Abstract
This research aims to tackle the limitations faced in surgical education nowadays, particularly in the complex field of spinal cord tumor removal surgery. An innovative flexible piezoresistive sensor designed to mimic a motor nerve was developed and integrated into a bionic spine surgery simulation system, allowing for the intraoperative nerve monitoring possible during simulated tumor removal surgeries. The motor nerve, fabricated using a combination of carbon nanotubes and silicone rubber, exhibited a strong correlation between applied force and resultant changes in resistance, as confirmed by experimental results. This creative system can play an important role in providing valuable feedback for training doctors, facilitating the assessment of surgical precision and success, and enabling doctors to take necessary precautions to minimize the risk of nerve damage in real surgical scenarios. Ultimately, this proposed system has the potential to elevate the standard of surgical education, foster skill development among doctors, and significantly contribute to enhanced patient care and recovery.
I. INTRODUCTION
Spinal cord tumors are anomalous formations that develop within the spinal canal, exerting pressure on the spinal cord. Treatment options for these tumors may include surgery, radiation therapy, or careful observation depending on the tumor type and the severity of symptoms before treatment. The prognosis for recovery varies from patient to patient, considering factors such as tumor type and symptom severity. Neurosurgeons must have extensive knowledge and excellent surgical skills to implement complex surgical procedures and treat spinal cord tumors effectively. Thus, simulators should be routinely used for training and skill assessment of neurosurgeon doctors. Cadavers can be used to help students familiarize with the structure of the spinal cord under the supervision of instructors. Unfortunately, cadavers are expensive and limited to obtain, require special preparation and storage spaces, lack pathology, and pose a potential biohazard risk to trainees.1,2 In addition, hospital administrators must ensure patient safety by restricting access to all but the most highly trained personnel so that many junior neurosurgeons have limited practical training opportunities.
Virtual reality (VR) devices are increasingly used to create computer-based models (e.g., visual and audio) as teaching aids for surgery students.3–5 In particular, developing VR simulators for spine-related training practices can accelerate and allow students to improve their abilities and reduce the risk of patient discomfort. Even as virtual reality simulators continue to advance, they are still expensive, lack the realism of natural or synthetic surgical training tools, and are hard to use for limited and complex surgical anatomy.6,7 Sharples et al. also pointed out through the research results that some operators will have symptoms of motion sickness during the operation with VR and those with severe symptoms cannot use the virtual reality screen for 30 min.8 It can be seen that students who are prone to motion sickness cannot gain operational experience from virtual reality surgical training. Additionally, students encounter challenges in adopting this technique due to its high cost and complex instructional process. Furthermore, the lack of exposure to the authentic environment of spinal cord tumor surgery hampers their learning experience. Consequently, the incorporation of physical medical simulators has become imperative to compensate for the limited practical opportunities available to students.
In the field of neurosurgery, spinal procedures are intricate and involve complex surgical interventions. The epidermis, subcutaneous tissue, muscles, vertebrae, ligaments, and other structures are incised and dissected to access deep nerves and intervertebral disk cartilage. Executing precise and accurate surgical techniques within a specific scope poses a significant challenge, testing the patience and surgical proficiency of both experienced doctors and aspiring students. It is classified as a critical training operation due to its complexity. Given the invasive nature of spinal surgery, even a minor lapse in attention can result in spinal nerve damage or, in severe cases, paralysis, adversely affecting the patient's physiological functions. To mitigate this risk and acquire essential knowledge, doctors must possess clinical and surgical experience. Notably, during the removal of spinal cord tumors, distinguishing between the motor nerve and the spinal cord is challenging, as they lack differentiation in color. Any slight damage to the motor nerve can lead to irreversible harm to the patient's recovery process.
Therefore, surgeons have utilized intraoperative neurophysiological monitoring (IONM) in spinal surgery to reduce the incidence of postoperative neurological complications. Among the relative techniques of IONM, the most reliable are somatosensory evoked potentials (SEP),9,10 motor evoked potentials (MEP),11–13 and electromyography (EMG).14,15 Applying IONM in surgery has been proven effective over a long period, but it is an expensive and complex equipment unsuitable for student surgical training, leading to tremendous obstacles for surgical education training. According to the experience shared by senior doctors, the practice models used in spinal surgery training held by hospitals are all based on humanoid physical simulation models. The nerve parts in the simulator do not have the same characteristics as the natural nerves, so it is easy for students to injure the nerves without knowing it when doing exercises, and this makes the surgical training less effective than expected.
