Neuromuscular assessment system

Abstract

A neuromuscular assessment system and method of operation may include: attaching a first electrode array to an active muscle configured to detect a first electromyogram (EMG) signal; attaching a second electrode array to an antagonist muscle configured to detect a second EMG signal, the active muscle and the antagonist muscle forming an active / antagonist muscle pair; decomposing the first EMG signal into a first motion unit spike sequence; decomposing the second EMG signal into a second motion unit spike sequence; correlating the first motion unit spike sequence and the second motion unit spike sequence to generate a correlation signal; determining synchronism and periodicity within the correlated signals; and generating a tremor score, the tremor score being a percentage of the correlated signal determined to have both synchronism and periodicity.

Description

[0001] Cross-reference to related applications

[0002] This application claims priority to U.S. Nonprovisional Patent Application No. 18 / 199,180, filed May 18, 2023, and to all common subject matter thereof. The entire contents of this application are incorporated herein by reference. Technical Field

[0003] This disclosure relates to health assessment systems, and more specifically to neuromuscular assessment systems with an electromyography (EMG) array configured to detect neurodriven limbic motor symptoms of Parkinson's disease and pre-Parkinsonian neurodegeneration. Background Technology

[0004] Parkinson's disease is a neuromotor disorder caused by the loss of dopamine-secreting neurons in the midbrain. Motor symptoms of Parkinson's disease can include resting tremor, bradykinesia, muscle rigidity, gait instability, balance problems, and coordination problems.

[0005] Early diagnosis is crucial. By the time motor symptoms are diagnosed, up to 70% of all dopamine neurons may have already been lost, at which point treatment becomes largely ineffective. Medical costs increase as the disease progresses, sometimes requiring expensive institutional care in worsening conditions.

[0006] For example, treatments such as levodopa and carbidopa are more effective when administered early in the disease. Non-pharmacological treatments, such as increasing physical activity, are also more readily available in the early stages of Parkinson's disease and may help slow disease progression.

[0007] Nearly one million people in the United States have Parkinson's disease, while seven million people worldwide suffer from it. The number of Parkinson's patients in the United States is projected to reach 1.2 million by 2030. In 2017, direct medical costs associated with Parkinson's disease amounted to $25.4 billion, and this is expected to increase over time.

[0008] Many previous developments in Parkinson's disease monitoring technology relied on monitoring obvious Parkinsonian motor symptoms corresponding to the visible and late stages of the disease. This was primarily due to the fact that previous developments in monitoring technology employed a limited number of EMG sensors placed individually on the patient's skin.

[0009] This previous monitoring technology could distinguish visually observable symptoms but failed to provide insight into premotor neurodegeneration, a characteristic of pre-Parkinson's disease conditions. Examplely, previous monitoring technologies employed a limited number of EMG sensors, individually placed on the patient's skin by a healthcare professional, to monitor muscle drive.

[0010] Because these sensors are limited in number and not arranged in any kind of array or with fixed positions relative to each other, the detection, triangulation, and measurement of premotor nerve drives for individual motor units are impossible to determine. Therefore, the technical processes implemented by previous monitoring techniques failed to account for the premotor nerve origins of Parkinson's disease symptoms, resulting in the inability to identify pre-Parkinson's conditions and providing only measurements of late-stage symptoms.

[0011] Using this prior surveillance technique, there is evidence that nearly half (47%) of Parkinson's disease diagnoses are incorrect when conducted in primary care settings. Experts whose specialties are not specific to a particular movement disorder have an error rate of approximately 25%, while movement disorder specialists have an error rate of only 6% to 8%.

[0012] Even when correctly diagnosed by healthcare professionals using prior monitoring techniques, up to 70% of neurons may be lost by the time movement disorder specialists identify Parkinson's disease, as the disease can be detected using prior monitoring techniques based on EMG signals corresponding to visible and late-stage Parkinson's symptoms.

[0013] Solutions have been sought for a long time, but previous developments have not taught or suggested any complete solutions, and solutions to these problems have long been beyond the reach of those skilled in the art. Therefore, there remains a considerable need for devices and methods that can detect and identify the neural origins of individual motor units for the generation, monitoring, and tracking of pre-Parkinson's disease information, which can enable early diagnosis. Summary of the Invention

[0014] A neuromuscular assessment system and method for generating, monitoring, and tracking pre-motor information within pre-motor neural origins are disclosed. The neuromuscular assessment system does not rely on late, pronounced motor symptoms, but is described as having technical features that enable the use of pre-motor neural origins, which provide early, distinct pre-motor characteristics of neurodegeneration.

[0015] A neuromuscular assessment system and method of operation are disclosed. The method of operation may include: attaching a first electrode array to an agonist muscle configured to detect a first electromyography (EMG) signal; attaching a second electrode array to an antagonist muscle configured to detect a second EMG signal, the agonist muscle and the antagonist muscle forming an agonist / antagonist muscle pair; decomposing the first EMG signal into a first motor unit spike sequence; decomposing the second EMG signal into a second motor unit spike sequence; correlating the first motor unit spike sequence and the second motor unit spike sequence to generate a correlation signal; determining synchronicity and periodicity within the correlation signal; and generating a tremor score, the tremor score being the percentage of the correlation signal determined to have both synchronicity and periodicity.

[0016] Other contemplated embodiments may include objects, features, aspects, and advantages other than or in lieu of those described above. These objects, features, aspects, and advantages of the embodiments will become more apparent and understandable from the following detailed description and accompanying drawings. Attached Figure Description

[0017] The neuromuscular assessment system is illustrated in the various figures of the accompanying drawings. These figures are intended to be exemplary and not limiting, wherein the same reference numerals are intended to refer to the same parts, and wherein:

[0018] Figure 1 This is a side view of a neuromuscular assessment system.

[0019] Figure 2 yes Figure 1 A block diagram of a neuromuscular assessment system.

[0020] Figure 3A It is used to calculate for Figure 1 The control process for resting tremor information in the neuromuscular assessment system.

[0021] Figure 3B It is used for calculation Figure 3A The flowchart for tremor scores.

[0022] Figure 4 yes Figure 1 Side view of the lower arm.

[0023] Figure 5 yes Figure 1 An isometric view of a neuromuscular assessment system.

[0024] Figure 6 It is used with Figure 1 A schematic diagram of the analog front end (AFE) used in neuromuscular assessment systems.

[0025] Figure 7 It is used with Figure 1 A schematic diagram of a surface electromyography (sEMG) amplifier used in neuromuscular assessment systems.

[0026] Figure 8A yes Figure 1 Graphical depiction of HD-sEMG signals from the combined sensor.

[0027] Figure 8B yes Figure 1 A graphical representation of the IMU signal from the combined sensor.

[0028] Figure 8C yes Figure 1Graphical representation of the optical volumetric (PPG) signal of the combined sensor.

[0029] Figure 8D yes Figure 8B A graphical representation of the IMU signal in the frequency domain.

[0030] Figure 9A It is used for Figure 3A A graphical depiction of the decomposition process of computational neural drive steps.

[0031] Figure 9B yes Figure 9A A magnified view of area 9B.

[0032] Figure 10 It is used for Figure 3A A graphical depiction of the cumulative spike sequence (CST) of computational neural driving steps.

[0033] Figure 11A It is aimed at Figure 8A A graphical depiction of CST autocorrelation of the left extensor carpi radialis (ECR).

[0034] Figure 11B yes Figure 8A Left ECR and Figure 8A A graphical depiction of the CST cross-correlation between the left radial flexor carpi radialis (FCR) and the left radial flexor carpi radialis (FCR).

[0035] Figure 11C yes Figure 8A Left ECR and Figure 8A A graphical depiction of the CST cross-correlation between the right ECRs.

[0036] Figure 11D yes Figure 8A Left ECR and Figure 8A A graphical depiction of the CST cross-correlation between the right FCRs.

[0037] Figure 11E yes Figure 8A Left FCR and Figure 8A A graphical depiction of the CST cross-correlation between the left ECR and the right ECR.

[0038] Figure 11F It is aimed at Figure 8A The left FCR is depicted by the CST autocorrelation graph.

[0039] Figure 11G yes Figure 8A Left FCR and Figure 8A A graphical depiction of the CST cross-correlation between the right ECRs.

[0040] Figure 11H yes Figure 8A Left FCR and Figure 8A A graphical depiction of the CST cross-correlation between the right FCRs.

[0041] Figure 11I It is aimed at Figure 8A The right ECR is depicted by the CST autocorrelation graph.

[0042] Figure 11J yes Figure 8A The right ECR and Figure 8A A graphical depiction of the CST cross-correlation between the right FCRs.

[0043] Figure 11K yes Figure 8A The right FCR and Figure 8A A graphical depiction of the CST cross-correlation between the right ECRs.

[0044] Figure 11L It is aimed at Figure 8A The CST autocorrelation graph of the right FCR is depicted.

[0045] Figure 12 It is used to calculate for Figure 1 The control process for gait freeze information in the neuromuscular assessment system.

[0046] Figure 13 It is a side view depicting the patient's gait.

[0047] Figure 14 yes Figure 1 Side view of the lower leg.

[0048] Figure 15A It is for when standing still Figure 1 Combined sensor Figure 1 A graphical representation of the combined sensor output signal.

[0049] Figure 15B It is targeted during gait freeze. Figure 1 Combined sensor Figure 1 A graphical representation of the combined sensor output signal.

[0050] Figure 16A It is used for patients exhibiting a normal gait. Figure 14 A graphical depiction of the EMG signal output channel of each of the combined sensors in the combined sensor array.

[0051] Figure 16B It is used for patients exhibiting gait freezing due to shuffling. Figure 14 A graphical depiction of the EMG signal output channel of each of the combined sensors in the combined sensor array.

[0052] Figure 17AIt is a graphical representation of the power spectral density estimation of heart rate variability in healthy patients.

[0053] Figure 17B It is a graphical representation of the power spectral density estimation of heart rate variability in Parkinson's disease patients.

[0054] Figure 18 It is a control flow for operating the neuromuscular assessment system 100. Detailed Implementation

[0055] In the following description, reference is made to the accompanying drawings, which form a part of this document, and embodiments in which the neuromuscular assessment system can be practiced are shown by way of illustration. It should be understood that other embodiments can be utilized and structural changes can be made without departing from the scope of the neuromuscular assessment system.

[0056] When the features, aspects, or embodiments of a neuromuscular assessment system are described in accordance with the steps of a process, operation, control flow, or flowchart, it should be understood that steps may be combined, performed in different orders, deleted, or additional steps may be included without departing from the neuromuscular assessment system as described herein.

[0057] The neuromuscular assessment system is described in sufficient detail to enable those skilled in the art to manufacture and use it, and many specific details are provided to give a thorough understanding of the neuromuscular assessment system; however, it is obvious that the neuromuscular assessment system can be practiced without these specific details.

[0058] To avoid obscuring the neuromuscular assessment system, some well-known assessment system configurations, algorithms, and descriptions are not disclosed in detail. Similarly, the accompanying drawings illustrating embodiments of the assessment system are semi-schematic and not drawn to scale, and specifically, some dimensions are shown in great exaggerated form for clarity.

[0059] As used herein, the term "coupled" or "coupling" should be understood as a physical connection between elements, whether direct or indirect. As used herein, "correlated signal" means autocorrelation signal, cross-correlation signal, or a combination thereof.

[0060] Now for reference Figure 1 The figure shows a side view of a neuromuscular assessment system 100. The neuromuscular assessment system 100 may include a component box 102 coupled to a combination sensor 104.

[0061] It is conceivable that the component box 102 and the combined sensor 104 can be formed as an integral unit, or coupled via cable or wireless communication. The following is about... Figure 2 , Figure 4 and Figure 5 The combined sensor 104 is described in more detail. As depicted, each combined sensor in the combined sensor 104 is communicatively coupled to the component box 102 via wireless transmission 106.

[0062] Wireless transmission 106 can be achieved through communication between a transponder built into the combined sensor 104 and a transponder built into the component box 102. Wireless transmission 106 can transmit the combined sensor output signal 108 to the component box 102. For example, it is conceivable that the combined sensor output signal 108 can be a digital representation of multiple sensor signals; such as regarding... Figure 2 and Figures 8A to 8D Those disclosed and described. When implemented using a communication cable between the combined sensor 104 and the component box 102, the combined sensor output signal 108 may include an I2C signal or an SPI interface signal.

