What is the Bereitschaftspotential?

rits i Human Motor Control Section, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, MD 20892-1428, USA Accepted 28 April 2006 Keywords: BP, Bereitschaftspotential; Pre-movement slow negativity; Early BP; Late BP (NS 0); Conscious will to move ments at a self-paced rate, and stored all the data on mag- netic tape. Then they made an off-line averaging of the EEG segment prior to the EMG onset by playing the tape backward. By using this chronologically reversed averaging * Corresponding author. Present address: Takeda General Hospital, Ishida, Fushimi-ku, Kyoto 601-1495, Japan. Tel.: +1 81 75 572 6468; fax: +1 81 75 571 8877. E-mail address: shib@kuhp.kyoto-u.ac.jp (H. Shibasaki). Clinical Neurophysiology 11. Introduction Kornhuber and Deecke (1964) made the first report of electroencephalographic (EEG) activity preceding volition- al movement in humans. Prior to that, Bates (1951) attempted to record the movement-related activity by photographic superimposition of multiple single-sweep EEG traces, but he could identify only the post-movement activity probably due to low signal-to-noise ratio. In the 1960’s, no computer software for making on-line back averaging was available. Therefore, Kornhuber and Dee- cke (1964) recorded EEG and electromyogram (EMG) simultaneously while the subjects were repeating move-Available online 28 July 2006 Abstract Since discovery of the slow negative electroencephalographic (EEG) activity preceding self-initiated movement by Kornhuber and Deecke [Kornhuber HH, Deecke L. Hirnpotentiala¨nderungen bei Willkurbewegungen und passiven Bewegungen des Menschen: Bereitschaftspotential und reafferente Potentiale. Pflugers Archiv 1965;284:1–17], various source localization techniques in normal subjects and epicortical recording in epilepsy patients have disclosed the generator mechanisms of each identifiable component of movement-related cortical potentials (MRCPs) to some extent. The initial slow segment of BP, called ‘early BP’ in this article, begins about 2 s before the movement onset in the pre-supplementary motor area (pre-SMA) with no site-specificity and in the SMA proper according to the somatotopic organization, and shortly thereafter in the lateral premotor cortex bilaterally with relatively clear somatotopy. About 400 ms before the movement onset, the steeper negative slope, called ‘late BP’ in this article (also referred to as NS 0), occurs in the contralateral primary motor cortex (M1) and lateral premotor cortex with precise somatotopy. These two phases of BP are differentially influenced by various factors, especially by complexity of the movement which enhances only the late BP. Event-related desynchronization (ERD) of beta frequency EEG band before self-initiated movements shows a different temporospatial pattern from that of the BP, suggesting different neuronal mechanisms for the two. BP has been applied for investigating pathophysi- ology of various movement disorders. Volitional motor inhibition or muscle relaxation is preceded by BP quite similar to that preceding voluntary muscle contraction. Since BP of typical waveforms and temporospatial pattern does not occur before organic involuntary movements, BP is used for detecting the participation of the ‘voluntary motor system’ in the generation of apparently involuntary movements in patients with psychogenic movement disorders. In view of Libet et al.’s report [Libet B, Gleason CA, Wright EW, Pearl DK. Time of conscious intention to act in relation to onset of cerebral activity (readiness-potential). The unconscious initiation of a freely voluntary act. Brain 1983;106:623–642] that the awareness of intention to move occurred much later than the onset of BP, the early BP might reflect, physiologically, slowly increasing cortical excitability and, behaviorally, subconscious readiness for the forthcoming movement. Whether the late BP reflects conscious preparation for intended movement or not remains to be clarified. � 2006 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.Invited What is the Bere Hiroshi Shibasak1388-2457/$32.00 � 2006 International Federation of Clinical Neurophysiolo doi:10.1016/j.clinph.2006.04.025eview chaftspotential? *, Mark Hallett www.elsevier.com/locate/clinph 17 (2006) 2341–2356gy. Published by Elsevier Ireland Ltd. All rights reserved.