At the same time, the research and development of flexible electronic devices have gained considerable interest. Conductive polymer composites (CPCs) are employed as piezoresistive sensors, herein, due to their excellent electrical conductivity and softness, and they can transmit signals because of resistance changes in terms of pressure,16 tension,17 and temperature.18 Conductive polymer composites are typically constituted of carbon nanotubes (CNTs),19 carbon nanocapsules,20 or graphene nanoplates (GNPs),21 and polymer coating techniques22–25 are applied in this study to create a conductive motor nerve. Chen et al.22 utilized spraying technology to create a sensor by placing a CNT coating between two layers of polymer. The researchers investigated the conductivity of various CNT concentrations and observed a positive correlation between conductivity and CNT concentration. Additionally, they examined the impact of different CNT concentrations on the sensor's surface light transmittance and sensitivity, finding an opposite relationship between these factors. Kong et al.24 used polydimethylsiloxane (PDMS) and AgNW to create a sensor with a spiral structure, which was used to detect the human body motion of small force changes such as joint bending or extension and used a tensile test to detect the mechanical properties of the sensor and electrical properties. Dong and Xie25 mixed CNTs with polymers to form a conductive layer sandwiched between polymer layers, stretching the conductive layers for different structures. When the sensor was stretched, the CNTs were longitudinally squeezed to increase density and resistance. To improve the sensitivity and stability of the sensor, Guo et al.26 added silicone fluid (SF) to the original multi-walled carbon nanotube (MWCNT)-PDMS to improve the dispersion of MWCNTs to increase the tunnel resistance of the surface. However, the relationship between compressive strain and electrical properties of flexible composite sensors is lacking in these studies. In addition, in order to simulate the force exerted by doctors' surgical instruments on the flexible sensor, Jushiddi et al.27 studied the contact stress of a needle inserted into a gel under constant speed, constant needle diameter, needle tip shape, and gel elasticity, and changing bevel angle. Wang et al.28 used needle characteristics (such as diameter and tip type), insertion speed, insertion angle, and plug-and-contact pressure and showed that changing the needle bevel angle from 27° to 18° resulted in peak force (i.e., puncture force) decreases while needle deflection increases.
The main objective of this study was to develop a bionic spine equipped with a piezoresistive sensor to facilitate the training of neurosurgery doctors in the precise removal of spinal cord tumors. The training process involves several crucial steps, such as bone grinding, meningeal incision, spinal cord removal, and tumor extraction. To achieve this, an additive manufacturing technique and a novel fabrication process for creating the piezoresistive sensor were used, and the complete system includes the bionic spine, spinal cord tumor module, bionic motor nerve, and an intraoperative monitoring system. In the construction of the motor nerve, a blend of carbon nanotubes and silicone rubber were used as the primary materials. This bionic motor nerve enables the generation of a warning signal whenever a doctor unintentionally makes contact with the motor nerve using forceps, imitating a realistic scenario during the surgical procedure.
II. METHODOLOGY AND FABRICATION PROCESS
A. Materials
Long multi-wall carbon nanotube (LMWCNT) was purchased from CONJUTEK (New Taipei City, Taiwan). Isopropyl alcohol (IPA) was purchased from Echo Chemical Co., Ltd. (Miaoli, Taiwan), PMMA was purchased from Hoya Ltd. (Taipei, Taiwan). Filter paper was purchased from Hong I Instruments Co., Ltd. (Taipei, Taiwan). Silicone rubber was purchased from Emperor Chemical Co., Ltd. (Taipei, Taiwan), and electronic parts were purchased from Jin Hua Electronic Co., Ltd. (Taipei, Taiwan).
