[page 49↓]

4.  EMG-EMG-frequency analysis in limb dystonia

Dystonia is a syndrome characterised by prolonged muscle contractions causing sus­tained twisting movement and abnormal mobile and/or fixed postures often along with tre­mor or myoclonus. It may be generalised or focal and is usually classified into primary and secondary causes. Dystonia of the extremities presents a particular problem with major functional limitation and different aetiologies and occurs in patients with primary and se-condary dystonia. Limb dystonia is a hallmark of idiopathic torsion dystonia, with muta­tions in the DYT1 gene representing the commonest genetic basis (Németh, 2002). How­ever, it may also be seen with structural lesions involving the basal ganglia due to trauma, stroke and malignancy (Marsden et al., 1985; Pettigrew and Jankovic, 1985; Kostic et al., 1996; Nardocci et al., 1996) and in the syndrome of fixed dystonia (Marsden et al., 1984; Jankovic and van der Linden, 1988; Bhatia et al., 1993), which can also occur in the con­text of reflex sympathetic dystrophy (Schwartzmann and Kerrigan, 1990) or the causalgia-dystonia syndrome (Bhatia et al., 1993).

Here it was investigated whether the character of EMG discharge may provide a clue to both pathophysiology and diagnosis in patients with dystonia in whom upper and lower limbs are affected. Some EMG features have been previously reported in dystonic patients (Yanagisawa and Goto, 1971), but these accounts remain largely descriptive and provide relatively little insight into pathophysiology or diagnosis. Thus dystonia is associated with sustained EMG activity of co-contracting muscles lasting up to a few seconds. In addition, shorter regular or irregular bursts of muscle activity in the range from 50 to 200 ms can give rise to additional clinical symptoms such as tremor or myoclonus, depending on the frequency, rhythmicity and duration of these bursts (Yanagisawa and Goto, 1971; Obeso et al., 1983; Jedynak et al., 1991). In the case of fixed dystonia, EMG-bursts between 4 and 8 Hz have also been reported (Marsden et al., 1984; Jancovic and van der Linden, 1988).

Recently, attention has focussed on more sophisticated analyses of EMG discharge in dystonia. These have the potential to disclose the character of the descending discharges [page 50↓] responsible for the abnormal muscle activity. For example, frequency analysis can differen­tiate idiopathic dystonic torticollis from voluntary torticollis, as patients with dystonic tor­ticollis exhibit an abnormal synchronised drive to agonistic sternocleidomastoid and sple­nius capitis muscles with a frequency of 4-7 Hz (Tijssen et al., 2000). The same abnormal 4-7 Hz drive has also been reported in patients with complex cervical dystonia (Tijssen et al., 2002) and the report of coherence between pallidal oscillations and dystonic neck mus­cle activity at similar frequencies in a patient with familial myoclonic dystonia (Liu et al., 2002) serves to highlight the saliency of EMG-EMG coupling in this frequency band. Ne­vertheless, the extent to which this abnormal descending drive characterises involuntary dystonic limb contraction remains unkown. In the upper limb, for example, the available evidence points to an abnormal corticomuscular drive in the 15-30 Hz band leading to co-contraction between antagonistic muscles, with the exception of writer’s cramp where a discrete peak in EMG-EMG coherence may be seen at 11-12 Hz (Farmer et al., 1998), while, in the absence of tremor, EMG-EMG coherence between agonist muscles is normal (Cordivari et al., 2002)

The above observations lead to the hypothesis that the nature of the descending drive to muscles in dystonia may vary depending on aetiology and the muscles under consideration. Hence, in this study the pattern of EMG-EMG coherence in the dystonic upper and lower limb are being investigated in patients with dystonia due to a variety of known aetiologies to establish whether an abnormal 4-7 Hz drive is present in any or all of these conditions.