technique, they successfully identified two components, one each before and after the EMG onset. Those were the Ber- eitschaftspotential (BP) or readiness potential (RP), and reafferente Potentiale (Kornhuber and Deecke, 1964, 1965). Later they found two more components just before the movement onset: pre-motion positivity (PMP) and motor potential (MP) (Deecke et al., 1969). Since then, a number of studies on the movement-related cortical poten- tials (MRCP) have been reported both in terms of physio- logical findings and clinical application, but the physiological significance of each identifiable component, peak time of each component with respect to the movemen onset obviously differs depending on how to define the movement onset. In case of finger movements, the onse of the mechanogram is about the same time as the EMG peak, but would lag the EMG onset by about 30 ms. These days it is most common to use EMG onset as the fiducia point. BP starts about 2.0 s before the movement onset. It is maximal at the midline centro-parietal area, and symmet rically and widely distributed over the scalp regardless o the site of movement. The onset of BP with respect to the movement onset significantly differs among different con ditions of movement and among subjects. For example in the experimental setting in which the subject is request ed to repeat the same movement at a self-paced rate o once every 5 s or longer, the BP commonly starts much Post-movement components MP RAP N2 (?) N2 (?) P2 N�10 N+50 P+90 N+160 P+300 isMP ppMP fpMP MF MEFI MEFII MEFIII PMF MP PMPP MEPI MEPI N-10 (MP) N+50 (fpMP) Early BP Late BP 2342 H. Shibasaki, M. Hallett / Clinical Neurophysiology 117 (2006) 2341–2356among others that of BP, has not been fully clarified yet. Libet et al. (1983), by employing their novel technique in which the subjects were requested to remember the time of their actual awareness of intention to move by watching a clock, reported that the intention to move occurred much later than the onset of BP. Their report has brought up a continuing question as to the physiological implication of the BP (Klein, 2002; Eagleman, 2004). In 2003, a compre- hensive book entirely devoted to ‘‘The Bereitschaftspoten- tial’’ was published (Jahanshahi and Hallett, 2003). Since, in that book, each specific aspect of BP was discussed in detail, it seems now appropriate to integrate various aspects of the BP by further updating more recent findings with special emphasis placed on the information obtained by epicortical recording, the issue of voluntary motor inhi- bition, praxis movement, and the physiological implication of the BP. 2. Components of MRCP Different terminologies have been proposed for identifi- able components of MRCP (Table 1). Shibasaki et al. (1980a), based on the scalp distribution of averaged data across 14 subjects, identified 8 components, 4 each before and after the movement onset (BP, NS 0, P�50, N�10, N+50, P+90, N+160 and P+300 for finger movements) (Fig. 1). In this terminology, each component, except for BP and NS 0, was named according to the surface polarity (P, positive; N, negative) and the mean time interval in ms between the peak of each component and the peak of the averaged, rectified EMG, the interval being designated negative if the peak occurred before the EMG peak, and positive if it occurred after the EMG peak (Table 1). The Table 1 Terminology of movement-related cortical potentials Pre-movement components Kornhuber and Deecke (1965) BP PMP Vaughan et al. (1968) N1 P1 Shibasaki et al. (1980a)a BP NS 0 P�50 Dick et al. (1989) NS1 NS2 Lang et al. (1991) BP1 BP2 Tarkka and Hallett (1991) BP NS 0 PMP Kristeva et al. (1991)b RF Cui and Deecke (1999) BP1 BP2a Peak of each component, except for BP and NS0, was measured from the peak of averaged, rectified EMG. b Based on movement-related magnetic fields.t t l - f - , - f IFig. 1. Waveforms and terminology of movement-related cortical poten- tials (MRCPs) from a single normal subject. Self-initiated left wrist extension. Average of 98 trials. Reference (Ref): linked ear electrodes (A1A2). Early pre-movement negativity (early BP) starts 1.7 s before the onset of the averaged, rectified EMG of the left wrist extensor muscle, and is maximal at the midline central electrode (Cz) and widely and symmetrically distributed on both hemispheres. Later negative slope (late BP) starts 300 ms before the EMG onset and is much larger over the right central region (contralateral to the movement). A negative peak localized at the contralateral central area (C2) is N�10 or motor potential (MP). Another negative peak occurring shortly after N�10 is localized over the midline frontal region and corresponds to N+50 or the frontal peak of motor potential (fpMP).