[0063] Component box 102 may include a non-transitory computer-readable medium 110 usefully associated with processor 112 for storing combined sensor output signals 108 and instructions 114 for operation and control of the neuromuscular assessment system 100. Processor 112 may be a processor contained within component box 102, but may also include a distributed processing network.

[0064] The component box 102 and the combined sensor 104 may further include a power source 116, such as a battery or an adapter for a public power outlet. In one contemplated embodiment, the power source 116 in the component box 102 may be used to provide power to the combined sensor 104 when coupled to a communication cable.

[0065] The combination sensors 104 can be paired and attached to the agonist / antagonist muscle pair of the patient 120. Illustratively, for example, a pair of combination sensors 104 is shown attached to the lower arm 122 of the patient 120. Specifically, two of the combination sensors 104 are attached to the left arm 124, and two of the combination sensors 104 are attached to the right arm 126. Regarding... Figure 4 The attachment of the combined sensor 104 to the right arm 126 of the patient 120 is described and illustrated in more detail.

[0066] For clarity, the combined sensor 104 is described as attaching superiorly to the forearm 122 of the patient 120 to the agonist / antagonist pair of the extensor carpi radialis (ECR) and flexor carpi radialis (FCR). Therefore, the neuromuscular assessment system 100 can utilize the combined sensor 104 to monitor the left ECR, right ECR, left FCR, and right FCR.

[0067] Furthermore, a pair of combined sensors 104 are shown attached to the lower leg 128 of the patient 120, which is related to... Figure 5 A more detailed description is provided. For clarity, the combined sensor 104 is depicted as being placed on the gastrocnemius muscle and the tibialis anterior muscle.

[0068] It is conceivable that the combined sensor 104 could also be attached to other agonist / antagonist muscle pairs on the patient's upper arm 130, thigh 132, hand, or foot. It is also conceivable that the agonist / antagonist muscle pairs could be located in the lower arm 122 or calf 128, and be associated with other muscles located in the lower arm or calf. Figure 4 or Figure 14 The text describes different muscle combinations.

[0069] Now for reference Figure 2 , which shows Figure 1 A block diagram of a neuromuscular assessment system 100. The neuromuscular assessment system 100 is shown as having representations of... Figure 1 The sensor frame 202 within the combined sensor 104 provides the combined sensor output signal 108 to the signal frame 204. Inventively, the sensor frame 202 may include a high-density surface EMG (HD-sEMG 206). The HD-sEMG 206 can provide the HD-sEMG signal 208.

[0070] Such as about Figure 4 and Figure 9A As shown in more detail, the HD-sEMG 206 can provide... Figure 9A Multiple EMG channels, wherein each of the EMG signal channels is configured in Figure 1 Array of patients on the skin of 120 Figure 4 An electrode is generated in one of the electrodes 406. As used herein, an array of electrodes 406 means at least a 2×1 array or a 1×2 array, and is envisioned to include an array of dozens or hundreds of electrodes 406.

[0071] When an electrode or electrode array is described as being attached to or connected to muscle, it should be understood that the electrode or electrode array is indirectly attached to or connected to muscle through the skin. Multiple EMG channels contained within the HD-sEMG signal 208 can... Figure 1 The processor 112 analyzes and provides neural drive signals 210.

[0072] Neural drive signals 210 can be spatial triangulations of muscle activity and activation over time, such as regarding Figure 9A The motor unit spike sequence 912 is depicted and described. Neural driving signals 210 can be further processed, as described below regarding... Figure 10 and Figures 11A to 11L The described method derives additional signals useful in generating and identifying pre-Parkinsonian neurodegenerative information, which can enable early diagnosis.

[0073] Information about Parkinson's disease and its pre-Parkinsonian neurodegenerative processes can be generated by at least two physical conditions, including resting tremor and gait freezing. Resting tremor is a muscle tremor of approximately 4 to 6 Hz and is specific to... Figures 3A to 11L The basis of the description.

[0074] Gait freeze is an aberrant gait pattern, characterized by a temporary pause in movement during a single step, and is specifically targeted at… Figures 12 to 16B This forms the basis for the description. Therefore, the neural driving signal 210 can be evaluated to provide the neural origin of the resting tremor signal 212 and the neural origin of the gait freeze signal 214.

[0075] For example, such as regarding Figures 8A to 11L As described, the neural origin of the resting tremor signal 212 can be calculated and determined from the neural driving signal 210. Furthermore, as regarding... Figures 15A to 16B As described, the neural origin of the gait freeze signal 214 can be calculated and determined from the neural drive signal 210.

[0076] As used herein, neural drive signal 210, resting tremor signal 212, and gait freeze signal 214 should be understood as pre-Parkinson's disease information. Sensor box 202 may include other sensors besides HD-sEMG 206, such as inertial measurement unit (IMU 216) and PPG sensor 218. IMU 216 can provide IMU signal 220.

[0077] IMU signal 220 may include magnetic readings, accelerometer readings, and gyroscope readings. IMU signal 220, and specifically accelerometer readings, can be used by signal block 204 with little modification, since the combined acceleration of sensor 104 is useful in determining both gait freeze and resting tremor.

[0078] However, it has been found that the neural driving signal 210, which provides the neural origin of the gait freeze signal 214 and the neural origin of the resting tremor signal 212, can be used to generate and identify pre-Parkinsonian neurodegenerative information before the patient 120 exhibits even visible tremor or gait abnormalities. Furthermore, it has been found that the neural driving signal 210 is sensitive enough to generate pre-Parkinsonian neurodegenerative information even before tremor or gait abnormalities appear in the acceleration readings of the IMU signal 220.

[0079] The incredible sensitivity and discriminative power provided by the neural drive signal 210 arises from spatial triangulation of muscle-driven activity, which is depicted as a motor unit spike sequence 912 and depends on the HD-sEMG signal 208, which can be detected by the combined sensor 104 and analyzed within the signal frame 204, even when other physical characteristics, including tremor and gait abnormalities, are not performed by the patient 120 at a level detectable using the IMU 216 or PPG sensor 218.

[0080] PPG sensor 218 can provide PPG signal 222. PPG sensor 218 can generate a waveform based on changes in the attenuation of light energy as it passes through the tissue of patient 120. (See also: ...) Figure 8C As described, analysis of the PPG signal 222 within signal block 204 can provide an indication of blood pressure 224 by estimating the systolic and diastolic pulse waves. Furthermore, the time interval between systolic pulses can be determined, from which heart rate 225 and heart rate variability 226 can be derived.

[0081] While heart rate variability 226 is not specific to Parkinson's disease, it is associated with it and provides a heart rate variability threshold 227 to determine the level of heart rate variability present in Parkinson's disease. Similarly, initial orthostatic hypotension (IOH), which occurs when blood pressure 224 drops during the first thirty seconds after standing, is associated with Parkinson's disease.

[0082] For example, when the systolic pulse wave 822 in Figure 8 detects a drop in blood pressure 224 below the IOH threshold 228 within 30 seconds of standing, patient 120 has IOH. As used herein, IMU signal 220, blood pressure 224, heart rate 225, and heart rate variability 226 should be interpreted as pre-Parkinson's disease information. In general, neural drive signal 210, the neural origin of resting tremor signal 212, the neural origin of gait freeze signal 214, IMU signal 220, blood pressure 224, heart rate 225, and heart rate variability 226 are considered neuromuscular signals in this paper.

[0083] Analysis of the combined sensor output signal 108, performed within signal box 204 and on processor 112, can be provided to result box 230. Result box 230 can provide and display Parkinson's disease predisposition score 232 and Parkinson's disease progression score 234.

[0084] The Parkinson's Disease Propensity Score 232 and the Parkinson's Disease Progression Score 234 can be displayed visually on a monitor or printed output, via an audio device such as a speaker, or via a tactile or sensory device. The Parkinson's Disease Propensity Score 232 can be displayed as follows: Figure 3A The tremor score 236, as described and calculated, and as per... Figure 12 The described gait freeze score 238, or Parkinson's disease tendency score 232, can be the average of the tremor score 236 and the gait freeze score 238.

[0085] Tremor score 236 and gait freeze score 238 can be tracked over time, and changes in these two scores can be provided to patients as Parkinson's Disease Progression Score 234 120. As used in this article, Parkinson's Disease Propensity Score 232, Parkinson's Disease Progression Score 234, tremor score 236, and gait freeze score 238 should be understood as pre-Parkinson's disease information.

[0086] Now for reference Figure 3A The diagram shows the calculation for... Figure 1 The control flow for resting tremor information in the neuromuscular assessment system 100. Example, the control flow may include calculating... Figure 2 The tremor score is 236. Calculating resting tremor information can be understood as generating pre-Parkinson's disease information. The neuromuscular assessment system 100 can begin calculating resting tremor information by collecting data in step 302.

[0087] Step 302 of collecting may include... Figure 2 The HD-sEMG 206 is placed on the agonist / antagonist muscle pairs from both limbs, as per [reference needed]. Figure 1 As described. For monitoring resting tremor, the lower arm FCR and ECR functioned well.

[0088] Resting tremor is a symptom of Parkinson's disease. Resting tremor is a low-frequency component of neural drive, characterized by bilateral asymmetrical and alternating contractions between agonist and antagonist muscles. Resting tremor can provide information from... Figure 2 The relative sinusoidal displacement readings of the IMU 216. Resting tremor is periodic and synchronous between agonist / antagonist pairs, while voluntary contractions are asynchronous.

[0089] Previous studies failed to identify resting tremor before it could be observed through body movement because there was no... Figure 9AThe motion unit spike sequence 912 level or using Figures 11A to 11L The relevant methods for evaluating neural drive are limited. This leads to the inability to identify and differentiate resting tremor until many early treatments become more difficult, expensive, or infeasible.

[0090] Collection step 302 may include Figure 2 The collection of HD-sEMG signals 208, and for example in Figure 8A As further illustrated, when collection step 302 is continuously running, the HD-sEMG signal 208 can be collected or stored. Figure 1 In the non-transitory computer-readable medium 110.

[0091] It has been found that HD-sEMG data provides richer information than conventional sEMG by adding spatial dimensions to EMG signals, thus allowing for spatial triangulation of muscle activity and activation over time. This can be calculated in step 304 of the computational neural drive. Figure 2 210.

[0092] The neural drive signal 210 can be the motor unit spike sequence 912. That is, the neural drive calculation step 304 can retrieve the HD-sEMG signal 208 from the HD-sEMG 206 and calculate the motor unit spike sequence 912.

[0093] The operation of computational neural drive step 304 is about Figure 9A A more detailed description was provided, and it involved decomposing the HD-sEMG signal 208 into... Figure 9A The individual motor unit action potentials (MUAP 920) are abstracted to only target specific motor unit action potentials. Figure 5 The peak time of MUAP 920 for each motion unit 508 is retained in the motion unit peak sequence 912. Once the motion unit peak sequence 912 has been calculated, it can be summed to calculate and determine the peak time. Figure 10 The cumulative peak sequence (CST 1002).

[0094] Can target Figure 1 The combined sensor 104 is used to calculate CST 1002 for each of the combined sensors. Continuing from... Figure 1 For the example of the left and right lower arms 122, the computational neural drive step 304 can compute the CST for each of the left ECR, right ECR, left FCR and right FCR.

[0095] For example, computational neural drive step 304 can apply a low-pass filter to frequencies above 20 Hz. The low-pass filter of computational neural drive step 304 can be used to target resting tremor, which is a low-frequency event in the range of 4 Hz to 6 Hz.

[0096] Once CST 1002 is determined in the neural drive calculation step 304, the neuromuscular assessment system 100 can perform the CST comparison step 306. The CST comparison step 306 can calculate the relevant signal for CST 1002.

[0097] For example, it can be targeted at Figure 11A The left ECR depicted in the image is CST 1002. Figure 11I The CST1002 of the right ECR depicted in the middle Figure 11F The left FCR depicted in CST 1002 and Figure 11L The CST 1002 of the right FCR depicted in the figure is used to calculate the autocorrelation. Alternatively, the CST comparison step 306 can also correlate the motion unit spike sequence 912 instead of CST 1002.

[0098] Furthermore, CST comparison step 306 can identify the relevant signals as cross-correlations of CST 1002 for each agonist / antagonist pair in the agonist / antagonist pair. Additionally, CST comparison step 306 can also autocorrelate and cross-correlate each motor unit spike sequence in the motor unit spike sequence 912. Importantly, cross-correlation between agonist / antagonist muscles on the same limb can provide an indication of Parkinson's disease and pre-Parkinson's symptoms before any observable symptoms appear.

