| [page 7↓] |
Since the synchronisation of muscle through the central nervous system had first been demonstrated in humans (McLachlan and Leung, 1991; Farmer et al., 1993) and other primates (Murthy and Fetz, 1992; Murthy and Fetz 1996a; Murthy and Fetz 1996b; Sanes and Donoghue, 1993) about a decade ago frequency analysis of the motor system has increasingly received recognition as a new tool to investigate the human motor system. However, the fact that muscle discharge tends to be rhythmic has been known for almost 200 years. William Wollaston, using a precursor of the stethoscope, was the first to describe this in 1810 (Wollaston, 1810). He determined the rhythm to be in the beta-frequency band by comparing the pitch of the sound picked up over his muscles with that from a horse drawn carriage driven over the cobbled streets of London at different speeds. A century later, the pioneering German neurophysiologist, Hans Piper (Fig. 1), delineated a further modulation of motor unit discharge in the low gamma-frequency band at around 40 Hz (Piper, 1907; Piper, 1912). But only the past decade the has seen steadily growing interest in this field, with specific attention being turned to whether specific patterns of oscillatory drives to muscle may be of pathophysiological and/or diagnostic significance.
Fig. 1: Hans Edmund Piper, German physiologist, born 1877, died 1915. Read biology in Kiel, Munich, Berlin and Freiburg; PhD in Freiburg in 1902. Research assistant at the Institute of Physiology in Berlin, later in Kiel. In 1908 he became head of the department for physics at the Institute of Physiology in Berlin, 1909 promotion to professor. Initially he focussed his research on embryology, his later work encompassed mostly physiological topics, in particular optics, acoustics. the physiology of muscles and nerves and a theory on electrical currents in the retina where he developed the “Piper’s law”. (From: Abeßer, Elke/Schubert, Ernst. Das Berliner Physiologische Institut der Humboldt-Universität. 100 Jahre nach seiner Gründung. Wissenschaftliche Schriftenreihe der Humboldt-Universität zu Berlin. Berlin 1977, p. 29) | ||
|
[page 8↓]This line of inquiry is being taken up in the following experiments both in diseased patients and healthy subjects to further assess the relevance of frequency analysis of the motor system. Thus, this work aims at delineating the usefulness of the technique for diagnostic purposes and at providing some new insights into the pathophysiology of some movement disorders such as cortical myoclonus (Chap. 3), dystonia (Chap. 4), and the myoclonus of corticobasal degeneration (Chap. 5). Particular attention is being paid as to whether intermuscular frequency analysis is an asset to the spectrum of methodologies within the scope of frequency analysis of the motor system. Further, it is explored in healthy subjects whether intermuscular frequency analysis is helpful in assessing physiological subcortical oscillations, such as those elicited by the acoustic startle response, as an innovative means to identify non-invasively subcortcal drives within the motor system (Chap. 6).
Frequency analysis is a useful way of analysing neuronal synchrony and is based on the cross-correlation between two separate signals in the time and frequency domain. The principal measure of the linear dependence or correlation between two signals in the frequency domain is coherence. It is mathematically bounded between zero and one, where one indicates a perfect linear relationship and zero indicates that the two signals are not linearly related at that frequency. Thus oscillatory coupling between motor elements of the central nervous system and EMG discharge is most clearly measured as coherence between the motor cortex and muscles whereas the phase informs on the temporal relationship between the two signals.
The human central nervous system drives muscle discharges at a number of frequencies and, although the physiological function of these oscillations is far from clear (Farmer, 1998a; Brown, 2000; Brown and Marsden, 2001), one of the interests from the clinical point of view is that these different activities may be characteristic of functional activities in distinct circuits. The different physiological oscillatory drives to spinal motoneurones [page 9↓]are summarised in table 1.1.
