如何选择脑电系统：面向科研、脑机接口与临床实验室的选购指南

How to Choose an EEG System: Buyer’s Guide for Research, BCI and Clinical Labs

Short answer: The right EEG system depends on four questions. Where will you record: in a shielded lab or in real-world settings? How many channels does your analysis need: 8 for most BCIs, 32 to 64 for ERP and cognitive studies, 128 to 256 or more for source localization? Which electrodes fit your participants and setup time: wet, dry or hybrid? And do you need real-time access or synchronization with TMS, fNIRS or eye tracking?

As a rule of thumb, choose a stationary high-channel amplifier such as g.HIamp for high-density mapping, invasive recordings and closed-loop stimulation; a wireless research headset such as g.Nautilus for mobile and multimodal studies; a certified medical device (g.HIamp or g.Nautilus PRO FLEXIBLE) if you work with patients; and a low-cost 8-channel Unicorn device for teaching, prototyping and app development.

Buying an EEG system is a decision your lab will live with for ten years or more. This guide shows how to choose an EEG system step by step. It explains how an EEG amplifier works, which specifications matter and which are marketing, how electrodes, caps and software affect your daily work, and how to plan a complete lab. It then profiles every system in the g.tec portfolio, compares them side by side and maps each one to the applications it is built for, including the cases where a system is not the right choice.

What an EEG System Consists Of

Knowing how to choose an EEG system starts with knowing what it consists of. An EEG system is more than an amplifier, so compare complete setups when you compare offers. A typical research setup has four building blocks.

1. The Amplifier

The biosignal amplifier is the heart of the system. It measures the voltage differences between electrodes on the scalp, which are only a few microvolts in size, amplifies them, filters them and converts them into digital data. Its specifications define the upper limit of what you can measure: how many channels, how fast, how precisely and with how much noise. Amplifiers come as stationary lab devices connected by USB or as small wireless units that are worn on the head.

2. Electrodes and Cap

Electrodes pick up the signal at the scalp; the cap holds them in standardized positions. Together they determine signal quality, preparation time, participant comfort and hygiene. A perfect amplifier cannot compensate for poor electrode contact, so electrodes deserve as much attention as the amplifier.

3. Software

Software records, displays and stores the data, presents stimuli, processes signals in real time and exports data for analysis. Many labs also need programming interfaces (APIs) to integrate the amplifier into their own code or existing toolchain.

4. Synchronization and Triggers

EEG is almost always analyzed relative to events: a stimulus on a screen, a button press, a TMS pulse or a movement. Digital trigger inputs, trigger boxes and network streaming protocols align these events with the EEG. When you combine several devices, precise synchronization becomes a requirement rather than a nice-to-have.

How an EEG Amplifier Works

You do not need to be an engineer to buy an EEG system, but understanding the signal chain helps you read data sheets critically, which is an essential part of knowing how to choose an EEG system.

1. Pick-up at the electrode: Each electrode measures the electrical potential at its position. Active electrodes contain a small pre-amplifier directly in the electrode, which makes the signal less sensitive to cable movement and electrical interference. Passive electrodes pass the raw signal through the cable to the amplifier.
2. Reference and ground: EEG is always a voltage difference. In a monopolar (referential) montage, every channel is measured against a common reference electrode; in a bipolar montage, channels are measured between pairs of electrodes. A ground electrode connects the participant to the amplifier’s reference potential and helps suppress interference that appears on all channels at once. Many g.tec amplifiers let you switch between monopolar and bipolar configurations in software.
3. Input stage: A high input impedance prevents the amplifier from loading the electrode-skin contact, which keeps the signal stable even when skin impedance is not perfect. g.HIamp and g.USBamp specify an input impedance of more than 1,000 GΩ.
4. Input range: The input range defines the largest voltage the amplifier can measure without clipping. A wide range matters because electrodes produce DC offsets and because events such as TMS pulses or movement create large artifacts. If the amplifier saturates, data are lost until it recovers. g.HIamp offers ±340 mV and g.USBamp ±250 mV, and g.Nautilus PRO FLEXIBLE lets you select ranges up to ±2.25 V.
5. Analog-to-digital conversion: The analog signal is digitized. All current g.tec amplifiers use 24-bit converters and internal oversampling, which means the signal is sampled much faster internally (for example 1.024 MHz per channel in g.Nautilus) and then digitally filtered down to the output sampling rate. This improves the signal-to-noise ratio and prevents aliasing.
6. Digital filtering and transfer: Band-pass and notch filters (for 50 or 60 Hz mains interference) can be applied before the data are sent over USB or radio to the recording computer.
7. Software: Recording, visualization, real-time processing and storage complete the chain.

