Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Estimation of the time fluctuation of polysynaptic responses evoked by constant spinal cord stimulation

Estimation of the time fluctuation of polysynaptic responses evoked by constant spinal cord... DE GRUYTER Current Directions in Biomedical Engineering 2022;8(3): 01-04 Jose L. Vargas Luna*, Anna Pataraia, Richard Crevenna, Winfried Mayr and Milan Dimitrijevic Estimation of the time fluctuation of polysynaptic responses evoked by constant spinal cord stimulation https://doi.org/10.1515/cdbme-2022-2001 Keywords: Spinal Cord Stimulation, Polysynaptic responses, EMG, spinal cord injury, temporal centroid Abstract: Polysynaptic activity is necessary to engage the neural circuitry that controls the motor behaviour and is crucial to the recovery after spinal cord injury (SCI) [1]. Polysynaptic 1 Introduction responses can be elicited with spinal cord stimulation, and there are few reports on these types of responses in human Spinal Cord Stimulation (SCS) is getting traction as more electrophysiology, most of them describing them as constant and more reports show the potential benefits of motor recovery to fixed stimuli. However, we have observed that the responses in spinal cord injured people [2]. In a previous report, we have fluctuate in amplitude and time. Amplitude variations can be shown that SCS eliciting long-latency responses facilitates the analysed with statistical methods. However, time fluctuations initiation of rhythmic or sustained activity in muscles [3]. are more complex due to the multiple parameters involved. These long-latency responses are projections of the spinal This work describes a methodology to analyse the time cord's polysynaptic activity to the muscles, e.g. lower limb fluctuation in the responses of each pulse. It is based on the muscles, and are observed in the electromyographic (EMG) calculation of the signal temporal centroid, representing the activity as polysynaptic responses. They are composed of whole activity, weighted by its latency, in a single time value. multiple asynchronous potentials that form discharges or This value is then used to model the temporal fluctuation with groups. Regardless of their importance, polysynaptic linear regression. The methodology was verified with an responses are not well characterised in their morphology and electromyography dataset from a discomplete spinal cord spatio-temporal pattern, and the factors of influence are not injured patient with spinal cord stimulation. The parameter is sufficiently studied and understood in detail with in-vivo able to follow small changes in the responses' distribution. models yet. Examples of how the temporal centroid and linear model Electrophysiology of the human body is associated with identify the fluctuations are presented. Once fitted in a linear complex dynamic processes, and while nothing is really model, the fluctuation coefficient describes time fluctuations constant, it is common practice to presume stable responses to in the interneuron processing and, together with amplitude repeated electrical stimuli. Even though all controllable metrics, can characterise changes in polysynaptic responses variables are fixed (intensity, frequency, electrode position, during the application of fixed stimulation parameters. posture, anatomy, SCI), changes in the central state of excitability can often induce amplitude and time fluctuations. This static assumption is established in clinical practice and helps to assess electrophysiological behaviour and identify ______ features of pathophysiological mechanisms. However, it fails *Corresponding author: Jose Luis Vargas Luna: Department of to describe both monosynaptic and polysynaptic response Physical Medicine, Rehabilitation and Occupational Medicine, Medical University of Vienna, Währinger Gürtel 18-20, Vienna, fluctuations. For monosynaptic responses, at least small Austria, e-mail: jose.vargasluna@meduniwien.ac latency fluctuations and substantial amplitude changes have Anna Pataraia, Richard Crevenna, Winfried Mayr: Department been reported earlier [4]. On the other hand, there is less of Physical Medicine, Rehabilitation and Occupational Medicine, information in polysynaptic responses, which can substantially Medical University of Vienna, Vienna, Austria vary in all their components—latency, amplitude, duration and Milan Dimitrijevic: Department of Rehabilitation and Physical Medicine, Baylor College of Medicine, Houston, United States and grouping. Foundation for Movement Recovery, Oslo, Norway. Open Access. © 2022 The Author(s), published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. 