1. Overview: The Biophysics of Sub-Scalp vs Scalp Continuous EEG
Continuous electroencephalographic monitoring in ambulatory epilepsy has long been constrained by the mechanical and impedance fragility of standard scalp electrochemistry. Traditional silver/silver-chloride ($\text{Ag/AgCl}$) surface electrodes require conductive gels that rapidly dehydrate over 24 to 72 hours, precipitating massive skin-electrode impedance spikes, motion artifacts, and cutaneous breakdown. To circumvent these temporal limitations, implantable continuous EEG monitoring ($\text{iCEM}$) systems, such as the Minder® subgaleal device and UNEEG 24/7 EEG SubQ (EpiSight), surgically position multi-contact platinum-iridium electrode arrays directly within the subgaleal space beneath the scalp and temporoparietal musculature.
While subgaleal placement provides structural stability and shields the recording dipole from ambient electromagnetic powerline interference ($50\text{ Hz}$ in Australia / $60\text{ Hz}$ in the US) and epidermal shearing, it introduces profound trade-offs in spatial sampling, foreign body encapsulation impedance, and analog hardware filtering. This interactive simulation laboratory provides a rigorous, physics-based environment to investigate how spatial distance, glial scar development, and amplifier bandwidth govern the diagnostic fidelity of sub-scalp neural interfaces relative to full International 10–20 scalp arrays.
3. Technical Details: Mathematical Modeling & Biophysical Equations
The simulation engine calculates signal propagation through multilayered volume conductors (brain parenchyma, cerebrospinal fluid, skull bone, galea aponeurotica, and epidermis) using dipole forward solutions governed by Poisson's equation for quasi-static electromagnetic fields:
$$\nabla \cdot (\sigma \nabla \Phi) = -I_v$$
Where $\Phi$ represents electric potential, $\sigma$ is the tensor conductivity of each biological tissue layer, and $I_v$ represents the primary neural current source density. The lead field potential $V_i$ at electrode contact $i$ generated by a focal current dipole moment $\mathbf{p}$ located at source coordinates $\mathbf{r}_0$ is computed via:
$$V_i = \mathbf{L}_i(\mathbf{r}_0) \cdot \mathbf{p} = \frac{1}{4 \pi \sigma_{eff}} \frac{\mathbf{p} \cdot (\mathbf{r}_i - \mathbf{r}_0)}{\|\mathbf{r}_i - \mathbf{r}_0\|^3}$$
Glial Encapsulation & Thermal Noise Physics
Over extended in-vivo residence, host immune responses deposit a collagenous fibrous capsule around subgaleal contacts. We model this progressive impedance increase $Z(t)$ as an asymptotic logistic saturation function:
$$Z(t) = Z_0 + \frac{Z_{max} - Z_0}{1 + e^{-k (t - t_{mid})}}$$
Where $Z_0 = 5\text{ k}\Omega$, $Z_{max} = 52\text{ k}\Omega$, $k = 0.22\text{ week}^{-1}$, and $t_{mid} = 8\text{ weeks}$. The corresponding root-mean-square (RMS) thermal Johnson-Nyquist noise voltage $V_n$ generated across bandwidth $\Delta f = 70\text{ Hz}$ at body temperature ($T = 310.15\text{ K}$) is dynamically calculated via:
$$V_n = \sqrt{4 k_B T \cdot \text{Re}\{Z(t)\} \cdot \Delta f}$$
Where $k_B = 1.380649 \times 10^{-23}\text{ J/K}$ is the Boltzmann constant. At $5\text{ k}\Omega$, $V_n \approx 0.41\text{ }\mu\text{V}_{\text{rms}}$, whereas at $50\text{ k}\Omega$, thermal noise exceeds $1.31\text{ }\mu\text{V}_{\text{rms}}$, degrading the Signal-to-Noise Ratio ($\text{SNR}$):
$$\text{SNR}_{\text{dB}} = 20 \log_{10}\left( \frac{V_{\text{signal}}}{V_n + V_{\text{bio-noise}}} \right)$$
First-Order High-Pass Hardware Filter Model
The analog front-end differential amplifier's high-pass RC filter is modeled via its continuous transfer function $H(s)$ and discrete bilinear transformation:
$$|H(f)| = \frac{\frac{f}{f_c}}{\sqrt{1 + \left(\frac{f}{f_c}\right)^2}}$$
Where $f_c = 1.5\text{ Hz}$ for the Minder iCEM hardware, contrasted with $f_c = 0.15\text{ Hz}$ for standard clinical scalp EEG amplifiers (e.g., Compumedics Siesta). At $f = 0.5\text{ Hz}$ (delta frequency), the implant attenuates signal power by over $-10.3\text{ dB}$ ($>68\%$ voltage loss), eliminating post-ictal diagnostic biomarkers.
