMR learning was associated with commission error reduction, while beta1 learning displayed the opposite association. Learning on both protocols was positively correlated with increased target P300 ERPs in an auditory oddball task, which indexes integration of task-relevant stimuli in working memory (Donchin and Coles, 1988). These findings could be interpreted as supporting SMR's role in improving impulse control, and beta1 training as increasing impulsive response tendencies, while both protocols may be associated with improved integration of relevant environmental stimuli.
However, conclusions drawn from this study were limited by the fact that subjects were trained on both protocols, resulting in correlational analyses for distinguishing between the impact of the two protocols. Furthermore, possible practice effects on the attention task were not controlled for, and the attention test employed did not provoke a substantial amount of omission errors, preventing the authors drawing any conclusions regarding the beta band protocols’ effects on inattentiveness. The current experiment was devised to test for protocol-specific effects of SMR and beta1 neurofeedback by directly contrasting attention performance and target P300 ERPs between the protocols in an independent-groups design. On the basis of the previous findings (Egner and Gruzelier, 2001), it was hypothesized that SMR training would result in reduced commission errors and increased d′ scores and P300 amplitudes, while beta1 training would result in increased commission error incidence and P300 amplitudes.
In order to control for possible practice and motivational factors affecting the attention measures, a control group engaging in a training regime of equal duration and experimenter-contact was included in the study. As the investigation was carried out as part of a large-scale project at a music conservatoire (Royal College of Music, London), the control group was involved in an ‘Alexander technique’ training program. The Alexander technique refers to a system of kinaesthetic education aimed at avoiding excessive postural tension, and constitutes the most widely practised behavioural training in professional orchestral musicians (Watson and Valentine, 1987). This intervention was not expected to affect attention performance.
Section snippets
Subjects
Participants were 25 students (7 males, 18 females; mean age 21.7 years, SD 2.24) from the Royal College of Music (London). The subjects volunteered for participation, gave their informed consent, and the investigation received ethical approval from the Riverside Research Ethics Committee (ref. RREC 2224). The subjects did not receive monetary reward for participation. Participants were screened with a health-related questionnaire in order to exclude subjects with a history of mental or
TOVA
The hypothesized commission error changes in the beta1 and SMR groups, and d′ increase in the SMR group, were assessed by planned comparisons (paired t tests) at one-tailed P<0.05 levels. In order to determine differential effects of the neurofeedback protocols on variables for which no a priori hypotheses existed, 3×2 (group×time) mixed-subjects analyses of variance (ANOVAs) were applied, followed by post hoc comparisons assessing within-group changes for each group (paired t tests), and
Discussion
Of the hypothesized cognitive-behavioural effects of the SMR and beta1 neurofeedback protocols, the expected commission error decrease in the SMR group and commission error increase in the beta1 group were not confirmed by the current data. The expected d′ improvements in the SMR group on the other hand were confirmed in both the TOVA and the divided attention task. The expected target P300 increments subsequent to beta band training were found exclusively in the beta1 group, with no changes
Acknowledgments
This research was funded by the Leverhulme Trust and generously supported by the Royal College of Music.
References (43)
- et al.
EOG-correction of blinks with saccade coefficients: a test and revision of the aligned-artefact average solution
Clin Neurophysiol
(2000) - et al.
Frontal interhemispheric asymmetry: self-regulation and individual differences in humans
Neurosci Lett
(1997) - et al.
Cortical-subcortical EEG correlates of suppressed motor behavior during sleep and waking in the cat
Electroenceph clin Neurophysiol
(1972) - et al.
Cognitive and biological determinants of P300: an integrative review
Biol Psychol
(1995) - et al.
Cortical self-regulation in patients with epilepsies
Epilepsy Res
(1993) - et al.
Suppression of seizures in an epileptic following sensorimotor EEG feedback training
Electroenceph clin Neurophysiol
(1972) - et al.
Quantitative analysis of training, sleep EEG and clinical response to EEG operant conditioning in epileptics
Electroenceph clin Neurophysiol
(1980) - et al.
Effects of sensorimotor EEG feedback training on seizure susceptibility in the rhesus monkey
Exp Neurol
(1978) EEG sensorimotor rhythm biofeedback training: Some effects on the neurologic precursors of learning disabilities
Int J Psychophysiol
(1984)Brainwave signatures – an index reflective of the brain's functional neuroanatomy: further findings on the effect of sensorimotor rhythm biofeedback training on the neurologic precursors of learning disabilities
Int J Psychophysiol
(1985)
The effect of training distinct neurofeedback protocols on aspects of cognitive performance
Int J Psychophysiol
Operant control of slow cortical potentials: a tool in the investigation of the potentials’ meaning and its relation to attentional dysfunction
Biofeedback on event-related potentials of the brain
Int J Psychol
Clinical-psychological treatment of epileptic seizures: a controlled study
A spelling device for the paralysed
Nature
Is the P300 component a manifestation of context updating?
Behav Brain Sci
Electrocortical frequencies in hyperactive, learning-disabled, mixed, and normal children
Biol Psychiatry
Learned self-regulation of EEG frequency components affects attention and event-related brain potentials in humans
Neuroreport
Testing efficiency and independence of attentional networks
J Cogn Neurosci
Neurofeedback treatment for attention-deficit/hyperactivity disorder in children: a comparison with methylphenidate
Appl Psychophysiol Biofeedback
Signal detection theory and psychophysics
Cited by (318)
The effect of musk incense stick aroma inhalation on different features of electroencephalogram signals and working memory for use in neurofeedback training
2023, Biomedical Signal Processing and ControlA comparison between classical and new proposed feature selection methods for attention level recognition in disordered children
2022, Alexandria Engineering JournalEnhancing Overt and Covert Attention Using a Real-Time Neurofeedback Game With Consumer-Grade EEG
2024, IEEE Transactions on Cognitive and Developmental SystemsFeasibility of a novel neurofeedback system: a parallel randomized single-blinded pilot study
2023, Scientific Reports