In wearable health monitoring and brain-computer interfaces, the transition from rigid, laboratory-grade equipment to flexible, ungelled sensors represents a massive leap in user comfort. However, removing conductive gels and moving out of shielded rooms introduces a severe physical challenge: signal degradation from environmental noise and motion artifacts.
Traditional sensing architectures rely on passive electrodes that capture weak biopotentials and send them along analog wires to a central processing unit. This approach is highly vulnerable to interference. To solve this, a modern paradigm shifts signal conditioning, amplification, and even digitization directly to the sensor site. By modifying the signal before it ever encounters a cable, active sensing fundamentally changes how wearable devices combat noise.
The Physics of Cable Noise and High Impedance
To understand why active sensing works, we must first look at why passive wearable sensors fail. When using dry-contact or ear-based sensors, the electrical impedance at the boundary between the skin and the electrode is exceptionally high.
When a microvolt-scale physiological signal is captured at a high-impedance node, it must travel through a wire to reach an external amplifier. As Simon L. Kappel et al. observe in Physiological artifacts in scalp EEG and ear-EEG, physical movements—such as head rotations or body motion—cause the cables to flex and deform. This movement alters the cable's internal capacitance and triggers triboelectric effects.
Because the source impedance at the electrode is so high, the tiny currents induced by these capacitive shifts cannot easily dissipate. Instead, they manifest as large voltage fluctuations that completely overwhelm the fragile physiological signal.
Transforming Impedance at the Electrode Site
The core intuition of active sensing is to neutralize this vulnerability by transforming the signal's impedance right at the point of contact. Instead of routing a high-impedance analog signal down a long wire, an active electrode embeds a miniature, low-noise buffer circuit—typically a voltage follower—directly onto the sensor headstage.
Passive Sensing:
[High-Impedance Sensor] ===(High-Impedance Analog Cable: Vulnerable to Noise)===> [Amplifier/ADC]
Active Sensing:
[High-Impedance Sensor -> Buffer Circuit] ===(Low-Impedance Analog Cable: Immune to Noise)===> [ADC]
This architecture changes the electrical dynamics of transmission:
- High Input Impedance: The buffer circuit features an exceptionally high input impedance that matches or exceeds the skin-electrode interface, allowing it to capture the weak biopotential without distorting or loading the signal.
- Low Output Impedance: The circuit outputs the exact same voltage signal, but at a very low impedance.
As Asma M. Naim et al. explain in Low-cost Active Dry-Contact Surface EMG Sensor for Bionic Arms, converting the signal to a low-impedance output before it travels through any leads dramatically mitigates its sensitivity to external noise. Because a low-impedance line can rapidly dissipate the stray charges induced by cable movement, the system achieves a high common-mode rejection ratio (CMRR).
This allows researchers like Roozbeh Jafari in A 16-channel bluetooth enabled wearable EEG platform with dry-contact electrodes for brain computer interface and Amir Aminifar et al. in EEG glasses for real-time brain electrical activity monitoring to deploy dry-contact sensors that maintain high signal-to-noise ratios (SNR) without requiring conductive gels or heavy shielding.
Localized Amplification and At-Source Digitization
Active sensing extends beyond impedance buffering; the entire front-end signal conditioning chain can be brought directly to the sensing site.
In Design and Verification of a Dry Sensor-Based Multi-Channel Digital Active Circuit for Human Brain Electroencephalography Signal Acquisition Systems, Chi‐Hsien Liu et al. demonstrate that amplifying and filtering EEG signals directly at each individual dry sensor site minimizes signal degradation. When a signal is magnified at the source, any environmental noise picked up along the downstream cable becomes negligible relative to the amplified signal.
To eliminate analog cable noise entirely, the signal can be digitized immediately at the recording interface.
Analog vs. Digital Headstages:
Analog: [Sensor] ---> [Analog Cable (Picks up EMI/Noise)] ---> [Central ADC]
Digital: [Sensor -> ADC] ===(Digital Cable (Immune to EMI/Noise))===> [Processor]
As Jeffrey M. Weiss et al. show in Demonstration of a portable intracortical brain-computer interface, minimizing the distance between the recording site and the point of digitization prevents electromagnetic interference from corrupting high-density neural recordings. Because digital signals represent data as discrete states rather than continuous voltages, small fluctuations induced by the environment cannot alter the underlying data stream. This enables high-fidelity performance outside strict laboratory environments.
In-Situ Processing and Intelligent Hardware
Once active circuitry is embedded at the sensor, the boundary between sensing and computing begins to blur. Instead of merely cleaning a signal for a central processor, the active sensor hardware can execute localized, intelligent operations to reduce system latency and data complexity.
The authors of Multi-sensing based target tracking by using decision-making strategy with spatial and temporal properties, Liu Xiu-ju et al., demonstrate this by integrating specialized physical circuitry and decision-making logic directly into a sensor head. This hardware-level conditioning eliminates common-mode signals and filters data before it ever reaches the primary processing layer, bypassing the latency typically introduced by purely algorithmic software filtering.
Taking this concept to its biological conclusion, Xiaosong Wu et al. introduce an "in-sensor" reservoir computing architecture in Wearable in-sensor reservoir computing using optoelectronic polymers with through-space charge-transport characteristics for multi-task learning. By performing contrast enhancement and noise reduction directly within the sensing material, the hardware mimics biological sensory organs like the retina. This localized pre-processing simplifies the data structure before it is passed to a classification readout, maximizing both processing speed and energy efficiency.
Key Foundational Work
A 16-channel bluetooth enabled wearable EEG platform with dry-contact electrodes for brain computer interface (2013) — contributed 2 concept elements to this direction.
Physiological artifacts in scalp EEG and ear-EEG (2017) — contributed 2 concept elements to this direction.