Interface geometry governs probe insertion mechanics and electrode deviation in layered brain-mimicking phantoms for deep brain stimulation Authors: Siyu Chen, Chengyuan Wu, Ethan Anders, Rungun Nathan, Olivia M. Bailey, Qifu Wang, Curtis L. Johnson, Qianhong Wu Published: 2026 ABSTRACT Parkinson’s disease (PD) is a progressive neurological disorder often treated with Deep Brain Stimulation (DBS), a surgical procedure that involves implanting electrodes to improve motor function. Accurate electrode placement is critical for treatment efficacy but can be influenced by tissue deformation, insertion forces, and the mechanical properties of brain tissue. In this study, we investigated the role of layer interface angle as a controlled biomechanical factor affecting insertion forces and electrode deviation. A two-layer agar gel phantom was constructed to produce a prescribed stiffness transition between a lower-stiffness and a higher-stiffness gel layer, and its small-amplitude dynamic shear-stiffness contrast was characterized using magnetic resonance elastography (MRE) at 100 Hz. A stage-wise analytical model was formulated to describe the Z-direction axial force–depth response \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{z}\left(z\right)$$\end{document}Fzz during indentation, fracture, relaxation, and frictional sliding, and a hybrid physics-guided residual neural network (PGNN) was explored as a secondary residual-correction approach for the remaining axial-force model discrepancies. The lateral interface-force change \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{iy}$$\end{document}Fiy was obtained directly from the measured Y-direction force signal, while electrode deviation was quantified in a separate series of phantom experiments. Insertion experiments at interface angles of 0°, 15°, 30°, 45°, 60°, and 75° demonstrated that the lateral (Y-direction) interface force increased from 0.886 mN at 0° to 7.72 mN at 75°. For the deviation analysis, the mean value of the 0° group was defined as the zero reference; relative to this baseline, the largest group mean deviation was 0.0529 mm at 75°. These findings highlight the role of interface geometry in probe trajectory stability and provide a mechanistic baseline for future studies of probe insertion in two-layer media.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s10237-026-02122-1. Introduction Parkinson's disease (PD) is a progressive neurodegenerative disorder that significantly affects the quality of life of millions of individuals worldwide. It is characterized by motor symptoms such as tremor, rigidity, and bradykinesia, which worsen over time and impose a substantial burden on patients, caregivers, and healthcare systems. Recent estimates indicate that nearly 90,000 people are diagnosed with PD annually in the USA, with the global burden of the disease resulting in 5.8 million disability-adjusted life years (DALYs) in 2019, marking an 81% increase since 2000 (Ray Dorsey et al. 2018; Marras et al. 2018). While pharmacological therapies, such as levodopa, are effective in managing early-stage symptoms, their efficacy diminishes as the disease progresses, leaving patients with unmet clinical needs (Beckers et al. 2022). Deep Brain Stimulation (DBS) has served as a transformative surgical intervention for advanced PD patients. By implanting electrodes into specific brain regions to modulate neural activity, DBS provides significant improvements in motor symptoms and overall quality of life (Miocinovic et al. 2013). One commonly used technique involves inserting a rigid cannula along the planned trajectory, with or without the use of microelectrode recordings for physiological confirmation. Another method employs a radiofrequency (RF) probe with dynamic impedance monitoring to establish a low-resistance channel, allowing direct electrode insertion after probe withdrawal without a cannula (Foltynie et al. 2011) (Fayed et al. 2024) (Ziechmann et al. 2026), as illustrated in Fig. 1a. Electrode placement accuracy and deviation have been shown to be comparable between these two approaches (Rajabian et al. 2023). However, the success of DBS is highly dependent on the accuracy of electrode placement, and enhanced targeting strategies have been a key driver in the evolution of DBS applications (Burchiel et al. 2015). Suboptimal electrode placement can compromise therapeutic efficacy and increase the risk of adverse effects, such as dysarthria and motor deficits (Kluin et al. 2022).Fig. 1a Steps of a typical DBS procedure. Left: A probe is first inserted into the brain, passing through multiple structures to reach the target. Middle: The probe is then withdrawn, creating a channel. The shape, size, and location of this channel may be influenced by local tissue properties, and deflection can occur especially at interfaces. Right: The DBS electrode is inserted along the preformed channel toward the target region. However, deformation of the channel may lead the electrode to deviate from the target. b The trajectory of the electrode reaching the DBS target passes through multiple