[0099] Illustratively, Figure 11J The right ECR and right FCR shown and their positions are shown in the figure. Figure 11K The inverted terms in the diagram are shown as periodic and synchronous signals indicating the presence of Parkinson's disease or pre-Parkinson's symptoms. Similarly, Figure 11B The left ECR and left FCR shown and their positions are shown in the figure. Figure 11E The inverted terms in the table are shown as those that do not exhibit periodic or synchronous signals indicating the absence of Parkinson's disease or pre-Parkinson's disease symptoms.

[0100] It is worth noting that cross-correlation between left and right ECR or left and right FCR often shows little or no correlation, because the resting tremor in Parkinson's disease does indeed manifest as tremor together. Once Figures 11A to 11L Once each of the autocorrelation and crosscorrelation is determined, the neuromuscular assessment system 100 can perform reporting step 308.

[0101] Reporting step 308 can identify important pre-Parkinsonian neurodegenerative information in CST 1002 and motor unit spike sequence 912. For example, this pre-Parkinsonian neurodegenerative information may include synchronicity 310 and periodicity 312 calculated from CST 1002 or motor unit spike sequence 912.

[0102] Synchronicity 310 can refer to the degree of coordinated firing of motor units, determined by the density of CST 1002 or the number of individual MU spikes over a period of time. One method for determining synchronicity 310 can be based on the cross-correlation of CST 1002 between agonist / antagonist pairs. Figure 11J and Figure 11K The high degree of synchronization between the relevant signals is shown 310.

[0103] Figure 11J and Figure 11K The cross-correlation is shown as the degree of synchronization between periodic peaks and periodic peaks. The periodic peaks will... Figure 11J A smaller peak is located to the right of a larger peak. The distance between the peaks indicates the synchronization amount or phase shift between the cross-correlated inputs CST 1002.

[0104] In one proposed method, a synchronization threshold 316 can be used to determine whether related signals are synchronized. If the horizontal timing distance between peaks is less than the synchronization threshold 316, the signals can be considered synchronized. If the distance between peaks is greater than or equal to the synchronization threshold 316, the signals can be considered asynchronous.

[0105] Synchronization 310 provides an indication of when the motor unit spike sequence 912 or CST 1002 begins and ends relative to each other. Parkinson's tremor is synchronous, while voluntary muscle movements are asynchronous. Other methods for determining synchronization 310 are envisioned, including coherence analysis or phase analysis of the cross-correlation waveforms of CST 1002 or motor unit spike sequence 912, to determine the temporal shift of the signal phase relative to each other.

[0106] One proposed method for determining the periodicity 312 of a cross-correlation signal could include calculating the power spectral density for each cross-correlation. Peaks would be observed at periodic frequencies within the correlated signal.

[0107] If no identifiable frequency is found in the power spectral density, the signal can be considered aperiodic; otherwise, if a frequency with a correlated power spectral density is found, periodicity is found. As will be understood, when the correlated CST 1002 or the motion unit spike sequence 912 is periodic, oscillations will be observed and identified in the correlation coefficient based on hysteresis.

[0108] When both synchronicity 310 and periodicity 312 are detected, resting tremor can be identified within the correlated signal, regardless of whether the correlation is between CST 1002 and the motor unit spike sequence 912. The tremor amplitude 318 can also be calculated during reporting step 308.

[0109] The flutter amplitude 318 can be the amplitude of the peaks in the autocorrelation or cross-correlation waveforms. In other words, the vertical distance from the minimum trough to the maximum peak can be considered as the flutter amplitude 318.

[0110] During reporting step 308, the neuromuscular assessment system 100 can further determine the tremor frequency 320. The tremor frequency 320 can be determined by acquiring the power spectral density of the relevant signal. The power spectral density of the relevant signal may have a peak at a periodic frequency. A dominant peak is selected for the tremor frequency 320.

[0111] During reporting step 308, the neuromuscular assessment system 100 may further determine an antagonistic tremor score 322. The antagonistic tremor score 322 may be determined as the percentage of relevant signals of the agonist / antagonist pair exhibiting resting tremor.

[0112] In other words, the percentage of relevant signals for an agonist / antagonist pair identified as having both synchronicity 310 and periodicity 312 will be the antagonist tremor score 322. For example, this calculation is performed in... Figure 3B Described in the text.

[0113] During reporting step 308, the neuromuscular assessment system 100 may further determine a bilateral tremor score 326. The bilateral tremor score 326 may be determined as the percentage of relevant signals from similar muscles on both the left and right sides of the body that exhibit resting tremor.

[0114] In other words, the percentage of relevant signals for similar muscles from both the left and right sides of the body can be, for example, the cross-correlation between the left ECR 802 and the right ECR 806 in Figure 8. The percentage of relevant signals for similar muscles from both the left and right sides of the body, identified as having both synchronicity 310 and periodicity 312, would be the bilateral tremor score 326. For example, this calculation is performed in... Figure 3B Described in the text.

[0115] During reporting step 308, the neuromuscular assessment system 100 may further determine a tremor score 236. The tremor score 236 may be determined as the percentage of the total cross-correlation signal exhibiting resting tremor.

[0116] In other words, the percentage of cross-correlated signals identified as having both synchronicity 310 and periodicity 312 would be the tremor fraction 236. For example, this calculation is performed in... Figure 3B The description is in the middle. The tremor fraction over time, 236, can be reported as... Figure 2 The Parkinson's disease progression score is 234.

[0117] It should be understood that synchronicity 310, periodicity 312, tremor amplitude 318, tremor frequency 320, antagonistic tremor score 322, bilateral tremor score 326, and tremor score 236 are considered to be neurodegenerative information in the early stages of Parkinson's disease.

[0118] Another aspect of the neuromuscular assessment system 100 can be a convolutional neural network 332. The convolutional neural network 332 can be a continuous learning algorithm with a sequence of spikes 912 of individual motor units as input and a tremor score 236 as output.

[0119] It is conceivable that the convolutional neural network 332 may include hidden layers employing a sigmoid activation function, which provides results from 0 to 1 and can be used to predict probabilities such as the total tremor score 330. Furthermore, the hidden layers may employ a hyperbolic tangent (tanh) activation function for the hidden states.

[0120] Furthermore, the hidden layers can employ a softmax activation function, which is envisioned to determine non-co-occurrences in the MUAP 920 of Figure 9. Even further, the mean absolute error activation function can be used to calculate the error. Recurrent cells appear in the hidden layers.

[0121] Now for reference Figure 3B The diagram shows the calculations. Figure 3A The flowchart for tremor scores. Figure 1 Neuromuscular assessment system 100 and Figure 3A The report step 308 is shown together, thereby comparing agonist / antagonist pairs on the same limb in agonist comparison step 340.

[0122] The agonist comparison step 340 includes agonist / antagonist pairs from the same limb. Figure 10 The cross-correlation of CST 1002 or the motor unit peak sequence 912 in Figure 9. Specifically, in the agonist comparison step 340, the first left extensor muscle is correlated with the first left flexor muscle.

[0123] Similarly, the second agonist / antagonist pair and the third agonist / antagonist pair are shown as being related to the left side. The first right extensor is related to the first right flexor, and the second agonist / antagonist pair and the third agonist / antagonist pair are shown as being related to the right side.

[0124] Determined to have Figure 3A Synchronization 310 and Figure 3AThe correlation signal of both periodicity 312 and periodicity 312 will be considered as resting tremor and marked with 1. The correlation signal that does not have synchronicity 310 or periodicity 312 will not be considered as resting tremor and will be marked with 0. Here, three of the six correlation signals are marked as showing tremor, and therefore the antagonistic tremor score 322 is calculated as the percentage of correlation signals that are determined to have both synchronicity and periodicity, or 0.5% or 50%.

[0125] The neuromuscular assessment system 100 further illustrates in reporting step 308 that, simultaneously in bilateral comparison step 344, signals from similar muscles on both the left and right sides of the body are correlated. Continuing with the same 3×3 example above, bilateral comparison step 344 may be a cross-correlation of CST 1002 or motor unit spike sequence 912 for similar muscles from opposite sides of the body.

[0126] Specifically, in the bilateral comparison step 344, the first left extensor muscle is associated with the first right extensor muscle. Similarly, the second and third extensor muscles are shown as associated between the left and right sides of the body. The first left flexor muscle is associated with the first right flexor muscle. Similarly, the second and third flexor muscles are shown as associated with the left and right sides of the body.

[0127] Correlated signals identified as having both synchronicity 310 and periodicity 312 are considered resting tremor. Here, four of the six related signals are identified as exhibiting tremor, and therefore the bilateral tremor score 326 is calculated as the percentage of related signals identified as having both synchronicity and periodicity, or 0.667 or 67%.

[0128] Once the bilateral tremor score 326 and the antagonistic tremor score 322 have been calculated, the total tremor score 330 can also be calculated. Examplely, the total tremor score 330 can be the total percentage of the cross-correlation signals determined from both the agonist comparison step 340 and the bilateral comparison step 344, which are both synchronous and periodic.

[0129] Here, the total flutter score 330 is determined as the total percentage of the cross-correlated signal having both synchronicity and periodicity, either 0.58 or 58%. The total flutter score 330 over time can be reported as... Figure 2 The Parkinson's disease progression score is 234.

[0130] Now for reference Figure 4 , which shows Figure 1 A side view of the forearm 122. The forearm 122 is depicted as the right forearm with a combined sensor 104 attached thereto. It is conceivable that the combined sensor 104 may be attached to an agonist / antagonist pair of muscles on the palm, wrist, forearm, or a combination thereof.

[0131] Specifically, the combined sensor 104 may include a first combined sensor 402 and a second combined sensor 404. Each combined sensor in the combined sensor 104 may include an HD-sEMG 206, an IMU 216, and a PPG sensor 218.

[0132] The HD-sEMG 206 may include multiple electrodes 406, wherein each electrode 406 has a known and fixed distance 408 between it. The electrodes 406 may be arranged in a grid configuration. Here, the electrodes 406 are shown in a five-by-thirteen grid configuration.

[0133] Electrode 406 can be disposed within electrode carrier 410. It is conceivable that electrode carrier 410 can be a polyimide material with a thickness of approximately 50 μm.

[0134] It can be further envisioned that electrode 406 could be a silver or silver chloride material that provides an impedance of approximately five thousand ohms between skin 412 and electrode 406. In the worst-case scenario, the impedance can reach one megaohm while maintaining functionality.

[0135] The electrode 406 itself is envisioned to have a small diameter of less than three to five millimeters. Furthermore, the fixed distance 408 between the electrodes 406 should be less than eight to ten millimeters.

[0136] It is conceivable that, when targeting resting tremor, the combined sensor 104 could be placed on multiple different agonist / antagonist pairs on the lower arm 122. Here, the first combined sensor 402 is shown attached to the skin 412 above the FCR 414, while the second combined sensor 404 is attached to the skin 412 above the ECR 416.

[0137] Component box 102 can be communicatively coupled to combined sensor 104, and can... Figure 1 The combined sensor output signal 108 is transmitted to the component box 102. Alternatively, it is conceivable that the component box 102 can be integrally formed with one or more of the combined sensors in the combined sensor 104, which can reduce the amount of wiring required.

[0138] HD-sEMG 206 can provide Figure 9A Multiple EMG channels, where each EMG signal channel is composed of multiple EMG channels. Figure 1 Electrodes 406 are generated on the skin of patient 120 at a fixed distance 408, closely spaced. Multiple EMG channels contained within the HD-sEMG signal 208 can... Figure 1 The processor 112 analyzes and provides neural drive signals 210.

[0139] It has been found that using multiple electrodes 406 arranged in a grid at a fixed distance 408 enables [the following]: Figure 9B The spatial dimension of each MUAP in MUAP 920 generates neural driving signals 210. This finding further enables the evaluation of information about Parkinson's disease and its pre-Parkinson's disease stages before symptoms become observable by physicians.

[0140] Now for reference Figure 5 , which shows Figure 1 An isometric view of the neuromuscular assessment system 100. The neuromuscular assessment system 100 is shown as having a combined sensor 104 near muscle fibers 502. It should be understood that the combined sensor 104 is coupled to... Figure 1 The patient's skin 120 and indirectly coupled to muscle fibers 502.

[0141] Muscle fiber 502 can be Figure 4 ECR 416 or Figure 4 The muscle fibers within FCR 414. Combined sensor 104 with... Figure 4 An alternative configuration is shown in which the electrodes 406 are distributed along two rows or alternatively there may be two electrodes 406.

