- Email: [email protected]
ERS in the context of brain–computer interface (BCI) developments
Neuper & Klimesch (Eds.) Progress in Brain Research, Vol. 159 ISSN 0079-6123 Copyright r 2006 Elsevier B.V. All rights reserved
CHAPTER 28
Future prospects of ERD/ERS in the context of brain–computer interface (BCI) developments Gert Pfurtscheller1, and Christa Neuper2 1
Laboratory of Brain– Computer Interfaces (BCI-Lab), Institute for Knowledge Discovery, Graz University of Technology, Krenngasse 37. A-8010 Graz, Austria 2 Institute of Psychology, University of Graz, Universita¨tsplatz 2/III, A-8010 Graz, Austria
Abstract: ERD/ERS patterns characterize the dynamics of brain oscillations time-locked but not phaselocked to an externally or internally triggered event. Recent studies have shown that ERD/ERS phenomena in narrow frequency bands are remarkably stable over time and across different testing situations. The high reproducibility of ERD/ERS promotes the usefulness of this biometric measure in assessing individual characteristics. In addition to the spatio-temporal patterns of (de)synchronization processes the most reactive frequency components are especially highly subject-specific and, therefore, open up new possibilities for user authentication and person identification. In contrast, ERD/ERS research will continue to be useful in clinical brain–computer interface (BCI) implementation. Promising novel applications of an ERD/ ERS based BCI may contribute to enhanced functional recovery and rehabilitation in patients suffering from chronic stroke. According to current therapeutic strategies, feedback-regulated motor imagery could be used to enhance antagonistic ERD/ERS patterns and therewith, support activation of the stroke affected and inhibition of the non-affected, contralesional hemisphere. Keywords: ERD; ERS; person identification; user authentication; stroke rehabilitation Another important aspect of BCI technology is its usability for neurofeedback applications. In this context, it is relevant that electrophysiological brain signals can be brought under operant (or self-) control. It has been convincingly shown that by self-regulation of specific EEG frequency components (e.g., sensorimotor rhythms) and the volitional control of slow cortical potentials, a reduction of epileptic seizures and an effective treatment of attention deficit hyperactivity disorder are possible (reviewed in Chapter 27). A further promising use of BCI feedback has been recently uncovered in chronic stroke patients (see Chapter 24). Combining BCIs with virtual reality training environments would allow, e.g., online-monitoring of the electrophysiological activity coupled with feedback control of the training environment.
EEG based brain–computer interfaces (BCIs) represent an additional mode of communication between human thought and the environment. Previous studies on the efficacy of such communication (in terms of transferable information) have revealed that the detection of different types of cognitive processes in real-time is basically feasible (e.g., Obermaier et al., 2001; for a review, see Wolpaw et al., 2002). Recently, sophisticated application interfaces have been developed that allows control of a robot, prosthesis, or wheelchair by mere thought. In this way, BCIs have a great potential for patients with severely affected motor functions to improve their independence and quality of life (see Chapters 24 and 25, this volume). Corresponding author. E-mail: [email protected] DOI: 10.1016/S0079-6123(06)59028-4
433
434
In fact, the integration of brain interfaces in virtual reality applications is already being investigated (e.g., Pfurtscheller et al., 2006a, b). In the near future, the abovementioned and novel fields of BCI applications based on ERD/ERS measurements may gain importance. This technology opens up new vistas in biometric assessment and person identification. In the following sections, we will outline current concepts in the context of (i) user authentication and person identification and (ii) virtual reality based stroke rehabilitation.