interfaces. Taking the Globus Pallidus Internus (GPi) as an example, the trajectory intersects with an average of four distinct tissue interfaces, each with different angles a Steps of a typical DBS procedure. Left: A probe is first inserted into the brain, passing through multiple structures to reach the target. Middle: The probe is then withdrawn, creating a channel. The shape, size, and location of this channel may be influenced by local tissue properties, and deflection can occur especially at interfaces. Right: The DBS electrode is inserted along the preformed channel toward the target region. However, deformation of the channel may lead the electrode to deviate from the target. b The trajectory of the electrode reaching the DBS target passes through multiple interfaces. Taking the Globus Pallidus Internus (GPi) as an example, the trajectory intersects with an average of four distinct tissue interfaces, each with different angles Electrode placement during DBS is influenced by several factors, including anatomical variations, accuracy of the stereotactic system, and intraoperative factors such as global brain shift (Li et al. 2016; Ko et al. 2017). Real-time assessment of electrode location during surgery has been recognized as a key technical challenge (Vissani et al. 2020). Biomechanical factors could also influence electrode placement accuracy. Electrode–tissue interactions, including tissue deformation and probe–tissue interaction, have been shown to significantly contribute to deviations from the intended electrode trajectory (Jiang et al. 2014; Chapelle et al. 2021). During the DBS procedure, the electrode must traverse multiple tissue structures and their interfaces, as illustrated in Fig. 1b. These tissues exhibit distinct biomechanical properties, which can introduce mechanical asymmetry during insertion and contribute to systematic deviations in electrode trajectory. As a probe is inserted or removed, the brain’s viscoelastic properties, anisotropy, and material heterogeneity may generate resistance and reaction forces that impact insertion accuracy. Following the extraction of the probe, the electrode is inserted along the channel created by the probe to reach the planned target. We hypothesize that electrode deviation from the target is influenced largely by biomechanical factors present during probe or cannula insertion, and that these effects can be isolated and quantified using controlled phantom experiments. These deviations may be particularly pronounced at tissue interfaces where mechanical properties differ, such as the boundary between gray and white matter (Budday et al. 2015). Recent advancements in DBS technology, such as directional leads, adaptive stimulation systems, and advanced imaging techniques, have improved clinical outcomes by increasing tolerance to electrode deviations (Ramirez-Zamora et al. 2018). However, a gap remains in understanding the biomechanical interactions that occur during electrode insertion. Studies have suggested that the probe–tissue interaction and interface angle could influence electrode trajectory (Siyu Chen et al. 2024); yet these factors are not well characterized from a mechanics perspective. In a clinical study of 21 DBS electrode trajectories, the angle between the planned trajectory and the normal to an MRE-defined stiffness interface was significantly correlated with target deviation, with larger angles associated with greater deviations (r=0.57, p=0.007) (Wu et al. 2025). This finding provides clinical evidence of an association between interface geometry and electrode deviation, while the underlying mechanics remain inadequately characterized. The present study, therefore, isolates interface angle under controlled phantom conditions and addresses three core questions:What mechanical processes characterize the axial force–depth response during probe insertion,How interface angle affects the directly measured lateral interface-force change \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{iy}$$\end{document}Fiy, andHow interface angle is associated with electrode deviation measured in a separate experimental series. What mechanical processes characterize the axial force–depth response during probe insertion, What mechanical processes characterize the axial force–depth response during probe insertion, How interface angle affects the directly measured lateral interface-force change \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{iy}$$\end{document}Fiy, and How interface angle affects the directly measured lateral interface-force change \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{iy}$$\end{document}Fiy, and How interface angle is associated with electrode deviation measured in a separate experimental series. How interface angle is associated with electrode deviation measured in a separate experimental series. To address these objectives, we constructed a two-layer agar gel phantom with a prescribed contrast between lower- and higher-stiffness layers, enabling controlled and systematic investigation of interface-driven biomechanical responses. This simplified experimental