B. Manufacturing of bionic spine
A bionic spine serves as protection for motor nerves and is crucial in the initial stages of spine surgery for surgeons to navigate. In this study, a highly accurate bionic spine simulator was developed. The process of fabricating the bionic spine using fused deposition modeling (FDM) 3D printing is depicted in Figs. 1(a)–1(d). Beginning with Fig. 1(a), a thoracic spine image was consulted with Dr. Lin at Shuanghe Hospital in Taipei, Taiwan, and translated into a detailed computer-aided design (CAD) model. Subsequently, Meshmixer, a modeling software by Autodesk, was employed to refine the surface, enhancing its smoothness, as illustrated in Fig. 1(b). To simulate a realistic spine, only two vertebrae were manufactured, considering the average length of the human spine to be about 65–70 cm. Pins and hooks were meticulously incorporated into each half of the vertebral model to facilitate rapid assembly. Utilizing a fused deposition 3D printer, specifically the Fortus 360mc model by Stratasys based in Eden Prairie, MN, USA, the thoracic spine was printed using polylactic acid (PLA) filament, as shown in Fig. 1(c), with the final printed product showcased in Fig. 1(d).
FIG. 1.
(a) Digital Imaging and Communications in Medicine (DICOM) of the patient's spine; (b) two vertebrae extracted from patient's spine; (c) FDM 3D printer; (d) printed vertebral made of the PLA material.
C. Manufacturing of bionic motor nerve
In order to create bionic motor nerves, multi-walled carbon nanotubes (MWCNTs) are used, with a diameter of 10–20 nm, a length of 10–30 nm, and a surface area ratio of 200 m2/g. The manufacturing process of bionic motor nerves is shown in Fig. 2. First, shown in Fig. 2(a), a certain amount of carbon nanotubes was dispersed in the isopropyl alcohol (IPA) solution to prepare a nanocarbon powder dispersion with a content of 0.3 wt. %. Second, a spray device was used to evenly distribute the dispersion on the surface of the 65 × 80 mm filter paper. Maintain a spraying distance of 8 cm and continue spraying until the entire process is completed, shown in Fig. 2(b). After that, the carbon nanotube coating must dry naturally and the IPA must evaporate completely before the coating can conduct electricity, shown in Fig. 2(c). At the same time, 25 g of silicone rubber was evenly dispersed on the PMMA substrate using spin coating, shown in Fig. 2(d). After waiting for 10 min at 40 °C, the silicone rubber will partially solidify and reach a semi-solid state. Subsequently, the dried elongated carbon nanotube layer was fixed on the silicone rubber [Fig. 2(e)]. Carefully remove the filter paper from one side of the carbon nanotube, roll up the silicone rubber, and finally complete the structure of the flexible piezoresistive sensor [Fig. 2(f)].
FIG. 2.
(a) Carbon nanotubes in IPA solution, (b) application of spray coating, (c) natural drying of the sprayed coating, (d) spin coating of silicone rubber, (e) transfer printing of the CNT layer onto the surface of silicone rubber, (f) final sensor.
D. Fabrication of spinal cord and tumor
The fabrication process described in Ref. 28 was employed to create a spinal cord model with an embedded tumor. Initially, 3D printing was utilized to produce the top and bottom molds required for casting the spinal cord. Subsequently, the completed bionic motor nerve and tumor were placed inside the mold, and the upper and lower molds were closed. A mixture comprising glycerin, jelly wax, and thermoplastic rubber at a specific ratio was then poured into the mold at room temperature, as shown in Fig. 3(a). Once the mixture solidified, the upper and lower molds were opened, and the spinal cord model was removed, as shown in Fig. 3(b). Finally, the spinal cord model was coated with bionic meninges made of silicone rubber, as shown in Fig. 3(c). Figure 3(d) presents the complete spinal cord tumor model, which is subsequently inserted into the spine bone model crafted from 3D printing, shown in Fig. 3(e). Figure 3(f) illustrates a commercially available spine medical simulator as comparison.
FIG. 3.
(a) The mold was filled with the bionic motor nerve and tumor, and the jelly wax mixture was cast, (b) after the mixture solidified, the spinal cord model was removed from the mold, (c) the spinal cord model was then covered with bionic meninges, (d) the complete spinal cord tumor model, (e) the spinal cord inserted into the spine model, (f) commercially available spine medical simulator as a comparison.
E. Characterization and measurement
The measurements of the sensor are categorized into three main areas: spray uniformity, mechanical properties, and electrical properties. The uniformity of spraying is represented by measuring the transmittance of laser light with a wavelength of 532 nm through the filter paper and the conductive coating layer. The spot size is 15 mm, measured by a power meter detector, which receives laser light and converts optical energy into power. Mechanical properties and electrical properties were measured using JISC Force Gauge HF-10 and KeithLink 2410 Source Meter to investigate electrical property change caused by compression force deforming the piezoresistive sensor.