ERD/ERS patterns as a biometric measure Beside event-related potentials, as e.g., the P300 waveform (Thie and Fried, 2005) also the dynamics of brain oscillations (ERD/ERS) may be suitable as biometric measure for purposes of user authentication or person identification. Studies addressing the reliability and stability of EEG measures have demonstrated that band power features in specific sites, particularly measures of alpha and lower beta power, are highly reliable within adult subjects over extended periods of time (e.g., Pollock et al., 1991; Kondacs and Szabo, 1999). Recently, Neuper et al. (2005a) reported surprisingly high long-term stability and consistency also for ERD/ERS values obtained while subjects performed a cognitive task (see also Chapter 11). The most stable EEG feature across three measurements within a period of 2 years, however, was the individual alpha-peak frequency. The excellent reliability of the individual alpha frequency (i.e., Cronbach’s alpha of 0.9) found in that study corroborates the suggestion that genetic factors might explain individual differences in alpha-frequency peak (Posthuma et al., 2001). Hence, in the case of ERD/ERS it is not only the spatio-temporal pattern of the brain’s response to an internally paced (e.g., thought) or externally paced event (e.g., stimulus), but also particularly the frequency of the induced oscillations that may be of importance. The frequency of such induced oscillations depends on the neural network properties and feedback loops. If such a specific neural network is activated, the frequency of the induced oscillatory response (‘resonance-like’ frequency)
can be related to the number of cells or cell assemblies involved. In general, rapidly oscillating cell assembling comprised fewer neurons as compared with slowly oscillating assemblies (Singer, 1993). Two types of oscillations can be differentiated in sensorimotor areas: one type is modified during execution of a sensorimotor task, the other is dominant after cessation of a motor task or after somatosensory stimulation. An example of the former is, e.g., the induced hand area mu rhythm during foot motor imagery and examples for the latter are the stimulation-induced beta bursts (beta ERS) (for a more detailed description see Chapters 2, 14, and 16, this volume). The frequency of this stimulation induced beta ERS between 15 and 35 Hz is non-modifiable by the individual and therefore, among other features, promising for person identification. An example is given in Fig. 1. Beta ERS was obtained in three normal subjects after periodic mechanical stimulation of the tip of different fingers (Pfurtscheller et al., 2001). The frequencies are not only different in the three persons studied, but also display characteristic differences between individual fingers. This observation raises the possibility that the beta ERS induced by tactile (or electrical nerve) stimulation may be suitable as one type of non-modifiable marker for biometric identification measurable by a BCI.
Fig. 1. Induced-beta ERS for different fingers in three subjects. For subjects P1 and P2 significant frequency differences are found between individual fingers. For the analysis of single EEG trials the Matching Pursuit algorithm was used (modified from Pfurtscheller et al., 2001).
435
In contrast to the beta ERS occurring in response to somatosensory stimulation, mentally induced oscillations are consciously modifiable and controllable by the user and therefore, may be exploited for user authentication (Palaniappan and Ravi, 2003). A prerequisite for the use of intentionally produced oscillations as ‘‘pass-thought’’ in a user authentication system is the stability and reproducibility of the dominant frequency. An
example of repeated frequency measurement in one subject during motor imagery with mentally induced beta ERS is given in Fig. 2. Measurements over a period of 9 months revealed stable frequencies of around 16.5 Hz.It is notable that these frequency-specific oscillations were induced with intend by mental imagery. A BCI capable of detecting specific bioelectrical ‘‘brain patterns’’ such as, e.g., mentally induced brain oscillations with a characteristic frequency in real time may therefore introduce a new form of biometric authentication, which in the future may be offered in addition to finger prints, exploration of the iris, and hand-written signature verification. The concept of user authentication with a BCI is displayed in Fig. 3. Such a system is able to give permission to enter, e.g., a highly protected environment after EEG recording and feature extraction.
EEG-based BCI and rehabilitation after stroke Fig. 2. Frequency of motor imagery induced beta oscillations (mean7SD) in one subject in follow-up measurements within 9 months.
There is growing knowledge that the brain is capable of significant functional recovery from
pass thought stored “pass thought” feature consciously modifiable brain pattern (e.g., imagery of right hand movement)
yes match ?