[0142] The combined sensor 104 is shown as further including a PPG sensor 218 and an IMU 216. The combined sensor 104 can be provided via a wireless transmission 106 to the component box 102. Figure 1 The combined sensor output signal 108. More specifically, the combined sensor 104 can provide a signal with... Figure 9A HD-sEMG signals 208 from multiple EMG channels.

[0143] The combined sensor 104 can also provide PPG signals 222 and IMU signals 220 to the component box 102 via wireless transmission 106. The muscle fiber 502 can be innervated by motor neurons 504 extending between the spinal cord 506 and the muscle fiber 502 in the patient 120.

[0144] Motor neurons 504 and muscle fibers 502 may together comprise a motor unit 508. Each innervation of the motor unit 508 may be... Figure 9B One of the MUAPs in the MUAP 920. The MUAP 920 for each motion unit 508 is described as follows. Figure 9A The motor unit peak sequence 912, and the motor unit peak sequence 912 for multiple motor units 508 within the muscle can be Figure 10 CST 1002.

[0145] It can be envisioned that the electrodes 406 should be small (less than three to five millimeters in diameter), spaced eight to ten millimeters apart, and placed distal to the nerve innervation area.

[0146] Now for reference Figure 6 The diagram shows the method for use with Figure 1 A schematic diagram of the analog front end 602 (AFE) used in conjunction with the neuromuscular assessment system 100. Figure 4 or Figure 5 Electrode 406 can be directly or indirectly coupled to input 604 of AFE 602.

[0147] It is conceivable that each electrode in electrode 406 will be coupled to a respective input in input 604. Therefore, it is conceivable that AFE 602 has enough inputs 604 to accommodate the array (e.g., Figure 4 Direct connection of each electrode in electrode 406 of the array disclosed herein.

[0148] Input 604 can be coupled to amplification. Figure 9A The multi-stage amplifier 606 is used for the EMG channel. Before amplification, the EMG signal channel can be DC filtered by a capacitor between electrode 406 and multi-stage amplifier 606. The output of multi-stage amplifier 606 can be an amplified EMG signal that can be input into analog multiplexer 608.

[0149] The output of the analog multiplexer 608 can be digitized using the analog-to-digital converter 610. The digitized EMG signal can be stored in a digital register 612, which can be... Figure 1 The non-transitory computer-readable medium 110. The AFE 602 can be controlled by a digital controller 614 having a serial peripheral interface that provides synchronous serial communication.

[0150] Now for reference Figure 7 The diagram shows the method for use with Figure 1 A schematic diagram of the sEMG amplifier 702 used in the neuromuscular assessment system 100. The sEMG amplifier 702 can be used for input... Figure 6 The AFE 602 has a low input bias current buffer because if there is poor contact between electrode 406 and patient 120, the input bias current of the AFE 602 can polarize the electrode.

[0151] The sEMG amplifier 702 may include a junction field-effect transistor input operational amplifier (JFET input operational amplifier 704) with an input bias current of less than 1 picoampere. It is envisioned that the sEMG amplifier 702 has sufficient JFET input operational amplifiers 704 to accommodate an array (e.g., Figure 4Direct connection of each electrode in electrode 406 of the array disclosed herein.

[0152] The second JFET input instrumentation amplifier 706 can have an input bias current of less than 20 picoamps. It can be assumed that the input impedance for the sEMG amplifier 702 should be lower than... Figure 1 The maximum expected resistance between electrode 406 and patient 120 is at least 100 times greater.

[0153] Use silver or silver chloride electrodes, provided the skin is properly prepared. Figure 1 The impedance between the patient's skin (120°) and electrode (406) at 10 Hz is typically about 5 kilohms, but in the worst case, it is about 1 megahm.

[0154] Electrodes 406 should be small (diameter less than three to five millimeters), arranged at intervals of less than eight millimeters, and placed... Figure 5 The distal side of the nerve innervation area of ​​muscle fiber 502.

[0155] Here, the JFET input operational amplifier 704 is envisioned to have an input impedance greater than 100 megohms, common-mode rejection greater than 100 dB, a gain of one kilohertz, and a bandwidth greater than one kilohertz. The driven right leg 708 can drive the common-mode voltage back to the patient's skin at 120°. The feedback loop of the driven right leg 708 increases the common-mode rejection by an amount equal to the sum of the closed-loop gain of the driven right leg 704.

[0156] Overall reference Figures 8A to 8D These figures are used as Figure 1 Examples of multiple combined sensor output signals 108 are provided for the early identification of resting tremor. The combined sensor output signals 108 will include signals for... Figure 1 left arm 124 and Figure 1 On the right arm 126 Figure 1 The signal of one of the combined sensors in the combined sensor 104.

[0157] More specifically, the combined sensor output signal 108 will include signals targeting... Figure 4 The first combined sensor 402 and Figure 4 The signals from the second combined sensor 404, these combined sensors such as Figure 4 The depicted coupling to the right arm 126 and will further include coupling to Figure 1 Patient 120's left arm 124 Figure 4 ECR 416 and Figure 4 The FCR 414 has two additional combined sensors 104. For ease of description, when referring to the electrode array, ellipses are used to depict further readings from the individual combined sensors 104, and when referring to... Figure 2 IMU signal 220 and Figure 2 When the PPG sensor 218 is used, an ellipse is used to depict more readings from different combination sensors 104 on the same arm.

[0158] Now for reference Figure 8A , which shows Figure 1 A graphical depiction of the HD-sEMG signal 208 of the combined sensor 104. The HD-sEMG signal 208 may include HD-sEMG signals for the electrode array of the combined sensor 104, which includes the left ECR 802, left FCR 804, right ECR 806, and right FCR 808.

[0159] For clarity, a complete array of HD-sEMG signals 208 for each of ECR ​​and FCR is not shown. The vertical axis may be a measurement in volts, with each partition providing one millivolt. The horizontal axis may be a measurement in time, with each partition providing two hundred milliseconds.

[0160] Resting tremor involves a low-frequency component of neural drive, which is bilaterally asymmetrical and asynchronous through alternation between agonist and antagonist muscles in the same limb. Less common resting tremor can be synchronous between agonist and antagonist muscles.

[0161] This can manifest as tremors in one arm in the later stages of Parkinson's disease. Bilateral asymmetry is depicted by HD-sEMG signals 208 targeting the right ECR 806 and right FCR 808, which show little activity, while the left ECR 802 and left FCR 804 show activity.

[0162] Furthermore, alternating characters are depicted within a first time window 810 and a second time window 812. Examplely, time window 810 shows alternating activation as activation in the left ECR 802 and minimal activation in the left FCR 804, while time window 812 shows alternating activation as activation in the left FCR 804 and minimal activation in the left ECR 802.

[0163] In the later stages of Parkinson's disease, bilateral asymmetry and alternation of agonist / antagonist muscles can be observed in EMG readings or visually. Figure 1 Patients were observed in 120 cases. However, early and pre-Parkinson's disease symptoms are not so easy to identify, but can be detected using the method proposed in this article.

[0164] In other words, Figure 4 The array of electrodes 406 can provide spatial dimensions to the data, thereby allowing for the generation of... Figure 2The neural driving signals 210 generate information about pre-Parkinsonian disease and early pre-Parkinsonian disease, which is not available through previous developments in the field.

[0165] Now for reference Figure 8B , which shows Figure 1 A graphical depiction of the IMU signal 220 of the combined sensor 104. The IMU signal 220 may include signals corresponding to those coupled to the sensor respectively. Figure 1 left arm 124 and Figure 1 The right arm 126 Figure 1 The combined sensor 104 has a left acceleration 814 and a right acceleration 816.

[0166] For clarity, only a single IMU signal 220 is shown for the left arm 124 and the right arm 126. The vertical axis can be a measurement of physical displacement, with each segment providing one millimeter. The horizontal axis can be a measurement of time, with each segment providing two hundred milliseconds.

[0167] The bilateral asymmetry of the resting jitter is depicted by IMU signal 220, which shows a relative sinusoidal displacement on the left acceleration 814, while almost no displacement is recorded for the right acceleration 816. It should be understood that the resting jitter will exhibit the approximate sinusoidal displacement shown in the left acceleration 814; and can be interpreted together with the right acceleration 816 as a bilateral asymmetry between the two sensors.

[0168] Now for reference Figure 8C , which shows Figure 1 A graphical depiction of the PPG signal 222 from the combined sensor 104. The waveform of the PPG signal 222 is generated by the attenuation changes of light energy as it passes through biological tissue, is scattered and reflected within the biological tissue, or is absorbed by the biological tissue before being detected by the optical sensor.

[0169] PPG signal 222 may include signals corresponding to those coupled to respectively. Figure 1 left arm 124 and Figure 1 The right arm 126 Figure 1 The combined sensor 104 contains the left PPG signal 818 and the right PPG signal 820.

[0170] For clarity, only a single PPG signal 222 is shown for the left arm 124 and the right arm 126. The vertical axis can be a measurement in volts, with each segment providing one-tenth of a millivolt. The horizontal axis can be a measurement in time, with each segment providing two hundred milliseconds.

[0171] Figure 2 Analysis of the PPG signal 222 within signal frame 204 can provide insights into the systolic pulse wave 822 and diastolic pulse wave 824. Figure 2The blood pressure reading is 224. Furthermore, the time interval between systolic pulse waves 822 can be determined, from which the following can be derived. Figure 2 Heart rate 225 and Figure 2 Heart rate variability 226.

[0172] For example, a heart rate of 225 can be calculated by dividing sixty by the time in seconds between the systolic pulse wave 826 and the heart rate variability 226. Heart rate variability 226 can be calculated from a heart rate of 225 or directly from the systolic pulse wave 822. Heart rate variability 226 is an early indicator of Parkinson's disease, but it is not specific enough on its own. Therefore, it is conceivable to correlate heart rate variability 226 with... Figure 2 The HD-sEMG signal 208 is combined to improve specificity.

[0173] Now for reference Figure 8D , which shows Figure 8B A graphical representation of the IMU signal 220 in the frequency domain. For ease of description, only the IMU signal 220 is shown. Figure 8B The left acceleration 814 and the right acceleration 816 are shown.

[0174] The vertical axis can be a measurement of acceleration or gravity, with each section providing 6 × 10⁻⁴ g². The horizontal axis can be a measurement of frequency, with each section providing two hertz.

[0175] The bilateral asymmetry of resting tremor was characterized by IMU signal 220, which showed acceleration spikes between 4 Hz and 6 Hz. This corresponds to Figure 1 The left arm 124 exhibits a relative sinusoidal displacement during a static tremor, while almost no spike is observed in the right acceleration 816.

[0176] The sine addition score can be correlated with the Parkinson's disease propensity score 232, tremor score 236, gait freeze score 238, and Parkinson's disease progression score 234 (all within...). Figure 2 (The values ​​are added together.) For example, if a sine wave is detected to show an acceleration spike between 4 Hz and 6 Hz, the sine wave addition could be ten percentage points.

[0177] Now for reference Figure 9A , which shows the use of Figure 3A A graphical depiction of the decomposition process 900 of the computational neural drive step 304. The decomposition process 900 first includes receiving data for... Figure 1 The combined sensor 104 Figure 4 Each electrode in electrode 406 Figure 2 The HD-sEMG signal 208. For clarity, ellipses are used to depict the additional HD-sEMG signals 208 from the other electrodes 406 in the array used for the various combined sensors 104.

[0178] Example, for instance, surface EMG readings from each of the electrodes in electrode 406 can be EMG channels. Here, a first EMG channel 902, a second EMG channel 904, a third EMG channel 906, and an nth EMG channel 908 are shown. It is conceivable that dozens or hundreds of EMG channels can be provided by each of the combined sensors in combined sensor 104.

[0179] Action potentials of each motor unit ( Figure 9B The MUAP 920 passes through the motion unit, such as Figure 5 The motion unit 508. MUAP 920 is detected using different electrodes 406 within the array of combined sensors 104, and due to the interaction between electrodes 406... Figure 4 The fixed distance 408 will be used to detect MUAP 920 by utilizing the difference in waveform shape between electrodes 406.

[0180] Compared to MUAP 920 located further away from electrode 406, MUAP 920 located closer to electrode 406 will be detected with greater intensity. These MUAP 920s of varying intensities from multiple different motion units 508, along with noise, can be mixed together and detected by a single electrode 406 as one of the channels from the first EMG channel 902 to the nth EMG channel 908.