Two further terms appear in data sheets. DC coupling means the amplifier does not remove very slow signal components in hardware, so slow cortical potentials and other near-DC activity can be recorded; g.Nautilus and the Unicorn devices are DC-coupled. Applied part type (BF or CF) and safety class describe electrical safety for medical devices: CF is the stricter classification required for applications with direct cardiac contact and is the category of g.HIamp, which also supports invasive ECoG recordings.

Five Questions to Answer Before You Buy

The fastest way to work out how to choose an EEG system for your lab is to answer these five questions before you request a quote.

1. What is your research question or application? ERP studies, BCI, source localization, sleep, epilepsy research and neurofeedback have very different requirements. Write down the analyses you plan to run, because they determine channel count, sampling rate and electrode type.
2. Who are your participants? Healthy adults, children, patients or animals? Patient work usually requires a certified medical device. Children and long sessions favor fast, comfortable electrodes and well-fitting caps in small sizes.
3. Where will you record? A shielded lab, a hospital ward, a flight simulator, a sports field or at home? Mobility requires a wireless, lightweight amplifier with enough battery life for your longest session.
4. Do you need real-time processing? BCIs, neurofeedback and closed-loop stimulation need low-latency raw data access, not just recording to file.
5. Which other signals will you combine with EEG? EMG, ECG, GSR, respiration, fNIRS, eye tracking or TMS each need compatible hardware and precise synchronization.

Add two practical questions: how many people will operate the system, and how many recordings per week you plan. A system used daily by several students needs fast preparation and robust, washable components more than a system used for occasional pilot studies.

How to Choose an EEG System: Key Technical Criteria

Channel Count

More channels are not automatically better. They increase spatial resolution, but also cost, preparation time and the risk of bad channels.

• 4 to 8 channels: most BCI paradigms (P300, SSVEP, motor imagery), neurofeedback, teaching and prototyping.
• 16 to 32 channels: standard clinical montages (for example the 10-20 system with 19 channels), ERP studies and mobile cognitive research.
• 64 channels: cognitive neuroscience, ICA-based artifact removal and basic source analysis.
• 128 to 256 channels and more: high-density source localization, functional brain mapping, high-gamma and ECoG. g.HIamp supports up to 256 channels per device and up to 1,024 synchronized channels with several devices. Read more in Why Ultra-High-Density EEG Is Changing Brain Mapping.

Plan for growth: a system that can be extended by stacking amplifiers or adding channels protects your investment when your research questions change.

Sampling Rate

The sampling rate must be at least twice the highest frequency you want to analyze (the Nyquist criterion), and in practice several times higher. For conventional EEG up to roughly 100 Hz, 250 to 500 Hz is sufficient. Higher rates matter for EMG, high-gamma activity, brainstem responses, TMS artifacts and invasive recordings. Stationary amplifiers such as g.HIamp and g.USBamp sample at up to 38.4 kHz per channel. Wireless headsets typically use 250 to 500 Hz to save radio bandwidth and battery.

Resolution and Input Noise

All current g.tec systems use 24-bit ADCs, so resolution is rarely the limiting factor. More informative is the input noise, which determines how small a signal you can detect. Compare noise in µV RMS over the same frequency band, typically 1 to 30 Hz: g.USBamp < 0.4 µV RMS, g.HIamp < 0.5 µV RMS and g.Nautilus < 0.6 µV RMS. For ERP components of a few microvolts, these values leave ample headroom.

Input Range and Artifact Recovery

Check the input range if you plan EEG-TMS, electrical stimulation or recordings with movement. A narrow input range saturates during large artifacts and loses data. Ask how quickly the amplifier recovers after a TMS pulse and whether it supports a hold or blanking input; g.HIamp, for example, has a HOLD input for artifact suppression.

