Orthostatic Myoclonus—Clinical and Electrophysiological Features in a Large Retrospective Cohort Authors: Sai A. Nagaratnam, Niroshan Jeyakumar, Gobishan Jeyakumar, Florence Chang, Duncan Wilson, Paloma Becker, Hugo Morales‐Briceno, Neil Mahant, Victor S.C. Fung Published: 2026 ABSTRACT AbstractBackgroundOrthostatic myoclonus is characterized by irregular, lower limb myoclonic bursts during stance and is a major cause of postural instability and falls. However, studies are limited, and little is known about its pathophysiology.ObjectivesWe sought to define the clinical and electrophysiological features of orthostatic myoclonus in a large, single‐center cohort.MethodsWe included 42 participants (24 males, 18 females) with a mean age of 74 years (range, 46–93) from Westmead Hospital presenting with orthostatic myoclonus from 2007 to 2023. Medical records were retrospectively reviewed for demographic details, symptoms, co‐morbidities, and treatment. Lower limb surface electromyography (EMG) was analyzed using a custom‐designed algorithm to automatically identify myoclonic bursts and measure their duration, synchronicity, and rhythmicity. Differences in burst parameters between muscles and associations between burst parameters and clinical characteristics were statistically evaluated.ResultsMean burst durations during standing were 77 to 90 ms across lower limb muscles. Maximum burst activity and bilateral synchronicity occurred in tibialis anterior. Only 12% of participants exhibited any rhythmicity. A total of 79% of participants had a coexistent neurological disorder including 26% with parkinsonism. There was no significant association between parkinsonism and burst parameters. However, there was a significant, inverse correlation between the presence of neuropathy or radiculopathy and synchronous activity (P = 0.02).ConclusionsWe provide a computationally robust clinical and electrophysiological analysis of orthostatic myoclonus in a large cohort. Our findings support the theory of a subcortical generator arising from protean secondary causes and subject to peripheral modulation. Further work is needed to clarify treatment outcomes. Orthostatic myoclonus is a recently described hyperkinetic movement disorder associated with gait instability. 1 , 2 It is defined by the presence of irregular, myoclonic bursts in the lower limb muscles on surface electromyography (EMG) triggered or greatly enhanced by standing and is distinct from the rhythmic pattern of orthostatic tremor. 1 , 2 , 3 Clinical diagnosis may be challenging because of symptomatic and phenomenological overlap with classic or pseudo‐orthostatic tremor, which highlights the need for electrophysiological confirmation. Although orthostatic tremor is characterized by highly regular EMG bursting and a feeling of unsteadiness on standing rather than while walking, orthostatic myoclonus is typically associated with unsteadiness during both gait and stance and, therefore, carries a significant risk of falls. In fact, orthostatic myoclonus may be one of the most common causes of unsteadiness in the elderly and is likely underdiagnosed. 4 Despite this, orthostatic myoclonus remains poorly understood. Notably, it is frequently observed in association with other neurological disorders such as Parkinson's disease and atypical parkinsonism, subcortical microvascular disease, and lumbosacral radiculopathy. 5 , 6 However, the clinical significance of these associations, their relationship to the electrophysiological properties of orthostatic myoclonus and the underlying pathophysiology driving orthostatic myoclonus remain unclear. Consequently, evidence to guide classification, prognostication, and treatment has been lacking. To date, studies of orthostatic myoclonus have been limited by small sample sizes and descriptive analysis, which, may, in part, be because of the practical limitations of manual EMG interpretation. In this study, we address these limitations by developing a novel, standardized computational methodology to analyze myoclonic bursts on EMG, facilitating subsequent quantitative, statistical analysis of various electroclinical associations in a large, single‐center cohort of patients with orthostatic myoclonus. Our computational scheme introduces the first pragmatic, standardized definition of a myoclonic burst applied iteratively for efficient, large‐scale data analysis. Methods Participants The Westmead Hospital Movement Disorders Laboratory database was searched to identify all patients between November 2007 and April 2023 diagnosed with orthostatic myoclonus by a movement disorders specialist. Diagnosis required the presence of irregular, short‐duration bursts in the lower limb muscles on surface EMG, triggered or greatly enhanced by standing, as described by Glass et al. 1 Patients with clearly rhythmic and coherent activity suggestive of orthostatic tremor, including typical orthostatic tremor with clear, narrow spectral peaks in the 13 to 18 Hz range as well as “pseudo‐orthostatic tremor” with broader spectral peaks spanning lower frequency ranges, were excluded. 