[0181] It is conceivable that the HD-sEMG signal 208 can be decomposed using a variety of available methods, ranging from older stencil-driven decomposition to decomposition by... Figure 1 Command 114 controls Figure 1 The processor 112 contains a modern recurrent neural network. The decomposition algorithm 910 can take the HD-sEMG signal 208 as input and return a sequence of motion unit spikes 912.

[0182] The horizontal axis of the HD-sEMG signal 208 and the motion unit spike sequence 912 can be time in seconds or milliseconds. The vertical axis of the HD-sEMG signal 208 can be voltage, while the vertical axis of the motion unit spike sequence 912 can be dimensionless and only represent MUAP 920 existing at a specific time along the horizontal direction.

[0183] The motion unit spike sequence 912 is depicted as MU1 to MU6 corresponding to the six motion units 508, the number of which is exemplary but can actually include any number of motion units 508 activated by the detectable MUAP 920. Ellipses are used to depict other possible motion unit spike sequences 912.

[0184] As an example, it can be envisioned that a gated recurrent neural network following a gradient convolution kernel algorithm could be employed in the decomposition process 900, which relies on retrospective computation based on past and current values ​​to decompose and identify MUAP 920.

[0185] The architecture of the gated recursive decomposition algorithm 910 can be described as follows:

[0186] Equation 1

[0187] Equation 2

[0188] Equation 3

[0189] Equation 4

[0190] Where xt and ht-1 are the input vector at time t and the hidden vector from the previous time step, respectively. rt, zt, and nt are the reset, update, and new gate vectors, respectively. The output of cell ht is the new hidden vector and also the new output vector. The gate is parameterized by weights W and bias b. σ and tanh are the sigmoid and hyperbolic tangent activation functions, respectively. It is the Hadamard product.

[0191] Common elements are decomposed and identified, and each decomposed signal corresponds to a motion unit in motion unit 508. Multiple electrodes 406 provide spatial resolution, while temporal resolution is derived from the decomposition process 900.

[0192] Furthermore, motor unit activation is essentially a point process and is extremely unlikely to occur simultaneously with another source. The softmax activation function can take advantage of this property:

[0193] softmax(yi) = exp(yj ) / ∑j (yj ) Equation 5

[0194] The softmax function can transform a vector into a probability distribution and can be used as the final activation function in the decomposition algorithm 910, which provides the probability distribution over the predicted output class. The sigmoid and hyperbolic tangent functions can also be used in the hidden layers of the decomposition algorithm 910.

[0195] The motor unit spike sequence 912 (MU6 to MU1) can be a set of MUAPs 920 generated by a motor unit 508 and shown as motor unit action potential spikes 914. The motor unit action potential spikes 914 are located at their occurrence time or separated by their firing intervals 916. Thus, temporal resolution is maintained within the motor unit spike sequence 912, and spatial resolution is provided by identifying each motor unit in the different motor units 508 innervated by each action potential spike in the action potential spikes 914.

[0196] The decomposition algorithm 910 requires HD-sEMG signal 208 as input and returns a motor unit spike sequence 912 as output. This process cannot actually be performed in the human brain; for example, the human brain is not equipped to receive HD-sEMG signal 208, and the computation time would limit any real-world application of the decomposition algorithm 910. In fact, the decomposition algorithm 910 cannot be fully executed in the human brain.

[0197] Furthermore, the decomposition algorithm 910 requires manipulation of computer data structures (e.g., HD-sEMG signals 208) and the output of the motor unit spike sequence 912, which cannot actually be fully executed in the human brain. Furthermore, Figure 10 The autocorrelation and cross-correlation of the motor unit spike sequences 912 and CST 1002 require manipulation of computer data structures, which cannot actually be performed entirely in the human brain.

[0198] Previous developments in Parkinson's disease surveillance have been plagued by deep-seated technical problems, namely that they were unable to use, or did not utilize, neural drive data, such as the motor unit spike sequences 912 and CST 1002 used to generate information about Parkinson's disease or pre-Parkinson's disease. These technical problems were amplified because previous developments did not analyze neural drive data.

[0199] Here, the use of neurally driven data provides a technology-driven solution to address this technical problem, and also offers a new ability to generate pre-Parkinson's disease information from early neurally driven data without waiting for the appearance of obvious symptoms, a long-standing need in the monitoring industry.

[0200] Neural drive data should be considered as functional data identified by the neuromuscular assessment system 100 for the purpose of generating pre-Parkinson's disease information and neural drive data. Neural drive data also determines the technical capabilities and internal operations of the neuromuscular assessment system 100 in generating pre-Parkinson's disease information.

[0201] The format of neural drive data, as a spike sequence or cumulative spike sequence, inherently includes or reflects the corresponding technical features of the neuromuscular assessment system 100. That is, the neural drive data reflects the physical collection via an electrode array and the decomposition of the raw signal into spike sequences.

[0202] Now for reference Figure 9B , which shows Figure 9A A magnified view of area B. Area 9B can depict MUAP 920.

[0203] MUAP 920 can be a single motion unit action potential distorted by noise 922 and other MUAPs 924, which are present and can be generated by... Figure 4 One of the electrodes in electrode 406 is detected. The horizontal axis can provide partitioned time measurements in milliseconds, while the vertical axis can provide partitioned voltage measurements in millivolts.

[0204] Now for reference Figure 10 , which shows the use of Figure 3A A graphical depiction of the cumulative spike sequence (CST) 1002 of the computational neural drive step 304. Once the motor unit spike sequence 912 of Figure 9 has been computed, the motor unit spike sequence 912 can be summed to compute and determine CST 1002.

[0205] Motor unit spike sequences 912 and CST 1002 will be understood in this paper as neural drive data. CST 1002 can be the total neural drive to the muscle, which is the sum of the individual motor unit spike sequences entering CST 1002.

[0206] The CST 1002 can then be low-pass filtered to maintain frequencies below a low-pass filter threshold (such as below 20 Hz). The low-pass filter can isolate stationary jitter, which is a low-frequency event.

[0207] Now for reference Figures 11A to 11L It describes what can be done Figure 10 The CST 1002 is used to calculate all combinations of autocorrelation and cross-correlation for the left ECR 802, left FCR 804, right ECR 806 and right FCR 808 in Figure 8. Figures 11A to 11L The vertical axis can be a dimensionless number between -1 and +1, indicating the degree of correlation. Figures 11A to 11L The horizontal axis can be a measurement of time lag from -1 second to +1 second.

[0208] Importantly, if Parkinson's disease or pre-Parkinson's symptoms are present, the interaction between agonist / antagonist pairs in the same arm can be periodic and synchronous. The interaction between agonist / antagonist pairs is as follows: Figure 11B , Figure 11E , Figure 11J and Figure 11K As shown.

[0209] When the innervation of the agonist / antagonist muscles does not begin or end simultaneously, the innervation pattern of the input CST 1002 can have a certain degree of asynchrony, which is also described as a phase shift in the input CST. Figure 3A The synchronicity threshold of 316 explains this. The innervation pattern of input CST 1002 can have a periodic correlation when the innervation of the agonist / antagonist pair occurs repeatedly over time.

[0210] Now for reference Figure 11A The document shows the target Figure 8A A graphical depiction of the CST autocorrelation of the left ECR 802. For the left ECR 802... Figure 10 The CST 1002 is autocorrelated with itself and exhibits a classic peak of autocorrelation.

[0211] Now for reference Figure 11B , which shows Figure 8A The left ECR 802 and Figure 8A A graphical depiction of the CST cross-correlation between the left FCR 804. As previously mentioned, if Parkinson's disease or pre-Parkinson's symptoms are present, the cross-correlation between agonist / antagonist pairs in the same arm can be periodic and synchronous.

[0212] Here, the cross-correlation between CST 1002 of the left ECR 802 and CST 1002 of the left FCR 804 did not show a peak, indicating that there is no correlation between the input CST 1002. Resting tremor in Parkinson's disease typically does not occur simultaneously in... Figure 1 The symptoms were observed on both sides of patient 120.

[0213] Therefore, the lack of periodicity and synchronicity in the innervation of agonist / antagonist muscles on one side of Patient 120's body may not rule out Parkinson's disease symptoms, as periodicity and synchronicity could exist on the other side of Patient 120's body. The relationship between CST 1002 at left ECR 802 and CST 1002 at left FCR 804 is... Figure 11B and Figure 11E The mirror image between the cross-correlation.

[0214] Now for reference Figure 11C , which shows Figure 8AThe left ECR 802 and Figure 8A A graphical depiction of the CST cross-correlation between the right ECR 806 and the left ECR 802 at CST 1002 and the right ECR 806 at CST 1002 is depicted as having little correlation because the resting tremor in Parkinson's disease is asymmetrical and usually does not occur simultaneously at the right ECR 806. Figure 1 Patient 120 showed bilateral manifestations. Inverse CST cross-correlation between right ECR 806 and left ECR 802 could also be calculated, and a similar correlation was demonstrated to be lacking, but it is not reproduced for the sake of description.

[0215] Now for reference Figure 11D , which shows Figure 8A The left ECR 802 and Figure 8A A graphical depiction of the CST cross-correlation between the right FCR 808 and the left ECR 802 at CST 1002 and the right FCR 808 at CST 1002 is depicted as having little correlation because the resting tremor in Parkinson's disease is asymmetrical and usually does not occur simultaneously at the right FCR 808. Figure 1 Patient 120 showed bilateral manifestations. Inverse CST cross-correlation between right FCR 808 and left ECR 802 could also be calculated, and a similar correlation was demonstrated to be lacking, but it is not reproduced for the sake of description.

[0216] Now for reference Figure 11E , which shows Figure 8A The left FCR 804 and Figure 8A A graphical depiction of the CST cross-correlation between the left ECR 802. As previously mentioned, if Parkinson's disease or pre-Parkinson's symptoms are present, the cross-correlation between agonist / antagonist pairs in the same arm can be periodic and synchronous.

[0217] Here, the cross-correlation between CST 1002 of left FCR 804 and CST 1002 of left ECR 802 did not show a peak, indicating no correlation between the input CST 1002. Resting tremor in Parkinson's disease typically does not occur simultaneously in... Figure 1 The symptoms were observed on both sides of patient 120.

[0218] Therefore, the lack of periodic and synchronous innervation of the agonist / antagonist muscles on one side of Patient 120's body does not rule out Parkinson's disease symptoms, as periodic and synchronous correlations could exist on the other side of Patient 120's body. The relationship between CST 1002 of left FCR 804 and CST 1002 of left ECR 802 is... Figure 11B and Figure 11E The mirror image between the cross-correlation.

[0219] Now for reference Figure 11FThe document shows the target Figure 8A A graphical depiction of the CST autocorrelation of the left FCR 804. For the left FCR 804... Figure 10 The CST 1002 is autocorrelated with itself and exhibits a classic peak of autocorrelation.

[0220] Now for reference Figure 11G , which shows Figure 8A The left FCR 804 and Figure 8A A graphical depiction of the CST cross-correlation between the right ECR 806 and the left FCR 804 at CST 1002 and the right ECR 806 at CST 1002 is depicted as having little correlation because the resting tremor in Parkinson's disease is asymmetrical and usually does not occur simultaneously at the right ECR 806. Figure 1 Patient 120 showed bilateral manifestations. Inverse CST cross-correlation between right ECR 806 and left FCR 804 could also be calculated, and a similar correlation was demonstrated to be lacking, but it is not reproduced for the sake of description.

[0221] Now for reference Figure 11H , which shows Figure 8A The left FCR 804 and Figure 8A A graphical depiction of the CST cross-correlation between the right FCR 808 and the left FCR 804 at CST 1002 and the right FCR 808 at CST 1002 is depicted as having little correlation because the resting tremor in Parkinson's disease is asymmetrical and usually does not occur simultaneously at CST 1002 and CST 1002. Figure 1 Patient 120 showed bilateral manifestations. Inverse CST cross-correlation between right FCR 808 and left FCR 804 can also be calculated, and a similar correlation is demonstrated to be lacking, but it is not reproduced for the sake of description.

[0222] Now for reference Figure 11I The document shows the target Figure 8A A graphical depiction of the CST autocorrelation of the right ECR 806. For the right ECR 806... Figure 10 The CST 1002 is autocorrelated with itself and exhibits multiple classical peaks of autocorrelation, with the input CST 1002 having periodic properties.

[0223] Now for reference Figure 11J , which shows Figure 8A The right ECR 806 and Figure 8A A graphical depiction of the CST cross-correlation between the right FCR 808. As previously mentioned, if Parkinson's disease or pre-Parkinson's symptoms are present, the cross-correlation between agonist / antagonist pairs in the same arm can be periodic and synchronous.