4. Economic Teardown & Competitor Landscape (UNEEG SubQ vs. Minder)
To evaluate the commercial viability of ambulatory sub-scalp EEG devices, one must analyze competitors, device pricing, and trial data integrity:
| Device & Manufacturer |
Lead Placement & Architecture |
Device Price / System Cost |
Primary Clinical Function |
Regulatory Status |
UNEEG 24/7 EEG SubQ (EpiSight)
UNEEG Medical A/S (Denmark) |
Unilateral 3-contact subQ lead overlying frontotemporal or parietal area (2-Ch bipolar) |
~£13,500 GBP (~$17,500 USD)
(NICE UK MIB277 guidance) |
Passive Diagnostic: 24/7 loop recording & circadian seizure counting |
CE Marked (Europe)
FDA Breakthrough Designation (2024) |
Minder iCEM®
Epiminder Pty Ltd (Australia) |
Bilateral 4-contact subgaleal lead ear-to-ear across vertex line (2-Ch bipolar) |
~$25,000 USD Target ASP
(US Reimbursement target ~$27,700) |
Passive Diagnostic: Continuous ambulatory loop recording |
FDA De Novo Cleared (April 2025)
Publicly traded: `ASX: EPI` |
NeuroPace RNS System
NeuroPace Inc. (USA) |
Dual 4-contact depth leads or cortical strip leads directly in epileptogenic focus |
~$35,000 – $45,000 USD
($80,000+ total surgical bill) |
Active Therapeutic: Closed-loop electrical stimulation to abort seizures |
FDA Approved (PMA)
Reimbursed therapeutic device |
Cochlear Implant
Cochlear Ltd (Australia) |
22-contact intracochlear array stimulating auditory nerve fibers |
~$24,649 USD Device ASP
($35,000–$50,000 total procedure) |
Active Therapeutic: Restores sensory hearing function |
Global Standard of Care
36% shareholder of Epiminder |
The Diagnostic vs. Therapeutic Price Disparity
A cochlear implant or NeuroPace RNS system delivers **active, life-changing therapeutic intervention** (restoring a physical sense or delivering electrical pulses to halt cortical seizures). In contrast, both the Minder iCEM ($25,000\text{ USD}$) and UNEEG SubQ ($17,500\text{ USD}$) are **passive diagnostic loop recorders**. They do not treat, halt, or prevent seizures. Pricing a passive 2-channel recorder at nearly the same cost as an active therapeutic stimulator represents a significant commercial premium for diagnostic data.
Trial Data Integrity & The US DETECT Study
The initial pilot study evaluating the Minder system (Ganguly et al., 2026; Halliday et al., 2025) contained notable data discrepancies, including a $5 + 4 = 10$ mathematical addition error in Supplemental Table 2, an abstract claiming $N=26$ when 2 patients had no Video-EEG monitoring ($N=24$), and reliance on an automated baseline algorithm (*Persyst 14*) known independently to miss $\sim30\%$ of scalp EEG seizures.
Following Epiminder's **A$125 million IPO** on the Australian Securities Exchange in December 2025 (valuing the company at A$325 million), the company is deploying capital into the **US DETECT trial (`NCT07110337`)**—a 210-patient randomized controlled trial across 20+ major US medical centers (including Mayo Clinic, Harvard, and Stanford) aimed at securing US Medicare reimbursement.