There are now a number of ways to create a window into the structure and function of the human nervous system. Computed tomography (CT) and magnetic resonance imaging (MRI) can identify fine structural abnormalities in the brain and are used routinely in medical practice. However, there are a number of contraindications to CT scanning and MRI, and both are costly to administer. Other imaging approaches (positron emission tomography and magnetic resonance spectroscopy) can detect changes in brain metabolism or metabolite concentrations. Radiation exposure and cost are 2 limitations of positron emission tomography, and the use of magnetic resonance spectroscopy faces the same contraindications as MRI. Functional brain imaging approaches, such as functional (f)MRI and magenetoencephalography, are also costly and technically challenging, whereas functional near-infrared spectroscopy is less costly and more clinically feasible than other approaches; however, it offers an indirect measure of brain activity.
Electroencephalography (EEG) is a low-cost, noninvasive imaging technique that measures neuroelectric activity in the brain with excellent temporal resolution. Electroencephalograph has almost no contraindications, making it an attractive approach for clinical populations. In the following sections, we introduce the basic principles of EEG before highlighting clinical applications of EEG, using stroke as an example. Then, we present a recently developed paradigm to study brain function using simultaneous EEG recording with a form of noninvasive brain stimulation, transcranial magnetic stimulation (TMS). We conclude by discussing the potential benefits of combining EEG with TMS and highlight other emerging EEG-related rehabilitation technologies that can improve our understanding of brain function in neurologic clinical populations that can leveraged into improving rehabilitation outcomes.
A BRIEF HISTORY OF EEG
Berger1 initially collected the first EEG recordings in human subjects in 1924. This work utilized 3 surface electrodes placed on the scalp to noninvasively document rhythmic oscillations in electrical potentials.2 These potentials are generated from a mixture of synchronous excitatory and inhibitory postsynaptic currents in similarly oriented neuronal cell populations.3,4 In the intervening 90 years, there have been major technological advances improving the quality of EEG recordings; expanded electrode configurations now allow for quantification of neural activity from various brain regions and analysis of neural networks, and computational advances now enable the processing and storage of large data sets. Furthermore, EEG can be combined with other imaging approaches to increase the amount and type of available data for analysis. For example, combining EEG and MRI offers the excellent temporal resolution provided by EEG with the high spatial resolution of MRI.
Since its inception, EEG has been used in a myriad of research applications to measure specific aspects of brain activity. Electrical oscillations, a fundamental organizing property of brain activity,3 and responses to environmental stimuli can be examined at rest and during behavior. Electroencephalography has been used to investigate cognitive processes underlying perception,5 decision making,6 and action.7 Clinically, EEG measures can inform the diagnosis, prognosis, and treatment response for many disorders including epilepsy,8 schizophrenia,9 altered consciousness,10 sleep,11 tumors,12 traumatic brain injury,13 and stroke.14
BASIC EEG METHODOLOGY
Electroencephalography utilizes noninvasive electrodes covering the scalp to quantify neuroelectric currents in the brain. Electric potentials detected by scalp electrodes depend both on the strength of the electric field resulting from local field potentials summed within populations of thousands of neurons, biological properties of the conducting media, and distance between the current generators and scalp.3 Capturing electrical activity through an external electrode exploits the principles of the Ohm's Law—current measured through a conductor is dependent on the voltage applied and the impedance or resistance.3 Current simply refers to the flow of an electric charge, whereas impedance reflects the degree of hindrance to the time-varying flow of current. Therefore, minimizing exogenous sources of impedance (hair, skin cells, oils, etc) during electrode preparation is important. To ensure controlled and reproducible scalp recordings, an electrode (Cz) is commonly placed over the center of the head (vertex). Electrode arrangements, known as montages, are highly customizable based on the population studied, information desired, and recording conditions (Figure 1). Although full montage preparation is typical, EEG setups containing only a ground electrode, a reference electrode, and a single active electrode can be used. Conversely, EEG recordings can be made from montages up to 256 electrodes, greatly increasing the amount of information available for analysis.