Stationary or Wireless EEG

Stationary amplifiers offer the highest channel counts and sampling rates and the most inputs and outputs for triggers and stimulation. Wireless headsets let participants move freely, which is essential for real-world, sports, VR, aviation and pediatric studies, and they shorten setup. When you compare wireless systems, check:

• Range: g.Nautilus transmits in the 2.4 GHz band with up to 10 m indoor range; Unicorn devices use Bluetooth.
• Battery life: g.Nautilus runs more than 10 hours with up to 32 channels and more than 6 hours with 64 channels.
• Weight and size: g.Nautilus weighs under 110 g (8 to 32 channels) or under 140 g (64 channels); the Unicorn BCI Core-8 weighs 23 g.
• Robustness: g.Nautilus is waterproof and charges contactlessly via Qi.
• Synchronization: how triggers reach the wireless system, for example through 8 digital trigger inputs on the g.Nautilus base station.

Real-Time Access and Latency

If you build BCIs or closed-loop experiments, ask for measured latencies, not just “real-time capable”. g.HIamp provides raw data block access in < 2 ms with g.HIsys, EEG-to-TMS trigger latencies of < 4 ms in selected closed-loop configurations and EEG-to-stimulation command latencies of < 5 ms with g.Estim PRO or g.Estim FES. Also check the APIs (Python, MATLAB and Simulink, C, C#, .NET) and support for Lab Streaming Layer (LSL). The details are in Real-Time EEG System and Closed-Loop BCI.

Synchronization With Stimuli and Other Devices

Timing errors between stimulus and EEG blur event-related potentials. Hardware triggers recorded synchronously with the EEG are the most precise solution: g.HIamp has 2 × 8 digital trigger inputs and g.USBamp 8 trigger inputs plus 4 digital outputs. When several g.HIamp amplifiers are combined, they are synchronized in hardware with less than 1 µs jitter. For multimodal setups without shared hardware, LSL synchronizes data streams over the network.

Open-Source Tools and File Formats

Your data should flow into the tools your lab already uses, such as MNE-Python, EEGLAB, FieldTrip, BCI2000 and OpenViBE, and into standard formats such as EDF, GDF and BIDS. See Raw EEG Data Access with g.tec for every streaming and export route and EEG Interoperability for tool-by-tool instructions.

Certification: Research or Medical Device

Research devices are suitable for studies with healthy volunteers under ethics approval. For clinical use or patient studies, many institutions require a CE-certified medical device and, in the US, FDA clearance. g.HIamp and g.Nautilus PRO FLEXIBLE are CE-certified and FDA-cleared medical devices that comply with the IEC 60601 family of safety standards. Clarify the requirements with your ethics committee before you buy, because changing the device later can delay a study by months.

Support, Training and Service

An EEG system is only as good as the people operating it. Ask about onboarding, workshops, application support, repair times and software updates. g.tec runs regular workshops, the BCI & Neurotechnology Spring School and local support through representatives in many countries.

The g.tec EEG Portfolio in Detail

The following profiles summarize what each system is designed for, its key specifications and when another system is the better choice.

g.HIamp: High-Density and Multimodal Lab Amplifier

g.HIamp is g.tec’s flagship stationary amplifier for high-channel and high-bandwidth recordings. It is available with 16, 32, 80, 144 or 256 channels per device, and several devices can be synchronized for up to 1,024 channels. It supports active and passive electrodes as well as ECoG grids and strips and is used for high-density EEG, functional brain mapping, EEG-TMS and closed-loop stimulation.

