A Control-Referenced Tri-Channel OECT Receiver for Hybrid Molecular Communication Toward Brain Organoid Interfaces

Hongbin Ni, Özgür B. AkanarXiv 2026 · DOI · arXiv

TL;DR

This paper introduces a novel tri-channel organic electrochemical transistor (OECT) receiver designed for brain organoid interfaces. By incorporating a matched control channel, the receiver significantly improves the accuracy of decoding hybrid molecular signals, particularly in the presence of common-mode noise, paving the way for more robust brain-computer interfaces.

What Problem Does This Paper Address?

Brain organoids, self-organized neural tissues, are increasingly used to study brain development and function. Understanding their state requires sensing neuromodulators like dopamine and serotonin, which play crucial roles in neural signaling. Existing methods for detecting these molecules face challenges: electrophysiology can't directly sense neuromodulators, while techniques like fast-scan cyclic voltammetry, microdialysis, and optical reporters have limitations in selectivity, invasiveness, speed, or phototoxicity. This paper aims to develop a low-bias, label-free receiver compatible with long-term organoid experiments that can specifically detect and decode neuromodulator signals. A key challenge is dealing with common-mode drift and low-frequency noise inherent in electrolyte-based sensing, which can corrupt signal detection.

Methods

The core innovation is a tri-channel OECT receiver. This receiver features:

  • Two selective channels: One functionalized with an aptamer to detect dopamine (DA) and another for serotonin (5-HT). Aptamers are DNA or RNA molecules that can bind to specific targets.
  • One control channel: This channel is identical to the selective channels in terms of its hydrogel and electrolyte but lacks the aptamer. Its purpose is to act as a matched reference to capture common-mode noise and drift affecting all channels.

The receiver uses a hybrid modulation scheme, where one bit of information is encoded in the molecule identity (DA vs. 5-HT) and another bit in the signal amplitude. The OECTs transduce the binding of neuromodulators into electrical currents. The paper develops a physics-based model that accounts for:

  • Molecular transport: Restricted diffusion and clearance in the organoid's extracellular space.
  • Aptamer binding kinetics: Langmuir binding dynamics.
  • OECT transduction: Converting binding events into drain currents.
  • Correlated electrical noise: Including thermal, flicker (1/f), and drift (1/f²) noise, with low-frequency components shared across channels.

Crucially, control referencing is implemented by subtracting the signal from the control channel from the selective channels before amplitude decisions are made. This is a key differentiator from prior work that often focused on single-axis receivers or lacked a matched control.

Datasets

The paper does not use pre-existing datasets in the traditional machine learning sense. Instead, it relies on custom Monte Carlo simulations to generate data for evaluating the receiver's performance. The simulations are parameterized using values from existing literature for physical properties (e.g., diffusion coefficients, aptamer affinities, OECT characteristics) and specific "scenario baselines" for parameters like clearance rates and noise correlations. No new benchmark datasets are introduced.

Key Insights and Ideas

  • The control channel is not a third sensor, but a nuisance estimator: Its primary role is to cancel out common-mode noise and drift, particularly benefiting the amplitude decoding part of the hybrid signal.
  • Hybrid detection is limited by amplitude, not identity: For the hybrid scheme, the accuracy of identifying the molecule type (DA vs. 5-HT) is generally high, but the accuracy of decoding the signal amplitude is the bottleneck. The control channel significantly improves this amplitude decoding.
  • Control referencing is context-dependent: It's most beneficial when low-frequency noise dominates and the signal is weak (e.g., at longer distances). At short distances or high signal strengths, it can even slightly degrade performance by adding uncorrelated noise.
  • Organoid-to-gate distance is critical: Performance degrades rapidly as the gap between the organoid and the receiver increases beyond ~45-60 µm.
  • Device co-design and timing are crucial: Optimizing OECT parameters (gain, capacitance) and symbol timing is essential for maximizing the receiver's operating envelope, especially when dealing with inter-symbol interference (ISI).

Key Results

  • Improved SER for Hybrid detection: At a representative distance of 45 µm, Hybrid detection with control referencing (Hybrid+CTRL) reduced the Symbol Error Rate (SER) from 3.71 × 10⁻² to 1.09 × 10⁻² at a molecule budget of 1.40 × 10⁴ molecules/symbol. This was primarily due to a 93.1% reduction in the amplitude component's error.
  • Lower Limit of Detection (LoD): Hybrid+CTRL achieved a lower LoD of 11866 molecules/symbol compared to CSK-4+CTRL (17177 molecules/symbol) at 45 µm, representing a 30.9% reduction. This advantage increased with distance.
  • Optimal symbol period for ISI: When inter-symbol interference (ISI) was considered, the optimal symbol period shifted from short durations (no ISI) to intermediate durations (around 10² seconds), suggesting a slow chemical-state readout application rather than a high-rate data link.
  • Wider feasible operating envelope: Control referencing expanded the range of usable OECT device parameters (gain and capacitance) for Hybrid detection, making the receiver more robust to variations.
  • Sensitivity to noise correlation: Control referencing was most beneficial when the pre-subtraction correlation between channels was high (ρ ≈ 0.9), indicating strong common-mode noise.

Limitations and Future Work

  • Idealized transport model: The paper uses an idealized point-source emitter and assumes equal spacing for all three channels. Future work could explore non-uniform spacing and more complex organoid-emitter geometries.
  • Negligible cross-reactivity: The model assumes minimal cross-reactivity between DA and 5-HT aptamers. Future work should incorporate more realistic cross-reactivity models.
  • Scenario-based performance: The absolute SER and LoD values are scenario-based forecasts, dependent on the chosen baseline parameters (e.g., clearance rate, noise correlations). The more transferable insight is the regime-dependent rule for when control referencing is beneficial.
  • No adaptive recalibration: The current model does not account for adaptive recalibration of thresholds or device mismatch over time.
  • Future directions: The authors suggest extending the framework to unequal channel spacing, control mismatch, richer cross-reactivity models, and collaborative experimental validation of tri-channel OECT front-ends. They also highlight the potential for this architecture to support multimodal signaling, combining species, amplitude, and timing information.

Summary updated

BibTeX
@misc{ni2026control,
  author = {Hongbin Ni and Özgür B. Akan},
  title = {{A Control-Referenced Tri-Channel OECT Receiver for Hybrid Molecular Communication Toward Brain Organoid Interfaces}},
  year = {2026},
  eprint = {2604.10798},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2604.10798},
}
Figure