Electroencephalography recordings can be studied using various approaches that depend on the research or clinical question of interest and the experimental testing conditions employed. There are 2 primary EEG experimental paradigms—the event-related potential (ERP) technique and the continuous recording of ongoing brain activity. Event-related potentials are the result of specific cortical activation patterns associated with a specific event.15 These events can be external sensory stimulations such as a vibratory stimulus applied to the finger or can be internal such as the onset of a movement. Early ERP components are largely seen in primary processing regions and reflect the arrival of information in the area covered by the electrode, whereas the later components are thought to reflect higher-order processing occurring primarily in secondary cortical areas or association areas.16 Electroencephalography can also monitor and quantify spontaneous oscillatory activity when recorded continuously. This approach is commonly used to evaluate patterns of brain activity during different arousal states (eg, altered consciousness or sleep) or in response to interventions over time.3,17
Multiple frequency bands associated with cortical functioning have been documented (Table 1). Delta waves (0.5-4 Hz), theta waves (4-7 Hz), and alpha waves (8-13 Hz) are considered to be indicators of widespread cortical processing.18 These waves are seen across numerous cortical regions and relate to the integration of activity among brain regions. Modulations in frequency band amplitude are thought to be indicative of patterns of cortical activation and deactivation, such that regions showing related frequency changes likely represent an information-processing network underlying behavior.4 Alterations in frequency band amplitude may be associated with event-related changes; however, they are not noted until hundreds of milliseconds after a stimulus, suggesting that they provide a clearer measure of sustained task-related changes in activation patterns, rather than immediate responses to stimuli.19
Frequency oscillations can be examined in continuous data to determine the state of cortical activity at rest, or they can be documented in relation to a specific event. Phase synchronization of such oscillations may be examined in the following 3 ways: (1) phase coupling over distance, (2) phase coupling between frequencies, or (3) phase locking due to an external stimulus. Each measure provides unique information on frequency rhythms at immediate time points.36
Coherence examines whether shifts in oscillations in separate cortical areas are correlated with one another. Evidence of phase coupling between distant brain regions likely suggests an interconnected neural network.36 For example, coherence may provide a useful method to investigate changes in the motor preparation network, as opposed to simply quantifying local motor cortex excitability.
Phase coupling can also occur between oscillations of differing frequencies. Functionally, this phenomenon may be important for communication between separate, yet interrelated networks.37 Communication between 2 such networks is inferred from a constant phase lag between oscillations measured at a specific time point.37–39 If communication between networks is expected to be altered following a treatment, phase coupling could address this question.
Phase locking can also occur when recording ERPs. Event-related potentials may result from phase resetting of brain oscillations.40,41 Therefore, ERPs provide an immediate glimpse into the effect of peripheral stimuli on underlying brain activity. Event-related potentials are typically modality specific; for example, somatosensory-evoked potentials (SEPs) are elicited by peripheral nerve stimulation and represent afferent information processing in the contralateral sensorimotor cortex (Figure 2). This brief overview of common data acquisition and analysis approaches highlights the versatility of EEG and demonstrates the importance of closely aligning EEG recording parameters to the specific clinical or research question to be addressed.
PROS AND CONS OF EEG TO MEASURE CHANGES IN BRAIN ACTIVITY
Although EEG has a number of strengths that make it an attractive imaging technology, traditional EEG has a number of potential pitfalls that need to be considered. The chief limitation of EEG is low spatial resolution. Electrode locations provide gross localization for the distribution of cortical activity, and spatial accuracy can potentially be improved by increasing the number of recording electrodes42,43; however, true 3-dimenstional spatial localization with EEG is not possible, an important consideration for brains that have reorganized as a result of lesions or damage. Pairing EEG techniques with MRI, which has high spatial resolution (mm) can, in part, mitigate this limitation.44
Another challenge of EEG technology is that the strength of the signal dissipates with distance from current-generating sources and is influenced by tissue characteristics within the brain. Source estimation on the basis of the scalp potentials is imperfect; activity of a single source can be represented across many electrodes. Therefore, scalp signals are not equal to source signals. This situation is akin to the cocktail party analogy, where a microphone placed in a room will pick up sound from a mixture of sources, making difficult to interpret the precise location of individual sources of sound. Similarly, a single electrode on the scalp will pick up activity from a multitude of sources (cortical electrical activity, subcortical activity, external noise, etc), leading to difficulty in accurately localizing the source of activity. Multiple algorithms and computational approaches have now been developed that can improve source localization.45–47 Recognizing that there are an infinite number of current-generating dipoles within the brain that could explain the voltages measured on the scalp by EEG is important.42 As a result, localizing observed EEG activity to the actual current-generating source within the brain with absolute certainty is impossible.
Finally, EEG may be influenced by external factors such as participant alertness and environmental noise. Monitoring and maintaining a desired arousal level and eliminating as many unwanted sources of electrical noise as possible during recording become critical. Capitalizing on the strengths and minimizing the weaknesses of EEG are important considerations of this technology in both research and clinical applications. The following section discusses the utility of EEG in characterizing brain behavior after stroke as an example.
USING EEG TO MEASURE BRAIN ACTIVITY AFTER STROKE
Functional impairments following stroke are a result of direct ischemic loss of neurons combined with maladaptive brain reorganization.48 Measuring these changes in brain organization and activity has become possible through advances in neuroimaging and neurostimulation techniques. Yet, there is still a limited understanding of the relationships between reorganization of individual brain regions with one other and within coherent networks of functional brain activity. For decades, EEG has been used as a prognostic, diagnostic, and therapeutic tool during stroke recovery49 (Table 2). Somatosensory-evoked potentials are frequently measured acutely as a means to assess cortical responses to external perturbations to predict motor recovery following stroke.50,51 Abnormal SEPs also correlate with increased length of acute rehabilitation stay52 and poorer functional outcomes.50,52 However, while SEPs alone may have limited prognostic utility for determining long-term functional outcomes after stroke, the combination of EEG measures with motor-evoked potentials (MEPs) using noninvasive brain stimulation53 and clinical variables50 may improve predictive accuracy after stroke.