The characteristic experiments are categorized into compression and needle insertion experiments, aiming to evaluate both the mechanical and electrical properties of composite materials compared to pure materials. In the compression experiments, standard 3 mm pins are utilized, while the needle insertion experiments employ standard 22G, 23G, and 24G needles with diameters of 0.7, 0.65, and 0.55 mm, respectively. Both experiments are conducted at a test speed of 70 mm/min, and the force and displacement are recorded, and each test is conducted in five cycles.
F. Building intraoperative monitoring system
The nerve function monitoring system utilizes the voltage divider rule and Arduino series electronic components to establish a connection with bionic motor nerves. The system then transmits the signals to a monitoring window via Bluetooth. The flow chart diagram consists of three main parts: an input signal, signal processing, and an output signal, as shown in Fig. 4(a). To enhance the signal quality, a moving average filter, which acts as a low-pass filter, is incorporated into the signal processing program. This filter helps to minimize noise interference. Moreover, considering the portability of the monitoring system device, Bluetooth is employed to transmit the signals received by the piezoresistive sensor to a tablet. On the tablet, there is a Bluetooth signal display that can provide a visual representation in real time. The monitoring system is shown in Fig. 4(b).
FIG. 4.
(a) Flow chart diagram of the monitoring system; black indicates the physical device and blue indicates the programming; (b) real photo of the intraoperative neuromonitoring system.
III. RESULTS AND DISCUSSION
A. Uniformity of CNT conductive layer
To assess the uniformity of this coating across a large area, a non-destructive detection method involving laser light penetration is employed. Since the conductive layer of CNTs cannot be directly measured alone, it must be affixed to the filter paper during measurement. First, the intensity of laser light penetrating a single filter paper is established as the benchmark. Then, the laser intensity for the conductive layer adhered to the filter paper's surface is measured. Finally, the laser light intensity difference is calculated for further analysis. Figure 5 shows the experiment result, and it is clear that the laser intensity difference varies from 2.06% to 2.30%, showing a uniform coating layer.
FIG. 5.
The CNT conductive coating is divided into nine grids of the light transmittance distribution map.
B. Compression experiment of flexible piezoresistive sensors
Figures 6(a) and 6(b) show the cross-sectional views of a flexible piezoresistive sensor. To ensure that the CNT coating maintains its conductivity, it is necessary to ensure that the CNT coating was closely attached to the silicone rubber layer. Figure 6(a) is the cross-sectional view, showing the successful and uniform attachment between the silicone layer and the CNT layer, while Fig. 6(b) shows the enlarged view of Fig. 6(a).
FIG. 6.
(a) Cross-sectional view of the piezoresistive element and (b) enlarged cross-sectional view of the piezoresistive element.
Figure 7 shows the mechanical property comparison between the pure silicone and CNT-silicone material. It is clear that both show extremely similar behavior, indicating that the CNT conductive layer has a negligible influence on the mechanical properties of pure silicone rubber.
FIG. 7.
The distance–force diagrams of pure silicone rubber and CNT-silicone rubber sensors.
C. Needle inserting experiments
Needle insertion experiments on silicone rubber and CNT-silicone rubber were conducted using standard 22G, 23G, and 24G needles with a needle moving speed of 70 mm/min. In the experiment, the results of measuring the electrical properties during the process of needle insertion into CNT-silicone rubber are shown in Fig. 8. The process is divided into five stages: the deformation phase, insertion phase I, insertion phase II, Insertion phase III, and extraction phase.
-
(a)
Deformation phase (from A to B), the piezoresistive sensor undergoes compression and deformation due to the downward pressure of the needle. This results in compression with the CNT coating of the motor nerve and close to the rupture of the meninges layer, leading to the first peak resistance value (B).
-
(b)
In insertion phase I (from B to C), the needle penetrates the meninges layer and moves toward the center of the motor nerve. The elasticity of the silicone rubber surrounding the penetration point allows the motor nerve to try to rebound to its original state, causing a decrease in the resistance.
