Why do raw EEG signals show high values?

When working with raw EEG data, it is common to observe signal values that appear higher than expected. This behavior is normal and does not indicate a problem with the device or the recording.

Hardware filtering and the DC offset 

The point is that all EEG amplifiers must apply an upper cut-off frequency to prevent aliasing. This is a mandatory requirement for all signal acquisition systems and applies to all EEG devices on the market. In Bitbrain EEG systems, the upper cut-off frequency is set to either 40 Hz or 70 Hz, depending on the model.

Many EEG manufacturers also apply a lower cut‑off frequency at the hardware level, typically between 0.1 and 1 Hz. This means the signal is already high‑pass filtered before the user accesses it and does not include the DC offset or very‑low‑frequency activity. Bitbrain systems take a different approach by delivering the raw EEG signal from DC (0 Hz) up to the hardware upper cut‑off frequency, also referred to as a full‑band EEG signal. This provides researchers with full control over preprocessing and enables analysis of very‑low‑frequency activity below 1 Hz. As a consequence, the raw signal preserves the DC offset, which can result in higher‑than‑expected values when visualizing unfiltered data.

The EEG signal is transmitted raw via LSL and stored raw in CSV format, preserving the full‑band signal and the DC component. When exporting data to EDF format, the signal is stored with a 0.1 Hz high‑pass filter applied, meaning the DC component is removed at that stage. More details about the different export formats can be found in this article: Export format differences.

How should raw EEG data be interpreted? 

EEG signals are measured in microvolts, and when working with raw, full‑band data, it is standard practice to apply data processing before interpretation. In most EEG analysis pipelines, the first step is to apply a high‑pass filter, commonly around 1 Hz, to remove the DC component and slow drifts. Once this filter is applied, the EEG signal appears within the expected physiological ranges and can be analyzed as usual.

Summary 

Bitbrain systems provide raw EEG data by transmitting and storing the full‑band signal, including the DC offset. This design choice offers greater flexibility for research, particularly for low‑frequency analysis, but may result in higher than expected values when viewing unfiltered data. Applying a standard high‑pass filter (around 1 Hz) before analysis is sufficient to remove the DC component and interpret the EEG signal correctly.