4.1. Patients and methods

4.1.1. Patients

Frequency analysis of EMG from the lower limb was performed in three groups of pa­tients: 12 symptomatic carriers of the DYT1-gene (5 men, 7 women; mean age 41 yrs, range 21-63 yrs; table 4.1), three patients with symptomatic dystonia due to trauma or in­[page 51↓]farct leading to structural brain lesions (3 women, mean age: 36 yrs, range: 28-44 yrs; table 4.2) and 11 patients with the syndrome of fixed dystonia (2 men, 9 women; mean age: 38 yrs, range: 22-61 yrs; table 4.3). The findings were compared to those in 15 healthy sub­jects (6 men, 9 women; mean age: 36 yrs, range: 23-60 yrs). In addition, the results in sym­ptomatic carriers of the DYT1-gene were compared to those in four asymptomatic gene carriers (2 men, 2 women; mean age: 51 yrs, range: 43-57 yrs; table 4.1). To compare cohe­rence patterns between upper and lower extremities six out of the 12 symptomatic DYT1 patients and a further three patients with symptomatic dystonia (giving a total of 1 man, 5 women; mean age: 36yrs, range: 24-53 yrs, table 4.2) were investigated. All subjects gave written informed consent and the study was approved by the Joint Research Ethics Com­mittee of the National Hospital for Neurology and Neurosurgery and the Institute of Neuro­logy.

The 12 symptomatic carriers of the DYT1 gene had generalised dystonia, although in two patients clinical manifestations were exclusively task specific (cases 9 and 10). The four carriers of the DYT1 gene had no clinically apparent involvement (Burke-Fahn-Mars­den scale score = 0; Burke et al., 1985). Some cases were related (cases 2, 3 and 13; cases 6, 9 and 15). The six patients with symptomatic dystonia had cerebral lesions, of whom four had imaging evidence of involvement of the basal ganglia (table 4.2), but only three had major leg involvement. The majority of the 11 patients with fixed dystonia had a hi­story of peripheral trauma and complex regional pain syndrome (table 4.3). All of the latter patients presented with one leg affected, except case 33 in whom all four limbs were in a fixed dystonic posture. Three of these patients had clinically apparent regular or jerky tre­mor. All patients with fixed dystonia were DYT1 gene negative.


[page 52↓]

Table 4.1: Clinical details of the DYT1 gene carriers

Case

Sex

Age

(yrs)

Disease duration (yrs)

Burke-Fahn-Marsden rating scale-score

Drugs

symptomatic

1

f

57

46

40

Clonazepam

2

f

36

11

12

Benzhexol

3

m

34

30

41

Benzhexol, Baclofen

4

f

22

14

40

none

5

m

45

35

45

Botulinum toxin

6

f

21

15

43

Baclofen, Sertraline, Botulinum toxin

7

f

33

22

29

none

8

m

31

15

29

none

9

m

54

~45

6

none

10

f

63

~50

5

none

11

m

41

31

48

Botulinum toxin, Trihexyphenydil

12

f

51

35

37

none

asymptomatic

13

f

56

-

0

none

14

m

57

-

0

none

15

m

49

-

0

none

16

f

43

-

0

none

Table 4.2: Clinical details of patients with symptomatic dystonia

Case

Sex

Age (yrs)

Aetiology

Imaging abnormalities**

Drugs

17*

f

28

perinatal ischemia

head of caudate nucleus, lentiform nucleus (MRI)

Botulinum toxin

18*

f

44

perinatal ischemia

inferior parietal lobe (MRI)

Botulinum toxin

19*

f

37

postpartal ischeamia

globus pallidus, putamen (MRI)

Botulinum toxin

20

m

30

posttraumatic

globus pallidus (MRI)

Botulinum toxin

21

f

53

perinatal

parietal atrophy (CT)

Botulinum toxin

22

f

24

posttraumatic

lentiform nucleus; frontal and parietal white matter (MRI)

Botulinum toxin

*patients in whom both the upper and lower extremities were studied
**MRI signal change unless otherwise stated


[page 53↓]

Table 4.3: Clinical details of patients with fixed dystonia

Case

Sex

Age (yrs)

Disease duration (yrs)