framework is designed to isolate fundamental interface-driven mechanical effects, rather than replicate the full complexity of in vivo brain tissue. A stage-wise analytical force model was formulated to characterize the measured Z-direction axial force–depth response \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{z}\left(z\right)$$\end{document}Fzz during indentation, fracture, viscoelastic relaxation, and frictional sliding, and a hybrid physics-guided residual neural network (PGNN) was incorporated as a residual correction to this axial-force model. The lateral force \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\boldsymbol{F}}}_{{\boldsymbol{y}}}$$\end{document}Fy and the derived interface-force change \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{iy}$$\end{document}Fiy were quantified directly from the force-transducer measurements across the prescribed interface angles. Electrode deviation was quantified in a separate series of phantom experiments using high-speed imaging. This combined design enables the axial-force response, lateral interface force, and electrode deviation to be evaluated as complementary outcomes with interface angle as the common controlled variable. Together, these experimental and modeling approaches allow us to characterize axial insertion mechanics and determine the angle-dependent trends in lateral force and electrode deviation under controlled phantom conditions, providing insights that advance the mechanistic understanding of probe–tissue interactions in two-layer media and may inform future studies of insertion mechanics in more complex tissue environments. Materials and methods We first investigated the mechanical properties of agar gel at different concentrations using Magnetic Resonance Elastography (MRE). A custom-designed system was constructed in which a probe, driven by a motor, was inserted into a two-layer gel with interface angles set at 0°, 15°, 30°, 45°, 60°, and 75°. Reaction forces on the probe were measured. After the probe was removed, a DBS electrode was inserted. High-speed imaging captured the probe insertion, removal, and electrode placement processes, enabling quantification of electrode deviation from the intended trajectory. Experimental materials and MRE test Agar gel (Sigma-Aldrich 05039) was selected as a transparent and mechanically tunable phantom material that facilitates real-time visualization during the experiments. The previous studies have demonstrated the utility of agar-based gels in optically accessible brain phantom experiments (Albor-Ramírez et al. 2023) and in controlled investigations of brain tissue injury patterns (Falland-Cheung et al. 2016). To establish a prescribed stiffness contrast motivated by the mechanical heterogeneity among brain regions (Hiscox et al. 2020), 0.6% w/w agar gel was used as the lower-stiffness outer layer and 0.7% w/w agar gel as the higher-stiffness inner layer. These designations refer to the relative stiffness of the two layers within the phantom and do not imply full mechanical equivalence to gray matter, white matter, or deep brain nuclei. Preparation followed a standardized protocol: Agar gel and water were weighed by mass fraction, mixed, heated to 90–95 °C, poured into containers, and refrigerated for 24 h to solidify (McIlvain et al. 2019). To characterize the small-amplitude, frequency-specific dynamic response of the gel phantoms, MRE was employed to quantify stiffness and damping ratio by imaging the displacement fields induced by harmonic vibrations (Muthupillai et al. 1995). MRE operates by generating shear waves within the material, acquiring MR images of wave propagation, and processing these images to produce quantitative stiffness maps known as elastograms. This technique has demonstrated excellent repeatability and interobserver consistency (Murphy et al. 2019), and has shown promising applicability in a variety of contexts, including brain imaging (Nanjappa and Kolipaka 2021). Vibrations at 100 Hz were applied to agar gel samples using a pneumatic actuator system (Resoundant; Rochester, MN). MRE data were acquired on a Siemens 3T Prisma MRI scanner using an EPI sequence with 2.5 mm isotropic spatial resolution (Chaze et al. 2019). To ensure the consistency and stability of the agar gel phantoms used for MRE, all samples were prepared on the same day by a single experimenter following a standardized protocol (McIlvain et al. 2019). Each sample was created using 1000 mL of water and cast into a disposable square plastic container (dimensions: 156 mm × 156 mm × 86 mm; volume: 1180 mL). Each phantom was allowed to solidify within its container for 24–48 h and was scanned in situ during MRE acquisition. These 100-Hz MRE measurements were used to characterize the dynamic stiffness contrast between the gel formulations and were not used to infer their finite deformation, rupture, or frictional behavior during probe insertion. Experimental setup Force measurement experiment Figure 2 provides the schematic of the force measurement setup. To minimize noise caused by vibrations from the driving mechanism, the