• Sampling rate up to 38.4 kHz per channel, 24-bit
• Noise < 0.5 µV RMS (1 to 30 Hz), input range ±340 mV, input impedance > 1,000 GΩ
• 2 × 8 digital trigger inputs, HOLD input, multi-device sync with < 1 µs jitter
• Verified TMS integrations with MagVenture, Magstim and Nexstim
• PRO version: CE-certified and FDA-cleared medical device, applied part CF; RESEARCH version available

Choose it if you need more than 64 channels, invasive recordings, high sampling rates or the tightest synchronization with stimulation. Consider alternatives if you mainly record 8 to 32 channels or need mobility.

g.USBamp RESEARCH: Versatile 16-Channel Amplifier

g.USBamp RESEARCH is a compact 16-channel amplifier with one dedicated ADC per channel. Several units can be stacked and synchronized for higher channel counts. It records EEG together with EMG, ECG, EOG and physiological sensors and is widely used in BCI research labs.

• 16 monopolar or 8 bipolar channels per unit, stackable
• Sampling rate up to 38.4 kHz per channel, 24-bit
• Noise < 0.4 µV RMS (1 to 30 Hz), input range ±250 mV
• 8 digital trigger inputs, 4 digital outputs, 2 analog outputs, built-in calibration and impedance check
• Research device

Choose it if you build BCIs, run evoked-potential or stimulation studies and want a flexible, scalable lab amplifier. Consider alternatives for more than 64 channels (g.HIamp) or mobile studies (g.Nautilus).

g.Nautilus RESEARCH: Wireless EEG for Real-World Studies

g.Nautilus RESEARCH is a wireless EEG headset with 8, 16, 32 or 64 active channels. The amplifier sits on the cap, so participants can walk, cycle, fly a simulator or play in VR.

• 500 Hz sampling (250 Hz with 64 channels), 24-bit
• Noise < 0.6 µV RMS (1 to 30 Hz)
• g.SAHARA hybrid (dry or gel) or g.SCARABEO wet active electrodes
• Battery > 10 h (8 to 32 channels), waterproof, Qi charging, < 140 g

Choose it if movement, mobility or fast setup matters. Consider alternatives for high-frequency analyses above about 100 Hz or for patient studies requiring a medical device.

g.Nautilus MULTI-PURPOSE: Brain and Body Monitoring

g.Nautilus MULTI-PURPOSE adds 4 detachable biosignal channels to 8 to 64 EEG channels, and channels can be allocated to physiological signals as needed. It is designed for psychophysiology, human factors, aviation and sports science, where EEG is combined with ECG, EMG, GSR or respiration. A real-world example is described in Multimodal EEG for Flight Simulation Research.

g.Nautilus PRO FLEXIBLE: Wireless Medical EEG

g.Nautilus PRO FLEXIBLE is the CE-certified and FDA-cleared version for clinical research. It offers 8, 16, 19 or 32 channels with removable electrode strands that can be cleaned separately, an integrated 3-axis accelerometer and selectable input ranges up to ±2.25 V. It is used in neurorehabilitation, disorders of consciousness, epilepsy research, EEG-TMS and neurofeedback. Our article The Ultimate Guide to the Best EEG Device of 2026 looks at it in detail.

Multimodal Bundles: EEG With fNIRS and Eye Tracking

For combined recordings, g.tec offers pre-integrated systems with synchronized hardware: g.HIamp NIRx, g.Nautilus NIRx and g.Nautilus fNIRS for EEG-fNIRS, and g.HIamp Tobii Pro Fusion and g.Nautilus Tobii Pro Glasses 3 for EEG with eye tracking. The g.GAMMAcap accepts fNIRS modules alongside EEG electrodes, so both modalities share one cap.

Unicorn BCI Core-8: Miniature 8-Channel Amplifier

Unicorn BCI Core-8 is a 23 g amplifier with 8 DC-coupled channels at 24 bit and 250 Hz, with a 1,000 Hz option for ECG and EMG. It attaches magnetically to caps and base stations, transmits via Bluetooth 5 and runs more than 6 hours. With g.SAHARA hybrid or gold cup electrodes it is used for BCI prototyping, sleep recordings, VR and animal research, from dogs and horses to cats; see Animal EEG Through Fur.

Unicorn Hybrid Black: BCI Headset for Developers and Teaching

Unicorn Hybrid Black is an 8-channel Bluetooth headset with hybrid dry or wet electrodes, a 3-axis accelerometer and gyroscope, 56 g weight and more than 3 hours of battery life. It comes with the Unicorn Suite and APIs for Python, C, .NET, Android, Unity and MATLAB Simulink, with no software subscription. It is the entry point for BCI development, neurofeedback, games, art and teaching.