History of peripheral trauma

Drugs

23

f

61

11

yes

Botulinum toxin, Paroxetine

24

m

34

12

yes

Botulinum toxin

25

f

41

5

yes

Botulinum toxin

26

f

46

6

yes

Botulinum toxin

27

f

45

8

yes

Botulinum toxin

28

f

40

6

yes

none

29

f

23

2

yes

none

30

f

45

5

yes

none

31

m

37

7

yes

Botulinum toxin

32

f

22

8

yes

Benzhexol, Gabapentin

33

f

29

14

no

Botulinum toxin, Tetrabenazine, Buprenorphin, Baclofen, Lorazepam, Fluoxetine

4.1.2. Recordings

The core assessment was the determination of EMG-EMG coherence between the pro­ximal and distal parts of TA. TA was chosen as it is a superficial muscle that can easily be recorded using surface electrodes and the resulting EMG-EMG coherence has proven a ro­bust measure (Hansen et al., 2002). Additionally, in eight symptomatic DYT1 patients, five patients with fixed dystonia and five healthy subjects TA EMG was recorded simultaneous­ly with needle EMG from both heads of the ipsilateral GC. GC was chosen instead of the soleus muscle as e.g. in Hansen et al. (2002) because the former is more accessible com­pared to the latter. In the upper limb surface EMG was recorded from triceps, biceps, finger extensor, finger flexor, 1DI and APB.

EMG in TA was recorded using pairs of surface (Ag-Ag, 9mm diameter) electrodes placed 2 cm apart in the horizontal plane on the muscle belly of the proximal and distal portions of the muscle. Electrode pairs were separated by a distance of ~8-10 cm depending [page 54↓]on the subject’s height. To study GC concentric needle-EMG electrodes were placed in the lateral and medial heads of the muscle. For upper extremity muscles surface electrodes were placed at a distance of 2 cm on the muscle belly (except for 1DI and APB where one electrode was sited over the metacarpo-phalangeal joint).

Signals were amplified and digitised with 12-bit resolution by a CED 1401 analogue-to-digital converter. The sampling rate was 5000 Hz. Signals were displayed and stored on a PC by a software package (CED Spike 2, version 4). Needle EMG form GC was levelled at > 50uV, off-line to convert the multi-unit analogue EMG signal to a point process, which was then used for subsequent frequency analysis. Concentric needle recordings and level­ling were performed so as to avoid contamination of GC signals by volume conduction from TA and movement artefact.

To record lower limb EMG subjects were asked to sit on a chair with the hip and knee flexed at an angle of 90° and to perform sustained submaximal dorsiflexion of the ankle joint with the heel on the ground. All healthy controls, asymptomatic DYT1 carriers and most symptomatic DYT1 patients and patients with cerebral lesions were able to perform this task. The remainder and those patients with fixed dystonia found difficulty in perfor­ming the task due to existing back-ground contraction. Similarly, many of the dystonic patients could not exert a maximal voluntary contraction (MVC) of TA that was compara­ble to that in healthy subjects or asymptomatic DYT1 carriers due to additional co-contrac­tion. In addition, there was no certain equivalence of mean rectified EMG voltage between subjects given the very different interference pattern in some subjects. Accordingly, it was not possible to accurately quantify and compare the % of MVC exerted by each subject. EMG autospectra were therefore expressed as a percentage total EMG power prior to com­parison between subjects.

Some patients were asked to perform specific activation tasks (knee extension along with dorsiflexion of the ankle, standing, writing). To record upper limb EMG patients were asked either to write or to perform extension of the elbow, wrist extension and thumb abduction. Both tasks provoked dystonia in the arm.


[page 55↓]

4.1.3.  Analysis

Analysed record lengths were kept constant at 200 s for all muscles and all tasks. Sig­nals were down-sampled to 1 kHz and EMG was rectified. Rectification emphasises tremor peaks in spectra (Fig 1 F) and accentuates firing rate information, thereby improving EMG-EMG coherence estimates (Myers et al., 2003). The discrete Fourier transform and para­meters derived from it were estimated as outlined in chapter II by dividing the records into a number of disjoint sections of equal duration (1024 data points).