During stroke rehabilitation, the focus is often on successful performance of functional motor tasks to demonstrate recovery progress. Electroencephalography can be used to parse the potential deficits involved within individual processing stages of producing voluntary movement (ie, motor planning vs movement completion). For example, the contingent negative variation (CNV), a slow-wave ERP beginning approximately 2 seconds prior to a predictable cued stimulus, reveals cognitive preparation and planning prior to action.67,68 In healthy individuals, the CNV is primarily observed contralateral to the side of movement (ie, in the active hemisphere). Whereas individuals with chronic stroke with lesser functional recovery display a midline shift in the CNV during cued movement preparation compared with the nonparetic hand.61 This shift is thought to be associated with maladaptive reorganization after stroke.61 However, despite poor motor recovery, the neural mechanisms that underlie movement preparation of the paretic limb are maintained albeit reorganized. Therefore, rehabilitation interventions focusing on the preparatory phase of movement generation may encourage positive functional reorganization to support paretic extremity function.
Longitudinal EEG recording may be of value in assessing and predicting recovery trajectories. Impairments following stroke are multifaceted, and efficacy of rehabilitation is highly variable between individuals. The specific characteristics contributing to individual therapeutic response remain incompletely understood. Advances in the quantitative EEG (qEEG) analysis have resulted in an influx of promising results using qEEG as a prognostic marker of recovery in stroke. Specifically, qEEG overcomes the problem of interobserver variability by objectively quantifying EEG features using power spectrum analysis of frequency band content and by computational modeling.69,70 Although early EEG studies in individuals with stroke did not yield promising prognostic results, qEEG variables recorded that immediately following stroke were significantly correlated with residual disability and showed greater prognostic accuracy than the Canadian Neurological Scale Clinical stroke scale.14 More recent work demonstrated that following ischemic stroke, alpha and theta frequencies band power were significant predictors of short-term functional outcomes, whereas delta absolute power was a strong predictor of long-term functional outcomes.56 During the subacute phase of stroke recovery, qEEG measures were positively correlated with disability level (modified Rankin scale values) at 6 months poststroke57 and could identify individuals with a poor prognosis for functional recovery.55 A recent review concluded that qEEG is a sensitive approach to detect interhemispheric asymmetries and can inform clinical prognosis when collected during the subacute and acute phases following ischemic stroke.71 The assessment of qEEG within 72 hours of stroke diagnosis appears to be a useful tool to predict short- and long-term functional outcomes after stroke.71 The clinical utility of qEEG to denote poststroke markers of recovery will largely depend on the ability to implement recordings into routine clinical care. Recent work has shown that qEEG measures collected using a reduced electrode montage (4 electrodes vs 19 electrodes) were correlated with cognitive outcomes at 3 months poststroke.72 Although it may not always outperform clinical assessments or imaging data, EEG may be considered a potentially complementary measure to routine clinic management.
Following stroke, the regional connectivity of brain regions is also altered. Electroencephalography has been used to measure time-dependent functional brain connectivity. Gerloff et al73 demonstrated that following stroke cortico-cortical connectivity was reduced in the ipsilesional hemisphere and slightly increased in the contralesional hemisphere. Basic corticospinal commands are likely maintained by the ipsilesional cortex, whereas higher-order corticospinal commands, such as response selection and movement preparation, are mediated by the contralesional hemisphere.73 Computational EEG data modeling also suggest that there are connectivity differences in persons with stroke compared with neurologically healthy persons during anticipation and execution of self-initiated movements.59 Electroencephalography technology has advanced the understanding of the temporal aspects of cortical network connectivity following stroke that has clinical implications; however, localization of these changes to specific brain regions remains limited when EEG recordings are made in isolation.
In addition to using EEG to identify neural biomarkers of recovery after stroke, information from EEG can be harnessed for rehabilitation applications. Brain-computer interfaces (BCIs) are being used to augment recovery for individuals who typically demonstrate limited response to therapeutic interventions following stroke65 and spinal cord injury.74 Brain-computer interfaces allow individuals with limited motor ability to manipulate an upper-extremity orthosis by monitoring and quantifying EEG patterns in the ipsilesional hemisphere during imagined movements of the stroke-affected limb.62–65 Electroencephalography-driven robotics can supplement motor function following stroke with functional electric stimulation (FES). Functional electric stimulation utilizes focal activation of target muscle during the motor intention phase of movement using electromyography from the effector itself.66 In cases where sufficient electromyography activity is not present, EEG activity associated with movement intention can drive FES protocols.75 Although BCI technologies are still primarily in the development stage, the potential benefits for neurorehabilitation to address impairments after stroke and other neurological disorders present an important clinical application of EEG recording. In an effort to further maximize the utility of EEG, novel multimodal imaging approaches have been proposed that offer promise in their ability to identify new biomarkers of altered brain function and also recovery of function.