Unicorn BCI Core-4 Headband: Quick-Start Dry EEG

Unicorn BCI Core-4 Headband records 4 channels over parietal regions with dry electrodes and no gel. It is designed for quick BCI and neurofeedback applications, interactive installations and games.

g.tec EEG Systems Compared

The table below puts the criteria into practice and shows how to choose an EEG system from the g.tec portfolio at a glance.

*List prices in the g.tec online shop at the time of writing, excluding electrodes, caps and software unless stated. Prices vary by region and configuration. Use the configurator or contact your local g.tec representative for a quote.

Technical Specifications Side by Side

n/a: not specified in the public product data; ask g.tec for details.

Choosing EEG Electrodes and Caps

Electrodes are often underestimated when labs work out how to choose an EEG system, yet they have a larger influence on daily work than most amplifier specifications. Two decisions matter: active or passive, and wet, dry or hybrid.

Active or Passive Electrodes

Active electrodes amplify the signal directly at the scalp. This reduces cable artifacts and interference and tolerates higher skin impedance, so less skin preparation is needed. Passive electrodes are simpler, cheaper and suitable for shielded labs and short cables, but need lower impedance and more careful preparation. For mobile, multimodal and high-density recordings, active electrodes are the standard choice.

Wet, Dry or Hybrid Electrodes

Wet electrodes use conductive gel and still provide the lowest impedance and most stable signals, especially for long recordings and small ERP components. Dry electrodes contact the scalp through pins, need no gel and are ready in minutes, but are more sensitive to movement. Hybrid electrodes such as g.SAHARA can be used either way, and gel can be added only at selected positions.

EEG Caps and Sizes

The g.GAMMAcap holds electrodes in standardized 10-10 positions with 74 labeled and 86 intermediate positions, so you can build custom high-density montages. It is compatible with all g.tec EEG electrodes, is washable, and electrodes can stay mounted during cleaning, which shortens preparation. Order several sizes, because a poorly fitting cap is a common source of bad channels:

For TMS-EEG specifically, see Optimized TMS-EEG Recordings with 3 mm Ultra-Flat Electrodes.

Software: Recording, Real-Time Processing and Analysis

Software is the part of how to choose an EEG system that labs most often decide too late. Compare it as carefully as hardware, including licensing, operating system and how open the data formats are.

• g.tec Suite 2024: the base package with the g.NEEDaccess server and APIs for Python, MATLAB, C, .NET and network access, a montage creator, data conversion from HDF5 to EDF and GDF, and stimulation interfaces. It is discounted when bought together with an amplifier.
• g.Recorder: acquisition, visualization and storage in HDF5 with impedance check. The Professional version adds recording review, evoked-potential views, synchronized video and paradigm presentation via UDP.
• g.HIsys: real-time processing in MATLAB and Simulink with up to 38.4 kHz per channel and visualization of up to 1,024 channels. g.HIsys Professional adds BCI paradigms (P300, SSVEP, motor imagery, slow cortical potentials), stimulation and closed-loop tools, LSL and UDP interfaces and OSCAR LIVE for real-time artifact removal.
• g.BSanalyze: offline analysis with preprocessing, feature extraction, classification, statistics and batch processing for EEG, ECoG, ECG and more; the Professional version includes OSCAR PRO artifact correction.
• g.Pype: a Python SDK, installed via pip, for real-time pipelines with LSL and UDP output and compatibility with NumPy, SciPy, scikit-learn, PyTorch and MNE-Python.
• Unicorn Suite: recorder and interfaces for Unicorn devices, with lifetime software access and no subscription.

If you prefer open-source tools, every g.tec system can stream to LSL or export to EDF/GDF for MNE-Python, EEGLAB and FieldTrip. Developers find a complete workflow in The Complete Developer Guide to Brain-Computer Interface Development.

Which EEG System for Which Application

If you already know your research field, the quickest answer to how to choose an EEG system is to start from your application.