eurearlier as compared to the movement executed in more natural conditions, because, in such experimental condi- tions, the subject has a much longer time to prepare for the movement. As pointed out originally by Kornhuber and Deecke (1964, 1965) and in their subsequent reports (Deecke et al., 1969, 1976), and later by Kutas and Don- chin (1980), BP suddenly increases its gradient about 400 ms before the movement onset. Based on the clearly different scalp distribution of the late steeper slope from that of the early slow shift, Shibasaki et al. (1980a) desig- nated the late segment as Negative Slope (NS 0). The NS 0 was distinguished from the early BP based on abrupt increase of the gradient at the central electrode corre- sponding to the movement for each individual subject, instead of arbitrarily setting the time such as 500 ms before the movement onset for the distinction of the two slopes. In order to avoid confusion about the use of the term BP, however, in this review article we call the early segment ‘early BP’ and the late segment ‘late BP’, and just BP for the early BP and the late BP inclu- sive (Fig. 1). The late BP is maximal over the contralater- al central area (approximately C1 or C2 of the International 10–20 System) for the hand movement and at the midline (approximately Cz) for the foot movement (Shibasaki et al., 1981). For the study of BP in individual subjects, therefore, it is important to record EEG from multiple electrodes, including C1 and C2 for the study of hand movement, for identifying the abrupt increase of the gradient. Later, the late steeper slope was called NS2 to contrast it with the earlier NS1 by Dick et al. (1989), and BP2 to differentiate it from the earlier BP1 by the group of Deecke (Lang et al., 1991; Cui and Deecke, 1999). Initially, the late BP was thought to be more specific for the site of movement while the early BP was thought to represent the more general preparation for the forthcoming movement because of its diffuse distribution. However, as will be discussed later, the early BP might be also movement site-specific at least within the supplementary motor area (SMA) and the lateral pre- motor cortex. Its midline maximal, symmetric distribution is most likely due to the summation of electrical fields generated from homologous areas of both hemispheres via volume conduction. The asymmetric distribution of the late BP associated with unilateral hand movement has been studied as the lateralized readiness potential (LRP) (Coles et al., 1988). It is derived by subtracting the potential recorded at C4 from that at C3 for both the left-hand movement and the right-hand movement separately. In the field of behavioral psychology, LRP has been mainly obtained in a choice reaction time task, instead of a self-paced motor task, in order to investigate the time relationship between the stimulus evaluation system and the response activation sys- tem. For example, if the former can pass information to the latter before evaluation is completed, it is considered to sug- H. Shibasaki, M. Hallett / Clinical Ngest the evidence for ‘early communication’. Hence, the details of LRP will not be dealt with in this review article.PMP or P�50 is predominant over the hemisphere ipsi- lateral to the moving hand (Deecke et al., 1969; Shibasaki et al., 1980a). Based on the fact that this component was not seen with bilateral simultaneous hand movements, Shibasaki and Kato (1975) proposed that it might be relat- ed to the suppression of movement of the opposite hand in intended unilateral hand movement (suppression of the physiological mirror movements). However, PMP might be just a trough formed between two successive negative potentials, and this peak itself might not have any physio- logical significance. In fact, P�50 has not been identified by epicortical recording. As for MP or N�10, it is well localized to a small area of the contralateral central scalp, precisely corresponding to the movement site, and occurs immediately before the movement onset. N�10 in hand movement was named based on the mean time interval from its peak to the peak of the averaged, rectified EMG; the latter lagged the former by 10 ms. Thus, this component most likely represents the activity of pyramidal tract neurons in the primary motor cortex (M1). The isMP (initial slope of MP) as named by Tarkka and Hallett (1991) corresponds to the early part of MP (Table 1). As for the post-movement components of MRCP asso- ciated with hand movement, Shibasaki et al. (1980a) named the four peaks based on the time interval mea- sured from the peak of the averaged, rectified EMG to the peak of each identifiable peak (Table 1). The compo- nent N+50 is a prominent negative peak localized to the frontal region and corresponds to fpMP (frontal peak of motor potential) of Tarkka and Hallett (1991) (Fig. 1). The component P+90 is predominant over the parietal region, larger over the contralateral hemisphere. The peaks of N+50 and P+90 showed similar scalp distribu- tion to those of N70 and P65, respectively, of the aver- aged EEG responses to the passive finger movements, suggesting that these peaks are related to kinesthetic feedback (Shibasaki et al., 1980b). However, due to the different peak time after the movement onset between N+50 and P+90, it is still undetermined whether or not these two peaks are derived from a tangentially oriented dipole sitting in the postcentral gyrus. The component N+160 is localized to the contralateral parietal area, thus forming a localized positive–negative complex with P+90. The component P+300 corresponds to reafferente Potenti- ale of Kornhuber and Deecke (1964). Based on the recording of movement-related magnetic fields, BP, MP and four post-movement components were named readiness field (RF), motor field (MF) and move- ment-evoked field (MEF) I, II and III and post-movement field (PMF), respectively, by Kristeva et al. (1991) (Table 1). This review article will focus on the pre-movement com- ponents, especially the slow negative shifts (early BP and late BP) recorded under the self-paced conditions, and ophysiology 117 (2006) 2341–2356 2343those recorded under the cued condition will not be covered.