COMBINING EEG WITH TRANSCRANIAL MAGNETIC STIMULATION
A promising synergistic approach to characterize brain behavior is to combine EEG with TMS, a safe, noninvasive method used to stimulate cortical regions to measure levels of excitability. Transcranial magnetic stimulation essentially utilizes a magnetic field to readily carry electrical current through the skull painlessly and without current loss. When the coil is positioned on the scalp, the magnetic field induces small electrical currents in the underlying cortical tissue that can depolarize neurons transynaptically.76 When positioned over the primary motor cortex (M1), TMS can elicit a contralateral peripheral muscle response (a MEP). These MEPs provide information about excitability and integrity of corticospinal pathways. Using this methodology, ipsilesional M1 excitability is typically decreased following stroke,77,78 whereas contralesional M1 excitability remains largely unchanged.79,80 Importantly, this interhemispheric imbalance may interfere with recovery.81,82
Recent technologic advances now offer the capacity to combine TMS with EEG, to capitalize on the excellent temporal resolution and repeatability of EEG, in healthy individuals and clinical populations.83–85 Real-time integration of TMS and EEG provides the opportunity to directly characterize local and distributed cortical activity to empirically determine causal mechanisms of brain behavior in humans in vivo. This approach also provides the opportunity to stimulate any brain region and record the evoked activity using EEG, thus removing the need to elicit peripheral responses and infer cortical activity (Figure 3). Therefore, it is now possible to study brain behavior using TMS-elicited responses via EEG in any participant even when there is extensive damage to descending pathways. This ability could have important applications in individuals with severely disrupted sensorimotor input/output pathways. In persons with stroke, it can be difficult in many cases to generate measurable MEPs from ipsilesional M1 when the infarct location encroaches substantially on the descending corticospinal tract.86 It has previously been shown in individuals with chronic stoke where MEPs cannot be elicited from the ipsilesional M1, recovery is less complete and response to motor skill training is diminished.86 The ability to characterize brain behavior in this subset of persons with stroke using TMS-EEG may offer insights that could inform rehabilitation approaches for those with severe impairments after stroke.
The first successful report of TMS-evoked responses or potentials (TEPs) captured by EEG was in 1997.87 Since that time, significant technological advances now allow characterization of TEPs within 5 ms of stimulus delivery. Multiple reports have demonstrated a characteristic TEP waveform either recorded at the vertex (Cz) (Figure 4) or from the mean activity across electrodes. The characteristic positive and negative deflections observed in EEG recordings have been shown to be associated with brain activity in both cortical and subcortical regions in neurologically healthy individuals.42 Importantly, TEPs have been shown to be highly reproducible demonstrating excellent test-retest reliability in healthy participants.83 A TMS pulse can directly activate an area of cortical tissue approximately 1 cm2 at a depth of ∼1 to 2 cm depending on stimulation intensity. Thus, TEPs observed beyond local region of stimulation can be used to evaluate the spatiotemporal dynamics of intra-88 and interhemispheric89 connectivity as a result of activation a specific cortical region. This approach could be used to directly evaluate the status of specific neural pathways thought to play an important role in functional recovery following stroke. However, local cortical excitability and cortico-cortical connectivity have yet to be characterized using simultaneous TMS-EEG in participants with chronic stroke.
Using real-time TMS-EEG, one can understand differences in brain excitability and connectivity in the context of neurologic injury and/or disease and also evaluate changes in brain behavior in response to interventions (pharmacological, behavioral, and stimulation) or associated with spontaneous recovery. These 2 approaches can also be combined in an “off-line” design to circumvent the technological challenges of recording small amplitude EEG signals in the harsh TMS environment (for review, see Siebner et al43).
There are many challenges associated with combined TMS-EEG. Transcranial magnetic stimulation–evoked neuroelectric activity is susceptible to multiple artifact sources including the large electrical field induced by coil discharge, muscle activation near the stimulation site, auditory-evoked potentials, eye blink artifacts, and artifacts associated with recharging the TMS capacitor. Because of current methodological challenges and the cost of TMS stimulators, TMS-EEG will probably not soon become a part of routine clinical practice. Despite these challenges, combining these methodologies allows highly specific and repeatable perturbations to any cortical region using TMS and can be used in research investigations to evaluate (1) local cortical excitability, (2) spread of induced activation in time and space, (3) conduction times between the stimulated region and other cortical regions, and (4) changes in complex brain dynamics such as frequency band power or coherence using EEG. Furthermore, TMS-EEG can be used to evaluate the casual neural mechanisms underlying behavior. The ability to characterize the salient neural substrates of behavior may have important clinical implications for the design, delivery, and individualization of therapeutic interventions for individuals with neurologic conditions.
