The ultimate dream of technophiles is to do away with keyboards, touchscreens, and other traditional input devices and replace them with a direct brain-computer interface (BCI). We have computers with blazing-fast processors and immense amounts of memory these days, but using a keyboard to interface with them is a huge bottleneck. It’s like having a massive cloud computing cluster that is only accessible via a 56K dial-up modem.
Unfortunately, any type of BCI we can use at home is somewhere between completely unusable and really, really bad in terms of functionality. Not many people are willing to shell out tens of thousands of dollars or have electrodes implanted into their brain for better performance. So, it would seem, we have to settle for junky hardware if we want to experiment with BCI technology.
The board can be used with a traditional cap (📷: Cerelog)
That may not be the case for much longer, however. A company called Cerelog is working to develop an affordable, open source BCI that offers a research-grade electroencephalogram (EEG) system. It is built around an ESP32-S3 microcontroller to make the device, called the Cerelog ESP-EEG, accessible and hackable. But it is also equipped with industry-standard components, like the ADS1299 ADC, to provide precision EEG measurements.
The ESP-EEG is an eight-channel biosensing board that uses an ESP32-S3 to handle data acquisition, processing, and wireless connectivity. Paired with it is Texas Instruments’ ADS1299, a 24-bit, low-noise analog-to-digital converter widely regarded as the gold standard for EEG front ends. This combination allows the board to capture microvolt-level signals generated by the brain with far greater fidelity than typical hobbyist solutions.
One of the most important aspects of EEG hardware design is dealing with electrical noise, and this is where Cerelog focused much of its engineering effort. The human body readily picks up 50 or 60 Hz interference from power lines, which can easily swamp the faint signals researchers are trying to observe. Rather than relying on simple passive grounding, the Cerelog ESP-EEG implements a true closed-loop active bias system. The board continuously measures common-mode noise on the body, inverts it, and feeds it back through a dedicated bias electrode. The result is a dramatically lower noise floor and cleaner data that can reveal subtle brain activity such as alpha waves.
Visualizing the EEG data (📷: Cerelog)
From a usability standpoint, the board is designed to work with standard EEG electrodes, touch-proof adapters, and conductive gel, and it can be mounted on either a 3D-printed headset or used with a commercial EEG cap. For safety and signal integrity, it is intended to be used only with a battery-powered laptop, avoiding the risks and interference associated with mains power.
The Cerelog ESP-EEG integrates with BrainFlow for developers who want to write their own applications in Python or other languages, while a forked version of the OpenBCI GUI allows users to visualize brainwaves in real time without writing any code. Together, the hardware and software form a flexible platform that makes serious BCI experimentation far more accessible than it has ever been before.