ERP and Cognitive Neuroscience

Event-related potentials are small and averaged over many trials, so low noise, precise triggers and stable electrode contact matter most. 32 to 64 channels with wet active electrodes are the standard. A stationary g.USBamp setup or a g.Nautilus RESEARCH with g.SCARABEO electrodes both work; the wireless option is attractive if participants should sit naturally or move between tasks.

Brain-Computer Interfaces

BCIs need real-time data, low latency and convenient APIs more than many channels. A P300 speller works with 8 channels; a large study with 81 participants found that the great majority could use a P300 speller after a short calibration (Guger et al., 2009). For research, g.USBamp and g.Nautilus with g.HIsys or g.Pype are typical; for prototyping and products, the Unicorn devices offer the lowest entry cost.

EEG-TMS and Brain Stimulation

TMS pulses create artifacts thousands of times larger than EEG. You need a wide input range, fast recovery, flat electrodes that fit under the coil and precise trigger timing. g.HIamp with g.LADYbird TMS electrodes is the reference setup; g.Nautilus PRO FLEXIBLE supports EEG-TMS in wireless setups.

Multimodal Research

When EEG is combined with fNIRS, eye tracking, motion capture or physiological signals, synchronization is the main challenge. Pre-integrated bundles solve it in hardware; LSL solves it in software when devices come from different vendors. Read Brain and Body Monitoring for typical setups.

Clinical Research and Neurorehabilitation

Patient studies require medical devices, fast and hygienic preparation and robust workflows for staff who are not EEG specialists. g.Nautilus PRO FLEXIBLE was designed for this, with removable strands and selectable input ranges. For specific clinical applications, g.tec offers complete systems: recoveriX for stroke rehabilitation and mindBEAGLE for assessing and communicating with patients with disorders of consciousness.

Teaching and Product Development

For classes, hackathons and start-ups, cost per seat and ease of use count most. An 8-channel Unicorn Hybrid Black lets every student team record their own EEG and build a working BCI within a day.

Example Lab Setups

These examples show how the criteria combine in practice. Your configuration will differ, but they are a useful starting point for a quote.

• Cognitive ERP lab: g.USBamp RESEARCH stacked to 32 or 64 channels, g.GAMMAcap with g.SCARABEO electrodes in two or three sizes, g.Recorder and a trigger interface to the stimulus PC.
• Mobile and VR lab: g.Nautilus RESEARCH 32 with g.SAHARA hybrid electrodes, LSL for synchronization with the VR engine and eye tracker.
• Human factors and aviation lab: g.Nautilus MULTI-PURPOSE with EEG plus ECG, EMG and GSR, combined with eye tracking.
• TMS-EEG lab: g.HIamp with 64 g.LADYbird TMS electrodes, g.HIsys for closed-loop triggering and an integration with the lab’s TMS device.
• High-density mapping lab: g.HIamp with 128 or 256 channels, g.PANGOLIN grids or g.SCARABEO electrodes and g.BSanalyze for offline analysis.
• Clinical research unit: g.Nautilus PRO FLEXIBLE with g.SCARABEO and g.SAHARA strands for different patient groups.
• Teaching lab or start-up: several Unicorn Hybrid Black headsets or Unicorn BCI Core-8 amplifiers with g.Pype and Unity interfaces.

Planning Your EEG Lab

Knowing how to choose an EEG system is only half the work: the best amplifier cannot compensate for a poorly planned lab. Check these points before the system arrives:

• Room and electrical environment: keep distance from large electrical equipment, use grounded power and avoid cables running parallel to mains lines. Active electrodes reduce, but do not eliminate, interference.
• Stimulus presentation: decide how triggers get from the stimulus PC to the amplifier, and measure the real delay between a visual or auditory stimulus and its trigger, for example with a photodiode or microphone.
• Data management: plan file naming, storage, backups and conversion to BIDS from the first recording.
• Ethics and data protection: clarify device requirements, consent forms and pseudonymization early.
• Hygiene: washable caps, removable electrodes and enough caps per size for back-to-back sessions.
• Training: budget time for onboarding, and write a standard operating procedure for cap placement and impedance checks.