IMPLICATIONS FOR CLINICAL PRACTICE AND FUTURE DIRECTIONS
The capacity of EEG to measure electrical activity of the brain generated by postsynaptic potentials in neuronal populations continues to offer a powerful noninvasive, low-cost, widely applicable technique to study human brain behavior in a myriad of research paradigms and clinical populations. When combined with TMS, EEG has the potential to reveal causal mechanisms of altered cortical excitability and connectivity in neurologic disorders including stroke. It is inviting to speculate that an improved mechanistic understanding of the salient neural substrates underlying functional impairments and activity limitations observed in persons with neurological disorders will lead to new opportunities to develop targeted clinical interventions to restore previous levels of function. Using TMS-EEG to identify unique signatures of disordered cortical connectivity associated with certain stroke subtypes could inform clinical decision-making algorithms and foster the personalization of rehabilitation interventions. For example, distinct connectivity patterns consistent with a particular stroke subtype could be a negative prognostic indicator for response to a standard clinical intervention and may suggest an alternative treatment option (eg, noninvasive brain stimulation and pharmacological intervention). Improved characterization of altered brain behavior after stroke to provide better prognostic information demonstrates 1 potential future clinical application for EEG to aid in treatment selection and improved allocation of rehabilitation resources.
Combining EEG with BCI strategies offers additional avenues to support the recovery of motor function for individuals with severe motor impairments in situations where motor output pathways may no longer be intact.90,91 As a result, it could be possible to restore effective command of movement by using EEG-based signal detection of motor cortical network activity to bypass pathways that have been irreversibly damaged. This example demonstrates the potential clinical utility of EEG to noninvasively control a neuroprosthesis to perform movements in persons for whom independent motor control is not possible.
In summary, recent novel technological developments and sophisticated multimodal research approaches have created exciting new possibilities to capitalize on the strengths of EEG to improve our understanding of brain behavior. From a clinical perspective, furthering the science of brain behavior supporting motor recovery has the potential to improve rehabilitation outcomes by facilitating the development of sensitive measures to predict and monitor recovery trajectories, identification of the salient neural substrates underlying specific functional impairments, and selection of interventions on the basis of individual characteristics of abnormal brain behavior associated with neurologic disorders such as stroke.
We thank Drs Scott Makeig and Makoto Miyakoshi for consultation and assistance with aspects of data analysis. We also thank Sue Peters and Nick Snow for providing images for inclusion in this work.
1. Berger H. Uber das Elektroenkephalogramm des Menschen. Arch Psychiatr Nervenkr 1929;87:527–570.
2. Patterson MM, Thompson RF. Bioelectric Recording Techniques. New York, NY: Academic Press; 1973.
3. Nunez PL. Electric Fields of the Brain: The Neurophysics of EEG. New York, Oxford: Oxford University Press; 2006.
4. Peterson NN, Schroeder CE, Arezzo JC. Neural generators of early cortical somatosensory evoked potentials in the awake monkey. Electroencephalogr Clin Neurophysiol. 1995;96(3):248–260.
5. TallonBaudry C, Bertrand O, Delpuech C, Pernier J. Oscillatory gamma-band (30-70 Hz) activity induced by a visual search task in humans. J. Neurosci. 1997;17(2):722–734.
6. Cohen MX, Elger CE, Ranganath C. Reward expectation modulates feedback-related negativity and EEG spectra. Neuroimage. 2007;35(2):968–978.
7. Praamstra P, Boutsen L, Humphreys GW. Frontoparietal control of spatial attention and motor intention in human EEG. J Neurophysiol. 2005;94(1):764–774.
8. Litt B, Echauz J. Prediction of epileptic seizures. Lancet Neurol. 2002;1(1):22–30.
9. Spencer KM, Nestor PG, Niznikiewicz MA, Salisbury DF, Shenton ME, McCarley RW. Abnormal neural synchrony in schizophrenia. J Neurosci. 2003;23(19):7407–7411.
10. Ferrarelli F, Massimini M, Sarasso S, et al. Breakdown in cortical effective connectivity during midazolam-induced loss of consciousness. Proc Natl Acad Sci USA. 2010;107(6):2681–2686.
11. Riemann D, Spiegelhalder K, Feige B, et al. The hyperarousal model of insomnia: a review of the concept and its evidence. Sleep Med Rev 2010;14(1):19–31.
12. Fernandez-Bouzas A, Harmony T, Bosch J, et al. Sources of abnormal EEG activity in the presence of brain lesions. Clin Electroencephalogr. 1999;30(2):46–52.
13. Cruse D, Chennu S, Chatelle C, et al. Bedside detection of awareness in the vegetative state: a cohort study. Lancet. 2011;378(9809):2088–2094.