-
(c)
Insertion phase II (from C to D) involves the needle's insertion from the center of the motor nerve to the other side of the motor nerve. This proof involves the compression and stretch of the motor nerve, gradually resulting in the second peak resistance value (D).
-
(d)
Insertion phase III (from D to E) involves the completion of the needle's penetration through the sensor, especially the penetration of the meninges layer at the other side of the motor nerve, causing the fluctuation of the resistance.
-
(e)
Extraction phase (from E to F), the needle is withdrawn, which causes the meninges layer and conductive layer to return to its original shape, leading to a decrease in the resistance.
FIG. 8.
Resistance change during the process of needle insertion into CNT-silicone rubber.
Table I summarizes the destructive forces vs distance for each needle size on both materials, and Fig. 9(a) shows results with a 22G needle, Fig. 9(b) with a 23G needle, and Fig. 9(c) with a 24G needle. As shown in Fig. 9, the measured needle insertion force exhibited a continuous decrease from the first to the fifth cycle. This trend is attributed to the progressive reduction in the force required for needle insertion. Since this experiment is naturally destructive, the needle is inserted at the same position in each cycle. During the first cycle, the material's resistance is at its peak, resulting in a gradual force decrease in the subsequent cycles. Figure S1 in the supplementary material shows that the magnitude of the force diminishes along with the cycle numbers. Changes in the applied force deform the sensor, leading to resistance variations. Kong et al.24 state that stretching the spiral conductive coating causes network separation and disconnection through sliding. With increasing compression force or pulling force, disconnected areas of the conductive network grow until destruction, breaking the conductivity. However, with layers of silicone rubbers to sandwich the conductivity layer, an approximately linear relationship between resistance and cycles is still shown in Fig. S2 in the supplementary material among all cases.
TABLE I.
Needle insertion force and distance of pure silicone rubber and CNT-silicone rubber.
| Silicone rubber | CNT-silicone rubber | |||
|---|---|---|---|---|
| 22G (0.7 mm) | ||||
| Cycle | Distance (mm) | Force (N) | Distance (mm) | Force (N) |
| 1 | 7.75 | 1.13 | 7.66 | 1.51 |
| 2 | 7.88 | 0.94 | 7.77 | 1.19 |
| 3 | 7.79 | 0.87 | 7.78 | 1.15 |
| 4 | 7.95 | 0.75 | 7.75 | 1.11 |
| 5 | 7.94 | 0.74 | 7.66 | 1.11 |
| 23G (0.65 mm) | ||||
| Cycle | Distance (mm) | Force (N) | Distance (mm) | Force (N) |
| 1 | 7.68 | 1 | 7.75 | 1.43 |
| 2 | 7.92 | 0.88 | 7.84 | 1.14 |
| 3 | 7.86 | 0.82 | 7.82 | 1.05 |
| 4 | 7.92 | 0.8 | 7.92 | 1.01 |
| 5 | 7.85 | 0.78 | 7.86 | 0.98 |
| 24G (0.55 mm) | ||||
| Cycle | Distance (mm) | Force (N) | Distance (mm) | Force (N) |
| 1 | 7.81 | 1.22 | 6.68 | 1.26 |
| 2 | 7.9 | 1.15 | 7.76 | 1.02 |
| 3 | 7.65 | 0.94 | 7.76 | 0.9 |
| 4 | 7.85 | 0.89 | 7.82 | 0.88 |
| 5 | 7.72 | 0.88 | 7.68 | 0.84 |
FIG. 9.
The insertion force comparison between silicone rubber and CNT, and pure silicone: (a) with standard 22G needle, (b) with standard 23G needle, (c) with standard 24G needle.
The literature27,28 highlights that various parameters would influence contact force in needle insertion experiments, including insertion speed, needle diameter, needle tip shape, and changing insertion angle. This study only focuses on the influence of the needle diameter on the penetration force. Table I shows the results, a larger needle diameter corresponds to a larger contact area and puncture forces, aligning with literature findings.27,28 CNT-silicone rubber displays increased puncture resistance compared to pure silicone due to the CNT conductive coating situated between silicone insulating layers. Consequently, greater force is required for needle penetration in the composite material.