[0224] Here, the cross-correlation between CST 1002 of the right ECR 806 and CST 1002 of the right FCR 808 does indeed show multiple peaks over time, indicating the correlation between repetitive or periodic innervation recorded in the input CST 1002.

[0225] Furthermore, each of the multiple peaks also exhibits asymmetry, with smaller peaks located on the right and larger peaks on the left. This asymmetry indicates the amount of synchronization or phase shift between the cross-correlated inputs CST 1002, which is determined by... Figure 3A The synchronization threshold of 316 was measured. Therefore, the CST cross-correlation between the right ECR 806 and the right FCR 808 indicates information about Parkinson's disease or pre-Parkinson's disease. Figure 1 The neural drive of patient 120.

[0226] It has been found that the cross-correlation between CST 1002 values ​​of agonist / antagonist pairs provides a solution to long-standing technical problems in previous developments, which were limited to detecting and classifying late-stage Parkinson's disease symptoms when treatment options could become narrower and costs increased. In contrast, the cross-correlation between CST 1002 values ​​of agonist / antagonist pairs is qualitatively different because it evaluates the neural origin of resting tremor in Parkinson's disease with high sensitivity, enabling the generation of pre-Parkinson's information for resting tremor, even if previous developments could not detect or generate such information.

[0227] Although the evaluation of neural drive in patient 120 was shown and described in the evaluation of CST 1002, prior to being incorporated into CST 1002, the assessment of neural drive in patient 120 was not fully understood. Figure 9A The evaluation of the motor unit spike sequence 912 can provide this new qualitative capability, thereby assessing the neural drivers of Parkinson's disease and pre-Parkinson's symptoms in patients 120 with previously unknown sensitivity. The disclosed evaluation of the motor unit spike sequence 912, when incorporated into CST 1002, is for illustrative purposes only.

[0228] The synchronization amount between CST 1002 of the right ECR 806 and CST 1002 of the right FCR 808 will be Figure 11J and Figure 11K The cross-correlation is a mirror image between the two peaks, which is described by each of a plurality of peaks having smaller peaks on different sides of the larger peak.

[0229] Now for reference Figure 11K , which shows Figure 8A The right FCR 808 and Figure 8AA graphical depiction of the CST cross-correlation between the right ECR 806. As previously mentioned, if Parkinson's disease or pre-Parkinson's symptoms are present, the cross-correlation between agonist / antagonist pairs in the same arm can be periodic and synchronous.

[0230] Here, the cross-correlation between CST 1002 of the right FCR 808 and CST 1002 of the right ECR 806 does indeed show multiple peaks over time, indicating the correlation between repetitive or periodic innervation recorded in the input CST 1002.

[0231] Furthermore, each of the multiple peaks also exhibits asymmetry, with smaller peaks located on the left and larger peaks on the right. This asymmetry indicates the degree of synchronization or phase shift between the cross-correlated inputs CST 1002, which is determined by… Figure 3A The synchronization threshold 316 is measured. Therefore, the CST cross-correlation between the right FCR 808 and the right ECR 806 enables measurement based on... Figure 1 The patient's neural drive 120 generates information about Parkinson's disease or pre-Parkinson's disease.

[0232] It has been found that the cross-correlation between CST 1002 values ​​of agonist / antagonist pairs provides a solution to long-standing technical problems in previous developments, which were limited to detecting and classifying late-stage Parkinson's disease symptoms when treatment options may narrow and costs increase. In contrast, the cross-correlation between CST 1002 values ​​of agonist / antagonist pairs is qualitatively different because it evaluates the neural origin of resting tremor in Parkinson's disease with high sensitivity, enabling the generation of pre-Parkinsonian information for resting tremor, even if it could not be observed through previous developments.

[0233] Although the evaluation of neural drive in patient 120 was shown and described in the evaluation of CST 1002, prior to being incorporated into CST 1002, the assessment of neural drive in patient 120 was not fully understood. Figure 9A The evaluation of the motor unit spike sequence 912 can provide this new qualitative capability, thereby assessing the neural drivers of Parkinson's disease and pre-Parkinson's symptoms in patients 120 with previously unknown sensitivity. The disclosed evaluation of the motor unit spike sequence 912, when incorporated into CST 1002, is for illustrative purposes only.

[0234] The synchronization amount between CST 1002 of the right FCR 808 and CST 1002 of the right ECR 806 will be Figure 11J and Figure 11K The cross-correlation is a mirror image between the two peaks, which is described by each of a plurality of peaks having smaller peaks on different sides of the larger peak.

[0235] Now for reference Figure 11L The document shows the target Figure 8A A graphical depiction of the CST autocorrelation of the right FCR 808. For the right FCR 808... Figure 10 The CST 1002 is autocorrelated with itself and exhibits multiple classical peaks of autocorrelation, with the input CST 1002 having periodic properties.

[0236] Now for reference Figure 12 The diagram shows the calculation for... Figure 1 The control flow for gait freeze information in the neuromuscular assessment system 100. Calculating gait freeze information can be understood as generating pre-Parkinson's disease information. Gait freeze information may include a gait freeze score 238. The gait freeze score 238 can be expressed in a manner similar to... Figure 2 The tremor score of 236 is calculated in this way because... Figure 2 The HD-sEMG signal 208 can be decomposed into Figure 9A The motion unit spike sequence 912.

[0237] Because this procedure applies to both gait freeze score 238 and tremor score 236, the description of gait freeze score 238 will be returned to the reference. Figure 9A , Figure 10 and Figures 11A to 11L To avoid repeating these descriptions.

[0238] The neuromuscular assessment system 100 can begin calculating the gait freeze score 238 through collection step 1202. Collection step 1202 may include... Figure 2 The HD-sEMG 206 is placed on the agonist / antagonist pair from one leg, as per [reference needed]. Figure 1 As described.

[0239] To monitor gait freeze, the gastrocnemius and tibialis anterior muscles function well. Gait freeze is an abnormal gait pattern characterized by brief and temporary pauses in foot movement despite the intention to continue walking.

[0240] Due to multiple challenges, shortcomings in previously developed technologies have led to an inability to determine gait freeze. For example, Figure 9A The MUAP920 is active in both static standing and gait freeze, the movements in both are similar, and the return to a normal gait cycle is not a sufficient indication of the absence of gait freeze.

[0241] Previous techniques have drawbacks, including the inability to detect, isolate, and evaluate neural drives at the level of motor unit spike sequences using HD-sEMG signals. The collection step 1202 may include... Figure 2 The collection of HD-sEMG signals 208, and for example in Figure 15A As further illustrated, when collection step 1202 is continuously running, the HD-sEMG signal 208 can be collected or stored. Figure 1 In the non-transitory computer-readable medium 110.

[0242] It has been found that HD-sEMG data provides richer information than conventional sEMG by adding spatial dimensions to EMG signals, thus allowing for spatial triangulation of muscle activity and activation over time. This can be calculated in step 1204 of the computational neural drive mechanism. Figure 2 210.

[0243] The neural drive signal 210 can be the motor unit spike sequence 912. That is, the neural drive calculation step 1204 can retrieve the HD-sEMG signal 208 from the HD-sEMG 206 and calculate the motor unit spike sequence 912.

[0244] The operation of computational neural drive step 1204 is about Figure 9A A more detailed description is provided, involving the decomposition of the HD-sEMG signal 208 into MUAPs 920, which are abstracted to only target specific signals. Figure 5 The peak time of MUAP 920 for each motion unit 508 is retained in the motion unit peak sequence 912. Once the motion unit peak sequence 912 has been calculated, it can be summed to calculate and determine the peak time. Figure 10 CST 1002.

[0245] Can target Figure 1 The combined sensor 104 is used to calculate CST 1002 for each of the combined sensors. Continue Figure 1 In the example of the lower leg 128, the neural drive step 1204 can calculate the CST for both the gastrocnemius and tibialis anterior muscles.

[0246] Once CST 1002 is determined in the neural drive calculation step 1204, the neuromuscular assessment system 100 can perform the CST comparison step 1206. The CST comparison step 1206 can calculate the correlation between the CST 1002 of the gastrocnemius muscle and the CST 1002 of the tibialis anterior muscle. Alternatively, the CST comparison step 1206 can calculate the correlation for the motor unit spike sequence 912 instead of CST 1002.

[0247] Once the correlation with CST 1002 is calculated, the neuromuscular assessment system 100 can execute reporting step 1208. Reporting step 1208 can generate important pre-Parkinson's disease information for CST 1002 and motor unit spike sequence 912. Illustratively, gait freeze score 238 can be determined.

[0248] like Figure 16B As described, gait freeze can be identified by the asynchrony between agonist / antagonist pairs in CST 1002 or motor unit peak sequence 912. The asynchrony can be illustratively determined by the cross-correlation between CST 1002 of the gastrocnemius muscle and CST 1002 of the tibialis anterior muscle.

[0249] If gait freezing symptoms are present in Parkinson's disease or pre-Parkinson's disease, the cross-correlation between the CST 1002 of the gastrocnemius muscle and the CST 1002 of the tibialis anterior muscle can be asynchronous. The innervation pattern of the input CST 1002 can have an asynchronous correlation when the innervation of the agonist / antagonist muscles begins or ends at different times. This is also described as a phase shift in the input CST and can be determined using the asynchronous gait threshold 1212.

[0250] In other words, the cross-correlation between the CST 1002 of the gastrocnemius muscle and the CST 1002 of the tibialis anterior muscle can consist of smaller peaks adjacent to larger peaks. This indicates the asynchrony or phase shift between the input CST 1002 of the cross-correlation.

[0251] The asynchronous gait threshold 1212 can be a measure of the time or phase shift between correlated signals. For example, when correlated signals have, for instance, a threshold of 1212. Figure 11K In the bimodal periodic wave shown, the asynchronous gait threshold 1212 can be used to measure the distance between the smaller and larger peaks. If the time distance between the peaks is greater than the asynchronous gait threshold 1212, then the relevant signal can be determined to be asynchronous.

[0252] The gait freeze score 238 can be the percentage of asynchronous correlated signals. Therefore, if three out of four correlated signals show asynchrony, the gait freeze score 238 will be 75%.

[0253] It is conceivable that the combined sensor 104 can be attached to multiple pairs of agonist / antagonist muscles on the legs, and that autocorrelation and cross-correlation can be calculated for the HD-sEMG signals 208 from two pairs of combined sensors 104, such as... Figures 11A to 11L The gait freeze score of 238 over time can be reported as... Figure 2 The Parkinson's disease progression score was 234. Alternatively, the score could be adjusted each time the gait freeze score (238) and tremor score (236) were determined. Figure 2 The average of the gait freeze score 238 and tremor score 236 was used as the Parkinson's disease progression score 234.

[0254] Another aspect of the neuromuscular assessment system 100 can be a convolutional neural network 1220. The convolutional neural network 1220 can be a continuous learning algorithm with a sequence of spikes 912 of individual motor units as input and a gait freeze score 238 as output.

[0255] It is conceivable that the convolutional neural network 1220 may include hidden layers employing a sigmoid activation function, which provides results from 0 to 1 and can be used to predict probabilities such as the total tremor score of 330. Furthermore, the hidden layers may employ a tanh activation function for the hidden states.

[0256] Furthermore, the hidden layers can employ a softmax activation function, which is envisioned to determine non-co-occurrences in the MUAP 920 of Figure 9. Even further, the mean absolute error activation function can be used to calculate the error. Recurrent cells appear in the hidden layers.

[0257] Now for reference Figure 13 The image shows a side view depicting the gait 1302 of patient 120. Patient 120 can be seen in both a normal gait 1304 and a gait frozen 1306.

[0258] It is noteworthy that patient 120 can be seen in an intermediate standing phase 1308 between normal gait 1304 and gait freeze 1306. The intermediate standing phase 1308 of normal gait 1304 is shown, in which the patient's right leg 1310 is on the ground and the left leg 1312 is partially raised.

[0259] In the normal gait 1304, the intermediate stance phase 1308 is followed by the heel-off phase 1314, in which the left leg swings forward 1312 and the heel of the right leg 1310 leaves the ground. Conversely, in the gait freeze 1306, the intermediate stance phase 1308 is followed by the freeze phase 1316.

[0260] The freezing phase 1316 can be the position where the leg stops moving forward. After the freezing phase 1316, the gait freeze 1306 will present the heel lift-off phase 1314.