EEG
Graz-BCI preprocessing
feature extraction
Fig. 3. Layout of a BCI-system for user authentication.
login successful
436
neurological diseases such as stroke, provided that the appropriate therapies are used (for a recent review, see Hummel and Cohen, 2005; see also Chapter 15, this volume). Important in stroke rehabilitation is either to treat the paretic limb actively or passively, or if this is not possible, to imagine or simulate the movement of the affected limb (for a review, see Schaechter, 2004). It was shown that the frequent use of the paretic hand could recruit previously silent synapses and pathways (Papathanassiou et al., 2003) and activate neural circuitry in the affected sensorimotor cortices. The usefulness of the concept of motor imagery was demonstrated by Stevens and Stoykov (2003) in hemiparetic stroke patients. Patients underwent intensive training utilizing motor imagery consisting of imagined wrist movements. In the case of motor imagery, it is of importance to give feedback to the patient information about the successful execution of the motor imagery task. Virtual reality provides a powerful technology in stroke rehabilitation to give feedback about motor task performance and enhance the motivation to endure practice (for a review of virtual reality applications in motor rehabilitation see Holden, 2005). Especially in online experiments, the knowledge of whether or not motor imagery was performed correctly, plays a crucial role. One question is which form of visual feedback should be given during motor imagery training in the rehabilitation of paresis in stroke patients. From a pilot study there is evidence that the observation of a moving virtual body part has a greater impact on the neuronal activity in sensorimotor areas than, e.g., a moving geometrical object (Pfurtscheller et al., 2006b). Hence, one answer may be the use of a moving limb for feedback on motor imagery in virtual reality based stroke therapy (Teasell and Kalra, 2005; Holden, 2005). Interventional approaches in stroke rehabilitation have been proposed to suppress the afferent flow from the unaffected limb and to reduce herewith the activation of the intact hemisphere (Schaechter, 2004). As a consequence, the transcallosal inhibition of the affected hemisphere is reduced. One possibility of afferent flow suppression in patients with hemiplegia is, e.g., the fixation of the unaffected limb known as the
‘‘constraint-induced movement therapy’’. The decreased use of the unaffected limb during constrained induced movement therapy may contribute to a relative increase of activation in the representation area of the affected limb. A new EEG-based concept currently under investigation to enhance motor rehabilitation in patients with hemiparetic stroke is grounded on the following central themes:
Neurofeedback training utilizing motor imagery with focus on kinesthetic experiences involving the affected limb (kinesthetic motor imagery). The importance of such a kinesthetic type of motor imagery in BCI research was stressed by Neuper et al. (2005b; see also Chapter 25). Online EEG analysis, feature extraction and classification using an ERD/ERS-based BCI system. The output signal of the BCI is translated into the movement of a virtual hand, in a similar way to that reported in a recent ‘‘walking experiment’’ (Pfurtscheller et al., 2006b). If the motor imagery task is correctly executed, the corresponding virtual hand displays a closing and/or opening. The concurrent observation of the moving virtual body part may reinforce the activation of the sensorimotor cortex. To realize this EEG-based stroke rehabilitation concept building on a BCI involving motor imagery and providing feedback by presenting virtual body parts, it is essential to determine in each patient and EEG recording the most reactive sensorimotor frequency components (for details see Chapter 6). After intensive training of feedback-regulated motor imagery it can be expected that sensorimotor rhythms display a desynchronization (ERD) in the stroke affected hemisphere and a synchronization (ERS) in the intact hemisphere. This type of antagonistic ERD/ ERS pattern, also known as ‘‘focal ERD/surround ERS’’ (see Chapter 2), is characteristic in the able-bodied subject in association with right/left hand motor imagery (Pfurtscheller and Neuper, 1997). When we assume that ERD characterizes activation and ERS inhibition of sensorimotor networks (Hummel