Table 1.1: Physiological oscillatory drives synchronising motor units in humans
frequency range [Hz] |
origin |
task in which manifest |
detection |
References |
||||
~2 (“common drive”) |
unknown |
isometric contraction, slow movements |
EMG-EMG |
DeLuca and Erim, 1994; Kakuda et al., 1999 |
||||
6-12 |
unknown |
isometric contraction, slow movements |
MEG-EMG, EMG-EMG |
Vallbo and Wessberg, 1993; Conway et al., 1995b; Marsden et al., 2001a |
||||
12-18 |
brainstem |
galvanic stimulation of the inner ear |
EMG-EMG |
Sharott et al. 2003 |
||||
15-30 |
motor cortex |
submaximal voluntary contraction |
MEG-EMG EEG-EMG |
Conway et al., 1995b, Halliday et al., 1998 |
||||
30-60 (“Piper rhythm”) |
motor cortex |
strong voluntary contraction, slow movements |
MEG-EMG |
Brown et al., 1998 |
||||
60-90 |
brainstem |
eye movements |
EMG-EMG |
Brown and Day, 1997; Spauschus et al., 1999 |
||||
60-100 |
brainstem |
respiration |
EMG-EMG |
Carr et al., 1994 |
||||
The first is a low frequency drive at 2-3 Hz, that has been, in retrospect, rather confusingly termed “common drive,” even though there are many such drives (DeLuca et al., 1982). This rhythm can be picked up during isometric contraction or slow movements, even in muscles without muscle spindles (Kamen and DeLuca, 1992; DeLuca and Erim, 1994). The site of its generation is unclear. As it is preserved in patients with cortical or capsular strokes (Farmer et al., 1993) it is not likely to have an origin within the corticospinal system.
Oscillations in the 6-12 Hz range have been related to the pulsatile organisation of slow movements at ~10 Hz (Vallbo and Wessberg, 1993) - identical to physiological action tremor - and to the central component of physiological postural tremor (force tremor) (Conway et al., 1995a) as they prove to be unaffected by alterations of the limb mechanics (Halliday et al., 1999; Vallbo and Wessberg, 1996). The olivary-cerebellar system has been [page 10↓]suggested as a possible generator for the 6-12 Hz oscillations (Llinàs and Pare, 1995) based on findings in the animal harmaline-tremor-model (Llinàs and Volkind, 1973). Consistent with this, some studies have failed to show a cortical correlate at ~10 Hz (Kilner et al., 1999; Mima et al., 2000a). Nevertheless, the exclusivity of the subcortical generation of this drive has been challenged as other studies have detected significant cortico-muscular coherence at this frequency (Mima, 1999; Raethjen et al., 2000a; Salenius et al., 1997a) indicative of sensorimotor cortex involvement. In part, this variability in findings may be accounted for by task-dependency. Thus a recent MEG-EMG study found coherence at 6-12 Hz in force tremor with a source unequivocally originating in the primary motor cortex but no such coherence in action tremor (Marsden et al., 2001a).
In contrast, there is general agreement that motor unit synchronisation in the beta (15-30 Hz) and low gamma (30-60 Hz) bands is predominantly driven from the primary motor cortex, with less influential contributions possibly from supplementary motor and premotor cortices (Feige et al., 2000; Marsden et al., 2000a). Coupling between primary motor cortex and muscle has been demonstrated by both MEG (Conway et al., 1995b; Salenius et al., 1997a; Salenius et al., 1997b; Brown et al., 1998c; Gross et al., 2000) and surface EEG (Halliday et al., 1998; Mima et al., 1998a), although coherence in the gamma band is best seen with the former technique due to the low pass filtering charactaristics of the skull and scalp. Cortico-muscular coherence seems ubiquitous and is even demonstrated by those muscles with small representation in the motor cortex such as the paraspinal and abdominal wall muscles (Murayama et al., 2001). The coherence in the beta band appears during weak tonic contraction and is abolished by movement, whereas that in the gamma band is more obvious in strong contractions and may persist during slow movements (Baker et al., 1997; Brown et al., 1998c; Kilner et al., 1999). Oscillatory drives of motor cortex origin above 60 Hz have also been described through electrocorticographic recordings from the motor cortex (Marsden et al., 2000a) as an indication that the conduction properties of the scull prevents the full range of cortico-muscular coherence from being demonstrated when scalp EEG recordings are used.
