Dystonia is an involuntary movement disorder characterized by abnormal sustained or intermittent muscle contractions. ^{Reference Albanese, Bhatia and Bressman1,Reference Jinnah and Factor2} It is characterized by prolonged muscle contraction, co-contraction, and overflow. For a long time, dystonia was considered a psychiatric disorder. However, the evidence in the last thirty years indicates that dystonia most likely results from the dysfunction of a network, including basal ganglia, cerebellum, thalamus, and sensorimotor cortex. ^{Reference Jinnah, Neychev and Hess3,Reference Balint, Mencacci and Valente4} In the pathophysiology of dystonia, a defect in the striato-pallido-thalamic pathway causing disinhibition of excitatory thalamocortical neurons ^{Reference Hallett5} and sensory dysfunction, decreased inhibition, and abnormal plasticity ^{Reference Quartarone and Hallett6} have been suggested. While these findings are shared by focal dystonia and generalized dystonia, what is the reason for the variations in clinical phenomenology? Latorre et al. ^{Reference Latorre, Rocchi and Bhatia7} indicated that the various forms of dystonia (e.g., focal, generalized, or task-specific) probably reflect derangements at different levels of the network and may have an electrophysiological signature. For example, Sabbahi et al., using soleus H-reflex measures, identified neurophysiologic differences between generalized dystonia, cervical dystonia, and normal subjects. ^{Reference Sabbahi, Etnyre and Al-Jawayed8} Or, the recovery cycle of the blink reflex assessing brainstem excitability was abnormal due to reduced inhibition of the R2 component, mainly in patients with blepharospasm ^{Reference Berardelli, Rothwell, Day and Marsden9,Reference Schwingenschuh, Katschnig and Edwards10} and also in patients with segmental/generalized dystonia or torticollis, but not in those with focal arm dystonia, suggesting proximity is critical in dysfunction and one plausible explanation may be that each dystonia subtype might have an electrophysiological signature.

In this study, we hypothesized that “long latency reflexes” (LLRs), associated segmental reflex (SR), and mixed nerve silent periods (MnSPs) recorded on the distal upper extremity muscles may represent one of the electrophysiological signatures of the focal hand dystonia. Stimulation of a mixed nerve while recording on a distal hand muscle during slight contraction creates an M-response due to direct excitation of the motor axons, an SR (probably an H-reflex) mediated through Ia afferents, and up to three subsequent LLRs, I, II, and III. ^{Reference Deuschl and Hallett11} One of the most critical applications of testing LLRs is diagnosing and classifying myoclonus. ^{Reference Deuschl and Eisen12} In dystonia, one study reported high-amplitude LLR I and reduced or absent LLR II. ^{Reference Naumann and Reiners13} The MnSPs are electrophysiological inhibition periods mediated by descending volleys and activated by high-threshold cutaneous fibers of the mixed nerve. ^{Reference Leis14,Reference Leis, Ross and Emori15}

This was a cross-sectional study. We included ten patients with cervical dystonia, 12 with generalized dystonia, and four with focal hand dystonia who were admitted to our movement disorder outpatient clinic between January 2018 and January 2020. A control group of 38 healthy volunteers was also recruited. The gender and age of all participants were noted (Table 1). There was a significant age difference between groups due to the natural history of patients with focal hand dystonia and cervical dystonia. Patients with cervical dystonia were evaluated using the Toronto Western Spasmodic Torticollis Rating Scale (TWSTRS), and those with generalized dystonia or focal dystonia were assessed using the Burke-Fahn-Marsden dystonia rating scale. The institutional review board approved the study. We received informed consent to perform the analysis.

Table 1: Demographic findings of patients and healthy subjects

LLR = long latency reflex; MnSP = mixed nerve silent period; SR = segmental reflex.

We recorded LLRs, associated SR, and MnSPs in all participants. All studies were performed with a Neuropack Sigma MEB-5504k (Nihon Kohden Medical, Tokyo, Japan). All patients in the study had botulinum toxin treatment previously. However, we performed the electrophysiological investigations at least six months after the last botulinum toxin injections. The recordings were done according to the previously published methods. ^{Reference Deuschl and Hallett11,Reference Deuschl, Seifert and Heinen16} All measurements were performed for the symptomatic extremity in the focal hand dystonia group. The recordings were done on the right side for other patient groups and healthy subjects. We used a square wave electrical stimulus at 0.2 ms duration at 2 Hz. We have collected 20 trials for each analysis.