1 Simple statistical methods can analyse amplitude fluctuations (e.g. mean, variance). However, time fluctuations represent a more significant challenge since the polysynaptic part of the response can be constituted by multiple grouped potentials (bursts), varying delays, durations and amplitudes, which can increase or decrease independent of each other. Therefore, just comparing the PS response variables' mean and variability will not accurately reflect the time fluctuations in Figure 1: Representation of the mass distribution along with all the dimensions. This work introduces a methodology to the spatial distribution of an object (grey circles) and the analyse the time fluctuation based on the analogous concept of temporal distribution analogy of an EMG signal with centre of mass, the temporal centroid. It provides a multiple polysynaptic responses. representation of time fluctuation on polysynaptic responses evoked in a single case exploratory analysis. In the case of polysynaptic activity, we want to find a representative point for time distribution instead of a spatial one, the temporal centroid. The mass equivalent is the amplitude of the polysynaptic responses, and the response 2 Methodology latencies would represent the distance values. The polysynaptic bursts are identified as clusters of samples above the noise threshold with at least 10ms duration. 2.1 Estimation of time fluctuation Additionally, if two contiguous supra-threshold samples are As mentioned, the polysynaptic responses have more 5ms apart, then they are considered part of different clusters variables than the mono- or oligosynaptic responses. The [3]. This is necessary to calculate individual polysynaptic number of groups, their latency, amplitude, and duration are discharges' amplitude, latency, and duration. It is also useful among the variables of polysynaptic responses. This makes to visualise how the analysis principle is analogous to the mass them complex to analyse since the polysynaptic responses of problem (Figure 1); however, processing large datasets can two contiguous pulses are often not comparable. It is crucial become time-consuming. to remember that polysynaptic responses are mostly seen at This additional processing can be avoided if the low intensities and reflect the polysynaptic activity on the background noise remains constant for the period between neural networks of the spinal cord. This differs from mono- pulses. If this assumption is valid, and mono- and and oligosynaptic responses, which have more defined reflex oligosynaptic responses are excluded from the analysis, the pathways. If we analyse polysynaptic responses under this whole recorded signal can be used for the calculation. For this perspective, a variable representing the time component of the report, only samples between 70 and 420ms post-stimulation polysynaptic activity becomes useful for characterising are considered. With these considerations, eq. 1 can be adapted temporal features in comparable form. The concept is to eq. 2 analogous to the centre of mass of an object, which describes ∑ 𝒔 𝒕 the mass distribution of an object as a standardised interaction 𝒊 𝒊 𝒊 =𝟏 𝑻 = (2) point with external forces. ∑ 𝒔 𝒊 =𝟏 The centre of mass is defined as the equilibrium point where the weighted relative position of the distributed mass of Where T is the temporal centroid, and s is the sample c i an object sums zero [5]. Figure 1 describes the concept, where occurring at time ti. This form is much more efficient for different masses m to m are distributed along the length of 1 N calculation since it is done with matrix operations. If no the system with the variable distances of x1 to xN. Under these responses are detected on the analysed period, T will tend to conditions, the centre of mass is calculated with eq 1. the middle values of the time range, 245ms in this case (the middle between 70 and 420ms). ∑ 𝒎 𝒙 𝒊 𝒊 𝒊 =𝟏 After calculating the temporal centroid of the responses 𝑿 = (1) 𝒄𝒎 for each pulse, the data are approximated using a regression ∑ 𝒎 𝒊 =𝟏 for the linear model in eq. 3 𝑻 = 𝒇 ∗ +𝒄 (3) 𝑷𝑵 Where f is a fluctuation coefficient (slope) indicating how resolution by a CODAS ADC system (DATAQ Instruments, much the temporal centroid changes with each subsequent Akron, OH, USA). A more detailed description of the pulse, PN is the pulse number, and c is a constant. methodology can be found in our previous reports [3]. For the purpose of this work, all data were processed in MATLAB (The MathWorks Inc., Natick, MA, USA). 