force sensor was directly attached to the fixed probe. The two-layer agar gel was mounted on a platform that could be set in motion by the driving mechanism, allowing the insertion and removal of the probe from the agar gel. The maximum insertion depth for the force measurement experiments was 80 mm, and the insertion speed was 19.3 mm/s. Six experiments were performed for each interface angle (total of 36 experiments). In this study, we used the same 1.8-mm-diameter probe (Cosman G4, Boston Scientific) employed in DBS procedures at Thomas Jefferson Hospital (Fayed et al. 2024; Ziechmann et al. 2026). The probe, made of Nitinol (nickel–titanium alloy), has a 4 mm exposed tip and a total length of 29 cm and a Young’s modulus around 74.21 GPa (Massey et al. 2021), which is several orders of magnitude higher than that of the agar gel. Therefore, the probe can be reasonably regarded as a rigid body during its interaction with the gel. Force measurements were performed using an IP65 Nano17 transducer (ATI Industrial Automation, Apex, NC), which is capable of capturing forces and torques along three axes with a resolution as fine as 0.0001 N. Data acquisition and processing were conducted using NI-9205 and NI cDAQ-9174 system (National Instruments, Austin, TX) with a sampling rate of 3000 Hz for each input. A custom LabVIEW program synchronized data collection from multiple signal sources. The transducer directly recorded the Z-direction axial force \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{z}$$\end{document}Fz and the Y-direction lateral force \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{y}$$\end{document}Fy. The measured \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{z}\left(z\right)$$\end{document}Fzz response was used to identify and evaluate the stage-wise analytical and PGNN-corrected axial-force models, whereas the interface-force change \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{iy}$$\end{document}Fiy was calculated from the measured \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{y}$$\end{document}Fy signal and analyzed as an experimental outcome. To reduce random electrical noise, the input signal was filtered using a third-order Butterworth low-pass filter with an 80 Hz cutoff frequency.Fig. 2Schematic diagram of force measurement experiment Schematic diagram of force measurement experiment Force measurements were performed in two sequential stages. First, homogeneous single-layer agar phantoms were tested to identify the layer-specific parameters used in the analytical model. Single-layer phantoms were prepared using 0.6% w/w agar and 0.7% w/w agar, corresponding to the outer and inner layers of the subsequent two-layer configuration. Because these phantoms were mechanically homogeneous, the measured force–displacement response reflected the intrinsic insertion and withdrawal behavior of each individual material without the added complexity of an interface. Controlled insertion and withdrawal tests were performed for each concentration, with six repeated trials conducted per material. The resulting force–displacement curves were used to determine the layer-specific parameters for nonlinear indentation, post-fracture relaxation, cutting resistance, and probe–gel friction, which were then transferred to the two-layer model. Second, two-layer force measurement experiments were performed to evaluate interface mechanics under different interface angles. In these experiments, the phantom consisted of an outer 0.6% w/w agar layer and an inner 0.7% w/w agar layer, with interface angles set at 0°, 15°, 30°, 45°, 60°, and 75°. The maximum insertion depth was 80 mm, and the insertion speed was 19.3 mm/s. Six experiments were performed for each interface angle, for a total of 36 two-layer force measurements. In contrast with the single-layer tests, these experiments were designed to capture the additional force variations associated with interface crossing, including axial and lateral loading changes that arise from stiffness transition and interface geometry. Together, the single-layer calibration experiments and the two-layer force measurement experiments enabled separation of layer-specific material behavior from interface-driven force asymmetry. Electrode deviation experiment In a separate experimental series, we used the two-layer agar gel phantom to observe and quantify electrode deviation. The force measurement and electrode deviation datasets were acquired using separate phantoms and experimental trials, with interface angle serving as the common controlled variable for group-level comparison. The experimental setup, shown in Fig. 3, includes a linear driving mechanism that enables precise control of probe and electrode speed and displacement, thereby ensuring accurate and reliable insertions. For this experiment, the DBS electrode used was a 40 cm–long, 1.27 mm–diameter 8-channel directional lead (Infinity 6172, Abbott, Plano, TX) (Fayed et al. 2024)(Ziechmann et al. 2026). The distal tip of the electrode consists of four 1.5 mm tall rigid platinum–iridium electrodes spaced 0.5 mm apart on a flexible