3. Factors influencing BP The magnitude and time course of BP recorded in the self-paced condition are influenced by various factors such as level of intention, preparatory state, movement selection as to freely selected versus fixed, learning and skill acquisition, pace of movement repetition, praxis move- ment, perceived effort, force exerted, speed and precision of movement, discreteness and complexity of movement, and pathological lesions of various brain structures. Since this issue was extensively reviewed by Lang (2003), this review article refers to more recent findings and summarizes the current consensus on this issue in Table 2. As regards the force exerted in isometric muscle contrac- tion, Slobounov et al. (2004) found a larger amplitude of the last 100 ms segment of BP with greater perceived effort rather than the force itself, while other segments were not influenced. Since the exertion of stronger force is common- ly associated with the greater intention, motivation and 2344 H. Shibasaki, M. Hallett / Clinical Neureffort of the subject, the effects of these psychological and physical factors cannot be clearly distinguished from each other. Masaki et al. (1998), by comparing a target force production task with simple task, showed that the move- ment requiring precision in terms of force production was found to be preceded by a larger late BP than the simple task. The speed of movement also affects the onset and magnitude of BP; the faster the movement is executed, the later (closer to the movement onset) the BP begins. As for the site of movement, Jankelowitz and Colebatch (2002) found larger late BP in self-paced movement of the proximal than the distal part of upper extremity. The effect of discreteness and complexity of the move- ment on BP is noteworthy. Kitamura et al. (1993a) com- pared isolated extension of the middle finger with simultaneous extension of the middle and index fingers, Table 2 Differential influence of various factors on early and late BP in normal and pathological conditions Factors Early BP Late BP Level of intention Largera Preparatory state Earlier onseta Movement selection Larger No effect Learning Larger during learninga Praxis movement Start parietallya Force Largera Speed Later onseta Precision No effect Larger Discreteness No effect Larger Complexity No effect Larger Parkinsonism Small No change Cerebellar lesion Small Small Dystonia No change Small Hemiparesis recovery No change Involved Mirror movement No change Involved As for the factors in normal conditions, the effect is shown in a compar- ative form; the greater the factors, the larger or smaller the BP. a Late BP not clearly distinguished.and found larger amplitude of late BP, but not early BP, in the isolated movement of the middle finger as compared with the simultaneous movement of the two fingers. In spite of the fact that a larger number of muscles were involved in the two finger movement than in the single fin- ger movement, the late BP was larger in the latter, indicat- ing the importance of discreteness of the movement. In this case, the amplitude difference of late BP was seen only over the central region contralateral to the movement, suggest- ing the important role of M1 (see Section 4). This finding is consistent with the clinical notion that a fundamental symptom of the pyramidal tract lesion is impairment of fine finger movement. In addition to the factor of discreteness, active inhibition of one finger while moving the other finger might add cortical activity related to active motor inhibi- tion (Section 6). Another possibility is greater complexity of the isolated middle finger movement as compared with the simultaneous movement of two fingers. This is unlikely, however, because the amplitude difference was not seen over the midline central region where a complex movement would be expected to generate a larger amplitude of BP due to greater activation of SMA as compared to simple move- ment. The issue of complexity will be discussed further in the next paragraph. As for the complexity of the movement, Benecke et al. (1985) found larger BP before the sequential or simulta- neous performance of isotonic elbow flexion and isometric finger flexion than before either the isotonic elbow flexion or the isometric finger flexion alone. Simonetta et al. (1991) compared a simple movement with a motor sequence starting with the simple movement, and found earlier onset and larger amplitude of BP in the sequential movement than in the simple movement. In these two experiments, however, a larger number of muscles were involved in the complex movement than in the simple movement. In contrast, Kitamura et al. (1993b) compared the middle followed by index finger movement as a com- plex movement with the simultaneous movement of middle and index fingers as a simple movement, so that the mus- cles involved in the two conditions were matched. Further- more, the subjects were trained so that the duration of EMG discharge was the same in the two conditions. As the result, they found larger amplitude of late BP, not of early BP, in the sequential movement as compared with the simultaneous movement. Since this amplitude differ- ence was seen over the midline vertex as well as bilateral central regions, it was postulated that not only SMA but also bilateral sensorimotor cortices might play an impor- tant role in the preparation of the complex movement. This electrophysiological finding was later supported by the cerebral blood flow activation study with position emission tomography (PET) (Shibasaki et al., 1993). This observa- tion suggests that, whenever BP is employed for studying motor control mechanisms or its abnormality in patholog- ical conditions, the effect of task complexity or subjective ophysiology 117 (2006) 2341–2356difficulty to the subjects has to be taken into consideration as an important influencing factor.