437
et al., 2002, 2004), then correctly performed motor imagery of hand/wrist movement should result in activation of the affected and inhibition of the non-affected, contralesional hemisphere and therewith, mimics the constraint-induced movement therapy. In this direction novel BCI-based approaches are devoted to enhance neural plasticity after stroke (see also Chapter 24). The main goal is to stimulate cortical reorganization and compensatory activation of non-lesioned brain regions and reduction of contralesional hemispheric inhibition through, e.g., motor imagery involving the paralyzed limb. References Holden, M.K. (2005) Virtual environments for motor rehabilitation: review. CyberPsychol. Behav., 8: 187–211. Hummel, F., Andres, F., Altenmu¨ller, E., Dichgans, J. and Gerloff, C. (2002) Inhibitory control of acquired motor programmes in the human brain. Brain, 125: 404–420. Hummel, F.C. and Cohen, L.G. (2005) Drivers of brain plasticity. Curr. Opin. Neurol., 18: 667–674. Hummel, F., Saur, R., Lasogga, S., Plewnia, C., Erb, M., Wildgruber, D., Grodd, W. and Gerloff, C. (2004) To act or not to act. Neural correlates of executive control of learned motor behaviour. Neuroimage, 23: 1391–1401. Kondacs, A. and Szabo, M. (1999) Long-term intra-individual variability of the background EEG in normals. Clin. Neurophysiol., 110: 1708–1716. Neuper, C., Grabner, R.H., Fink, A. and Neubauer, A.C. (2005a) Long-term stability and consistency of EEG eventrelated (de-)synchronization across different cognitive tasks. Clin. Neurophysiol., 116: 1681–1694. Neuper, C., Scherer, R., Reiner, M. and Pfurtscheller, G. (2005b) Imagery of motor actions: differential effects of kinesthetic and visual-motor mode of imagery in single-trial EEG. Cogn. Brain Res., 25: 668–677. Obermaier, B., Neuper, C., Guger, C. and Pfurtscheller, G. (2001) Information transfer rate in a 5-classes brain–computer interface. IEEE Trans. Neural Syst. Rehabil. Eng., 9: 283–288.
Palaniappan, R. and Ravi, K.V.R. (2003) A new method to identify individuals using signals from the brain. Proc. Int. Conf. Inf. Comput. Secur., 1442–1445. Papathanassiou, I., Filipovic, S.R., Whurr, R. and Jahanshahi, M. (2003) Plasticity of motor cortex excitability induced by rehabilitation therapy for writing. Neurology, 61: 977–980. Pfurtscheller, G., Leeb, R., Keinrath, C., Friedmann, D., Neuper, C., Guger, C. and Slater, M. (2006a) Walking from thought. Brain Res., 1971: 145–152. Pfurtscheller, G. and Neuper, C. (1997) Motor imagery activates primary sensorimotor area in humans. Neurosci. Lett., 239: 65–68. Pfurtscheller, G., Scherer, R., Leeb, R., Keinrath, C., Neuper, C., Lee, F., Graimann, B., Schlo¨gl, A. and Bischof, H. (2006b) Viewing Moving Objects in Virtual Reality Can Change the Dynamics of Sensorimotor EEG Rhythms. Presence: Teleoperators and Virtual Environments. MIT Press, Cambridge, MA (in press) Pfurtscheller, G., Wo¨rtz, M., Krausz, G. and Neuper, C. (2001) Distinction of different fingers by the frequency of stimulus induced beta oscillations in the human EEG. Neurosci. Lett., 307: 49–52. Pollock, V.E., Schneider, L.S. and Lyness, S.A. (1991) Reliability of topographic quantitative EEG amplitude in healthy late-middle-aged and elderly subjects. Electroencephalogr. Clin. Neurophysiol., 79: 20–26. Posthuma, D., Neale, M.C., Boomsma, D.I. and de Geus, E.J.C. (2001) Are smarter brains running faster? Heritability of alpha peak frequency, IQ, and their interrelation. Behav. Genet., 31: 567–579. Schaechter, J.D. (2004) Motor rehabilitation and brain plasticity after hemiparetic stroke. Progr. Neurobiol., 73: 61–72. Singer, W. (1993) Synchronization of cortical activity and its putative role in information processing and learning. Annu. Rev. Physiol., 55: 349–374. Stevens, J.A. and Stoykov, M.E. (2003) Using motor imagery in the rehabilitation of hemiparesis. Arch. Phys. Med. Rehabil., 84: 1090–1092. Teasell, R.W. and Kalra, L. (2005) What’s new in stroke rehabilitation. Back to basics. Stroke, 36: 215–217. Thie, L. and Fried, L. (2005) Acclairisms: Questioning Biometric Technology Through an Airport Security Clearance System. ACM CHI, Portland, OR, 2005. Wolpaw, J.R., Birbaumer, N., McFarland, D., Pfurtscheller, G. and Vaughan, T. (2002) Brain–computer interfaces for communication and control. Clin. Neurophysiol., 113: 767–791.