Cortical oscillations coupled to motor unit discharge may arise intrinsically within the [page 11↓]cortex or may be under extrinsic, subcortical influence. The intrinsic generation of cortical oscillations may involve pacemaker cells, such as the “chattering cells” (Jefferys et al., 1996; Steriade et al., 1993), which fire rhythmically and may drive neuronal networks (Conners and Amitai, 2001) or result from network properties. The latter include recurrent circuits between excitatory and inhibitory cells and circuits involving the mutual inhibition of inhibitory neurons (Wilson and Bower, 1992; Jefferys et al., 1996).
Striking evidence in favour of a subcortical influence on cortical rhythmicity was initially found non-invasively in patients with Parkinson’s disease through muscle sound recordings using a stethoscope. In untreated patients the normal sound due to the Piper (around 40 Hz) rhythm of muscle was replaced by a 10 Hz rhythm, although the Piper drive could be restored by treatment with levodopa (Brown, 1997a). The implication was that the pattern of cortical drive to muscle was critically dependent on the effects of the basal ganglia on the motor areas of the cerebral cortex. This hypothesis was recently been confirmed by MEG-EMG studies (Salenius et al., 2002).
Further, through EMG-EMG frequency analysis in the striated ocular muscles (Brown and Day, 1997b; Spauschus et al., 1999) and respiratory muscles (Carr et al., 1994) high frequency drives >60 Hz have been idenfied which are of brainstem origin. On the other hand low frequency drives between 12 and 18 Hz of brainstem origin could be demontrasted by using galvanic stimulation of the inner ear (Sharott et al., 2003).
Pathological oscillatory drives manifest themselves either by a shift of the physiological peak(s) and/or by an the inflation of coherence at a given frequency. Usually, both aspects of pathological coherence coincide. In a few disorders of the motor system frequency analysis has already identified the abnormal features of the oscillatory drive from the central nervous system to muscle.
| [page 12↓] |
Frequency analysis shows the most diagnostic potential in cortical myoclonus. To date, the diagnosis of cortical myoclonus has relied on the detection of giant cortical sensory evoked potentials, which are not always present, and of a cortical correlate upon back-averaging (Shibasaki and Kuroiwa, 1975). Frequency analysis may, however, have several advantages over the time domain technique of back-averaging. High frequency myoclonic discharges with low amplitudes, such as in high frequency myoclons (“minipolymyoclonus”, Wilkins et al., 1985), do not preclude analysis as no arbitrary trigger level has to be chosen so that jitter is less, statistical evaluation of the results is possible and the technique is quick and automated, so that long sections of data may be analysed. Thus in a recent study it was possible to demonstrate cortical activity related to myoclonic jerking through frequency analysis in eight patients in whom classical back-averaging failed to show a cortical correlate in five (Brown et al., 1999). Three of the patients in this study also showed exaggerated coherence that encompassed not only the physiological frequency range between 15 and 60 Hz, but also much higher frequencies. This report described patients with large amplitude jerks of low frequency typical of post-anoxic myoclonus and progressive myoclonic epilepsy and ataxia. Recently significant coherence between EEG and EMG has also been reported in high frequency rhythmic myoclonus (Guerrini et al., 2001). Regardless of aetiology, phase spectra confirm that cortical activity precedes EMG by a delay appropriate for conduction in the fast conduction pyramidal pathway. However, it should be noted that occasional exceptions to this rule are met at low frequencies, where the cortical activity lags (Marsden et al., 2000b).
Patients with cortical myoclonus also have exaggerated coherence between ipsilateral muscles co-activated by myoclonic jerks (Brown et al., 1999). Thus, it has been suggested that EMG-EMG coherence analysis can be used as a surrogate marker of coherence between motor cortex and EMG, which will be analysed in chapter 3.
| [page 13↓] |
Cortico-muscular coherence in tremor with maximal coherence at the frequency of the tremor was first demonstrated in parkinsonian rest tremor using MEG (Volkmann et al., 1996). This finding has since been confirmed in studies of MEG/EEG-EMG coherence (Hellwig et al., 2000; Salenius et al., 2002), but the time delays between cortex and muscle are very variable, suggestive of efferent and afferent cortico-muscular drives in different patients (Hellwig et al., 2000). Some of this variability may be explained by the presence of two types of parkinsonian tremor with differing pathophysiological mechanisms (Lance et al., 1963). In higher frequency (7-10 Hz) parkinsonian action tremors cortical signals tend to lead EMG, whereas during low frequency (3-6 Hz) parkinsonian rest tremor EMG activity in the forearm precedes cortical activity, consistent with peripheral re-afference (Volkmann et al., 1996; Salenius et al., 2002).