D. Sensor electrical properties
To assess the impact of the sensor's deformation on its electrical performance stability, an experiment was conducted to examine voltage and current characteristics. The sensor was linked to a power supply and tested under both uncompressed and 3 mm downward compression conditions, with recorded voltages ranging from −5 to 5 V. Changes in current were observed, and the results are shown in Fig. 10. Regardless of the applied compression force, the relationship between the current and voltage displays a robust linear correlation. This suggests that the sensor's resistance adheres to Ohm's law both in its initial state and when subjected to a 3 mm compression. In other words, the sensor's resistance remains constant under a specific compressive strain, as shown in Fig. 10.
FIG. 10.
I–V characteristics for the CNT-silicone sensor under normal conditions or under 3 mm compression.
E. Application: bionic spine simulator test
In this study, a flexible piezoresistive sensor was integrated into a bionic spine simulator to enhance the surgical training of neurosurgeons, as illustrated in Fig. 11(a). This self-developed nerve function monitoring system connects to the bionic motor nerve within the spine, allowing medical students and doctors to practice surgical procedures with real-time feedback, effectively simulating a natural operation. For example, Fig. 11(b) shows a doctor using a scalpel to incise the meninges and begin tumor removal. In Fig. 11(c), forceps are used to remove the tumor while preserving motor nerve integrity and maintaining a stable waveform. Figure 11(d) demonstrates that if the motor nerve is accidentally touched, a warning light is triggered, and waveform fluctuations occur. A detailed demonstration of the device's operation is available in Video 1 in the supplementary material.
FIG. 11.
(a) Dr. Liu used the intraoperative neuromonitoring system as a demonstration to medical students, (b) Dr. Liu used scalpel to cut meninges and initiated tumor removal process, (c) the doctor used forceps to remove the tumor without touching the motor nerves; (d) when the motor nerve is touched, the warning light is triggered, and the waveform fluctuated.
The bionic spine model replicates the real-life scenario of spinal cord tumors growing and compressing the spinal cord and peripheral nerves, leading to abnormal limb paralysis. In actual organs, the spinal cord and nerves are similarly colored, making them indistinguishable to the human eye. Doctors use sharp forceps to remove tumors, which carries the risk of accidentally contacting nerves and potentially causing paralysis. To mitigate this risk, patients are connected to intraoperative neurophysiological monitoring (IONM) during surgery to determine whether the procedure can be safely continued. Given the high-risk nature of this surgical procedure, careful monitoring and extensive practice are crucial. Therefore, the flexible piezoresistive sensor integrated into the bionic spine simulator aims to train surgeons to remove tumors without touching the nerves. If a nerve is accidentally touched, warning lights and sounds are triggered, enhancing the surgeon's precision and skills for real surgeries.
IV. CONCLUSION
This study focuses on developing an innovative piezoresistive sensor that responds to applied pressure with changes in resistance. The results demonstrate a direct correlation between the force magnitude and resistance change, representing the compression properties of pure silicone rubber. The sensor exhibits good linearity in IV characteristics, confirming its stability. Needle insertion experiments, simulating medical procedures, reveal that the force of needle penetration causes resistance changes. Moreover, these flexible piezoresistive sensors can mimic bionic motor nerves, integrating into spinal simulation devices for neurosurgeon training. The system provides real-time signal feedback during intraoperative neurology, enhancing surgical abilities and confidence. The connected spine simulator, linked to an intraoperative neurological function monitoring system, facilitates the timely assessment of neuroma removal without nervous system damage. This approach contributes to improved patient safety and optimal postoperative recovery.
SUPPLEMENTARY MATERIAL
See the supplementary material for a detailed demonstration of the device's operation (Video 1).
ACKNOWLEDGMENTS
This work was supported from the Tri-Service General Hospital (No. TSGH-E 112228 to W.-H.L.), National Science and Technology Council (No. NSTC 112-2314-B-016-055 to W.-H.L.), Shuang Ho Hospital, Taipei Medical University (Nos. 111TMU-SHH-28 and 112TMU-SHH-17 to J.-C.L.), and the National Taiwan University of Science and Technology—Hwa Hsia University of Technology (Taiwan Tech-Hwa Hsia; No. 112TY5M08)
Contributor Information
Pin-Chuan Chen, Email: mailto:pcchen@mail.ntust.edu.tw.
Wei-Hsiu Liu, Email: mailto:liubear0812bear@yahoo.com.tw.