[0261] In the later stages of Parkinson's disease, gait freeze can be observed in EMG readings or by visually observing the patient. However, early and pre-Parkinson's symptoms are less easily identified, and previous developments have failed to provide technical solutions for early detection. However, as... Figure 12 The computational neural drive steps 1204 and Figure 12 The technical solution for determining asynchrony described in CST comparison step 1206 enables and allows for the early generation of pre-Parkinson's disease information based on gait freeze 1306.

[0262] Now for reference Figure 14 , which shows Figure 1 A side view of the lower leg 128. The lower leg 128 is depicted as a right lower leg with a combined sensor 104 attached thereto. For the depiction of the agonist / antagonist muscle pair discussed, the skin is not shown but the underlying muscle is depicted; however, it should be understood that the combined sensor 104 is attached to [the muscle] during use. Figure 1 Patient 120 Figure 4 The skin 412 or on the skin. It is conceivable that any agonist / antagonist muscle pair in the leg or even the foot could be used by attaching the combined sensor 104 to it. Figure 1 Neuromuscular assessment system 100.

[0263] Specifically, the combined sensor 104 may include a first combined sensor 1402 and a second combined sensor 1404. Each combined sensor in the combined sensor 104 may include an HD-sEMG 206, an IMU 216, and a PPG sensor 218.

[0264] The HD-sEMG 206 may include multiple electrodes 1406, wherein each electrode 1406 has a known and fixed distance 1408 between it. The electrodes 1406 may be arranged in a grid configuration. Here, the electrodes 1406 are shown in a two-by-three grid configuration.

[0265] Electrode 1406 can be disposed within electrode carrier 1410. It is conceivable that electrode carrier 1410 can be a polyimide material with a thickness of approximately 50 μm.

[0266] It can be further envisioned that electrode 1406 could be a silver or silver chloride material that provides an impedance of approximately five thousand ohms between the skin (not shown) and electrode 1406. In the worst-case scenario, the impedance can reach one megaohm while maintaining functionality.

[0267] The electrode 1406 itself is envisioned to have a small diameter of less than three to five millimeters. Furthermore, the fixed distance 1408 between the electrodes 1406 should be less than eight to ten millimeters.

[0268] It is conceivable that when gait is targeted to be frozen, the combined sensor 104 can be placed on multiple different agonist / antagonist pairs on the lower leg 128. Here, the first combined sensor 1402 is attached to the skin above the tibialis anterior muscle 1414 (not shown), while the second combined sensor 1404 is attached to the skin above the gastrocnemius muscle 1416 (not shown).

[0269] Wireless transmission 106 can Figure 1The combined sensor output signal 108 is transmitted to the component box 102. Alternatively, it is conceivable that the component box 102 can be integrally formed with one or more of the combined sensors in the combined sensor 104, which can reduce the amount of wiring required.

[0270] HD-sEMG 206 can provide Figure 15A And also Figure 9A The depicted multiple EMG channels, each generated by one of electrodes 1406 closely spaced at a fixed distance 1408 on the skin of patient 120. The multiple EMG channels contained within the HD-sEMG signal 208 can... Figure 1 The processor 112 analyzes the signals to generate neural drive signals 210.

[0271] It has been found that using multiple electrodes 1406 arranged in a grid at a fixed distance 1408 enables [the following]: Figure 9B Each MUAP in MUAP 920 identifies the spatial dimension of the neural driving signal 210. This finding further enables the neural driving signal 210 to be evaluated before symptoms become observable by a physician, thus generating information about Parkinson's disease and pre-Parkinson's disease.

[0272] Overall reference Figure 15A and Figure 15B , Figure 1 The combined sensor output signal 108 is shown in each graphic depiction, which is used for illustration. Figure 2 Each of the HD-sEMG signal 208, IMU signal 220, and PPG signal 222 is included. Importantly, the IMU signal 220 shows no difference between static standing and gait freeze. That is, Figure 2 The accelerometer inside the IMU 216 does not distinguish Figure 15A The static standing and Figure 15B The gait is frozen as shown.

[0273] This is Figure 15A and Figure 15B The combined sensor output signals 108 depicted are distinct, showing few distinguishing factors between them. However, it should be understood that the HD-sEMG signal 208 can be decomposed into a sequence of motion unit spikes 912 and further summed to provide Figure 10 The CST 1002, where CST is the nerve drive signal 210 for muscles, and can be as follows: Figures 11A to 11L The correlation shown is used to generate pre-Parkinson's disease information based on the differences between agonist / antagonist muscle pairs, which indicates gait freezing at the neural drive level of Parkinson's disease.

[0274] Now for reference Figure 15AThe diagram shows the effect of standing still. Figure 1 The combined sensor 104 Figure 1 A graphical depiction of the combined sensor output signal 108. The HD-sEMG signal 208 may include signals for use with... Figure 14 Gastrocnemius 1416 and Figure 14 HD-sEMG signal of the electrode array of the combined sensor 104 of the tibialis anterior muscle 1414.

[0275] For clarity, the complete set of HD-sEMG signals 208 for the entire array of combined sensor 104 is not shown. Ellipses are used to depict the remaining EMG signals from the other electrodes 1406 on the gastrocnemius muscle 1416 or tibialis anterior muscle 1414. The vertical axis of the HD-sEMG signals 208 may be a measurement in millivolts, while the horizontal axis may be a measurement in time, in seconds or milliseconds.

[0276] Gait freezing may involve a mismatch between agonist / antagonist muscle pairs during the gait cycle. This is, for example, in... Figure 16B The diagram shows that the tibialis anterior muscle exhibits repetitive contraction and stretching, while the gastrocnemius muscle remains essentially stretched.

[0277] Pre-Parkinson's disease information can be generated based on agonist / antagonist muscle pair mismatch during the gait cycle, which can be detected early in evaluating neural drive. When from... Figure 14 The neural drive was evaluated by decomposing the HD-sEMG signals of multiple electrodes 1406 208 into the motor unit spike sequence 912 in Figure 9.

[0278] The evaluation of motor unit spike sequences 912 from multiple electrodes 1406 allows for spatial triangulation of muscle activity and activation over time, and provides a technical solution and advantages over previous methods that do not utilize the dissociation, cross-correlation evaluation, or electrode arrays of the disclosed neural drives to provide HD-sEMG signals 208.

[0279] In the later stages of Parkinson's disease, agonist / antagonist asymmetry can be observed in EMG readings or visually. Figure 1 Patients were observed in 120 groups. However, early and pre-Parkinson's disease information is not so easy to generate, but it can be accomplished using the method proposed in this paper.

[0280] In other words, the array of electrodes 1406 can provide spatial dimensions to the data, thereby allowing for... Figure 2 The neural drive signals 210 are identified and compared to generate pre-Parkinson's disease information, which is not available through previous developments in the field.

[0281] The combined sensor output signal 108 may further include the PPG signal 222 of the combined sensor 104. The waveform of the PPG signal 222 is generated by the attenuation changes of light energy as it passes through biological tissue, is scattered and reflected within the biological tissue, or is absorbed by the biological tissue before being detected by the optical sensor.

[0282] For clarity, only a single PPG signal 222 for the combined sensor 104 is shown, and an ellipse is used to depict another PPG signal 222 from a pair of combined sensors 104, one connected to the gastrocnemius muscle 1416 and the other connected to the tibialis anterior muscle 1414.

[0283] The vertical axis can be a measurement of volts in millivolts, while the horizontal axis can be a measurement of time in seconds or milliseconds. Figure 2 Analysis of the PPG signal 222 within signal frame 204 can provide insights into the systolic pulse wave 1502 and diastolic pulse wave 1504. Figure 2 The blood pressure reading is 224. Furthermore, the time interval between systolic pulse waves can be determined, from which the following can be derived. Figure 2 Heart rate 225 and Figure 2 Heart rate variability 226.

[0284] For example, heart rate 225 can be calculated by dividing sixty by the time in seconds between systolic pulse waves 1502. Heart rate variability 226 can be calculated from heart rate 225 or directly from systolic pulse waves 1502. Heart rate variability 226 is an early indicator of Parkinson's disease, but it is not specific enough on its own. Therefore, it is envisioned to combine heart rate variability 226 with the HD-sEMG signal 208 to improve specificity.

[0285] The combined sensor output signal 108 is further shown as an IMU signal 220 having the combined sensor 104. The IMU signal 220 may include x-plane acceleration 1506, y-plane acceleration 1508, and z-plane acceleration 1510. The x-plane acceleration 1506, y-plane acceleration 1508, and z-plane acceleration 1510 correspond to the acceleration coupled to… Figure 1 One of the combined sensors in the combined sensor 104 of the lower leg 128, such as Figure 14 What is depicted.

[0286] For clarity, only a single IMU signal 220 for the lower leg 128 is shown, and an ellipse is used to depict another IMU signal 220 from a pair of combined sensors 104, one connected to the gastrocnemius muscle 1416 and the other to the tibialis anterior muscle 1414. The vertical axis can be a measurement of physical displacement in millimeters, and the horizontal axis can be a measurement of time in milliseconds.

[0287] Now for reference Figure 15B The diagram illustrates the effects of gait freeze on... Figure 1 The combined sensor 104 Figure 1 A graphical representation of the combined sensor output signal 108. Figure 15B With Figure 15A The combined sensor output signal 108 is depicted in a similar manner. The combined sensor output signal 108 is shown as having an HD-sEMG signal 208, an IMU signal 220, and a PPG signal 222.

[0288] Before extracting neural drivers by decomposing the HD-sEMG signal 208 into the motor unit spike sequence 912 in Figure 9, it was difficult to determine Figure 13 Gait freezing 1306, and previous development lacked only from Figure 2 The IMU 216 has the technological capability to generate pre-Parkinson's disease information between gait freeze and static standing. This is in... Figure 15A and Figure 15B The combined sensor output signals 108 depicted in the figure are distinct, showing few distinguishing factors between them.

[0289] The HD-sEMG signal 208 is decomposed into motion unit spike sequences 912 and then combined and provided. Figure 10 The CST1002 allows for cross-correlation of CST1002 between agonist / antagonist muscle pairs (such as the gastrocnemius and tibialis anterior). This cross-correlation of CST1002 can provide a measure of the difference in neural drive between two muscles directly related to the progression of Parkinson's disease.

[0290] Overall reference Figure 16A and Figure 16B , Figure 16A The typical gait illustrated demonstrates generally synchronized firing of the agonist / antagonist muscles, while Figure 16B The illustrated dragging gait freeze illustrates asynchronous signal divergence. While this can be seen in the HD-sEMG signal 208, the degree of divergence must be measured at the neural drive level; this may include decomposing the HD-sEMG signal 208 into the motor unit spike sequence 912 of Figure 9, and summing the motor unit spike sequence 912. Figure 10 The CST1002 of the agonist / antagonist muscle pair and the cross-correlation of CST1002.

[0291] Now for reference Figure 16A The image shown is from a patient exhibiting a normal gait. Figure 14 A graphical depiction of the EMG signal output channel of each combined sensor in the combined sensor 104. Illustratively, it can be seen from the attached... Figure 14The combined sensor 104 of the tibialis anterior muscle 1414 provides a first EMG channel 1602, which can be connected from the tibialis anterior muscle 1414. Figure 14 The combined sensor 104 of the gastrocnemius muscle 1416 provides a second EMG channel 1604.

[0292] The first EMG channel 1602 and the second EMG channel 1604 are shown with a vertical axis that provides a normalized amplitude from negative one to positive one, while a horizontal axis provides time in a one-second partition. The first EMG channel 1602 and the second EMG channel 1604 are substantially synchronized in their firing.

[0293] Now for reference Figure 16B The image shows data from patients exhibiting gait freezing with shuffling. Figure 14 A graphical depiction of one EMG signal output channel of each of the combined sensors in the combined sensor 104. Similar to the above. Figure 16A It can be attached to Figure 14 The combined sensor 104 of the tibialis anterior muscle 1414 provides a first EMG channel 1602, which can be connected from the tibialis anterior muscle 1414. Figure 14 The combined sensor 104 of the gastrocnemius muscle 1416 provides a second EMG channel 1604.

[0294] The first EMG channel 1602 and the second EMG channel 1604 are shown with a vertical axis that provides a normalized amplitude from negative one to positive one, while the horizontal axis provides time in a one-second partition. The first EMG channel 1602 and the second EMG channel 1604 diverge significantly during gait freeze.