Phase was formally assessed only where coherence was significant and extended over at least 5 consecutive data points in the frequency spectrum. The constant time lag between the 2 signals was calculated from the slope of the phase estimate after a line had been fitted by linear regression. The time lag was only calculated from the gradient if a linear relation­ship accounted for (r 2 ) > 71% of the variance (p < 0.05).

4.1.4. Statistics

To compare transformed EMG-EMG coherences between groups a repeated measures General Linear Model was performed using frequency as the main effect. Transformed co­herences were averaged across 4-7 Hz, 8-13 Hz and 14-30 Hz. These bands were selected as they have been associated with dystonia (Tijssen et al., 2000; Liu et al., 2002; Tijssen et al., 2002), physiological tremor (Vallbo and Wessberg, 1993; Halliday et al., 1999; Vallbo and Wessberg, 1996) and corticomuscular drives (Conway et al., 1995; Salenius et al., 1997; Brown et al., 1998; Halliday et al., 1998; Mima et al., 1998b; Gross et al., 2000), res­pectively. Where results were non-spherical, a Greenhouse-Geisser correction was used and when differences were significant at the group level post-hoc pair-wise comparison with Scheffé correction was carried out.


[page 56↓]

4.2.  Results

4.2.1. Clinical appearance and raw EMG in the lower limb

The 12 symptomatic DYT1 patients had varying degrees of clinical lower limb involve­ment. Four patients had periodic leg spasms along with a regular or jerky tremor (jerky because of the appearance of additional non-rhythmic, sometimes myoclonic elements) upon voluntary contraction, five had only a tremor and one patient had leg spasms but without clinically evident tremor. Two patients only had an abnormal gait (Table 1), which was not related to involuntary contractions in the lower leg.

Examples of the raw EMG recorded during active dorsiflexion of the ankle are shown in fig.4.1. Ten out of the 12 (83 %) DYT1 patients with leg involvement exhibited regular bursts in TA that ranged in duration from 50 (Fig. 4.1A) to 200 ms (Fig. 4.1B) in surface EMG. Burst frequency varied from 4 and 7 Hz between subjects. This clinically evident tremor was neurogenic and not due to mechanical resonance phenomena. Simultaneous needle EMG recordings from TA and GC indicated similar rhythmic bursts that were co-contracting with TA activity (Fig. 4.1A). Rhythmic EMG bursting was invariably accompa­nied by a clinically evident regular or jerky tremor in seven out of the 10 DYT1 patients with this EMG pattern when they performed dorsiflexion of the ankle while sitting. In cases 9 and 10 TA EMG showed a rhythmic bursting pattern, but this was only accompa­nied by clinically evident tremor when performing specific activation tasks (extension of the hip and knee, with dorsiflexion of the ankle joint and standing, respectively). In case 7 EMG bursts in TA only occurred when the patient wrote and these were not accompanied by clinically evident tremor. EMG had a normal interference pattern when case 7 per­formed other tasks. The only evidence of dystonia in the lower limb was an abnormal gait due to axial and thigh involvement in cases 11 and 12. Both these patients had a normal interference pattern in TA (Fig. 4.1C).

Conversely, asymptomatic DYT1 carriers, patients with symptomatic dystonia and six of the 11 patients with fixed dystonia had unremarkable EMG interference patterns. Five [page 57↓]patients with fixed dystonia (cases 23, 26, 29, 31, 32) had rhythmic EMG bursts in TA, with a frequency of ~8-10 Hz and an average duration of ~60 ms (Fig. 4.1D). These EMG bursts could be distinguished from those in the symptomatic DYT1 patients by their higher frequency.

