A coupled electro-mechanical approach for early diagnostic of carpal tunnel syndrome

Carpal tunnel syndrome (CTS) is a pathology affecting hand function caused by median nerve overload. Numbness in the fingers, a loss of sensory and motor function in the hand, and pain are all symptoms of carpal tunnel syndrome. The lack of numerical data about the median nerve mechanical strain inside the carpal tunnel is the main disadvantage of current clinical approaches employed in carpal syndrome diagnostics. Moreover, application of each diagnostic method alone often leads to misdiagnosis. We proposed a combined approach including hand motion capture, finite element modelling (FEM), and electromechanical simulations to evaluate median nerve compression and find a correlation with hand mobility. The hand motion capture provided the boundary conditions for FEM. After that, FEM simulations of finger flexion and hand flexion / extension were performed. Further, FEM results were put in the electrical model of nerve conduction based on the Hodgkin-Huxley model and extended cable equation. It was exhibited median nerve conduction reduced significantly throughout the flexion and extension of the hand that compared to finger flexion. During finger flexion and hand flexion and extension, the load distribution over each of nine finger flexor tendons was evaluated. The tendons of the index finger were found to have the highest Mises stress values. It was found how tendon and connective tissue contact types affected carpal tunnel pressure. The difference between the contact types was 31.7% for hand extension and 59.9% for hand flexion. The developed approach has the potential to become an alternative diagnostic method for CTS at early stages. Additionally, it can be employed as non-invasive procedure for evaluation of carpal nerve stress.

context, the studies by Guo [39] and Yao [44] have aimed at determining the load 111 on the transverse ligament. Ko and Brown [38] investigated the effect of fluid 112 within the carpal tunnel on median nerve stress using FSI calculations. Their study 113 shows that the effect of fluid is low compared to the mechanical contact of the from Main et al. study. However, the carpal tunnel has been simplified and fixed, 154 which corresponds to finger flexion, but does not accurately correspond to hand 155 flexion. In this study, this lack will be taken into account in the FEM. Nerve conduction modelling 159 The median nerve contains sensory, motor and autonomic fibers. Sensory nerve 160 fibers are known to lose their conductivity faster than motor fibers if the nerve is 161 under external mechanical stress [56][57][58][59]. The classical mathematical model 162 describing the conduction of an electrical signal along the axon of a neuron was 163 developed by Hodgkin and Huxley [60]. The model is based on differential 164 equations system that describe the membrane exchange of Na and K ions. Aim of study 176 The aim of this study is to develop a combined electro-mechanical approach 177 for early diagnosis of carpal tunnel syndrome based on modern biomechanical and 178 mathematical modelling methods as well as applying the developed approach and 179 obtaining new numerical results in the stages of the approach. 180 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 20, 2023. ; https://doi.org/10.1101/2023.06.16.23291511 doi: medRxiv preprint

182
This paper presents a combined electro-mechanical approach for early CTS 183 diagnosis, including hand motion capture, FEM of the patient-specific hand based 184 on MRI, CT, and mathematical modelling of nerve conduction (Fig 1) However, the resulting 3D models require post-processing. The post-211 processing of the tissues was performed using the Meshmixer software. The three-212 dimensional tendon geometry was built based on a set of cross-sectional planes of 213 the obtained DICOM file geometry. The resulting personal geometry of the left 214 hand and carpal tunnel after processing is shown in (Fig 2). Hand motion capture 223 The human hand is a unique manipulator that can handle many mechanical 224 tasks. Different people may perform certain mechanical tasks differently in their 225 daily lives. For example: how to hold a fork or a mug, how to type on a computer 226 keyboard, or how to hold a phone. Depending on age, anatomy, and gender, people 227 can flex their fingers and move their hands differently. To take individual hand and 228 finger movements into account, software was developed to capture hand movement 229 in real time and determine the coordinates of characteristic hand points (Fig 3) [69]. 230 Visual Studio Code and Python were used for software development. Learned free 231 artificial intelligence is used in the software to determine points on the hand using . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  (Table 1). 284 The median nerve was connected to the connective tissue by no separation CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 20, 2023. Boundary conditions 315 The different hand movements are dealt with in the fifth stage of the approach. 316 Human finger flexion is achieved by the finger flexor muscle actions and the effort 317 transferred via tendons to the median nerve that passes through the carpal tunnel. 318 The load from the tendons to the median nerve is transferred by the connective  377 where, g K and g Na are not constants but functions of the membrane potential, which 378 is due to the potential-dependent properties of the ion channels. In Hodgkin and 379 Huxley, the dependence functions of g K and g Na on membrane potential have been 380 determined on the basis of the analysis of the form of the experimental currents- . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 20, 2023.  385 where G K and G Na are maximum conductivities for potassium and sodium channels, 386 n and m -activation gate variables for potassium and sodium channels; h -387 inactivation gate variable for sodium channels. 388 The dynamics of gateway variables can be described by equations: 392 where α and β are the rate parameters for the transition of the gate particle to the 393 «open» and «close» state, respectively. These constants also depend on the 394 membrane potential and are described by equations: 395 (14) 396 The system of equations (5)-(11) is a H-H model and is solved using a self-397 developed code in Matlab using the finite difference method.

398
Cable equation CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 20, 2023. ; https://doi.org/10.1101/2023.06.16.23291511 doi: medRxiv preprint The cable theory is given as a second-order partial differential equation: 401 where R e and R i are extracellular and intracellular axial resistivity, respectively. The 402 parameters used in the nerve conduction model are listed in (Table 3). 403 Extended Cable Equation 404 The effect of changes in nerve geometry on action potential propagation is . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  (t=3s, t= 15s, t=25s, t=35s) during finger flexion are shown in (Fig. 5). The Mises stress is shown in (Table 4). Tendons were divided into two groups: deep and 450 superficial. The thumb tendon is shown in both groups. Stress distribution along the 451 tendon's midline as a function of distance is shown in (Fig 6). The distal plane of 452 the tendon is taken as the start of the distance report. contact is shown in (Fig 7, 8 at various modelling time t=1s, t=15s). Two types of . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 20, 2023. ; https://doi.org/10.1101/2023.06.16.23291511 doi: medRxiv preprint tendons and connective tissue connections were considered during wrist flexion and 467 extension. Stress along the midline of the tendon is shown in (Fig 9-12 to the median nerve being located near them and the load is transferred to the nerve. 565 The total tension of the deep tendons is 10% higher than the superficial tendons. 566 This is due to the deep tendons of the fingers are more stretched than the superficial 567 tendons [82,83]. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  (Table 5). However, the experimental technique of pressure 614 determination in the carpal tunnel allows to determine the fluid pressure within the 615 carpal tunnel. The technique used in this study allows to determine the stress of all . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 20, 2023. ; https://doi.org/10.1101/2023.06.16.23291511 doi: medRxiv preprint carpal tunnel tissues. In this case, the tendon and median nerve stress can be very 617 different. Tendon stress correlates better with experimental data than median nerve 618 stress. However, in some cases, a good correlation with median nerve tension was 619 also shown. In this work, a healthy patient without carpal tunnel syndrome was  Wrist extension 5300 -18600 3100 150 -5500 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 20, 2023. The construction of patient-specific geometry is associated with subjective is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 20, 2023.   . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 20, 2023.   CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 20, 2023. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 20, 2023. ; https://doi.org/10.1101/2023.06.16.23291511 doi: medRxiv preprint