1. 1. Long latency reflexes and segmental reflexes: We placed surface silver-silver chloride recording electrodes on the belly of the abductor pollicis brevis (APB) muscle. We stimulated the median nerve at the wrist. The SR and LLRs (LLRI, LLRII, LLRIII) were recorded at rest while the subjects performed a mild contraction of the APB muscle (approximately 25% of the maximum). Both auditory and visual feedback maintained the level of contraction. Figure 1 shows representative examples of SR and LLRs.

2. 2. Mixed nerve silent period: The MnSPs were measured while the recording electrodes were still on the APB muscle. The median nerve was stimulated with an electrical stimulus 25% above the supramaximal stimulus intensity threshold, producing a motor response at the wrist level while the subjects were performing a mild contraction of the APB muscle.

Figure 1: a . M-response (black arrow), segmental reflex (broad red arrow), and long latency reflex (thin red arrow) in a 26-year-old male patient with cervical dystonia. 1 b . Segmental reflex (broad red arrow) and long latency reflex (thin red arrow) in a 34-year-old female patient with generalized dystonia (gain 0.5 mV/div).

The onset latency of LLRs (distance from electrical stimulus artifact until the first negative deflection after segmental reflex) and the amplitude LLRs (peak-to-peak) were measured. The LLRs were classified into three groups based on Deuschl 1999: ^{Reference Deuschl and Eisen12} 35-46 ms for LLR I, 45-58 ms for LLR II, and > 68 ms for LLR III.

The minimum latency and the maximum amplitude of the segmental reflex were also measured. We calculated the percentage of presence of each wave as follows: the number of participants with LLRx100/ the total number of participants in the specific group. For MnSPs, the mean end latency was measured. Data were presented as mean ± standard deviation (SD) or percentages (n). We used multivariate analysis. Fixed factors were phenotypes (cervical dystonia, generalized dystonia, focal hand dystonia, healthy subjects) and age (<20 years, 20–40 years, >40 years) because there was a significant age difference between groups. Post hoc analysis was done using the Bonferroni test. The chi-square test has been used for the comparison of qualitative data. The data analysis was done using the SPSS 20.0 statistical package, and a p-value ≤ 0.05 was considered significant.

The mean TWSTRS scores of cervical dystonia patients were 21.0 ± 5.5 points. The neck and upper extremities were involved in seven patients with generalized dystonia. In other patients, axial muscles and lower extremities were also involved. The mean Burke-Fahn-Marsden score was 46.7 ± 9.1 points. The patients with focal hand dystonia did not have only task-specific symptoms and had symptoms provoked by many actions. The signs were mild in patients with focal hand dystonia.

The latencies of the SRs were similar between the diagnostic groups (p = 0.099). There was a significant difference in amplitudes of the SRs according to phenotypes (F = 2.248, p = 0.045, Table 2); however, not according to the age groups (F = 1.181, p = 0.324). Post hoc analysis showed that the mean amplitude of segmental reflexes was significantly lower on the symptomatic side of the focal dystonia group than the healthy subjects (p = 0.010). In the other groups, it was similar.

Table 2: Electrophysiological findings of patients and healthy subjects

LLR = long latency reflex; MnSP = mixed nerve silent period; SR = segmental reflex.

None of the participants in any of the groups had C reflex, i.e., LLR response at rest. Regarding the responses obtained during contraction, LLR II (69.2%) was commonly seen, followed by LLR III (23.1%) and LLR I (15.4%) among all patients with dystonia. The figures were very similar to those among healthy subjects. Comparing the whole group of dystonia patients with healthy subjects, the LLR amplitudes were not statistically different (p = 0.831).