insulated polycarbonate polyurethane substrate. In addition, a reducer with an inner diameter of 1.8 mm and length of 73 mm was used to maintain a straight trajectory and replicate surgical tools used in DBS implantation surgeries. The probe was first inserted into the agar gel through the reducer and then withdrawn to create a channel. The maximum insertion depth was 80 mm, and the movement speed was 20.0 mm/s. An industrial flare camera (IO Industries, Canada) was positioned above the agar gel to capture the insertion process, with an imaging resolution of 4096 × 3072 pixels and a frame rate of 100 FPS. The slider was then moved to allow the insertion of the electrode wire through the preformed channel, as shown in Fig. 3a, b. Agar gel samples were placed in a custom-designed container featuring side openings to facilitate probe insertion and removal. The samples consisted of two layers with varying interface angles, as illustrated in Fig. 3c. In three-dimensional vector form, the interface angle was defined as \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\alpha ={cos}^{-1}(\mid {\boldsymbol{t}}\cdot {\boldsymbol{n}}\mid ),$$\end{document}α=cos-1(∣t·n∣), where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\boldsymbol{t}}$$\end{document}t is the unit tangent vector along the probe trajectory, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\boldsymbol{n}}$$\end{document}n is the unit normal to the interface. For the planar, coplanar geometry used in the present experiments, this definition is equivalent to the two-dimensional construction shown in Fig. 3c, with \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\alpha ={0}^{\circ }$$\end{document}α=0∘ representing normal incidence. The angle ββ denotes the complementary in-plane angle between the probe trajectory and the interface line.Fig. 3a Probe inserted into agar gel b After removing the probe, the electrode is inserted into agar gel c Schematic diagram of the two-layer agar gel, where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\boldsymbol{\alpha}}$$\end{document}α is the interface angle between the probe trajectory direction and the interface normal (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\boldsymbol{\alpha}}={0}^{\circ }$$\end{document}α=0∘ at normal incidence), and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\boldsymbol{\beta}}$$\end{document}β is the complementary in-plane angle between the probe trajectory and the interface line d Schematic diagram of the deviation of the electrode insertion position from the position A. A demonstration video S1 is provided in the supplementary materials a Probe inserted into agar gel b After removing the probe, the electrode is inserted into agar gel c Schematic diagram of the two-layer agar gel, where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\boldsymbol{\alpha}}$$\end{document}α is the interface angle between the probe trajectory direction and the interface normal (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\boldsymbol{\alpha}}={0}^{\circ }$$\end{document}α=0∘ at normal incidence), and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\boldsymbol{\beta}}$$\end{document}β is the complementary in-plane angle between the probe trajectory and the interface line d Schematic diagram of the deviation of the electrode insertion position from the position A. A demonstration video S1 is provided in the supplementary materials A thin acrylic plate was positioned on top of the container, allowing unobstructed visualization while preventing sample deformation. To capture deviations in probe and electrode positions, a high-speed camera was mounted above the agar gel samples. If the position of the probe at target depth is defined as A, and the position of the electrode at target depth is defined as B, then the deviation is defined as the difference in the Y-direction between A and B (Fig. 3d). For subsequent comparisons, the mean measured deviation of the 0° group was defined as zero. The reported group means at 15°–75° therefore represent deviations relative to this zero-referenced baseline. These deviations were experimentally measured for different values of α. Ten independent electrode deviation trials were conducted at each of the six interface angles, yielding 60 deviation trials in total.Fig. 4Flowchart of the stage-wise axial-force model for probe insertion and withdrawal. The process is divided into six stages: pre- and post-fracture insertion for each gel layer, a transition phase during motion reversal, and withdrawal dominated by friction. Distinct scalar axial-force relations are applied to each stage according to the dominant interaction mechanisms Flowchart of the stage-wise axial-force model for probe insertion and withdrawal. The process is divided into six stages: pre- and post-fracture insertion for each gel layer, a transition phase during motion reversal, and withdrawal dominated by friction. Distinct scalar axial-force relations are applied to each stage according to the dominant interaction mechanisms Statistical analysis Statistical analysis was performed to quantify