CHAPTER 28
Future prospects of ERD/ERS in the context of brain–computer interface (BCI) developments Gert Pfurtscheller1, and Christa Neuper2 1
Laboratory of Brain– Computer Interfaces (BCI-Lab), Institute for Knowledge Discovery, Graz University of Technology, Krenngasse 37. A-8010 Graz, Austria 2 Institute of Psychology, University of Graz, Universita¨tsplatz 2/III, A-8010 Graz, Austria
Abstract: ERD/ERS patterns characterize the dynamics of brain oscillations time-locked but not phaselocked to an externally or internally triggered event. Recent studies have shown that ERD/ERS phenomena in narrow frequency bands are remarkably stable over time and across different testing situations. The high reproducibility of ERD/ERS promotes the usefulness of this biometric measure in assessing individual characteristics. In addition to the spatio-temporal patterns of (de)synchronization processes the most reactive frequency components are especially highly subject-specific and, therefore, open up new possibilities for user authentication and person identification. In contrast, ERD/ERS research will continue to be useful in clinical brain–computer interface (BCI) implementation. Promising novel applications of an ERD/ ERS based BCI may contribute to enhanced functional recovery and rehabilitation in patients suffering from chronic stroke. According to current therapeutic strategies, feedback-regulated motor imagery could be used to enhance antagonistic ERD/ERS patterns and therewith, support activation of the stroke affected and inhibition of the non-affected, contralesional hemisphere. Keywords: ERD; ERS; person identification; user authentication; stroke rehabilitation Another important aspect of BCI technology is its usability for neurofeedback applications. In this context, it is relevant that electrophysiological brain signals can be brought under operant (or self-) control. It has been convincingly shown that by self-regulation of specific EEG frequency components (e.g., sensorimotor rhythms) and the volitional control of slow cortical potentials, a reduction of epileptic seizures and an effective treatment of attention deficit hyperactivity disorder are possible (reviewed in Chapter 27). A further promising use of BCI feedback has been recently uncovered in chronic stroke patients (see Chapter 24). Combining BCIs with virtual reality training environments would allow, e.g., online-monitoring of the electrophysiological activity coupled with feedback control of the training environment.
EEG based brain–computer interfaces (BCIs) represent an additional mode of communication between human thought and the environment. Previous studies on the efficacy of such communication (in terms of transferable information) have revealed that the detection of different types of cognitive processes in real-time is basically feasible (e.g., Obermaier et al., 2001; for a review, see Wolpaw et al., 2002). Recently, sophisticated application interfaces have been developed that allows control of a robot, prosthesis, or wheelchair by mere thought. In this way, BCIs have a great potential for patients with severely affected motor functions to improve their independence and quality of life (see Chapters 24 and 25, this volume). Corresponding author. E-mail: [email protected] DOI: 10.1016/S0079-6123(06)59028-4
433
434
In fact, the integration of brain interfaces in virtual reality applications is already being investigated (e.g., Pfurtscheller et al., 2006a, b). In the near future, the abovementioned and novel fields of BCI applications based on ERD/ERS measurements may gain importance. This technology opens up new vistas in biometric assessment and person identification. In the following sections, we will outline current concepts in the context of (i) user authentication and person identification and (ii) virtual reality based stroke rehabilitation.
