Elsevier

Clinical Neurophysiology

Volume 129, Issue 9, September 2018, Pages 1884-1890
Clinical Neurophysiology

Do scalp-recorded slow potentials during neuro-feedback training reflect the cortical activity?

https://doi.org/10.1016/j.clinph.2018.05.015Get rights and content

Highlights

  • Slow potentials (SPs) from scalp and intracranial electrodes were simultaneously compared.

  • The coherence of SPs negatively correlated with the distance between subdural and scalp electrodes.

  • Neuro-feedback control by scalp SPs represent the cortical activity but not the galvanic skin response.

Abstract

Objective

Neuro-feedback (NFB) training by the self-regulation of slow potentials (SPs) <0.5 Hz recorded from the vertex scalp has been applied for seizure suppression in patients with epilepsy. However, SP is highly susceptible to artifact contamination, such as the galvanic skin response (GSR). This study aimed to evaluate the correlation between SPs recorded from the scalp and intracranial electroencephalography (EEG) by event-related coherence analysis.

Methods

The scalp and subdural SPs were simultaneously recorded during NFB training by the DC-EEG machine while undergoing invasive recordings before epilepsy surgery in 10 patients with refractory partial epilepsy. The SPs at the vertex electrode were used as a reference for coherence analysis.

Results

The coherence of SPs negatively correlated with the distance between the subdural and scalp electrodes. A significant negative correlation was noted between the linear subdural–scalp electrode distance and the coherence value (r =  − 0.916, p < 0.001).

Conclusion

Scalp-recorded SPs from the vertex area primarily reflect the cortical activity of high lateral convexity.

Significance

Our results strongly suggest that SPs in NFB recorded from the vertex scalp electrode is derived from the cortices of high lateral convexity but not from the artifacts, such as GSR.

Introduction

Neuro-feedback (NFB) is a therapy for treating various neurological diseases such as epilepsy, attention-deficit hyperactivity disorder, and migraine (Wyckoff and Strehl, 2011). The regulation of slow potentials (SPs) can be applied for alleviating the symptoms of brain disorders. The SPs recorded by the scalp electrodes area are generated in the cerebral cortex and are therefore called as slow cortical potentials (SCPs).

The proposed mechanism of SCP includes the direct current potential shifts of the large neuronal assemblies of the cortex. Negative SCP shifts increase the firing probabilities of a cell assembly, whereas positive SCP shifts inhibit this activity (Birbaumer et al., 1990). In epilepsy patients, negative SCPs are recorded at the time of seizure onset as ictal electrocortico-corticogram (ECoG) with subdural electrode changes (Ikeda et al., 1996). Thus, NFB training was attempted to reduce the seizure frequency by the reinforcement of positive SCP (positivation or positive shift). In fact, several studies have demonstrated considerable the clinical efficacy of this approach (Rockstrohet al., 1993, Kotchoubeyet al., 1999, Strehlet al., 2005).

However, a previous report revealed no direct relationship between the number of spikes at the seizure onset zone (SOZ) and the shifts of SCPs in the scalp vertex EEG (Fritz et al., 2011). Furthermore, SP is highly susceptible to artifact contamination from various sources such as eye blinks, body movement, and galvanic skin response (GSR). Therefore, there is considerable doubt on whether patients who learned NFB were actually trained with SCPs derived from the brain itself. Despite the considerable clinical value on the NFB of SPs for seizure control in epilepsy, the underlying mechanism remains unclear.

Thus, we aimed to investigate the consistency of SPs between the subdural ECoG and scalp electroencephalogram (EEG) by using simultaneous recordings in patients with refractory epilepsy who underwent chronic implantation of intracranial electrodes for presurgical evaluation. We hypothesized that SPs recorded from the scalp were well-correlated with the SCPs recorded from intracranial electrodes. In other words, SPs recorded from the scalp are SCPs. We also hypothesized that electrode pairs that are closer to each other produces SCPs with higher consistency irrespective of the anatomy. To prove this hypothesis, we first grouped the intracranial electrodes on the basis of their embedded anatomical regions and then analyzed the coherence of SPs between a scalp-recorded EEG and intracranial-recorded ECoG with respect to each anatomical region.

Section snippets

Subjects

The inclusion criteria consisted of patients with drug-resistant focal epilepsy who underwent extraoperative subdural electrocorticography recording for presurgical evaluation between 2011 and 2014. A total of 10 patients (4 women; age: 24–45 years), who provided their informed consent, were enrolled in our research (Table 1). Intracranial electrodes were placed over the cortical surfaces or within the cortices, including the lateral convexity and lateral and basal temporal areas (left

Results

After excluding a bad electrode, the number of analyzed ECoG channels ranged from 18 to 25 among the patients (Table 1). In Pt. 6, a defective electrode was found in an HEOG electrode. This fault caused several ECoG channels to develop considerable AC contamination. Hence, we excluded all data belonging to Pt. 6 from further analysis.

As shown in Table 1, most recorded intracranial electrodes in this study were subdural grid/strip, which attached over the cortical surface. On the other hand, no

Discussion

The results of this study suggest that SP coherence has a negative correlation with the distance between the subdural and scalp electrodes. In this study, electrodes in high lateral convexity, such as dPrCG, dPoCG, and middle frontal gyrus, which had the closest proximity to the scalp electrodes, revealed high coherence values. By contrast, the distantly positioned electrodes, such as those in the basal temporal and basal frontal regions, revealed comparatively lower coherence values (

Conclusions

This study proved that SPs recorded in the scalp vertex electrodes were primarily derived from high lateral convexity. We also confirmed that NFB participants regulate SPs derived from the cortices.

Acknowledgments

This work was supported by the Japan Ministry of Education, Culture, Sports, Science and Technology (MEXT) KAKENHI Grant Numbers 15H05874, 17H05907, Japan Society for the Promotion of Science (JSPS) KAKENHI Grant Numbers 26293209, 26282218, and the Research Grants from the Japan Epilepsy Research Foundation. The authors would like to thank Enago (www.enago.jp) for the English language review.

Conflict of interest

Department of Epilepsy, Movement Disorders and Physiology is an endowment department, supported with a grant from GlaxoSmithKline K.K., Nihon Kohden Corporation, Otsuka Pharmaceuticals Co., and UCB Japan Co., Ltd. K. Other authors have no conflict of interest to disclose.

References (24)

  • N. von Ellenrieder et al.

    Electrode and brain modeling in stereo-EEG

    Clin Neurophysiol

    (2012)
  • N. von Ellenrieder et al.

    Extent of cortical generators visible on the scalp: effect of a subdural grid

    NeuroImage

    (2014)
  • Cited by (1)

    • Cognitive effort decreases beta, alpha, and theta coherence and ends afterdischarges in human brain

      2019, Clinical Neurophysiology
      Citation Excerpt :

      However, epileptiform activity can be modified in at least some patients with generalized seizures (Beniczky et al., 2012; Lunardi et al., 2016). Operant conditioning, biofeedback, relaxation therapies, and other behavioral methods can reduce seizures and can alter EEG or evoked potential activity (Efron, 1957; Sterman et al., 1974; Sterman and Shouse, 1980; Kotchoubey et al., 2001; Nagai et al., 2004; Fumuro et al., 2013; Kotwas et al., 2016, Fumuro et al., 2018). These take time to learn, and often several training sessions, whereas our patients’ ADs were terminated quickly, and without prior training.

    View full text