3 Results 2.2 Validation dataset The methodology was applied to the whole dataset; the results are shown in Figure 2. The thin green line represents the temporal centroid at each pulse, while the thick line shows the 2.2.1 Subject linear regression of the data. The blue line on the left is proportional to the area under the curve of the polysynaptic The methodology was validated using the discharges. The monosynaptic responses are not clearly shown electromyographic responses of a discomplete SCI. The since the plot focuses on the polysynaptic response. subject was a 25-year-old male four years post-SCI. The injury level is C4 and has an AIS A classification. The data from this subject was retrospectively extracted from a database since it exemplifies different fluctuation patterns. The underlying study, from which the data was extracted, was approved by the local ethics committee, conducted in compliance with the Helsinki Declaration, and all participants gave their informed consent before the procedures. 2.2.2 Assessment methodology The spinal cord stimulation was applied with an implanted quadripolar epidural electrode (3487A, Medtronic, Minneapolis, USA). The electrode was placed in the posterior epidural space, centred over the vertebral levels T11-L1. The final position was such that the contact at the tip of the electrode evoked the largest responses in the quadriceps muscle groups. The stimulation was applied at 2.1pps (minimum repetition rate of the stimulator). The stimuli consisted of asymmetric biphasic pulses with a stimulating phase of 210µs, and an amplitude range from 1V to 10V or maximum comfortable intensity. The setup allowed the allocation of the cathode (+) and anode (-) to any of the four contacts or the stimulation case, providing multiple electrode configurations. Figure 2. Sequences of responses (first one on top) on the For example, 2-1+, means cathode in contact 2 and anode in triceps surae evoked by spinal cord stimulation at 2pps contact 1. using A) 2V and 0-C+; B) 4V and 2-1+; C) 4V with 0-2+; The responses evoked by the SCS were monitored on the and D) 5V and 3-C+. quadriceps, hamstrings, tibialis anterior and triceps surae Figure 2A shows that on the first pulses, the value is near bilaterally via surface electromyography (sEMG). In addition, 245ms, which is the expected behaviour when no polysynaptic an additional channel was placed on the lower trunk muscles responses appear. It can also be seen that after small activity to monitor the stimulus artefact, indicating the stimulation appears around 80ms, it rapidly steers the temporal centroid to onset. The sEMG was amplified via a Grass 12A5 system the left, showing that the value is quite sensitive. Furthermore, (Grass, Quincy, MA, USA), adjusted to a gain of 5000 and a the linear regression correctly follows the signal as stronger bandwidth of 50–800 Hz, and digitised at 2 kHz with a 12-bit activity appears at the end of the pulse, having a final value of 3 f = 1.43. The overall fluctuation factor (slope) can be used for time fluctuations, making it easier to characterise these types discriminating the presence of a fluctuation if adequately of responses and hopefully, helping to understand the adjusted to each setup. For this specific setup, deviations larger mechanisms that govern them. than 1ms per pulse clearly indicate fluctuation, although half that value could also work. Author Statement In the case that polysynaptic discharges are stable, the Research funding: The author state no funding is involved. temporal centroid also performed as expected (figure 2B), Conflict of interest: Authors state no conflict of interest. having a minimal fluctuation coefficient (f = 0.027) and only Informed consent: Informed consent has been obtained from negligible variability. the individual included in this study. Ethical approval: The Figure 2D shows the case when fluctuation is not visually research related to human use complies with all the relevant apparent. Nevertheless, this methodology can identify a national regulations, and institutional policies and was fluctuation caused by the decrease in amplitude of the performed in accordance with the tenets of the Helsinki discharges in the middle and end of the signal, having a strong Declaration, and has been approved by the local ethics f = -3.06. It is crucial to notice that, unlike Figure 2A, in Figure committee. The participant gave their informed consent before 2D the value is negative, indicating the direction of the the procedures. fluctuation. Finally, Figure 2C shows a case where there is a References fluctuation in amplitude (blue line on the left), which the temporal centroid cannot detect (f = 0.107). This shows that [1] Flynn JR, Graham BA, Galea MP, et al. The role of both metrics must be combined to characterise the fluctuation propriospinal interneurons in recovery from spinal cord of polysynaptic response. injury. Neuropharmacology [Internet]. 