Fig. 4.1: Raw EMG recordings of lo­wer leg muscles in different patient groups. Patients sitting with the ankle actively dorsiflexed. (A) and (B) sym­ptomatic DYT1 patients with syn­chronous muscle bursts of 50 (A, case 2) and 200 ms (B, case 3) duration at 4 Hz. Note that in A antagonistic mu­scles (GC, TA) are co-contracting. (C) Symptomatic DYT1 patient (case 11) with no involvement of the lower leg. The interference pattern is nor­mal. (D) Patient with fixed dystonia (case 31) with a ~8 Hz bursting pat­tern. Patient had a fixed posture with the ankle joint plantar-flexed at ~50° and was asked to try to dorsiflex the ankle. All recordings are surface EMGs, except for GC.


[page 58↓]

4.2.2.  Frequency analysis in the lower leg

The results from frequency analysis of TA EMG averaged across the subjects in each group are presented in Fig. 4.2. The power spectra of rectified TA EMG are dominated by a distinct peak at 4-7 Hz, only present among symptomatic DYT1 patients (Fig. 4.2A). This 4-7 Hz peak was present in all cases in whom the lower leg was clinically affected, but not the those cases (11 and 12) without leg involvement. Note that there is a distinct, but smaller peak at 8-10 Hz in patients with fixed dystonia that relates to the high frequency EMG bursting evident in five out of the 11 cases.

The averaged transformed EMG-EMG coherence in TA is illustrated in Fig. 4.2B. Again, the spectra are dominated by a distinct peak at 4-7 Hz, only present among sympto­matic DYT1 patients. The remaining groups, including the healthy controls, show a broad band of increased coherence over 8-30 Hz. That over the 15-30 Hz band is most likely due to rhythmic synchronised discharging in the corticospinal neurones (see Farmer, 1998; Brown, 2000 for reviews). Fig. 4.2C illustrates individual coherence spectra from a sym­ptomatic DYT1 patient. Note that this patient (case 9) only had clinical evidence of leg tre­mor during a specific task (extension of the hip and knee, with dorsiflexion of the ankle joint), which occasionally alternated with spasms without tremor. Nevertheless, an abnor­mal 4-7 Hz peak was seen regardless of task. None of the individual coherence spectra from asymptomatic DYT1 patients, patients with dystonia due to focal cerebral lesions, patients with fixed dystonia nor healthy controls showed a peak in the 4-7 Hz band.

Transformed coherences were entered into a general linear model with frequency band (3 levels: 4-7 Hz, 8-13 Hz and 14-30 Hz) as a main effect and grouped according to disease (5 groups: symptomatic DYT1 carriers, asymptomatic DYT1 carriers, symptomatic dysto­nia with structural lesions, fixed dystonia and healthy controls). There was a significant group x frequency interaction (F[8,80] = 6.566, p < 0.001). Post-hoc tests indicated that the latter interaction was due to the coherence in the 4-7 Hz band which was greater in sympto­matic DYT1 carriers than in patients with symptomatic dystonia and, more importantly, healthy subjects (Fig. 4.2D). In addition, if cases 11 and 12 were excluded (the two patients [page 59↓] whose only lower limb manifestation was an abnormal gait), then coherence in this group was significantly different from any other group including asymptomatic DYT1 carriers and patients with fixed dystonia.

Similarly, when power spectra were entered into a general linear model with the same frequency bands and disease groups there was a significant frequency x group interaction (F[8,170] = 8.133; p < 0.001; not shown). Here, post-hoc tests revealed significant differences between symptomatic DYT1 and asymptomatic DYT1 carriers and healthy sub­jects in the 4-7 Hz band. Again, if cases 11 and 12 were excluded, then DYT1 significantly differed from any other group in the 4-7 Hz band. Further, there was no significant frequen­cy x group interaction between symptomatic DYT1 carriers on medication and those with­out when entered into a separate general linear model.

The time differences between the EMG signals recorded at the rostal and caudal TA electrodes were calculated from phase spectra and are summarised for all subjects in Fig. 4.2E. The estimated delay for conduction between the proximal and distal portion of TA across the subject groups was 2-3 ms. This delay excludes a significant contribution from volume conduction to the observed coherence.