3 Forty‐two patients were identified and included in the final study. Clinical Data Electronic and paper medical records were reviewed to extract data on basic demographics, clinical signs and symptoms, comorbidities, medications, radiological features, laboratory markers, functional status (modified Rankin scale [mRS]), and treatment response. Fall frequency was scored based on previously published work; 0 = no falls in a quarter (3 months), 1 = once per quarter, 2 = once per week, 3 = 2 to 6 times per week, and 4 = daily falls. 7 Electrophysiology Signal Recording and Processing Surface EMG recordings from four bilateral lower limb muscles, tibialis anterior (TA), medial gastrocnemius (MG), quadriceps (QUAD), and biceps femoris (BF), were analyzed during two conditions: sitting at rest and standing with eyes open and feet together. The final 15‐second epoch of each recording was chosen to allow time for the participants to appropriately assume the correct condition being interrogated. One epoch per condition was analyzed in each patient. Raw EMG signals were processed in MATLAB (The MathWorks, Natick, MA). The sampling rate was 2 kHz. Raw signals were 20 to 200 Hz band‐pass filtered with a Butterworth bandpass filter applied forward and backward to eliminate any phase distortion. The filtered signals were then rectified and enveloped using a moving 50 ms root mean square (RMS) window to minimize noise and signal fluctuation. 8 Burst Identification There are no existing formal criteria for the definition of a myoclonic burst on EMG with reference to amplitude, which is necessary to distinguish true, individual bursts from electrical noise. 9 , 10 To facilitate standardized computational analysis, we proposed three different possible criteria: amplitude ≥25%, 50%, or 2 standard deviations above the RMS background for ≥20 ms. Relative amplitude criteria were chosen based on the consensus that myoclonic discharges typically stand out from background muscle activity, which can vary by muscle group and condition. The 20 ms minimum duration criterion was chosen to distinguish true muscle activity from shorter‐lasting electrical noise and is consistent with published reports of minimum burst durations in cortical myoclonus, the subtype of myoclonus in which the shortest bursts are seen. 10 , 11 We developed custom MATLAB code to automatically identify bursts meeting each criteria. We used EMG data from approximately 20% of the whole cohort (n = 10) to compare the performance of each of the three different burst criteria against manual burst identification performed by two, independent expert movement disorder neurologists. On visual inspection of annotated EMGs, the amplitude ≥25% above RMS background for ≥20 ms criteria had the greatest concordance with clinician‐marked bursts. As our goal was to create a pragmatic burst identification tool for efficient and standardized large‐scale analysis, rather than develop a diagnostic technique, we did not attempt to define traditional diagnostic performance metrics such as sensitivity and specificity. However, prospective studies using similar methodology should allow these parameters to be established. These criteria were applied to the processed EMG traces from the entire cohort to identify all myoclonic bursts for further analysis. Burst Parameters Burst duration, rhythmicity, and synchronicity were calculated. Rhythmicity was defined as having at least one frequency peak, between 0 and 20 Hz, which was ≥3 times the power of neighboring ±1 Hz frequencies on the Fast Fourier transform of the processed EMG signal. We modified the rhythmicity definition adopted by Glass et al 1 in their seminal paper on orthostatic myoclonus by retaining the relative power threshold of 3×, but applying it to neighboring frequencies to ensure identified frequency peaks are local maxima and, therefore, more likely to present true, distinct oscillators. 