14. Cuspineda E, Machado C, Aubert E, Galan L, Llopis F, Avila Y. Predicting outcome in acute stroke
: a comparison between QEEG and the Canadian Neurological Scale. Clin Electroencephalogr. 2003;34(1):1–4.
15. Blackwood DH, Muir WJ. Cognitive brain potentials and their application. Br J Psychiatry Suppl. 1990(9):96–101.
16. Sur S, Sinha VK. Event-related potential: an overview. Ind Psychiatry J. 2009;18(1):70–73.
17. Schiff ND, Nauvel T, Victor JD. Large-scale brain dynamics in disorders of consciousness. Curr Opin Neurobiol. 2014;25:7–14.
18. Knyazev GG. EEG delta oscillations as a correlate of basic homeostatic and motivational processes. Neurosci Biobehav Rev. 2012;36(1):677–695.
19. Woertz M, Pfurtscheller G, Klimesch W. Alpha power dependent light stimulation: dynamics of event-related (de)synchronization in human electroencephalogram. Brain Res Cogn Brain Res. 2004;20(2):256–260.
20. Steriade M. Cellular Substrates of Brain Rhythms. Philadelphia, PA: Lippincott Williams & Wilkins; 2005.
21. Bruns A, Eckhorn R. Task-related coupling from high- to low-frequency signals among visual cortical areas in human subdural recordings. Int J Psychophysiol. 2004;51(2):97–116.
22. Devrim M, Demiralp T, Ademoglu A, Kurt A. A model for P300 generation based on responses to near-threshold visual stimuli. Brain Res Cogn Brain Res 1999;8(1):37–43.
23. Schurmann M, Basar-Eroglu C, Kolev V, Basar E. Delta responses and cognitive processing: single-trial evaluations of human visual P300. Int J Psychophysiol. 2001;39(2–3):229–239.
24. Roehm D, Schlesewsky M, Bornkessel I, Frisch S, Haider H. Fractionating language comprehension via frequency characteristics of the human EEG. Neuroreport. 2004;15(3):409–412.
25. Kahana MJ, Seelig D, Madsen JR. Theta returns. Curr Opin Neurobiol. 2001;11(6):739–744.
26. Fell J, Klaver P, Elfadil H, Schaller C, Elger CE, Fernandez G. Rhinal-hippocampal theta coherence during declarative memory formation: interaction with gamma synchronization? Eur J Neurosci. 2003;17(5):1082–1088.
27. Klimesch W, Doppelmayr M, Stadler W, Pollhuber D, Sauseng P, Rohm D. Episodic retrieval is reflected by a process specific increase in human electroencephalographic theta activity. Neurosci Lett. 2001;302(1):49–52.
28. Klimesch W, Doppelmayr M, Yonelinas A, et al. Theta synchronization during episodic retrieval: neural correlates of conscious awareness. Brain Res Cogn Brain Res. 2001;12(1):33–38.
29. Lopes da Silva FH, Vos JE, Mooibroek J, Van Rotterdam A. Relative contributions of intracortical and thalamo-cortical processes in the generation of alpha rhythms, revealed by partial coherence analysis. Electroencephalogr Clin Neurophysiol. 1980;50(5–6):449–456.
30. Klimesch W. EEG alpha and theta oscillations reflect cognitive and memory performance: a review and analysis. Brain Res Brain Res Rev. 1999;29(2–3):169–195.
31. Klimesch W, Sauseng P, Hanslmayr S. EEG alpha oscillations: the inhibition-timing hypothesis. Brain Res Rev. 2007;53(1):63–88.
32. Salenius S, Hari R. Synchronous cortical oscillatory activity during motor action. Curr Opin Neurobiol. 2003;13(6):678–684.
33. Neuper C, Pfurtscheller G. Event-related dynamics of cortical rhythms: frequency-specific features and functional correlates. Int J Psychophysiol. 2001;43(1):41–58.
34. Bekisz M, Wrobel A. Attention-dependent coupling between beta activities recorded in the cat's thalamic and cortical representations of the central visual field. Eur J Neurosci. 2003;17(2):421–426.
35. Kahana MJ. The cognitive correlates of human brain oscillations. J Neurosci. 2006;26(6):1669–1672.
36. Sauseng P, Klimesch W. What does phase information of oscillatory brain activity tell us about cognitive processes? Neurosci Biobehav Rev. 2008;32(5):1001–1013.
37. Palva JM, Palva S, Kaila K. Phase synchrony among neuronal oscillations in the human cortex. J Neurosci Methods. 2005;25(15):3962–3972.
38. Schack B, Vath N, Petsche H, Geissler HG, Moller E. Phase-coupling of theta-gamma EEG rhythms during short-term memory processing. Int J Psychophysiol. 2002;44(2):143–163.
39. Schack B, Weiss S. Quantification of phase synchronization phenomena and their importance for verbal memory processes. Biol Cybern. 2005;92(4):275–287.