In simultaneous recordings of surface and needle EMG made from TA and GC, respec­tively, all but one of the symptomatic DYT1 patients showed strong coherence between the antagonistic muscles, with a peak at 4-7 Hz. Fig. 3F is a representative example showing a peak at 4-7 Hz that merged with a broader band of activity extending to 20 Hz. Coherence between TA and its partial antagonists over 8-30 Hz can be a normal finding (Hansen et al., 2002), but that at 4-7 Hz is pathological. Cumulant density estimates of TA-GC activity (not shown) confirmed that EMG was co-contracting among the patients. Note that fast vo­luntary alternating plantar and dorsi-flexion of the ankle may reach 5 Hz but is associated with out-of-phase coupling between TA and calf muscles (Hansen, 2002).


[page 60↓]

Fig. 4.2: Frequency analysis in the lower leg muscles. (A) Averaged %-power spectra of proximal and distal aspects of TA show a peak for symptomatic DYT1 patients at 4-7 Hz. (B) Averaged transformed EMG-EMG coherences for TA in the different groups. Note the distinct peak centred at ~5 Hz in the symptomatic DYT1 patients, which is absent in the other groups. (C) Individual coherence spectra in case 9 (symptomatic DYT1). Coherence at 4-7 Hz is present regardless of whether tremor was clinically evident or not. (D) Averaged transformed coherences in the 3 frequency bands. The 4-7 Hz activity was greater in symptomatic DYT1 pa­tients (asterix = p<0.05). (E). Time delay estimates (G) between proximal and distal aspects of TA, confir­ming conduction delay. Bars are 95%-confidence limits. (F) Abnormal coupling between TA and GC with a distinct peak at ~4 Hz in the same patient as Fig. 1A (levelled EMG). Error bars indicate the standard error of the mean.


[page 61↓]

4.2.3.  EMG-EMG Coherence in the upper extremity

The prominence of EMG-EMG coherence at 4-7 Hz in the lower limb of symptomatic patients carrying the DYT1 gene raised the question of whether this activity also dominated in the upper limb. This was studied in six of the symptomatic DYT1 patients. Only two of these patients showed a 4-7 Hz pattern in arm muscles similar to that in their lower limbs. These patients also had peaks in EMG-EMG coherence at 4-7 Hz both for TA and for selected muscle pairs of the upper limb (Fig. 4.3A), but there was no significant coherence between ispilateral upper and lower extremity muscles. Both patients had clinically mani­fest tremor in the upper limbs. The coherence pattern in the remaining four patients dif­fered between upper and lower limbs. Despite EMG-EMG coupling at 4-7 Hz in the lower limb these patients only had irregular low frequency spasms in the upper limb leading to EMG-EMG coherence < 3Hz (Fig. 4.3B). A similar pattern was found in the six patients with symptomatic dystonia due to focal cerebral lesions who had dystonic upper limbs.

Fig. 4.3: Coherence patterns in upper extremity muscles in symptomatic DYT1 patients. (A) Case 1. Inset: re­gu­lar bursting pattern in TA and fore­arm flexors (FF). Coherence spectra confirm exaggerated EMG-EMG co­he­rence in the 4-7 Hz range in upper and lower limb. FE = finger extensor. (B) Case 5. Inset: 7 Hz bursting pat­tern in TA is not reflected in upper ex­tremity muscles. Instead alternating bursts prevail in FF and FE at ~2-3 Hz. Coherence spectra confirm exag­gerated EMG-EMG coherence in the 4-7 Hz range in lower but not upper limb.


[page 62↓]

4.3.  Discussion

It can be shown that different patterns of EMG-EMG coherence prevail in different aetiological groups of patients with dystonic limbs. In addition, the pattern of coupling between muscles may differ between the upper and lower extremities in patients with the same aetiology. Differences in the pattern of synchronisation within and between muscles of the upper and lower limbs have been previously noted in healthy subjects (Nielsen and Kagamihara, 1994; Hansen et al., 2002). In so far as EMG-EMG coherence reflects the common drives to spinal motoneurones these observations imply that the nature of EMG-EMG coherence in dystonia may be constrained by the character of descending motor systems, both in terms of their anatomical distribution and their frequency characteristics.