Findings in essential and exaggerated physiological tremor have been more contradictory. A single channel MEG-EMG study failed to demonstrate cortico-muscular coherence at tremor frequency in essential tremor (Halliday et al., 2000). In contrast, a recent EEG-EMG study with extensive head coverage showed coherence between the contralateral sensorimotor cortex and the tremulous arm (Hellwig et al., 2001). The same authors could not, however, demonstrate EEG-EMG coherence at tremor frequency in enhanced physiological postural tremor although this is at odds with studies on physiological tremor using EMG-EMG coherence analysis in patients with mirror movement (Köster et al., 1998; Mayston et al., 2001) and with MEG-EMG coherence studies in physiological postural tremor (Marsden et al., 2001a).
In summary, there have been conflicting reports of coherence between cortex and tremor and at present EEG-/MEG-EMG coherence studies do not help differentiate different tremor types.
Parkinson’s disease is characterised by a reduction in the normal cortical oscillatory drive to muscles in the beta and gamma band. Instead, in untreated Parkinson’s disease MEG-EMG coherence tends to be at < 10Hz. Such synchronisation of muscle discharge at rest and action tremor frequencies leads to a sub-optimal unfused pattern of muscle activation, thereby slowing the onset of voluntary actions and decreasing contraction strengths (Brown et al., 1998a). Treatment with L-Dopa or therapeutic stimulation of the subthalamic nucleus restores the normal cortical drive and enables cortical motor elements to oscillate at higher frequencies (Salenius et al., 2002; Marsden et al., 2001b). Muscles can then be activated at high frequencies, improving bradykinesia and weakness. Motor cortical elements are also freer to form dynamic patterns of synchronised activity at frequencies above 20 Hz that might be important in higher-order aspects of motor control (Brown and Marsden, 1998b).
Patients with upper limb dystonia show abnormal coherence between extensor carpi radialis and flexor carpi radialis over 1-12 Hz and 14-32 Hz leading to the suggestion that cortical drives may be responsible for the co-contraction of antagonistic muscles in this condition (Farmer et al., 1998b). In contrast, in writer’s cramp the only abnormality was a discrete peak in EMG-EMG coherence at 11-12 Hz when tremor was present.
EMG-EMG frequency analysis has been used to distinguish idiopathic dystonic torticollis from voluntary torticollis in agonistic muscles. Patients with dystonic torticollis exhibit an abnormal synchronised drive in agonistic sternocleidomastoid and splenius capitis muscles between 4 and 7 Hz (Tijssen et al., 2000). The same common 4-7 Hz drive can also be found in complex cervical dystonia (Tijssen et al., 2002).
Transcranial magnetic stimulation and imaging studies have suggested that the ipsilateral motor cortex may show compensatory activity in stroke patients after recovery. Mima et al. explicitly tested this hypothesis in six patients with longstanding subcortical lacunar, pure motor strokes, but failed to find coherence between muscle and ipsilateral motor cortex (Mima et al., 2001b). Coherence between EMG and contralateral EEG was smaller for distal but not proximal muscles on the affected side, in line with the view that pyramidal pathways are differently organised to proximal and distal muscles (Turton and Lemon, 1999; Marsden et al., 1999).
In the future a specific clinical application of frequency analysis in patients with movement disorders treated with deep brain stimulation might be to identify the optimal electrode contact for stimulation. It has recently been shown that the degree of coherence between the local potential picked up by contacts on subthalamic nucleus macroelectrodes and EEG recorded over the midline scalp is correlated with the degree of clinical improvement derived from stimulation at that contact (Marsden et al., 2001b). A comparable finding for coherence between GPi and EEG in dystonia would be particularly useful as stimulation effects may be delayed for many months in this condition.
© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme. | ||
DiML DTD Version 3.0 | Zertifizierter Dokumentenserver der Humboldt-Universität zu Berlin | HTML generated: 14.07.2004 |