AUTHOR DECLARATIONS
Conflict of Interest
The authors have no conflicts to disclose.
Author Contributions
Sin-Syuan Wu: Data curation (lead); Methodology (equal); Validation (equal); Writing – original draft (lead); Writing – review & editing (lead). Meng Lun Hsueh: Conceptualization (equal); Formal analysis (equal); Investigation (equal). Pin-Chuan Chen: Conceptualization (lead); Funding acquisition (lead); Methodology (lead); Project administration (lead); Supervision (lead); Writing – original draft (lead); Writing – review & editing (lead). Wei-Hsiu Liu: Conceptualization (equal); Funding acquisition (lead); Methodology (lead); Supervision (equal). Jang-Chun Lin: Conceptualization (equal); Methodology (equal).
DATA AVAILABILITY
The data that support the findings of this study are available from the corresponding authors upon reasonable request.
REFERENCES
- 1.Weiss M. Y. et al. , “Design and validation of a cervical laminectomy simulator using 3D printing and hydrogel phantoms,” Oper. Neurosurg. 18(2), 202–208 (2020). 10.1093/ons/opz129 [DOI] [PubMed] [Google Scholar]
- 2.Ganju A. et al. , “The role of simulation in neurosurgical education: A survey of 99 United States neurosurgery program directors,” World Neurosurg. 80(5), e1–e8 (2013). 10.1016/j.wneu.2012.11.066 [DOI] [PubMed] [Google Scholar]
- 3.Bernardo A., “Virtual reality and simulation in neurosurgical training,” World Neurosurg. 106, 1015–1029 (2017). 10.1016/j.wneu.2017.06.140 [DOI] [PubMed] [Google Scholar]
- 4.Choudhury N. et al. , “Fundamentals of neurosurgery: Virtual reality tasks for training and evaluation of technical skills,” World Neurosurg. 80(5), e9–e19 (2013). 10.1016/j.wneu.2012.08.022 [DOI] [PubMed] [Google Scholar]
- 5.de Visser H. et al. , “Progress in virtual reality simulators for surgical training and certification,” Med. J. Aust. 194(4), S38–S40 (2011). [DOI] [PubMed] [Google Scholar]
- 6.Beier F. et al. , “An aneurysm clipping training module for the neurosurgical training simulator NeuroSim,” Medicine Meets Virtual Reality 19, 42–47 (2012). [PubMed] [Google Scholar]
- 7.Craven C. et al. , “Development of a modelled anatomical replica for training young neurosurgeons,” Br. J. Neurosurg. 28(6), 707–712 (2014). 10.3109/02688697.2014.913775 [DOI] [PubMed] [Google Scholar]
- 8.Sharples S. et al. , “Virtual reality induced symptoms and effects (VRISE): Comparison of head mounted display (HMD), desktop and projection display systems,” Displays 29(2), 58–69 (2008). 10.1016/j.displa.2007.09.005 [DOI] [Google Scholar]
- 9.Costa P. et al. , “Somatosensory- and motor-evoked potential monitoring during spine and spinal cord surgery,” Spinal Cord 45(1), 86–91 (2007). 10.1038/sj.sc.3101934 [DOI] [PubMed] [Google Scholar]
- 10.Stecker M. M. and Robertshaw J., “Factors affecting reliability of interpretations of intra-operative evoked potentials,” J. Clin. Monit. Comput. 20(1), 47–55 (2006). 10.1007/s10877-005-9006-8 [DOI] [PubMed] [Google Scholar]
- 11.Sala F. et al. , “Surgery for intramedullary spinal cord tumors: The role of intraoperative (neurophysiological) monitoring,” Eur. Spine J. 16, 130–139 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Gandhi R. et al. , “High-resolution direct microstimulation mapping of spinal cord motor pathways during resection of an intramedullary tumor,” J. Neurosurg. Spine 22(2), 205–210 (2015). 10.3171/2014.10.SPINE1474 [DOI] [PubMed] [Google Scholar]
- 13.Clark A. J. et al. , “Intraoperative neuromonitoring with MEPs and prediction of postoperative neurological deficits in patients undergoing surgery for cervical and cervicothoracic myelopathy,” Neurosurg. Focus 35(1), E7 (2013). 10.3171/2013.4.FOCUS13121 [DOI] [PubMed] [Google Scholar]