The most striking finding was an abnormal drive synchronising motor unit discharge at 4-7 Hz in the distal lower limb of symptomatic patients with the DYT1 gene mutation, ir­respective of whether patients were related with each other or not. This drive involved ago­nist and antagonist muscles leading to co-contracting EMG bursts. In most patients the pa­thological drive was manifest clinically as a jerky or regular tremor, though there were in­stances where EMG indicated an abnormal 4-7 Hz drive in the absence of clinical tremor. The advantage of frequency analysis was that it confirmed the homology of the tremor in the tremolous subjects, demonstrated that the same oscillatory drive could be subclinical, permitted quantitative assessment and comparison with other aetiological forms of dystonia and showed that EMG activity in symptomatic DYT1 carriers is dominated by a drive that leads to pathological synchronisation of motor unit discharge at 4-7 Hz. Coherence rather than power spectra were particularly important in demonstrating the latter.

The coherence at 4-7 Hz was not caused by electrical cross-talk. When detected in re­cordings from the proximal and distal aspects of TA it was associated with a phase diffe­rence that was appropriate for axonal conduction. When detected in recordings from TA and GC this was despite the use of concentric needles in GC and levelling of EMG dis­charge to select the largest amplitude EMG activity, both of which would have minimised the effects of cross-talk. In addition, whether detected within TA or between TA and GC, [page 63↓]coherence occurred in a discrete band, whereas cross-talk would have lead to more exten­sively elevated coherence.

The study did, however, have one particular limitation. For a variety of clinical reasons outlined in the Results we were unable to accurately match contraction strengths across patients. To limit the impact of this we compared normalised EMG power spectra. Never­theless, it must be acknowledged that the strength of contraction can change the relative di­stribution of EMG power and also the pattern of EMG-EMG coherence. The best docu­mented change with strong contractions is an increase in EMG activity in the Piper range of 40-60 Hz, due to increased corticomuscular drive in this band (Brown et al., 1998). However, differences in power and coherence in this band were not apparent between groups. Further, synchronisation at 4-7 Hz has not been demonstrated in lower leg muscles in healthy subjects, even when different levels of muscle contraction were assessed (Han­sen et al., 2002). Another way in which task execution may have altered the pattern of EMG is through fatigue. As discussed later this may have contributed to the altered inter­ference pattern in some patients with the syndrome of fixed dystonia, but is unlikely to ac­count for the segmented 4-7 Hz EMG pattern in the lower limb of symptomatic DYT1 carriers which was present when contractions were made from rest.

The abnormal drive was found in over 80 % of symptomatic DYT1 patients and in all of those with leg dystonia if gait abnormalities due to axial and proximal leg involvement were excluded. The 4-7 Hz activity was manifest in raw EMG and coherence spectra whether or not accompanied by a clinically evident regular or jerky tremor. However, it was absent in patients with leg dystonia due to other aetiologies. Together, these factors in­dicate that simple surface EMG recordings from TA may be helpful in suggesting a DYT1 mutation in patients with lower limb dystonia. Of course, a rhythmic synchronising of motor unit discharges at 4-7 Hz is not exclusively seen in dystonia, but may also be seen in Parkinson’s disease.