2011;60:809–822. Available from: http://dx.doi.org/10.1016/j.neuropharm.2011.01.016. [2] Darrow D, Balser D, Netoff TI, et al. Epidural Spinal Cord 4 Conclusion Stimulation Facilitates Immediate Restoration of Dormant Motor and Autonomic Supraspinal Pathways after Chronic Here, we present a methodology to estimate the time Neurologically Complete Spinal Cord Injury. J fluctuation of polysynaptic responses evoked by fixed spinal Neurotrauma. 2019;36:2325–2336. cord stimulation. It was designed to identify signals that [3] Vargas Luna JL, Brown J, Krenn MJ, et al. fluctuate in time by increasing the magnitude of the f value, Neurophysiology of epidurally evoked spinal cord reflexes and giving a smaller magnitude to signals that are stable over in clinically motor-complete posttraumatic spinal cord time or have a rhythmic activity (if the entire cycle is injury. Exp Brain Res [Internet]. 2021; Available from: included). https://link.springer.com/10.1007/s00221-021-06153-1. It is based on an easily calculated temporal centroid and [4] Minassian K, Jilge B, Rattay F, et al. Stepping-like movements in humans with complete spinal cord injury indicates the action potentials distribution estimated via the induced by epidural stimulation of the lumbar cord: EMG area in the studied muscle. This tool has shown to have electromyographic study of compound muscle action enough sensibility to detect changes in latency and the potentials. Spinal Cord [Internet]. 2004 [cited 2014 Dec distribution of the responses along with the whole pulse. 22];42:401–416. Available from: Additionally, the sign indicates the fluctuation direction. http://dx.doi.org/10.1038/sj.sc.3101615. Future developments would include the analysis of the [5] Wikipedia. Center of Mass [Internet]. [cited 2022 Mar 24]. model error, that can could indicate rhythmical activity, which Available from: the current model cannot identify if the cycles are complete. https://en.wikipedia.org/wiki/Center_of_mass. This tool, together with an amplitude-related variable like standard deviation or variance, can identify amplitude and http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Current Directions in Biomedical Engineering de Gruyter

Estimation of the time fluctuation of polysynaptic responses evoked by constant spinal cord stimulation

Download PDF
4 pages

Loading next page...
 
/lp/de-gruyter/estimation-of-the-time-fluctuation-of-polysynaptic-responses-evoked-by-x9C1JUM0bx
Publisher
de Gruyter
Copyright
© 2022 The Author(s), published by De Gruyter
eISSN
2364-5504
DOI
10.1515/cdbme-2022-2001
Publisher site
See Article on Publisher Site

Abstract

DE GRUYTER Current Directions in Biomedical Engineering 2022;8(3): 01-04 Jose L. Vargas Luna*, Anna Pataraia, Richard Crevenna, Winfried Mayr and Milan Dimitrijevic Estimation of the time fluctuation of polysynaptic responses evoked by constant spinal cord stimulation https://doi.org/10.1515/cdbme-2022-2001 Keywords: Spinal Cord Stimulation, Polysynaptic responses, EMG, spinal cord injury, temporal centroid Abstract: Polysynaptic activity is necessary to engage the neural circuitry that controls the motor behaviour and is crucial to the recovery after spinal cord injury (SCI) [1]. Polysynaptic 1 Introduction responses can be elicited with spinal cord stimulation, and there are few reports on these types of responses in human Spinal Cord Stimulation (SCS) is getting traction as more electrophysiology, most of them describing them as constant and more reports show the potential benefits of motor recovery to fixed stimuli. However, we have observed that the responses in spinal cord injured people [2]. In a previous report, we have fluctuate in amplitude and time. Amplitude variations can be shown that SCS eliciting long-latency responses facilitates the analysed with statistical methods. However, time fluctuations initiation of rhythmic or sustained activity in muscles [3]. are more complex due to the multiple parameters involved. These long-latency responses are projections of the spinal This work describes a methodology to analyse the time cord's polysynaptic activity to the muscles, e.g. lower limb fluctuation in the responses of each pulse. It is based on the muscles, and are observed in the electromyographic (EMG) calculation of the signal temporal centroid, representing the activity as polysynaptic responses. They are composed of whole activity, weighted by its latency, in a single time value. multiple asynchronous potentials that form discharges or This value is then used to model the temporal fluctuation with groups. Regardless of their importance, polysynaptic linear regression. The methodology was verified with an responses are not well characterised in