ERD/ERS patterns as a biometric measure Beside event-related potentials, as e.g., the P300 waveform (Thie and Fried, 2005) also the dynamics of brain oscillations (ERD/ERS) may be suitable as biometric measure for purposes of user authentication or person identification. Studies addressing the reliability and stability of EEG measures have demonstrated that band power features in specific sites, particularly measures of alpha and lower beta power, are highly reliable within adult subjects over extended periods of time (e.g., Pollock et al., 1991; Kondacs and Szabo, 1999). Recently, Neuper et al. (2005a) reported surprisingly high long-term stability and consistency also for ERD/ERS values obtained while subjects performed a cognitive task (see also Chapter 11). The most stable EEG feature across three measurements within a period of 2 years, however, was the individual alpha-peak frequency. The excellent reliability of the individual alpha frequency (i.e., Cronbach’s alpha of 0.9) found in that study corroborates the suggestion that genetic factors might explain individual differences in alpha-frequency peak (Posthuma et al., 2001). Hence, in the case of ERD/ERS it is not only the spatio-temporal pattern of the brain’s response to an internally paced (e.g., thought) or externally paced event (e.g., stimulus), but also particularly the frequency of the induced oscillations that may be of importance. The frequency of such induced oscillations depends on the neural network properties and feedback loops. If such a specific neural network is activated, the frequency of the induced oscillatory response (‘resonance-like’ frequency)
can be related to the number of cells or cell assemblies involved. In general, rapidly oscillating cell assembling comprised fewer neurons as compared with slowly oscillating assemblies (Singer, 1993). Two types of oscillations can be differentiated in sensorimotor areas: one type is modified during execution of a sensorimotor task, the other is dominant after cessation of a motor task or after somatosensory stimulation. An example of the former is, e.g., the induced hand area mu rhythm during foot motor imagery and examples for the latter are the stimulation-induced beta bursts (beta ERS) (for a more detailed description see Chapters 2, 14, and 16, this volume). The frequency of this stimulation induced beta ERS between 15 and 35 Hz is non-modifiable by the individual and therefore, among other features, promising for person identification. An example is given in Fig. 1. Beta ERS was obtained in three normal subjects after periodic mechanical stimulation of the tip of different fingers (Pfurtscheller et al., 2001). The frequencies are not only different in the three persons studied, but also display characteristic differences between individual fingers. This observation raises the possibility that the beta ERS induced by tactile (or electrical nerve) stimulation may be suitable as one type of non-modifiable marker for biometric identification measurable by a BCI.
Fig. 1. Induced-beta ERS for different fingers in three subjects. For subjects P1 and P2 significant frequency differences are found between individual fingers. For the analysis of single EEG trials the Matching Pursuit algorithm was used (modified from Pfurtscheller et al., 2001).
435
In contrast to the beta ERS occurring in response to somatosensory stimulation, mentally induced oscillations are consciously modifiable and controllable by the user and therefore, may be exploited for user authentication (Palaniappan and Ravi, 2003). A prerequisite for the use of intentionally produced oscillations as ‘‘pass-thought’’ in a user authentication system is the stability and reproducibility of the dominant frequency. An
example of repeated frequency measurement in one subject during motor imagery with mentally induced beta ERS is given in Fig. 2. Measurements over a period of 9 months revealed stable frequencies of around 16.5 Hz.It is notable that these frequency-specific oscillations were induced with intend by mental imagery. A BCI capable of detecting specific bioelectrical ‘‘brain patterns’’ such as, e.g., mentally induced brain oscillations with a characteristic frequency in real time may therefore introduce a new form of biometric authentication, which in the future may be offered in addition to finger prints, exploration of the iris, and hand-written signature verification. The concept of user authentication with a BCI is displayed in Fig. 3. Such a system is able to give permission to enter, e.g., a highly protected environment after EEG recording and feature extraction.
EEG-based BCI and rehabilitation after stroke Fig. 2. Frequency of motor imagery induced beta oscillations (mean7SD) in one subject in follow-up measurements within 9 months.
There is growing knowledge that the brain is capable of significant functional recovery from
pass thought stored “pass thought” feature consciously modifiable brain pattern (e.g., imagery of right hand movement)
yes match ?
EEG
Graz-BCI preprocessing
feature extraction
Fig. 3. Layout of a BCI-system for user authentication.