The abnormal 4-7 Hz drive to the dystonic lower limb in patients with the DYT1 gene mutation is similar to that previously reported in patients with idiopathic dystonic torti­collis (Tijssen et al., 2000; Tijssen et al., 2002). This, together with the coupling between [page 64↓] pallidal oscillations and dystonic neck muscle activity at a similar frequency in a patient with familial myoclonic dystonia (Liu et al., 2002) suggests that there is an abnormal syn­chronising drive in the theta (4-7 Hz) range in dystonia which may involve the globus pal­lidus. But whether this theta drive is exclusive to some types of dystonia rather than sug­gestive of extrapyramidal involvement per se seems unlikely. Strong coherence between activity in globus pallidus internus and arm muscles at tremor frequency (3-6 Hz) has also been reported in patients with Parkinson’s disease (Hurtado et al., 1999; Lemstra et al., 1999), so that the globus pallidus might be generally involved in different movement dis­orders associated with tremor and pathological synchronisation at 4-7 Hz. The implication here is that the 4-7 Hz drive is not necessarily responsible for dystonia, at least by itself, as similar drives may occur without dystonia, as for example in the resting leg tremor of Par­kinson’s disease, and dystonia may occur in the upper limbs of patients with the DYT1 gene in the absence of a 4-7 Hz drive to the affected muscles. Rather, the abnormal 4-7 Hz drive may be one product of basal ganglia, and particularly pallidal, dysfunction that is closely related to the aspect of basal ganglia dysfunction that causes dystonia, but which is not one and the same. Nevertheless, it is likely that the 4-7 Hz drive was closely associated with the development of dystonia in patients with the DYT1 gene as it was absent in asymptomatic carriers and present in nearly all affected carriers. The hypothesised relationship between the theta drive to muscle and the pallidum is particularly interesting given the greater success of pallidal stimulation in DYT1 dystonia compared to dystonia of other aetiologies (Coubes et al, 2000).

The exaggerated synchronising drive to lower leg muscles between 4 and 7 Hz in sym­ptomatic patients with DYT1 dystonia and clinical involvement of the affected leg was dis­tinct from the EMG-EMG coherence over 8-10 Hz evident in the lower leg of healthy sub­jects controls (Hansen et al., 2002) and exaggerated in some of our patients with fixed dys­tonia. The latter exaggeration may simply reflect the fact that this drive becomes more pro­minent with more prolonged contraction of lower leg muscles. Under this condition even healthy subjects may show a sufficiently pronounced burst-like EMG activity with a fre­quency of around 10 Hz that it becomes evident as a clear tremor (Hansen et al., 2002).


[page 65↓]

The finding of an abnormal common drive to spinal motoneurones innervating the lower limb in symptomatic patients with DYT1 but not in other forms of leg dystonia should also be placed in context of the other electrophysiological abnormalities identified in dystonia (see Berardelli et al., 1998 for review). Abnormalities in cortical (Ridding et al., 1995; Iko­ma et al., 1996), brainstem (Berardelli et al., 1985; Tolosa et al., 1988; Nakashima et al., 1989) and spinal inhibition (Panizza et al., 1990; Deuschl et al., 1992) have been found in dystonia, but these too cannot alone be responsible for the abnormal movement pattern as they may be seen outside of the clinically involved area, and, like the 4-7 Hz drive, may not be limited to patients with dystonia (Berardelli et al., 1998). However, these abnormalities of inhibition do not serve to distinguish different aetiological forms of dystonia.

It seems likely that frequency analysis may be able to differentiate between some aetiologies in dystonia. We have shown that symptomatic DYT1 leg dystonia is associated with an abnormal 4-7 Hz drive to spinal motoneurones, that is not present in asymptomatic DYT1 carriers, patients with symptomatic dystonia or patients with fixed dystonia, while Farmer et al. (1998) have previously shown that an 11-12 Hz drive may distinguish some cases of writer’s cramp from symptomatic hemidystonia or primary segmental dystonia in the upper limb. The difference in pathophysiological mechanisms in different aetiologies of dystonia is not limited to the pattern of common drive to muscles. Differences between idiopathic torsion dystonia (which includes DYT1) and symptomatic dystonia have also been reported in some PET studies (Ceballos-Baumann et al. 1995a; Ceballos-Baumann et al. 1995b). Whereas in patients with idiopathic dystonia the sensorimotor cortex is under-activated during voluntary movements, the reverse is seen in those with symptomatic dystonia. Both groups, however, show over-activity of other frontal lobe regions. Further studies are required to define pathophysiological differences between different dystonic aetiologies, especially when phenotypes may be so similar.


© 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 3.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated:
14.07.2004