the relationships between interface angle and the measured mechanical or geometric response variables. Pearson correlation analysis was used to evaluate the association between interface angle and the Z-axis force change at the interface \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\left({F}_{iz}\right)$$\end{document}Fiz, the Y-axis force change at the interface \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\left({F}_{iy}\right)$$\end{document}Fiy, and electrode deviation. In addition, one-way analysis of variance (ANOVA) was used to compare electrode deviation across interface angle groups. Statistical significance was defined as p<0.05p<0.05. Force and deviation data are reported as mean ± standard deviation unless otherwise stated. Stage-wise axial-force model for probe insertion and withdrawal To characterize the measured Z-direction axial force–depth response \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{z}\left(z\right)$$\end{document}Fzz, the probe insertion and withdrawal processes were divided into six distinct stages (Fig. 4): (1) pre-penetration of the first gel layer, characterized by pronounced nonlinearity before full penetration; (2) post-fracture insertion into the first gel layer, dominated by frictional and cutting forces; (3) pre-penetration of the second gel layer, again exhibiting nonlinear characteristics prior to full penetration; (4) post-fracture insertion into the second gel layer, governed by frictional and cutting forces; (5) a transitional phase during which the force direction reverses as motion shifts from insertion to withdrawal; and (6) probe withdrawal, primarily dominated by frictional interactions. Different axial-force relations were applied to each phase according to the dominant interaction mechanisms. The resulting model provides a stage-wise representation of the scalar axial force \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{z}\left(z\right)$$\end{document}Fzz. The lateral force \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{y}$$\end{document}Fy, the derived interface-force change \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{iy}$$\end{document}Fiy, and electrode deviation are evaluated as complementary experimental outcomes through the procedures described in Sects. 2.2.1 and 2.2.2. Insertion of the probe into the two-layer gel Stage 1: pre-penetrating indentation of the outer (first) gel layer As shown in Fig. 5a, during the pre-penetration stage, the probe tip indents the surface of the first gel layer before fracture occurs. In this stage, the agar gel exhibits pronounced nonlinear deformation behavior, and the reaction force increases rapidly with indentation depth. Such nonlinear stiffening is commonly observed in soft gels and biological tissues and cannot be fully captured by classical linear-elastic contact models such as the Hertz theory (K. L. Johnson 1985). The deviation from the Hertzian prediction mainly arises from large deformation, viscoelasticity, and intrinsic material nonlinearity (Efremov et al. 2015).Fig. 5Schematic illustration of the probe–gel interaction stages during insertion and withdrawal in a two-layer agar gel: a pre-penetration of the first layer; b post-fracture insertion with cutting and friction; c pre-penetration of the second layer; d post-fracture insertion into the second layer; and e withdrawal dominated by interfacial friction Schematic illustration of the probe–gel interaction stages during insertion and withdrawal in a two-layer agar gel: a pre-penetration of the first layer; b post-fracture insertion with cutting and friction; c pre-penetration of the second layer; d post-fracture insertion into the second layer; and e withdrawal dominated by interfacial friction To better describe this nonlinear force–displacement relationship, we adopted a Fung-type exponential relation as a compact phenomenological representation of the measured pre-puncture response (Fung 1967; Gefen et al. 2003). The relation was used to capture the observed nonlinearity under the present insertion conditions rather than as a complete constitutive law for brain tissue. The Fung model relates nominal stress σ to strain ε through1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\sigma = \frac{1}{\beta }\left( {e^{\beta k\varepsilon } - 1} \right),$$\end{document}σ=1βeβkε-1,where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\varepsilon =\left(z-{z}_{s}\right)/R$$\end{document}ε=z-zs/R is a dimensionless indentation measure. In the present N–mm implementation, kk is an effective stress-scale parameter with units of N/mm2, whereas ββ is an inverse-stress nonlinearity parameter with units of mm2/N. Consequently, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\beta k\varepsilon$$\end{document}βkε is dimensionless, and 1/β1/β has units of stress. For indentation, the zz is the indentation depth, and R is the probe radius. Assuming the nominal stress is given by \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\sigma =F/(\pi {R}^{2})$$\end{document}σ=F/(πR2), the relationship between the measured Z-direction axial force \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{z}$$\end{document}Fz and indentation depth zz can be derived as2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{z,1}\left(z\right)=-\frac{\pi {R}^{2}}{{\beta}_{1}}\left({e}^{\frac{{\beta}_{1}{k}_{1}\left(z-{z}_{s1}\right)}{R}}-1\right),$$\end{document}Fz,1z=-πR2β1eβ1k1z-zs1R-1,where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$R=0.9\hspace{0.17em}mm$$\end{document}R=0.9mm is the probe radius, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${k}_{1}$$\end{document}k1 and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\beta}_{1}$$\end{document}β1 are the effective stress-scale and inverse-stress nonlinearity parameters, respectively, which were obtained by fitting the experimental force–displacement data from single-layer 0.6% agar gel tests, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${z}_{s1}$$\end{document}zs1 denotes the onset position of this stage. These fitted parameters are specific to the present probe geometry, gel formulation, loading path, and experimental conditions. They were not derived from the MRE measurements and should not be interpreted as constitutive parameters of brain tissue. Stage 2: Post-penetrating insertion into the outer gel layer After the gel interface is punctured, as illustrated in Fig. 5b, the total insertion force can be decomposed into three distinct components: (1) an exponentially decaying term that accounts for the stress relaxation immediately following rupture (Abolhassani et al. 2007); (2) a depth-dependent frictional resistance distributed along the probe surface (Okamura et al. 2004); and (3) a constant cutting force acting at the probe tip (Barbé et al. 2007). Following the framework proposed in the previous soft-tissue insertion studies (Okamura et al. 2004; Barbé et al. 2007; Abolhassani et al. 2007), the cutting force is treated as a constant term, while the frictional resistance is considered proportional to the inserted length and dependent on the interfacial friction properties of the gel. The stress-relaxation behavior is modeled using an exponential decay, reflecting the time- and strain-dependent viscoelastic response of the agar matrix after fracture. Accordingly, the total force during this post-penetration stage is expressed as3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{z,2}\left(z\right)=-\left[{A}_{2}{e}^{-\frac{\left(z-{z}_{0}\right)}{{\tau}_{2}}}+{F}_{z, fric}^{\left(2\right)}\left(z\right)+{C}_{2}\right],$$\end{document}Fz,2z=-A2e-z-z0τ2+Fz,fric2z+C2,where zz is the insertion depth, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${z}_{0}$$\end{document}z0 is the transition point from Stage 1, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${A}_{2}$$\end{document}A2 is the amplitude of the relaxation term in N, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\tau}_{2}$$\end{document}τ2 is the relaxation length constant in mm, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{\mathrm{z},\mathrm{f}\mathrm{r}\mathrm{i}\mathrm{c}}^{\left(2\right)}\left(z\right)$$\end{document}Fz,fric2z is the frictional resistance, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${C}_{2}$$\end{document}C2 is the constant cutting-force term in N. Because the insertion speed was fixed, depth was used as a surrogate for time in the exponential relaxation term; \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\tau}_{2}$$\end{document}τ2 is, therefore, a relaxation length constant, and the corresponding effective relaxation time is \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\tau}_{2}/v$$\end{document}τ2/v. Continuity between the pre- and post-penetration stages is enforced at the transition point \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$({z}_{0}={z}_{e1})$$\end{document}(z0=ze1) through4\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${A}_{2}={-F}_{z,1}\left({z}_{0}\right)-{F}_{z,fric}^{\left(2\right)}\left({z}_{0}\right)-{C}_{2},$$\end{document}A2=-Fz,1z0-Fz,fric2z0-C2,where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${z}_{e1}$$\end{document}ze1 denotes the end position of Stage 1, corresponding to the fracture point of the outer layer, ensuring a smooth force transition across the fracture interface. The frictional component in this regime is modeled as a linearly increasing function of insertion depth,5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{z,fric}^{\left(2\right)}\left(z\right)=2\pi R{\mu}_{2}z,$$\end{document}Fz,fric2z=2πRμ2z,where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mu}_{2}$$\end{document}μ2 is the effective interfacial frictional shear-stress parameter (N/mm2) identified from the single-layer 0.6% agar experiments. Stage 3: Pre-penetrating indentation of inner gel layer As the probe proceeds into the second gel layer, it encounters a new pre-penetration phase before fracturing the inner interface, as illustrated in Fig. 5c. Similar to Stage 1, this stage is characterized by pronounced nonlinear stiffening as the gel resists indentation prior to rupture. However, due to the pre-compression and confinement effects induced