Common Mistakes When Buying an EEG System

1. Buying channels instead of answers: 256 channels do not help a P300 BCI, and 8 channels cannot answer a source-localization question. Start from the analysis.
2. Comparing amplifiers only: electrodes, caps, software and triggers often make the difference in daily work and cost.
3. Ignoring the participant: long gel preparation is a problem for children, patients and studies with many sessions.
4. Underestimating synchronization: unknown trigger delays add jitter that cannot be corrected afterwards.
5. Locking data into closed formats: make sure you can export to open formats and stream to LSL.
6. Choosing a research device for clinical work: check certification requirements before you buy.
7. Forgetting the people: training, support and documentation determine how quickly you get publishable data.

When a System Is Not the Right Choice

• g.HIamp is more than you need if you only record 8 to 32 channels for BCI or ERP work. A g.USBamp or g.Nautilus costs less and sets up faster. g.HIamp is also stationary and not designed for movement studies.
• g.USBamp needs stacking beyond 16 channels. For 128 channels or more, a single g.HIamp is the cleaner setup.
• g.Nautilus samples at up to 500 Hz. That is ideal for EEG, but not suited to high-gamma, ECoG or other high-frequency applications; use a stationary amplifier for those.
• g.Nautilus RESEARCH and MULTI-PURPOSE are research devices. For patient studies that require a certified medical device, choose g.Nautilus PRO FLEXIBLE or g.HIamp.
• Unicorn devices are excellent for development and teaching. With 4 to 8 channels at 250 Hz they do not replace a research-grade lab system for ERP or source analysis studies.

Budget and Total Cost of Ownership

Price is part of how to choose an EEG system, but the amplifier is only part of the investment. Plan for:

• Electrodes and caps in several sizes. Gel and wet electrodes are consumables.
• Software licenses for recording, real-time processing and analysis. Check whether licenses are perpetual or subscription-based and whether MATLAB is required.
• Synchronization hardware such as g.TRIGbox or g.STIMbox for stimulus presentation.
• Training and onboarding for your team, plus support and warranty.
• Scalability: can you add channels or modalities later without replacing the system?

Typical entry points: an 8-channel BCI development setup starts at around €1,200 (Unicorn Hybrid Black), a wireless research system at around €7,600 (g.Nautilus RESEARCH, 8 channels), and high-density stationary systems at around €40,000.

EEG Glossary

• Active electrode: electrode with an integrated pre-amplifier that reduces cable and interference artifacts.
• ADC resolution: number of bits used to digitize the signal; 24 bit is standard for research EEG.
• Bipolar montage: channels measured between pairs of electrodes instead of against a common reference.
• DC coupling: recording without a hardware high-pass filter, so very slow potentials are preserved.
• ECoG: electrocorticography, recordings from electrodes placed directly on the brain surface.
• ERP: event-related potential, the averaged brain response to a stimulus or event.
• Impedance: resistance of the electrode-skin contact; lower is better, especially for passive electrodes.
• Input range: the largest voltage the amplifier can record without clipping.
• LSL: Lab Streaming Layer, an open protocol for synchronized streaming of data from several devices.
• Monopolar montage: every channel measured against a common reference electrode.
• Nyquist criterion: the sampling rate must be at least twice the highest frequency of interest.
• SSVEP, P300, motor imagery: the three most common BCI paradigms, based on visual flicker responses, an attention-related potential and imagined movements.
• 10-20 / 10-10 system: international standards for electrode positions on the scalp.

References

• Jasper, H. H. (1958). The ten-twenty electrode system of the International Federation. Electroencephalography and Clinical Neurophysiology, 10, 371–375.
• Oostenveld, R., & Praamstra, P. (2001). The five percent electrode system for high-resolution EEG and ERP measurements. Clinical Neurophysiology, 112(4), 713–719.
• Wolpaw, J. R., Birbaumer, N., McFarland, D. J., Pfurtscheller, G., & Vaughan, T. M. (2002). Brain–computer interfaces for communication and control. Clinical Neurophysiology, 113(6), 767–791.
• Delorme, A., & Makeig, S. (2004). EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. Journal of Neuroscience Methods, 134(1), 9–21.
• Guger, C., Daban, S., Sellers, E., Holzner, C., Krausz, G., Carabalona, R., Gramatica, F., & Edlinger, G. (2009). How many people are able to control a P300-based brain–computer interface (BCI)? Neuroscience Letters, 462(1), 94–98.
• Gramfort, A., et al. (2013). MEG and EEG data analysis with MNE-Python. Frontiers in Neuroscience, 7, 267.
• Pernet, C. R., et al. (2019). EEG-BIDS, an extension to the brain imaging data structure for electroencephalography. Scientific Data, 6, 103.