1 , 12 Synchronicity was expressed quantitatively as the percentage of bursts occurring in muscle pairs with onset ≤15 ms of one another as defined by Glass et al 1 and calculated for homologous (eg, bilateral TA), agonist–antagonist (eg, TA‐MG), and proximal‐distal (eg, MG‐BF) muscle pairs. Statistical Analysis Statistical analysis was performed on RStudio version 2025.05.1 + 513. Repeated measures analysis of variance with Bonferroni‐corrected, post‐hoc pairwise t tests were used to evaluate differences in burst duration and level of synchronicity between different muscles and muscle pairs during standing. Multivariable logistic regression models were then used to examine the relationship between these electrophysiological parameters and three main clinical categories: parkinsonism, neuropathy/radiculopathy, and subcortical microvascular disease severity, the latter ranging from none to severe (Table 1). P < 0.05 was considered statistically significant. TABLE 1Baseline demographic and clinical characteristicsnDemographic informationAge at studyMean, 74Range, 46–93Age at onsetMean, 71Range, 45–89Male sex at birth2457%mRS012%11433%21536%31024%425%Neurological comorbiditiesParkinson's disease819%Atypical parkinsonism37%Normal pressure hydrocephalus513%Peripheral neuropathy1331%Lumbosacral radiculopathy921%Medication‐related myoclonus615%Other CNS disorder2048%Subcortical microvascular diseaseNil1331%Mild–moderate614%Moderate37%Moderate–severe25%Severe12%Total42Abbreviations: mRS, modified Rankin scale; CNS, central nervous system. Baseline demographic and clinical characteristics Abbreviations: mRS, modified Rankin scale; CNS, central nervous system. Results Clinical Features Table 1 summarizes the baseline demographic and clinical characteristics of the study cohort. Of 42, 33 (79%) participants had an underlying neurological disorder including Parkinson's disease, atypical parkinsonism, normal pressure hydrocephalus, epilepsy, stroke, hypoxic brain injury, cervical dystonia, cognitive impairment, depression, coeliac disease, neuropathy, or radiculopathy. Four participants had an identifiable antibody: two had anti‐GAD antibodies, one had anti‐IgLON5 antibodies, 13 and one had anti‐GFAP antibodies. With regards to symptoms, 28 of 42 (67%) participants reported subjective unsteadiness on standing, 24 of 42 (57%) reported awareness of shaking, 41 of 42 (98%) reported gait instability, and 18 of 42 (43%) reported associated falls. The mean fall frequency score was 1 (range, 0–4; ie, once every 3 months). There was a high rate of disability with 28 of 42 (67%) participants having a mRS ≥2. In terms of medications, 11 of 42 (26%) participants had a favorable response to medical therapy, which included benzodiazepines, dopaminergic medication, anti‐seizure medication, or neuropathic agents. Orthostatic myoclonus was temporally associated with medications in seven (17%). Monoamine reuptake blockers were most common: amitriptyline in three (7%) and mirtazapine in one. Other medications were chemotherapy, dopamine blockers, and sodium valproate. Individual participant data are provided in Tables S1 and S2. Electrophysiological Features By selection, all patients had an orthostatic pattern of short‐duration, irregular muscle bursts that were triggered or exaggerated by standing in comparison to the seated at rest or legs elevated positions. A typical recording of a single muscle and subsequent processing is shown in Figure 1. We have previously published an example of a full montage. 14 Mean burst duration during standing ranged between 77 and 90 ms and varied significantly between muscles (F = 4.32, P = 0.021) (Table 2) with significantly shorter burst durations in MG compared to QUAD (74 vs. 85 ms, P < 0.001) and BF (74 vs. 90 ms, P = 0.03). Five of 42 (12%) participants had evidence of rhythmicity, but only in one muscle per participant. Frequency peaks ranged from 1 to 15 Hz. (Table S3). (A) Post‐processed surface electromyography (EMG) recording of right tibialis anterior in one participant when seated at rest. Criteria‐defined myoclonic bursts are highlighted in red. (B) Post‐processed surface EMG recording of right tibialis anterior from the same participant when standing with eyes open and feet together. Criteria‐defined myoclonic bursts are highlighted in red. A typical orthostatic pattern of increased burst activity on standing is demonstrated. TABLE 2Burst duration during standingMuscle a Burst duration (ms)Tibialis anterior77 (58–233)Medial gastrocnemius74 (61–126)Quadriceps femoris85 (60–129)Biceps femoris90 (59–269) Note: Mean (range). a Average of left and right sides. Burst duration during standing Note: Mean (range). Average of left and right sides. During standing, there was a moderate, statistically significant difference in the number of bursts per 15‐second epoch between muscles (F = 27.5, P < 0.001). The greatest burst activity was seen in TA (TA, 38 vs. MG, 31; P = 0.05 vs. QUAD, 23; P < 0.001 vs. BF, 21; P < 0.001), followed by MG (MG vs. QUAD, P < 0.001; MG vs. BF, P < 0.001) and then QUAD and BF, with no significant difference between the latter two (P > 0.5). Table 3 summarizes the level of synchronicity between different muscle pairs during standing (representative Fig. S1). There was a small but significant difference between muscle pairs in the level of synchronicity (F = 3, P = 0.02). This effect was driven by significantly greater synchronicity between bilateral TA muscles compared to