ed p tom as pe t sid eurSo far, most studies of MRCPs have employed simple motor tasks such as finger extension, wrist extension, fist clenching, elbow movements, foot extension and tongue protrusion, which are quite different from the practical movements performed under more natural conditions. Fig. 2. MRCPs recorded from scalp electrodes in association with self-pac average across 8 subjects). Reference: linked ear electrodes. Both for the pan in black), BP starts at the superior parietal region predominantly on the left parietal region. About 1 s before the movement onset, a sharp negative slo predominantly on the left. Data of the left hemisphere shown on the lef permission]. H. Shibasaki, M. Hallett / Clinical NWheaton et al. (2005a) recently studied movements employed in daily life, or the so-called praxis movements, by recording MRCPs associated with pantomiming of common gestures or tool use in normal right-handed sub- jects. Under these conditions, BP was found to start at the parietal region, larger over the left, and then followed by the BP over the midline frontal region and bilateral cen- tral regions (Fig. 2). The initial activation of the left parie- tal cortex during preparation for praxis movements in the right-handed subjects is consistent with the clinical notion that ideomotor apraxia is caused by the lesion involving the left parietal lobe; in particular, the supramarginal gyrus and its projection fibers to the frontal lobe. 4. Generator sources of MRCP Various dipole source localization techniques have been applied to estimate the generator sources of MRCPs (Dee- cke and Kornhuber, 2003). In the case of hand movements, SMA and lateral precentral gyrus, both bilaterally, were estimated to be the main generator sources for early BP. Praamstra et al. (1996) estimated three dipole sources for explaining the early BP; one in the SMA and two others in bilateral M1. They further showed that only the current source identified in the SMA was affected by the mode of movement selection; larger before freely selected than fixed movement. Cui and Deecke (1999), based on the high-res-olution DC-EEG analysis, demonstrated that BP occurs earliest in the medial wall motor areas (SMA and cingulate motor areas), then in the contralateral M1, and lastly in the ipsilateral M1. Toma et al. (2002), by using principal component analysis and functional magnetic resonance raxis movements of the right hand in right-handed normal subjects (grand ime of tool use (transitive; shown in grey) and gesture (intransitive; shown early as 4 s before the movement onset, and then develops over the inferior is seen over the midline frontal region and bilateral sensorimotor regions e. *0.05 < p < 0.10, **p < 0.05. [cited from Wheaton et al. (2005a) with ophysiology 117 (2006) 2341–2356 2345imaging (fMRI)-constrained EEG dipole source analysis, determined the crown of the precentral gyrus bilaterally, specifically the hand area of area 6, as the main source of early BP, area 4 in addition to area 6 as the source of late BP, and area 3 as the source of fpMP or N+50. Most studies, including that of Toma et al., have localized the source of MP or N�10 in the M1 hand area. Deecke et al. (1982) reported the first MEG correlates of BP. They attributed BP to two current dipoles located in the SMA and the contralateral M1. Nagamine et al. (1996), by using the whole head MEG system, reported that the scalp distribution of magnetic field corresponding temporally to EEG BP is different from that of EEG BP, most likely due to the fact that MEG picks up only the dipole sources tangentially oriented to the head surface while EEG records both the radially and tangentially ori- ented dipole sources. The slow pre-movement shift on MEG (RF) begins much later than that of EEG and is dis- tributed almost exclusively over the contralateral central region (Fig. 3). This finding can be explained by postulat- ing that, at least as far as the lateral hemispheric convexity is concerned, early BP arises in the lateral premotor corti- ces (area 6) bilaterally, which is recorded by EEG but not by MEG because of its radial orientation, and late BP occurs in the contralateral M1 (area 4), which is recorded not only by EEG but also by MEG due to its tangential orientation (Nagamine et al., 1996). Furthermore, MEG