More publications that used g.tec systems are listed in the Recommended Citation List.

TALK TO AN EEG EXPERT

How to Choose an EEG System: Frequently Asked Questions

How do I choose an EEG system for my lab?

Start from your research question and participants, then decide on channel count, stationary or wireless recording, electrode type, real-time requirements and certification. Compare complete setups including electrodes, caps, software and triggers, and plan for training and future expansion.

How many EEG channels do I need?

For most BCI paradigms (P300, SSVEP, motor imagery) and neurofeedback, 8 channels are sufficient. ERP and cognitive studies typically use 32 to 64 channels. Source localization and high-density brain mapping require 128 to 256 channels or more.

Are dry EEG electrodes as good as wet electrodes?

Wet electrodes still provide the lowest impedance and the most stable signals, especially for long recordings and small ERP components. Active dry and hybrid electrodes such as g.SAHARA deliver good signal quality for BCI, mobile EEG and short sessions and save preparation time. Hybrid electrodes can be used dry or with gel.

What sampling rate do I need for EEG?

For standard EEG frequencies up to about 100 Hz, 250 to 500 Hz is sufficient. Sampling rates of several kHz are needed for EMG, high-gamma activity, brainstem responses and invasive recordings such as ECoG.

What is the difference between active and passive electrodes?

Active electrodes have a pre-amplifier in the electrode, which makes the signal less sensitive to cable movement and environmental noise and tolerates higher skin impedance. Passive electrodes are simpler and cheaper but need lower impedance and more careful preparation.

Should I buy a wireless or a stationary EEG system?

Choose a wireless system such as g.Nautilus if participants need to move, if you record outside the lab or if fast setup matters. Choose a stationary amplifier such as g.HIamp or g.USBamp for more than 64 channels, sampling rates above a few hundred hertz, invasive recordings or the tightest synchronization with stimulation.

Do I need a medically certified EEG device?

If you record from patients or use EEG clinically, you usually need a CE-certified medical device and, in the US, FDA clearance. For research with healthy participants, a research device is generally sufficient; check the requirements of your ethics committee. g.HIamp and g.Nautilus PRO FLEXIBLE are CE-certified and FDA-cleared.

Can I use g.tec systems with MNE-Python, EEGLAB or Lab Streaming Layer?

Yes. g.tec amplifiers provide APIs for Python, MATLAB and Simulink, C, C# and .NET, support Lab Streaming Layer and work with MNE-Python, EEGLAB, FieldTrip, BCI2000 and OpenViBE.

Which EEG system is best for brain-computer interface research?

For BCI research, real-time raw data access and low latency matter more than a very high channel count. g.USBamp and g.Nautilus are research systems for P300, SSVEP and motor imagery BCIs. For prototyping, teaching and app development, the 8-channel Unicorn Hybrid Black or Unicorn BCI Core-8 is a low-cost entry point.

Which EEG system is best for TMS-EEG?

TMS-EEG needs a wide input range, fast artifact recovery, flat electrodes that fit under the coil and precise trigger timing. g.HIamp with g.LADYbird TMS-EEG electrodes is designed for this and offers EEG-to-TMS trigger latencies below 4 ms in selected closed-loop configurations.

Can I record EEG in children?

Yes. g.GAMMAcap is available in three kids sizes from 32 to 48 cm head circumference. Wireless systems and fast electrodes such as g.SAHARA reduce preparation time and make recordings more comfortable for children.

How much does a research EEG system cost?

List prices range from around €1,200 for an 8-channel BCI development headset to around €7,600 for an entry-level wireless research system and around €40,000 or more for high-density stationary amplifiers. Electrodes, caps, software and synchronization hardware come on top.