[0295] When gait freeze exists, the divergence between the first EMG channel 1602 and the second EMG channel 1604 will move further apart over time. This can be achieved using methods such as those described above. Figure 12 The convolutional output described by the convolutional neural network 1220 is used to track this drift. Otherwise, for Figure 10 CST 1002 for the gastrocnemius 1416 and CST 1002 for the tibialis anterior 1414 can be cross-correlated to determine the degree of deviation or asynchrony between the gastrocnemius 1416 and the tibialis anterior 1414 at the level of neural drive, which is directly related to the progression of Parkinson's disease.

[0296] Now for reference Figure 17A The figure shows a graphical representation of the power spectral density estimate of heart rate variability in healthy patients. Heart rate and heart rate variability (HRV) are not specific to Parkinson's disease, but can be caused by a number of problems, including heart defects, lung defects, and even depression.

[0297] However, HRV exists in Parkinson's disease patients and will be caused by Figure 2The HRV indication determined by the PPG sensor 218 is consistent with Figure 2 The combination of neural driving signals 210 can provide a more powerful and specific tool for indicating Parkinson's disease and its progression.

[0298] Figure 2 The PPG signal 222 can be used Figure 1 The processor 112 processes the data to provide a power spectral density estimate. Inventively, the power spectral density estimate can be a frequency domain plot, where the vertical axis represents absolute power, which is the square of milliseconds divided by the number of cycles per second, with 10,000 cycles per partition. The horizontal axis can provide a frequency of 0.05 Hz per partition.

[0299] The power spectral density estimate can be determined by processor 112 using nonparametric or parametric methods. In illustrative terms, the Fast Fourier Transform is an example of a nonparametric method, while the autoregressive model estimation is an example of a parametric method.

[0300] The power spectral density estimate for healthy patients is shown as the standard deviation of the normal-to-normal RR interval (SDNN) with an RR interval of 876 ms and a normal-to-normal RR interval of 131 ms. SDNN can represent the median of heart rate variability.

[0301] Power spectral density estimation can depict the VLF range from 0.0033 Hz to 0.04 Hz, which consists of rhythms with periods between 25 seconds and 300 seconds. The VLF range can have a power of 2827 milliseconds squared per second of period.

[0302] Power spectral density estimation can depict the LF range from 0.04 Hz to 0.15 Hz, which consists of rhythms with periods between 7 and 25 seconds. The LF range can have a power of 891 milliseconds squared per second of period.

[0303] Power spectral density estimation can depict the HF range from 0.15 Hz to 0.40 Hz. The HF range can have a power of 423 milliseconds squared per second.

[0304] Now for reference Figure 17B The figure shows a graphical depiction of the power spectral density estimation of heart rate variability in patients with Parkinson's disease. Figure 2 The PPG signal 222 can be used Figure 1 The processor 112 processes the data to provide a power spectral density estimate. Inventively, the power spectral density estimate can be a frequency domain plot, where the vertical axis represents absolute power, which is the square of milliseconds divided by the number of cycles per second, with 10,000 cycles per partition. The horizontal axis can provide a frequency of 0.05 Hz per partition.

[0305] Power spectral density estimation for Parkinson's disease patients is shown as an RRI of 822 ms and an SDNN of 93 ms. The power spectral density estimation can depict a VLF range with power of 1225 ms squared per second, an LF range with power of 362 ms squared per second, and an HF range with power of 102 ms squared per second.

[0306] Figure 17B Power ratio of each frequency range in Parkinson's disease patients Figure 17A The levels are much lower in healthy patients. For example, when Parkinson's disease readings are compared with those of healthy patients, VLF drops to less than half, LF drops to less than half, and HF drops to less than a quarter.

[0307] When detected Figure 2 When HRV 226 is higher than the HRV threshold 227, the HRV score can be correlated with the Parkinson's disease propensity score 232, tremor score 236, gait freeze score 238, and Parkinson's disease progression score 234 (all within the range of HRV 226). Figure 2 The HRV scores are added together. For example, if HRV 226 is detected as higher than the HRV threshold of 227, the HRV score could be increased by ten percentage points.

[0308] Similarly, initial orthostatic hypotension (IOH) occurring within the first thirty seconds after standing can be detected by complex and continuous monitoring of a combined sensor 104 including a PPG sensor 218 and providing a PPG signal 222. For example, when indicating Figure 2 When the systolic pulse wave 822 of the blood pressure 224 in Figure 8 drops below the IOH threshold 228, the IOH score can be added to the Parkinson's disease tendency score 232, tremor score 236, gait freeze score 238 and Parkinson's disease progression score 234.

[0309] For example, if the systolic pulse wave 822, representing a blood pressure of 224, drops below the IOH threshold of 228, the IOH score could increase by ten percentage points. These cardiovascular parameters, along with others, are associated with Parkinson's disease and can be used to modify pre-Parkinson's information, enabling even earlier and more robust diagnoses.

[0310] Now for reference Figure 18The diagram illustrates a control flow for operating a neuromuscular assessment system 100. The control flow may include: in block 1802, attaching a first electrode from a first electrode array to an agonist muscle configured to detect a first electromyography (EMG) signal; in block 1804, attaching a second electrode from a second electrode array to an antagonist muscle configured to detect a second EMG signal, the agonist and antagonist forming an agonist / antagonist pair; in block 1806, decomposing the first EMG signal into a first motor unit spike sequence; in block 1808, decomposing the second EMG signal into a second motor unit spike sequence; in block 1810, correlating the first and second motor unit spike sequences to generate a correlated signal; in block 1812, determining synchronicity and periodicity within the correlated signal; and in block 1814, generating pre-Parkinson's disease information including a tremor score, the tremor score being a score of the correlated signal determined to have both synchronicity and periodicity.

[0311] Therefore, neuromuscular assessment systems have been found to offer important and previously unknown and unavailable solutions, capabilities, and functionalities. The resulting configurations are straightforward, cost-effective, uncomplicated, highly versatile, accurate, sensitive, and efficient, and can be achieved by adapting known components for easy, efficient, and economical manufacture, application, and utilization.

[0312] While the neuromuscular assessment system has been described in conjunction with a specific optimal model, it should be understood that many alternatives, modifications, and variations will be readily apparent to those skilled in the art based on the foregoing description. Therefore, it is intended to cover all such alternatives, modifications, and variations that fall within the scope of the appended claims. All matters set forth herein or illustrated in the accompanying drawings should be interpreted in an illustrative and non-limiting sense.

Claims

1. A method for operating a neuromuscular assessment system, the method comprising: A first electrode array is attached to the aortic muscle, and the first electrode array is configured to detect a first electromyography (EMG) signal. A second electrode array is attached to the antagonist muscle, the second electrode array being configured to detect a second EMG signal, the agonist muscle and the antagonist muscle forming an agonist / antagonist muscle pair; The first EMG signal is decomposed into a first motion unit spike sequence; The second EMG signal is decomposed into a second motion unit spike sequence; The first motion unit peak sequence and the second motion unit peak sequence are correlated to generate a correlation signal; Determine the synchronicity and periodicity within the relevant signals; as well as A jitter score is generated, which is determined as a percentage of the correlated signal having both the synchronicity and the periodicity.

2. The method according to claim 1, further comprising: Determine the asynchronicity within the relevant signals; Generate a gait freeze score, which is a second percentage of the correlated signal having the said asynchronicity; and in: Attaching the first electrode array and attaching the second electrode array includes attaching the first electrode array and the second electrode array to the agonist / antagonist muscle pair within the leg.

3. The method according to claim 1, wherein the method further comprises: Low-pass filtering is applied to the peak sequences of the first and second motion units.

4. The method according to claim 1, wherein: Decomposing the first EMG signal into the first motion unit spike sequence includes decomposing the first EMG signal into multiple motion unit spike sequences; and Further includes: The multiple motion unit spike sequences are summed to form a cumulative spike sequence.

5. The method according to claim 1, wherein: Determining the periodicity includes calculating the power spectral density of the correlated signal.

6. The method according to claim 1, wherein: Determining the synchronicity includes determining that one of the related signals includes time-separated peaks, the time of which is below a synchronicity threshold.

7. The method according to claim 1, wherein the method further comprises: The attachment is configured as an optical volumetric sensor for detecting cardiovascular parameters; as well as The tremor score is increased based on the cardiovascular parameters.

8. The method according to claim 1, wherein the method further comprises: A first inertial measurement unit (IMU) and a second IMU are attached, and the first IMU and the second IMU are configured to generate IMU signals; as well as The tremor score is increased based on the bilateral asymmetry of the IMU signal.

9. A non-transitory computer-readable medium usefully associated with a processor having instructions configured to perform the following operations: The first electrode array, which connects to the agonist muscle, generates the first electromyography (EMG) signal; A second EMG signal is generated from a second electrode array attached to the antagonist muscle, wherein the agonist muscle and the antagonist muscle form an agonist / antagonist muscle pair; The first EMG signal is decomposed into a first motion unit spike sequence; The second EMG signal is decomposed into a second motion unit spike sequence; The first motion unit peak sequence and the second motion unit peak sequence are correlated to generate a correlation signal; Determine the synchronicity and periodicity within the relevant signals; as well as A jitter score is generated, which is determined as a percentage of the correlated signal having both the synchronicity and the periodicity.

10. The computer-readable medium of claim 9, wherein the computer-readable medium further comprises: Instructions configured to determine the asynchronicity within the relevant signals from the agonist / antagonist muscle pair within the leg; and Instructions configured to generate a gait freeze score, the gait freeze score being a second percentage of the associated signal having the asynchronicity.

11. The computer-readable medium of claim 9, wherein the computer-readable medium further comprises: Instructions configured to perform low-pass filtering on the first motion unit spike sequence and the second motion unit spike sequence.

12. The computer-readable medium according to claim 9, wherein: The instructions configured to decompose the first EMG signal include instructions configured to decompose the first EMG signal into a plurality of motion unit spike sequences; and Further includes: Instructions configured to sum the plurality of motion unit spike sequences into a cumulative spike sequence.

13. The computer-readable medium according to claim 9, wherein: The instructions configured to determine the periodicity include instructions configured to calculate the power spectral density of the correlated signal.

14. The computer-readable medium according to claim 9, wherein: The instruction configured to determine the synchronicity includes an instruction configured to determine that one of the related signals includes a time-separated peak value, the time being below a synchronicity threshold.

15. The computer-readable medium of claim 9, further comprising: Instructions configured to generate cardiovascular parameters based on optical volumetric sensors; and Instructions configured to increase the tremor score based on the cardiovascular parameters.

16. The computer-readable medium of claim 10, further comprising: Instructions configured to generate IMU signals from the first inertial IMU and the second IMU; and The instruction is configured to increase the tremor score based on the bilateral asymmetry of the IMU signal.

17. A neuromuscular assessment system, the neuromuscular assessment system comprising: A first electrode array, configured to attach to an agonist muscle and configured to detect a first electromyography (EMG) signal; A second electrode array, configured to attach to an antagonist muscle and configured to detect a second EMG signal, wherein the agonist muscle and the antagonist muscle form an agonist / antagonist pair; and Processor, the processor being configured to: The first EMG signal is decomposed into a first motion unit spike sequence; The second EMG signal is decomposed into a second motion unit spike sequence; The first motion unit peak sequence and the second motion unit peak sequence are correlated to generate a correlation signal; Determine the synchronicity and periodicity within the relevant signals; as well as A jitter score is generated, which is determined as a percentage of the correlated signal having both the synchronicity and the periodicity.

18. The system according to claim 17, wherein: The processor is configured to determine asynchrony within the correlated signal and generate a gait freeze score, the gait freeze score being a second percentage of the correlated signal determined to have the asynchrony; and The agonist / antagonist pair is located in the leg.

19. The system according to claim 17, wherein: The processor is configured to perform low-pass filtering on the first motion unit spike sequence and the second motion unit spike sequence.

20. The system of claim 17, wherein the processor is configured to: The first electromyographic signal is decomposed into a sequence of multiple motor unit spikes; and The multiple motion unit spike sequences are summed to form a cumulative spike sequence.

21. The system according to claim 17, wherein: The processor is configured to calculate the power spectral density of the relevant signal.

22. The system according to claim 17, wherein: The processor is configured to determine that one of the related signals includes a time-separated peak value, the time being below a synchronicity threshold.

23. The system of claim 17, further comprising: An optical plethysmography sensor configured to detect cardiovascular parameters; and in: The processor is configured to increase the tremor score based on the cardiovascular parameters.

24. The system of claim 17, further comprising: A first IMU and a second IMU, configured to generate IMU signals; and in: The processor is configured to increase the tremor score based on the bilateral asymmetry of the IMU signal.

Images