by the rupture of the first layer, the deformation behavior of the second layer exhibits higher apparent stiffness and reduced compliance. The force–displacement relationship during this phase follows the same phenomenological exponential form used in Stage 1, with separate effective parameters identified for the second gel layer (Fung 1967; Gefen et al. 2003), which effectively captures the nonlinear stress–strain behavior of soft hydrogels:6\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$F_{z,3} \left( {\mathrm{z}} \right) = {\mathrm{F}}_{{{\mathrm{z}},2}} \left( {\mathrm{z}} \right) - \frac{{\pi R^{2} }}{{\beta_{3} }}\left( {{\mathrm{e}}^{{{\upbeta}_{3} {\mathrm{k}}_{3} \left( {{\mathrm{z}} - {\mathrm{z}}_{{{\mathrm{s}}3}} } \right)/{\mathrm{R}}}} - 1} \right),$$\end{document}Fz,3z=Fz,2z-πR2β3eβ3k3z-zs3/R-1, Where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${k}_{3}$$\end{document}k3 and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\beta}_{3}$$\end{document}β3 are the effective stress-scale and inverse-stress nonlinearity parameters for the second gel layer, respectively, obtained by fitting the corresponding single-layer 0.7% agar gel experiments, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${z}_{s3}$$\end{document}zs3 denotes the onset position of this stage. Stage 4: Post-penetrating insertion into the inner gel layer After the probe punctures the interface and enters the second gel layer, as shown in Fig. 5d, the insertion force again consists of three components: (1) an exponentially decaying term describing the stress relaxation immediately following fracture, (2) a depth-dependent frictional resistance distributed along the probe surface, and (3) a constant cutting force acting at the probe tip. Owing to the higher agar concentration of the second layer, both the frictional and cutting components exhibit greater magnitudes compared with those in the outer layer, while the relaxation process occurs over a shorter characteristic length. The total insertion force in this regime is expressed as7\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{z,4}\left(z\right)=-\left[{{A}_{4}e}^{-\frac{\left(z-{z}_{e3}\right)}{{\tau}_{4}}}+{F}_{z,fric}^{\left(4\right)}\left(z\right)+{C}_{4}\right],$$\end{document}Fz,4z=-A4e-z-ze3τ4+Fz,fric4z+C4,where zz is the insertion depth, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${z}_{e3}$$\end{document}ze3 denotes the transition point from Stage 3, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${A}_{4}$$\end{document}A4 is the amplitude of the relaxation term, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\tau}_{4}$$\end{document}τ4 represents the relaxation length constant, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{z,fric}^{(4)}(z)$$\end{document}Fz,fric(4)(z) is the frictional resistance, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${C}_{4}$$\end{document}C4 is the cutting force. Continuity with the preceding stage is enforced through8\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${A}_{4}={-F}_{z,3}\left({z}_{e3}\right)-{F}_{z,fric}^{\left(4\right)}\left({z}_{e3}\right)-{C}_{4},$$\end{document}A4=-Fz,3ze3-Fz,fric4ze3-C4,ensuring a smooth transition of the force curve across the inner layer fracture boundary. The frictional resistance is modeled as a linear function of insertion depth,9\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{z,fric}^{\left(4\right)}\left(z\right)={\mu}_{2}\cdot 2\pi R{z}_{e2}+{\mu}_{4}\cdot 2\pi R\left(z-{z}_{e2}\right).$$\end{document}Fz,fric4z=μ2·2πRze2+μ4·2πRz-ze2.where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${z}_{e2}$$\end{document}ze2 denotes the end position of Stage 2, corresponding to the length of probe insertion within the outer layer before entering the second layer. And \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\mu}_{4}$$\end{document}μ4 is the effective interfacial frictional shear-stress parameter (N/mm2) determined from experiments using single-layer 0.7% agar gels. Removal of the probe from a two-layer agar gel After the probe reached the target depth, it was gradually withdrawn from the two-layer agar gel, as shown in Fig. 5e. During this phase, the force response is dominated by interfacial friction between the probe shaft and the surrounding gel. Unlike during insertion, no further fracture or viscoelastic relaxation occurs, and the measured force primarily reflects sliding resistance along the wetted probe–gel interface. Under the fixed withdrawal speed used in this study, the withdrawal resistance was represented by a linear embedded length-dependent friction law (Barbé et al. 2007; Abolhassani et al. 2007), where the total resistance scales proportionally with the embedded length. To account for the different interfacial properties of the two-layer gel, the overall withdrawal force is modeled as10\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${F}_{z,6}\left(z\right)=\left\{\begin{array}{c}2\pi R{\mu}_{6}z, (z<{z}_{s3}) \\ 2\pi R{\mu}_{6}{z}_{s3}+2\pi R{\mu}_{7}\left(z-{z}_{s3}\right), (z\ge {z}_{s3})\end{array},\right.$$\end{document}Fz,6z=2πRμ6z,(z