their morphology and electromyography dataset from a discomplete spinal cord spatio-temporal pattern, and the factors of influence are not injured patient with spinal cord stimulation. The parameter is sufficiently studied and understood in detail with in-vivo able to follow small changes in the responses' distribution. models yet. Examples of how the temporal centroid and linear model Electrophysiology of the human body is associated with identify the fluctuations are presented. Once fitted in a linear complex dynamic processes, and while nothing is really model, the fluctuation coefficient describes time fluctuations constant, it is common practice to presume stable responses to in the interneuron processing and, together with amplitude repeated electrical stimuli. Even though all controllable metrics, can characterise changes in polysynaptic responses variables are fixed (intensity, frequency, electrode position, during the application of fixed stimulation parameters. posture, anatomy, SCI), changes in the central state of excitability can often induce amplitude and time fluctuations. This static assumption is established in clinical practice and helps to assess electrophysiological behaviour and identify ______ features of pathophysiological mechanisms. However, it fails *Corresponding author: Jose Luis Vargas Luna: Department of to describe both monosynaptic and polysynaptic response Physical Medicine, Rehabilitation and Occupational Medicine, Medical University of Vienna, Währinger Gürtel 18-20, Vienna, fluctuations. For monosynaptic responses, at least small Austria, e-mail: jose.vargasluna@meduniwien.ac latency fluctuations and substantial amplitude changes have Anna Pataraia, Richard Crevenna, Winfried Mayr: Department been reported earlier [4]. On the other hand, there is less of Physical Medicine, Rehabilitation and Occupational Medicine, information in polysynaptic responses, which can substantially Medical University of Vienna, Vienna, Austria vary in all their components—latency, amplitude, duration and Milan Dimitrijevic: Department of Rehabilitation and Physical Medicine, Baylor College of Medicine, Houston, United States and grouping. Foundation for Movement Recovery, Oslo, Norway. Open Access. © 2022 The Author(s), published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License. 1 Simple statistical methods can analyse amplitude fluctuations (e.g. mean, variance). However, time fluctuations represent a more significant challenge since the polysynaptic part of the response can be constituted by multiple grouped potentials (bursts), varying delays, durations and amplitudes, which can increase or decrease independent of each other. Therefore, just comparing the PS response variables' mean and variability will not accurately reflect the time fluctuations in Figure 1: Representation of the mass distribution along with all the dimensions. This work introduces a methodology to the spatial distribution of an object (grey circles) and the analyse the time fluctuation based on the analogous concept of temporal distribution analogy of an EMG signal with centre of mass, the temporal centroid. It provides a multiple polysynaptic responses. representation of time fluctuation on polysynaptic responses evoked in a single case exploratory analysis. In the case of polysynaptic activity, we want to find a representative point for time distribution instead of a spatial one, the temporal centroid. The mass equivalent is the amplitude of the polysynaptic responses, and the response 2 Methodology latencies would represent the distance values. The polysynaptic bursts are identified as clusters of samples above the noise threshold with at least 10ms duration. 2.1 Estimation of time fluctuation Additionally, if two contiguous supra-threshold samples are As mentioned, the polysynaptic responses have more 5ms apart, then they are considered part of different clusters variables than the mono- or oligosynaptic responses. The [3]. This is necessary to calculate individual polysynaptic number of groups, their latency, amplitude, and duration are discharges' amplitude, latency, and duration. It is also useful among the variables of polysynaptic responses. This makes to visualise how the analysis principle is analogous to the mass them complex to analyse since the polysynaptic responses of problem (Figure 1); however, processing large datasets can two contiguous pulses are often not comparable. It is crucial become time-consuming. to remember that polysynaptic responses are mostly seen at This additional processing can be avoided if the low intensities and reflect the polysynaptic activity on the background noise remains constant for the period between neural