bilateral medical gastrocnemius muscles (26% vs. 17%, P = 0.04) and between the anterior proximal‐distal pairs—QUAD and TA—compared to the posterior proximal‐distal pairs—BF and MG— (19% vs. 12%, P = 0.04). All other pairwise comparisons were not statistically significant. TABLE 3Synchronicity between muscle pairs during standingMuscle pairSynchronicity (%)Bilateral homologous a L + R TA26 (0–76)L + R MG17 (0–48)Agonist–antagonist b TA‐MG17 (0–76)QUAD‐BF14 (0–100)Proximal‐distal a QUAD‐TA19 (0–46)BF‐MG12 (0–29) Note: Mean (range).Abbreviations: TA, tibialis anterior; MG, medial gastrocnemius; QUAD, quadriceps; BF, biceps femoris; L, Left; R, Right. aBilateral recordings were not routinely obtained for QUAD and BF. b Average of left and right sides. Synchronicity between muscle pairs during standing Note: Mean (range). Abbreviations: TA, tibialis anterior; MG, medial gastrocnemius; QUAD, quadriceps; BF, biceps femoris; L, Left; R, Right. aBilateral recordings were not routinely obtained for QUAD and BF. Average of left and right sides. Association between Clinical and Electrophysiological Features The presence of a neuropathy or radiculopathy significantly reduced the level of synchronicity between bilateral TA muscles (B = −18, P = 0.02) (Table 4). There was no detectable relationship between the presence of parkinsonism or burden of subcortical microvascular disease and burst duration or synchronicity. TABLE 4Multivariable logistic regression of clinical categories predicted by demographic and electrophysiological parametersNeuropathy or radiculopathySubcortical microvascular disease severityParkinsonism B P B P B P DemographicAge0.250.03* 0.070.150.000.97Sex2.810.19−0.610.600.510.65Burst durationTA0.010.75−0.010.66−0.080.22MG0.020.850.200.050.001.00QUAD0.030.54−0.030.400.010.82BF−0.030.55−0.010.71−0.020.37SynchronicityL TA‐R TA−17.980.02* −0.760.82−5.990.10L MG‐R MG4.620.533.440.451.140.80TA‐MG14.900.07−0.870.780.610.84QUAD‐BF−25.120.052.310.410.930.75TA‐QUAD7.680.260.920.830.880.83MG‐BF−17.000.167.870.29−1.890.79Abbreviations: L, Left; R, RIght; TA, tibialis anterior; MG, medial gastrocnemius; QUAD, quadriceps; BF, biceps femoris.* P < 0.05. Multivariable logistic regression of clinical categories predicted by demographic and electrophysiological parameters Abbreviations: L, Left; R, RIght; TA, tibialis anterior; MG, medial gastrocnemius; QUAD, quadriceps; BF, biceps femoris. P < 0.05. Discussion In this study, we developed and applied a novel, standardized, myoclonic burst detection algorithm for surface EMG to characterize fundamental neurophysiological features in a large cohort of patients with orthostatic myoclonus. We found mean burst durations of 77 to 90 ms in all studied lower limb muscles, maximal synchronicity in homologous TA muscles, and only rare instances of rhythmicity, consistent with previous reports. 1 , 5 By confirming these findings using objective quantification in a larger cohort we help to establish these as key defining features of orthostatic myoclonus that can be used to support clinical diagnosis. However, compared to previous reports, our computational methodology has the added benefit of facilitating quantification of these burst parameters, permitting novel statistical analysis of inter‐muscle differences and electroclinical associations. The consensus classification of myoclonus includes burst durations for cortical–subcortical (50–100 ms) and subcortical origin myoclonus (normally >100 ms, but can be 50–100 ms). 10 , 15 In our study, we found that burst durations were variable in orthostatic myoclonus, being 50 to 100 ms on average, but at times >200 ms. This is compatible with the prevailing hypothesis that the generator of orthostatic myoclonus is of subcortical or cortical–subcortical origin based on the distribution of muscles affected and absence of cortical features such as stimulus sensitivity, giant somatosensory evoked potentials, and C‐reflexes. 4 , 5 Furthermore, subcortical microvascular disease was prevalent in our cohort, consistent with prior studies of orthostatic myoclonus, providing a possible corresponding structural substrate. In terms of the distribution of muscles affected, we found that orthostatic burst activity is greatest in TA and maximum synchronicity occurs in bilateral TA pairs, reinforcing earlier, qualitative observations. 1 , 2 , 5 , 6 The dominance of TA, which acts to dorsiflex the foot, raises the possibility that orthostatic myoclonus may represent a state of overactive anticipatory postural adjustments to perceived retropulsive forces. 16 A similar theory has gained some evidence in orthostatic tremor. 