40. Sayers BM, Beagley HA, Henshall WR. The mechanism of auditory evoked EEG responses. Nature. 1974;247(5441):481–483.
41. Başar E. Brain Function and Oscillations. New York, Berlin: Springer; 1998.
42. Luck SJ. An Introduction to the Event-Related Potential Technique. Cambridge, MA: MIT Press; 2005.
43. Siebner HR, Bergmann TO, Bestmann S, et al. Consensus paper: combining transcranial stimulation with neuroimaging
. Brain Stimul. 2009;2(2):58–80.
44. Dale AM, Halgren E. Spatiotemporal mapping of brain activity by integration of multiple imaging modalities. Curr Opin Neurobiol. 2001;11(2):202–208.
45. Delorme A, Makeig S. EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. J Neurosci Methods. 2004;134(1):9–21.
46. Pascualmarqui RD, Michel CM, Lehmann D. Low-resolution electromagnetic tomography—a new method for localizing electrical activity in the brain. Int J Psychophysiol. 1994;18(1):49–65.
47. Michel CM, Murray MM, Lantz G, Gonzalez S, Spinelli L, de Peralta RG. EEG source imaging. Clin Neurophysiol. 2004;115(10):2195–2222.
48. Taub E, Crago J, Burgio L, Groomes T, Cook EI, DeLuca S. An operant approach to rehabilitation
medicine: overcoming learned nonuse by shaping. J Exp Anal Behav. 1994;61:281–293.
49. Faught E. Current role of electroencephalography
in cerebral ischemia. Stroke
50. Feys H, Van Hees J, Bruyninckx F, Mercelis R, De Weerdt W. Value of somatosensory and motor evoked potentials in predicting arm recovery after a stroke
. J Neurol Neurosurg Psychiatry. 2000;68(3):323–331.
51. Gott PS, Karnaze DS, Fisher M. Assessment of median nerve somatosensory evoked potentials in cerebral ischemia. Stroke
52. Zeman BD, Yiannikas C. Functional prognosis in stroke
: use of somatosensory evoked potentials. J Neurol Neurosurg psychiatry. 1989;52(2):242–247.
53. Macdonell RA, Donnan GA, Bladin PF. A comparison of somatosensory evoked and motor evoked potentials in stroke
. Ann Neurol. 1989;25(1):68–73.
54. Giaquinto S, Cobianchi A, Macera F, Nolfe G. EEG recordings in the course of recovery from stroke
55. Stojanovic B, Djurasic L. Predictive importance of index of asymmetry in recovery following stroke
. Acta Chir Iugosl. 2013;60(1):101–104.
56. Cuspineda E, Machado C, Galan L, et al. QEEG prognostic value in acute stroke
. Clin EEG Neurosci. 2007;38(3):155–160.
57. Sheorajpanday RV, Nagels G, Weeren AJ, van Putten MJ, De Deyn PP. Quantitative EEG in ischemic stroke
: correlation with functional status after 6 months. Clin Neurophysiol. 2011;122(5):874–883.
58. Platz T, Kim IH, Pintschovius H, et al. Multimodal EEG analysis in man suggests impairment-specific changes in movement-related electric brain activity after stroke
. Brain. 2000;123(pt 12):2475–2490.
59. de Vico Fallani F, Astolfi L, Cincotti F, et al. Evaluation of the brain network organization from EEG signals: a preliminary evidence in stroke
patient. Anat Rec. 2009;292(12):2023–2031.
60. Wu W, Sun J, Jin Z, et al. Impaired neuronal synchrony after focal ischemic stroke
in elderly patients. Clin Neurophysiol. 2011;122(1):21–26.
61. Dean PJ, Seiss E, Sterr A. Motor planning in chronic upper-limb hemiparesis: evidence from movement-related potentials. PLoS One. 2012;7(10):e44558.
62. Birbaumer N, Cohen LG. Brain-computer interfaces: communication and restoration of movement in paralysis. J Physiol. 2007;579(pt 3):621–636.
63. Birbaumer N, Murguialday AR, Cohen L. Brain-computer interface in paralysis. Curr Opin Neurol. 2008;21(6):634–638.
64. Hinterberger T, Widman G, Lal TN, et al. Voluntary brain regulation and communication with electrocorticogram signals. Epilepsy Behav 2008;13(2):300–306.
65. Ramos-Murguialday A, Broetz D, Rea M, et al. Brain-machine interface in chronic stroke rehabilitation
: a controlled study. Ann Neurol. 2013;74(1):100–108.
66. Yeom H, Chang YH. Autogenic EMG-controlled functional electrical stimulation for ankle dorsiflexion control. J Neurosci Methods. 2010;193(1):118–125.
67. Rohrbaugh JW, Syndulko K, Lindsley DB. Brain wave components of the contingent negative variation in humans. Science 1976;191(4231):1055–1057.