networks of the spinal cord. This differs from mono- pulses. If this assumption is valid, and mono- and and oligosynaptic responses, which have more defined reflex oligosynaptic responses are excluded from the analysis, the pathways. If we analyse polysynaptic responses under this whole recorded signal can be used for the calculation. For this perspective, a variable representing the time component of the report, only samples between 70 and 420ms post-stimulation polysynaptic activity becomes useful for characterising are considered. With these considerations, eq. 1 can be adapted temporal features in comparable form. The concept is to eq. 2 analogous to the centre of mass of an object, which describes ∑ 𝒔 𝒕 the mass distribution of an object as a standardised interaction 𝒊 𝒊 𝒊 =𝟏 𝑻 = (2) point with external forces. ∑ 𝒔 𝒊 =𝟏 The centre of mass is defined as the equilibrium point where the weighted relative position of the distributed mass of Where T is the temporal centroid, and s is the sample c i an object sums zero [5]. Figure 1 describes the concept, where occurring at time ti. This form is much more efficient for different masses m to m are distributed along the length of 1 N calculation since it is done with matrix operations. If no the system with the variable distances of x1 to xN. Under these responses are detected on the analysed period, T will tend to conditions, the centre of mass is calculated with eq 1. the middle values of the time range, 245ms in this case (the middle between 70 and 420ms). ∑ 𝒎 𝒙 𝒊 𝒊 𝒊 =𝟏 After calculating the temporal centroid of the responses 𝑿 = (1) 𝒄𝒎 for each pulse, the data are approximated using a regression ∑ 𝒎 𝒊 =𝟏 for the linear model in eq. 3 𝑻 = 𝒇 ∗ +𝒄 (3) 𝑷𝑵 Where f is a fluctuation coefficient (slope) indicating how resolution by a CODAS ADC system (DATAQ Instruments, much the temporal centroid changes with each subsequent Akron, OH, USA). A more detailed description of the pulse, PN is the pulse number, and c is a constant. methodology can be found in our previous reports [3]. For the purpose of this work, all data were processed in MATLAB (The MathWorks Inc., Natick, MA, USA). 3 Results 2.2 Validation dataset The methodology was applied to the whole dataset; the results are shown in Figure 2. The thin green line represents the temporal centroid at each pulse, while the thick line shows the 2.2.1 Subject linear regression of the data. The blue line on the left is proportional to the area under the curve of the polysynaptic The methodology was validated using the discharges. The monosynaptic responses are not clearly shown electromyographic responses of a discomplete SCI. The since the plot focuses on the polysynaptic response. subject was a 25-year-old male four years post-SCI. The injury level is C4 and has an AIS A classification. The data from this subject was retrospectively extracted from a database since it exemplifies different fluctuation patterns. The underlying study, from which the data was extracted, was approved by the local ethics committee, conducted in compliance with the Helsinki Declaration, and all participants gave their informed consent before the procedures. 2.2.2 Assessment methodology The spinal cord stimulation was applied with an implanted quadripolar epidural electrode (3487A, Medtronic, Minneapolis, USA). The electrode was placed in the posterior epidural space, centred over the vertebral levels T11-L1. The final position was such that the contact at the tip of the electrode evoked the largest responses in the quadriceps muscle groups. The stimulation was applied at 2.1pps (minimum repetition rate of the stimulator). The stimuli consisted of asymmetric biphasic pulses with a stimulating phase of 210µs, and an amplitude range from 1V to 10V or maximum comfortable intensity. The setup allowed the allocation of the cathode (+) and anode (-) to any of the four contacts or the stimulation case, providing multiple electrode configurations. Figure 2. Sequences of responses (first one on top) on the For example, 2-1+, means cathode in contact 2 and anode in triceps surae evoked by spinal cord stimulation at 2pps contact 1. using A) 2V and 0-C+; B) 4V and 2-1+; C) 4V with 0-2+; The responses evoked by the SCS were monitored on the and D) 5V and 3-C+. quadriceps, hamstrings, tibialis anterior and triceps surae Figure 2A shows that on the first pulses, the value is near bilaterally via surface electromyography (sEMG). In addition, 245ms, which is the expected behaviour when no polysynaptic an additional channel was placed on the lower trunk muscles responses appear. It can also be seen that after small activity to monitor the stimulus artefact, indicating the stimulation appears around 80ms, it