login successful
436
neurological diseases such as stroke, provided that the appropriate therapies are used (for a recent review, see Hummel and Cohen, 2005; see also Chapter 15, this volume). Important in stroke rehabilitation is either to treat the paretic limb actively or passively, or if this is not possible, to imagine or simulate the movement of the affected limb (for a review, see Schaechter, 2004). It was shown that the frequent use of the paretic hand could recruit previously silent synapses and pathways (Papathanassiou et al., 2003) and activate neural circuitry in the affected sensorimotor cortices. The usefulness of the concept of motor imagery was demonstrated by Stevens and Stoykov (2003) in hemiparetic stroke patients. Patients underwent intensive training utilizing motor imagery consisting of imagined wrist movements. In the case of motor imagery, it is of importance to give feedback to the patient information about the successful execution of the motor imagery task. Virtual reality provides a powerful technology in stroke rehabilitation to give feedback about motor task performance and enhance the motivation to endure practice (for a review of virtual reality applications in motor rehabilitation see Holden, 2005). Especially in online experiments, the knowledge of whether or not motor imagery was performed correctly, plays a crucial role. One question is which form of visual feedback should be given during motor imagery training in the rehabilitation of paresis in stroke patients. From a pilot study there is evidence that the observation of a moving virtual body part has a greater impact on the neuronal activity in sensorimotor areas than, e.g., a moving geometrical object (Pfurtscheller et al., 2006b). Hence, one answer may be the use of a moving limb for feedback on motor imagery in virtual reality based stroke therapy (Teasell and Kalra, 2005; Holden, 2005). Interventional approaches in stroke rehabilitation have been proposed to suppress the afferent flow from the unaffected limb and to reduce herewith the activation of the intact hemisphere (Schaechter, 2004). As a consequence, the transcallosal inhibition of the affected hemisphere is reduced. One possibility of afferent flow suppression in patients with hemiplegia is, e.g., the fixation of the unaffected limb known as the
‘‘constraint-induced movement therapy’’. The decreased use of the unaffected limb during constrained induced movement therapy may contribute to a relative increase of activation in the representation area of the affected limb. A new EEG-based concept currently under investigation to enhance motor rehabilitation in patients with hemiparetic stroke is grounded on the following central themes:
Neurofeedback training utilizing motor imagery with focus on kinesthetic experiences involving the affected limb (kinesthetic motor imagery). The importance of such a kinesthetic type of motor imagery in BCI research was stressed by Neuper et al. (2005b; see also Chapter 25). Online EEG analysis, feature extraction and classification using an ERD/ERS-based BCI system. The output signal of the BCI is translated into the movement of a virtual hand, in a similar way to that reported in a recent ‘‘walking experiment’’ (Pfurtscheller et al., 2006b). If the motor imagery task is correctly executed, the corresponding virtual hand displays a closing and/or opening. The concurrent observation of the moving virtual body part may reinforce the activation of the sensorimotor cortex. To realize this EEG-based stroke rehabilitation concept building on a BCI involving motor imagery and providing feedback by presenting virtual body parts, it is essential to determine in each patient and EEG recording the most reactive sensorimotor frequency components (for details see Chapter 6). After intensive training of feedback-regulated motor imagery it can be expected that sensorimotor rhythms display a desynchronization (ERD) in the stroke affected hemisphere and a synchronization (ERS) in the intact hemisphere. This type of antagonistic ERD/ ERS pattern, also known as ‘‘focal ERD/surround ERS’’ (see Chapter 2), is characteristic in the able-bodied subject in association with right/left hand motor imagery (Pfurtscheller and Neuper, 1997). When we assume that ERD characterizes activation and ERS inhibition of sensorimotor networks (Hummel
437
et al., 2002, 2004), then correctly performed motor imagery of hand/wrist movement should result in activation of the affected and inhibition of the non-affected, contralesional hemisphere and therewith, mimics the constraint-induced movement therapy. In this direction novel BCI-based approaches are devoted to enhance neural plasticity after stroke (see also Chapter 24). The main goal is to stimulate cortical reorganization and compensatory activation of non-lesioned brain regions and reduction of contralesional hemispheric inhibition through, e.g., motor imagery involving the paralyzed limb. References Holden, M.K. (2005) Virtual environments for motor rehabilitation: review. CyberPsychol. Behav., 8: 187–211. Hummel, F., Andres, F., Altenmu¨ller, E., Dichgans, J. and Gerloff, C. (2002) Inhibitory control of acquired motor programmes in the human brain. Brain, 125: 404–420. Hummel, F.C. and Cohen, L.G. (2005) Drivers of brain plasticity. Curr. Opin. Neurol., 18: 667–674. Hummel, F., Saur, R., Lasogga, S., Plewnia, C., Erb, M., Wildgruber, D., Grodd, W. and Gerloff, C. (2004) To act or not to act. Neural correlates of executive control of learned motor behaviour. Neuroimage, 23: 1391–1401. Kondacs, A. and Szabo, M. (1999) Long-term intra-individual variability of the background EEG in normals. Clin. Neurophysiol., 110: 1708–1716. Neuper, C., Grabner, R.H., Fink, A. and Neubauer, A.C. (2005a) Long-term stability and consistency of EEG eventrelated (de-)synchronization across different cognitive tasks. Clin. Neurophysiol., 116: 1681–1694. Neuper, C., Scherer, R., Reiner, M. and Pfurtscheller, G. (2005b) Imagery of motor actions: differential effects of kinesthetic and visual-motor mode of imagery in single-trial EEG. Cogn. Brain Res., 25: 668–677. Obermaier, B., Neuper, C., Guger, C. and Pfurtscheller, G. (2001) Information transfer rate in a 5-classes brain–computer interface. IEEE Trans. Neural Syst. Rehabil. Eng., 9: 283–288.