17 These retropulsive forces may arise from protean disorders of gait and balance in the elderly. Accordingly, like others, we found that most participants in our cohort had an underlying neurological disorder, such as subcortical microvascular disease or parkinsonism, 5 , 6 and nearly all participants had an accompanying subjective sense of unsteadiness (Table S2). Of course, whether the latter reflects a primary impairment of postural stability, the myoclonus itself, or a combination of the two cannot be distinguished. Some participants responded to anti‐myoclonic medications, such as benzodiazepines and levetiracetam, suggesting that the myoclonus itself may be the primary cause of symptoms in these individuals. Conversely, the response to increased dopaminergic medication in some participants with parkinsonism and dopamine blocker‐related orthostatic myoclonus suggests that postural instability from the underlying parkinsonian disorder may be more important in these individuals. However, participants with parkinsonism did not differ from those without parkinsonism in their electrophysiological characteristics, supporting the idea of orthostatic myoclonus as a common, maladaptive downstream response to postural instability of diverse origin. The association between orthostatic myoclonus and centrally acting medications may instead suggest a role for dopaminergic, serotoninergic, and/or noradrenergic pathways in orthostatic myoclonus pathophysiology. Importantly, however, our data also suggest that orthostatic myoclonus may be peripherally modulated because we found reduction in bilateral synchronicity in TA in the presence of a concomitant neuropathy or lumbosacral radiculopathy. It is unclear whether this might represent modulation by homonymous or heteronymous spinal circuits or at a higher level. Although our study has strengths, including the comparatively large cohort size and standardized, computational approach, the retrospective design remains a limitation. Data on symptoms and medication response was incomplete owing to the retrospective design and would be valuable to collect prospectively to examine associations with burst parameters. Furthermore, there remains no gold standard for the precise definition of a myoclonic burst on surface EMG, and although our method still displayed concordance with clinician ratings, it has not been validated in other myoclonic disorders. However, the standardization, efficiency, and reproducibility of this approach are major advantages over manual analysis and have potential to be applied to the neurophysiological analysis of other movement disorders. We provide a computationally robust clinical and electrophysiological analysis of orthostatic myoclonus in a large cohort, which may aid diagnosis in this rare movement disorder. Our findings may shed new light on the complex relationship between the clinical and electrophysiological features of this condition, supporting the existence of a subcortical, central generator arising from protean secondary causes and subject to peripheral modulation. Further prospective studies are needed to established better treatment paradigms for this debilitating condition. Author Roles (1) Research project: A. Conception, B. Organization, C. Execution; (2) Statistical Analysis: A. Design, B. Execution, C. Review and Critique; (3) Manuscript Preparation: A. Writing of the First Draft, B. Review and Critique. S.A.N.: 1A, 1C, 2A, 2B, 2C, 3A, 3B N.J.: 1C, 2A, 2B, 2C, 3A, 3B G.J.: 1C, 2A, 2B, 2C, 3B F.C.: 1C, 2C, 3B D.W.: 1C, 2C, 3B P.B.: 1C, 2C, 3B H.M.B.: 1C, 2C, 3B N.M.: 1C, 2B, 2C, 3B V.S.C.F.: 1A, 1B, 1C, 2C, 3B Disclosures Ethical Compliance Statement: The Western Sydney Local Health District Human Research Ethics Committee approved the study. Written consent was not required for this retrospective study. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this work is consistent with those guidelines. Financial Disclosures and Conflicts of Interest: No specific funding was received for this work. The authors declare that there are no conflicts of interest relevant to this work. Author disclosures are available in the Supporting Information. Financial Disclosures for the Previous 12 Months: V.F. receives a salary from NSW Health, has received unrestricted research grants from The Michael J. Fox Foundation, AbbVie, and Merz, and receives royalties from Health Press and Taylor and Francis Group. N.M. has received honoraria from AbbVie, Ultrabionics, and Medtronic. The other authors declare that there are no additional disclosures to report. Supporting information Table S1. Demographic information and clinical features of study participants. Table S2. Symptoms and treatment response in study participants. Table S3. Peak frequencies in study participants with burst rhythmicity. Figure S1. Example of synchronicity: Surface EMG from right and left tibialis anterior (TA) of one patient during stance are shown with multiple synchronous bursts (detected bursts shown in red). Data S1. COI Disclosure.