68. Walter WG, Cooper R, Aldridge VJ, McCallum WC, Winter AL. Contingent negative variation: an electric sign of sensorimotor association and expectancy in the human brain. Nature 1964;203:380–384.
69. Geocadin RG, Ghodadra R, Kimura T, et al. A novel quantitative EEG injury measure of global cerebral ischemia. Clin Neurophysiol. 2000;111(10):1779–1787.
70. Finnigan SP, Rose SE, Walsh M, et al. Correlation of quantitative EEG in acute ischemic stroke
with 30-day NIHSS score: comparison with diffusion and perfusion MRI. Stroke
71. Finnigan S, van Putten MJ. EEG in ischaemic stroke
: quantitative EEG can uniquely inform (sub-) acute prognoses and clinical management. Clin Neurophysiol. 2013;124(1):10–19.
72. Schleiger E, Sheikh N, Rowland T, Wong A, Read S, Finnigan S. Frontal EEG delta/alpha ratio and screening for post-stroke
cognitive deficits: the power of four electrodes. Int J Psychophysiol. 2014;94(1):19–24.
73. Gerloff C, Bushara K, Sailer A, et al. Multimodal imaging of brain reorganization in motor areas of the contralesional hemisphere of well recovered patients after capsular stroke
. Brain. 2006;129(pt 3):791–808.
74. Vuckovic A. Hybrid brain-computer interface and functional electrical stimulation for sensorimotor training in participants with tetraplegia: A proof-of-concept study. J Neuro Phys Ther, 2015;39:3–14.
75. Takahashi M, Takeda K, Otaka Y, et al. Event related desynchronization-modulated functional electrical stimulation system for stroke rehabilitation
: a feasibility study. J Neuroeng Rehabil. 2012;9:56.
76. Day BL, Dick JPR, Marsden CD, et al. Interaction between electrical and magnetic stimulation of the human brain. J Physiol 1987;384:74P.
77. Thickbroom GW, Byrnes ML, Archer SA, Mastaglia FL. Motor outcome after subcortical stroke
: MEPs correlate with hand strength but not dexterity. Clin Neurophysiol. 2002;113(12):2025–2029.
78. Liepert J, Storch P, Fritsch A, Weiller C. Motor cortex disinhibition in acute stroke
. Clin Neurophysiol. 2000;111:671–676.
79. Liepert J, Hamzei F, Weiller C. Motor cortex disinhibition of the unaffected hemisphere after acute stroke
. Muscle Nerve. 2000;23:1761–1763.
80. Butefisch CM, Netz J, Wessling M, Seitz RJ, Homberg V. Remote changes in cortical excitability after stroke
. Brain 2003;126(pt 2):470–481.
81. Murase N, Duque J, Mazzocchio R, Cohen LG. Influence of interhemispheric interactions on motor function in chronic stroke
Ann Neurol. 2004;55:400–409.
82. Ward NS, Cohen LG. Mechanisms underlying recovery of motor function after stroke
. Arch Neurol. 2004;61(12):1844–1848.
83. Lioumis P, Kicic D, Savolainen P, Mäkelä JP, Kähkönen S. Reproducibility of TMS-evoked EEG responses. Hum Brain Mapp. 2009;30(4):1387–1396.
84. Komssi S, Aronen HJ, Huttunen J, et al. Ipsi- and contralateral EEG reactions to transcranial magnetic stimulation
. Clin Neurophysiol. 2002;113(2):175–184.
85. Ferrarelli F, Massimini M, Peterson MJ, et al. Reduced evoked gamma oscillations in the frontal cortex in schizophrenia patients: a TMS/EEG study. Am J Psychiatry. 2008;165(8):996–1005.
86. Stinear CM, Barber PA, Smale PR, Coxon JP, Fleming MK, Byblow WD. Functional potential in chronic stroke
patients depends on corticospinal tract integrity. Brain. 2007;130(pt 1):170–180.
87. Ilmoniemi RJ, Virtanen J, Ruohonen J, et al. Neuronal responses to magnetic stimulation reveal cortical reactivity and connectivity. Neuroreport. 1997;8(16):3537–3540.
88. Farzan F, Barr MS, Hoppenbrouwers SS, et al. The EEG correlates of the TMS-induced EMG silent period in humans. Neuroimage. 2013;83:120–134.
89. Hoppenbrouwers SS, Farzan F, Barr MS, et al. Personality goes a long a way: an interhemispheric connectivity study. Front Psychiatry. 2010;1:140.
90. Leamy DJ, Kocijan J, Domijan K, et al. An exploration of EEG features during recovery following stroke
—implications for BCI-mediated neurorehabilitation therapy. J Neuroeng Rehabil. 2014;11(1):9.
91. Machado S, Araujo F, Paes F, et al. EEG-based brain-computer interfaces: an overview of basic concepts and clinical applications in neurorehabilitation. Rev Neurosci. 2010;21(6):451–468.