rapidly steers the temporal centroid to onset. The sEMG was amplified via a Grass 12A5 system the left, showing that the value is quite sensitive. Furthermore, (Grass, Quincy, MA, USA), adjusted to a gain of 5000 and a the linear regression correctly follows the signal as stronger bandwidth of 50–800 Hz, and digitised at 2 kHz with a 12-bit activity appears at the end of the pulse, having a final value of 3 f = 1.43. The overall fluctuation factor (slope) can be used for time fluctuations, making it easier to characterise these types discriminating the presence of a fluctuation if adequately of responses and hopefully, helping to understand the adjusted to each setup. For this specific setup, deviations larger mechanisms that govern them. than 1ms per pulse clearly indicate fluctuation, although half that value could also work. Author Statement In the case that polysynaptic discharges are stable, the Research funding: The author state no funding is involved. temporal centroid also performed as expected (figure 2B), Conflict of interest: Authors state no conflict of interest. having a minimal fluctuation coefficient (f = 0.027) and only Informed consent: Informed consent has been obtained from negligible variability. the individual included in this study. Ethical approval: The Figure 2D shows the case when fluctuation is not visually research related to human use complies with all the relevant apparent. Nevertheless, this methodology can identify a national regulations, and institutional policies and was fluctuation caused by the decrease in amplitude of the performed in accordance with the tenets of the Helsinki discharges in the middle and end of the signal, having a strong Declaration, and has been approved by the local ethics f = -3.06. It is crucial to notice that, unlike Figure 2A, in Figure committee. The participant gave their informed consent before 2D the value is negative, indicating the direction of the the procedures. fluctuation. Finally, Figure 2C shows a case where there is a References fluctuation in amplitude (blue line on the left), which the temporal centroid cannot detect (f = 0.107). This shows that [1] Flynn JR, Graham BA, Galea MP, et al. The role of both metrics must be combined to characterise the fluctuation propriospinal interneurons in recovery from spinal cord of polysynaptic response. injury. Neuropharmacology [Internet]. 2011;60:809–822. Available from: http://dx.doi.org/10.1016/j.neuropharm.2011.01.016. [2] Darrow D, Balser D, Netoff TI, et al. Epidural Spinal Cord 4 Conclusion Stimulation Facilitates Immediate Restoration of Dormant Motor and Autonomic Supraspinal Pathways after Chronic Here, we present a methodology to estimate the time Neurologically Complete Spinal Cord Injury. J fluctuation of polysynaptic responses evoked by fixed spinal Neurotrauma. 2019;36:2325–2336. cord stimulation. It was designed to identify signals that [3] Vargas Luna JL, Brown J, Krenn MJ, et al. fluctuate in time by increasing the magnitude of the f value, Neurophysiology of epidurally evoked spinal cord reflexes and giving a smaller magnitude to signals that are stable over in clinically motor-complete posttraumatic spinal cord time or have a rhythmic activity (if the entire cycle is injury. Exp Brain Res [Internet]. 2021; Available from: included). https://link.springer.com/10.1007/s00221-021-06153-1. It is based on an easily calculated temporal centroid and [4] Minassian K, Jilge B, Rattay F, et al. Stepping-like movements in humans with complete spinal cord injury indicates the action potentials distribution estimated via the induced by epidural stimulation of the lumbar cord: EMG area in the studied muscle. This tool has shown to have electromyographic study of compound muscle action enough sensibility to detect changes in latency and the potentials. Spinal Cord [Internet]. 2004 [cited 2014 Dec distribution of the responses along with the whole pulse. 22];42:401–416. Available from: Additionally, the sign indicates the fluctuation direction. http://dx.doi.org/10.1038/sj.sc.3101615. Future developments would include the analysis of the [5] Wikipedia. Center of Mass [Internet]. [cited 2022 Mar 24]. model error, that can could indicate rhythmical activity, which Available from: the current model cannot identify if the cycles are complete. https://en.wikipedia.org/wiki/Center_of_mass. This tool, together with an amplitude-related variable like standard deviation or variance, can identify amplitude and

Journal

Current Directions in Biomedical Engineering – de Gruyter

Published: Sep 1, 2022

Keywords: Spinal Cord Stimulation; Polysynaptic responses; EMG; spinal cord injury; temporal centroid

There are no references for this article.