- 14.Harper C. M., “Intraoperative cranial nerve monitoring,” Muscle Nerve 29(3), 339–351 (2004). 10.1002/mus.10506 [DOI] [PubMed] [Google Scholar]
- 15.Lall R. R. et al. , “Intraoperative neurophysiological monitoring in spine surgery: Indications, efficacy, and role of the preoperative checklist,” Neurosurg. Focus 33(5), E10 (2012). 10.3171/2012.9.FOCUS12235 [DOI] [PubMed] [Google Scholar]
- 16.Georgopoulou A. and Clemens F., “Piezoresistive elastomer-based composite strain sensors and their applications,” ACS Appl. Electron. Mater. 2(7), 1826–1842 (2020). 10.1021/acsaelm.0c00278 [DOI] [Google Scholar]
- 17.Gao X. et al. , “Flexible stannum-doped SrTiO3 nanofiber membranes for highly sensitive and reliable piezoresistive pressure sensors,” ACS Appl. Mater. Interfaces 13(44), 52811–52821 (2021). 10.1021/acsami.1c17789 [DOI] [PubMed] [Google Scholar]
- 18.Shafique S. et al. , “Carbon nanotubes, orange dye, and graphene powder based multifunctional temperature, pressure, and displacement sensors,” J. Mater. Sci.: Mater. Electron. 31(11), 8893–8899 (2020). 10.1007/s10854-020-03424-5 [DOI] [Google Scholar]
- 19.Nag A. and Mukhopadhyay S. C., “Fabrication and implementation of carbon nanotubes for piezoresistive-sensing applications: A review,” J. Sci.: Adv. Mater. Devices 7(1), 100416 (2022). 10.1016/j.jsamd.2021.100416 [DOI] [Google Scholar]
- 20.Yang H. et al. , “Monotonic strain sensing behavior of self-assembled carbon nanotubes/graphene silicone rubber composites under cyclic loading,” Compos. Sci. Technol. 200, 108474 (2020). 10.1016/j.compscitech.2020.108474 [DOI] [Google Scholar]
- 21.Mu C. et al. , “Flexible strain/pressure sensor with good sensitivity and broad detection range by coupling PDMS and carbon nanocapsules,” J. Alloys Compd. 918, 165696 (2022). 10.1016/j.jallcom.2022.165696 [DOI] [Google Scholar]
- 22.Chen J. W., Zhu Y. T., and Jiang W., “A stretchable and transparent strain sensor based on sandwich-like PDMS/CNTs/PDMS composite containing an ultrathin conductive CNT layer,” Compos. Sci. Technol. 186, 107938 (2020). 10.1016/j.compscitech.2019.107938 [DOI] [Google Scholar]
- 23.Khalid H. R. et al. , “Facile synthesis of sprayed CNTs layer-embedded stretchable sensors with controllable sensitivity,” Polymers 13(2), 311 (2021). 10.3390/polym13020311 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Kong W.-W. et al. , “Highly stretchable and durable fibrous strain sensor with growth ring-like spiral structure for wearable electronics,” Composites, Part B 225, 109275 (2021). 10.1016/j.compositesb.2021.109275 [DOI] [Google Scholar]
- 25.Dong R. and Xie J. B., “Stretchable strain sensor with controllable negative resistance sensitivity coefficient based on patterned carbon nanotubes/silicone rubber composites,” Micromachines 12(6), 716 (2021). 10.3390/mi12060716 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Guo D. J., Pan X. D., and He H., “A simple and cost-effective method for improving the sensitivity of flexible strain sensors based on conductive polymer composites,” Sens. Actuators, A 298, 111608 (2019). 10.1016/j.sna.2019.111608 [DOI] [Google Scholar]
- 27.Jushiddi M. G. et al. , “Simulation of biopsy bevel-tipped needle insertion into soft-gel,” Comput. Biol. Med. 111, 103337 (2019). 10.1016/j.compbiomed.2019.103337 [DOI] [PubMed] [Google Scholar]
- 28.Wang Y. et al. , “Experimental study of the optimum puncture pattern of robot-assisted needle insertion into hyperelastic materials,” Proc. Inst. Mech. Eng., Part H 235(1), 28–43 (2021). 10.1177/0954411920950904 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available from the corresponding authors upon reasonable request.