Palaniappan, R. and Ravi, K.V.R. (2003) A new method to identify individuals using signals from the brain. Proc. Int. Conf. Inf. Comput. Secur., 1442–1445. Papathanassiou, I., Filipovic, S.R., Whurr, R. and Jahanshahi, M. (2003) Plasticity of motor cortex excitability induced by rehabilitation therapy for writing. Neurology, 61: 977–980. Pfurtscheller, G., Leeb, R., Keinrath, C., Friedmann, D., Neuper, C., Guger, C. and Slater, M. (2006a) Walking from thought. Brain Res., 1971: 145–152. Pfurtscheller, G. and Neuper, C. (1997) Motor imagery activates primary sensorimotor area in humans. Neurosci. Lett., 239: 65–68. Pfurtscheller, G., Scherer, R., Leeb, R., Keinrath, C., Neuper, C., Lee, F., Graimann, B., Schlo¨gl, A. and Bischof, H. (2006b) Viewing Moving Objects in Virtual Reality Can Change the Dynamics of Sensorimotor EEG Rhythms. Presence: Teleoperators and Virtual Environments. MIT Press, Cambridge, MA (in press) Pfurtscheller, G., Wo¨rtz, M., Krausz, G. and Neuper, C. (2001) Distinction of different fingers by the frequency of stimulus induced beta oscillations in the human EEG. Neurosci. Lett., 307: 49–52. Pollock, V.E., Schneider, L.S. and Lyness, S.A. (1991) Reliability of topographic quantitative EEG amplitude in healthy late-middle-aged and elderly subjects. Electroencephalogr. Clin. Neurophysiol., 79: 20–26. Posthuma, D., Neale, M.C., Boomsma, D.I. and de Geus, E.J.C. (2001) Are smarter brains running faster? Heritability of alpha peak frequency, IQ, and their interrelation. Behav. Genet., 31: 567–579. Schaechter, J.D. (2004) Motor rehabilitation and brain plasticity after hemiparetic stroke. Progr. Neurobiol., 73: 61–72. Singer, W. (1993) Synchronization of cortical activity and its putative role in information processing and learning. Annu. Rev. Physiol., 55: 349–374. Stevens, J.A. and Stoykov, M.E. (2003) Using motor imagery in the rehabilitation of hemiparesis. Arch. Phys. Med. Rehabil., 84: 1090–1092. Teasell, R.W. and Kalra, L. (2005) What’s new in stroke rehabilitation. Back to basics. Stroke, 36: 215–217. Thie, L. and Fried, L. (2005) Acclairisms: Questioning Biometric Technology Through an Airport Security Clearance System. ACM CHI, Portland, OR, 2005. Wolpaw, J.R., Birbaumer, N., McFarland, D., Pfurtscheller, G. and Vaughan, T. (2002) Brain–computer interfaces for communication and control. Clin. Neurophysiol., 113: 767–791.