The LLR I (during muscle activity) was obtained in two (20%) patients with cervical dystonia and two (16%) patients with generalized dystonia. None of the patients from the focal extremity dystonia group had an LLR I response, whereas there was an LLR I response in four (10.5%) of 38 healthy individuals (p = 0.703). The LLR II was obtained in eight (80%) patients with cervical dystonia, eight patients with generalized dystonia (66%), and two (50%) patients with focal hand dystonia, and there was LLR II response in 31(81%) healthy individuals (p = 0.419). The LLR III response was obtained from three out of 10 (30%) patients with cervical dystonia, two (16%) patients with generalized dystonia, one (25%) patient with focal extremity dystonia, and 10 (26%) healthy individuals (p = 0.893). The amplitude of LLRs was not different among patients with different phenotypes (F = 0.352, p = 0.788) or according to other age groups (F = 0.833, p = 0.484). Table 2 shows all the electrophysiological findings. The mean end latencies of MnSPs were not different among groups (F = 0.097, p = 0.961).

The significant finding in this study was the smaller SR amplitude in the group with focal hand dystonia and no change in LLR or SR in generalized dystonia.

Several physiological conditions modulate the SR. The H-reflex changes under certain pathological conditions. For example, the soleus H-reflex was suppressed throughout all phases of the contralateral rhythmic ankle movement. ^{Reference Mori, Horino and Matsugi17} Again, the amplitude of the H-reflex of the flexor carpi radialis muscle increases in association with teeth clenching even before the onset of the EMG activity of the masseter muscle. ^{Reference Takahashi, Ueno and Ohyama18} Soleus H-reflex is inhibited during gait in Parkinson’s disease, ^{Reference Hiraoka, Matsuo and Abe19} whereas the amplitudes of H-reflexes are increased in spasticity. ^{Reference Voerman, Gregoric and Hermens20} The excitability of H-reflex depends on the excitability of the lower and upper motor neuron pools, and basal ganglia, cerebellum, or sleep states modulate it. ^{Reference Matsugi, Mori and Uehara21,Reference Brylowski, Levitan and LaBerge22} In our study, the amplitude of the SRs was low only in the group with focal hand dystonia, suggesting reduced excitability in this motoneuron pool. Regarding H-reflex, there was no change in its amplitude or H/M ratio in a group of patients with cervical or generalized dystonia in a previous study. ^{Reference Boček, Štětkářová and Fečíková23} If the SRs are representative of H-reflex, to see no change in patients with cervical dystonia or generalized dystonia is compatible with the previous reports. Although the studies on this subject are limited if the cause of dystonia is related to pathophysiology, we would anticipate an increase in the excitability of the motoneuron pool compared to what we found in our study. Furthermore, SR is facilitated by muscle contraction, and we have recorded the reflex at the same limb that had dystonia, in other words, contractions. However, recordings did not coincide with the contractions, and reduced amplitudes may be related to the reduction of excitability after the powerful contractions. In a previous study, the shortest F-wave latency, the mean F-wave latency, and F-wave persistence of untreated muscles were measured before, one week, and five weeks after the treatment with botulinum toxin in patients with spasmodic torticollis and writer’s cramp. ^{Reference Wohlfarth, Schubert and Rothe24} The latencies were slightly prolonged one week after the treatment and returned to baseline five weeks later. The F-wave persistence was reduced one week after the treatment. The authors concluded that although there was a decreased excitability of alpha-motoneurons supplying non-treated muscles, it was transient. We also performed recordings in a relatively remote, toxin-naive muscle six months after the botulinum toxin injections. However, patients in this study received the treatment more than once. In focal hand dystonia, muscles located in the forearm are injected. Usually, it should not affect remote sites. Distal hand muscles may be affected by diffusion. Therefore, botulinum toxin could have led to, more or less, permanent excitability changes. Thus, hypoexcitability may result from phasic relation to powerful contractions or botulinum toxin injections. It is still interesting that there were no changes in other groups. Classically, other groups received higher doses of botulinum toxin, and some cases of generalized dystonia were also injected toxin in the upper extremities. At this point, we should acknowledge the limitation that there was a small number of patients in the focal dystonia group.

The LLRs were used to understand the underlying mechanism of movement disorders such as Parkinson’s disease, essential tremor, myoclonus, or dystonia. ^{Reference Dhar, Kamble and Pal25} Naumann and Reiners showed alterations in LLRI response in patients with cervical dystonia and upper limb dystonia. ^{Reference Naumann and Reiners13} Although increased amplitudes in LLRI responses were on the affected side, they showed that bilateral abnormality of LLRI response could be present. LLR II, occurring at approximately 50 ms, was obtained bilaterally in all controls but was reduced or absent in some patients, mainly on the clinically affected side. Therefore, there was a differential involvement in LLR I and II in this study. Another important finding in this study was more minor LLR II amplitudes after botulinum toxin injections. However, we were unable to replicate their results. There were two differences between this study and our study. First, these authors used a different statistical analysis than ours. They classified the LLR responses as normal or abnormal. For LLR I, they considered it abnormal when the response did not occur or when there was an increase in the amplitude. Second, they analyzed idiopathic dystonia patients without subgrouping. We also added an analysis comparing the entire group with healthy subjects and found no difference. LLRs were also recorded in patients with DYT11 myoclonus-dystonia syndrome and were normal. ^{Reference Marelli, Canafoglia and Zibordi26}

Electrical stimulation of peripheral nerves may elicit different kinds of reflex responses. The cutaneous silent period (CSP) is a robust and reproducible nociceptive electromyographic suppression mediated at the spinal level by small-diameter A-delta afferents. ^{Reference Kofler, Leis and Valls-Solé27} A recent review nicely presented all studies regarding CSP in focal dystonia, functional dystonia, and change of CSP after treatment with botulinum toxin or pallidal stimulation. ^{Reference Kofler, Leis and Valls-Solé28} The CSP in abductor pollicis brevis following D2 stimulation on the affected and the contralateral side showed delayed end latencies in patients with various forms of focal dystonia CSP. ^{Reference Pullman, Ford, Elibol, Uncini, Su and Fahn29,Reference Espay, Morgante, Purzner, Gunraj, Lang and Chen30} Notably, there were similar abnormalities in both organic and psychogenic dystonia. ^{Reference Espay, Morgante, Purzner, Gunraj, Lang and Chen30} Interestingly, the CSPs remained constant over one year in two dystonia patients, before and during relief from a sensory trick in one patient, and before and after botulinum toxin treatment in the same patient. ^{Reference Pullman, Ford, Elibol, Uncini, Su and Fahn29} In patients treated with bilateral pallidal stimulation, there was a trend towards shorter CSP duration due to later CSP onset, which did not change by switching the neurostimulator on. ^{Reference Boček, Štětkářová and Fečíková23} The MnSP is made up of three different parts, including the collision of antidromic with orthodromic motor impulses, Renshaw cell inhibition activated by an antidromic motor volley, and activation of high-threshold cutaneous fibers within the mixed nerve. ^{Reference Mori, Horino and Matsugi17,Reference Takahashi, Ueno and Ohyama18} The last half of the MnSP corresponds entirely to cutaneous afferent impulses, which, in isolation, produce a complete silent period between 70 and 120 ms after digital stimulation in contrast to the first half, where there are effects of voluntary and descending volleys as well as Renshaw cell inhibition. ^{Reference Takahashi, Ueno and Ohyama18} However, the presence of Renshaw cell inhibition in the distal hand muscles is still controversial. ^{Reference Edgley, Williams and Baker31} This is why we used the end latency, which represents the clear part of the silent period. The studies of MnSPs in dystonia are more limited compared to CSP. The MnSPs were abnormal in paroxysmal kinesigenic dyskinesia during the attack, in whom CSPs were normal. ^{Reference Cogez, Etard, Derache and Defer32} We find MnSPs in dystonia comparable to healthy individuals.

Besides the abovementioned limitations, there were certain other limitations of the study. The patients were not toxin-naïve, and we could not appoint the muscles and doses injected. There were significant age differences observed within the four groups. Additionally, it should be noted that the number of subjects was not evenly distributed among these groups. We acknowledge that these are the main limitations of our study. However, the age difference and uneven distribution of numbers of subjects across groups originate from the natural history of these disorders. We conducted the statistical analysis according to different age groups. Naturally, isolated hand tremor is less frequent than cervical dystonia. Additionally, we excluded patients with secondary causes, spasticity or tremor, which is another reason for reducing numbers.

In conclusion, we analyzed different levels of the nervous system using electrophysiology in various dystonia phenotypes and determined only low-amplitude segmental reflex in focal hand dystonia. Considering the small number of patients in the focal hand dystonia group, we suggest that there was reduced spinal excitability in this group, which may be related to botulinum toxin treatment. We could not show significant changes in the LLRs or MnSPs in any dystonia. Therefore, using SRs, LLRs, and MnSPs, we could not offer an electrophysiological signature specific to dystonia.