.svg)
.svg)
EEG with Eye Tracking – Integration Guide
Three clocks, one 8-bit TTL word, and the photodiode calibration that decides whether your ERPs survive review. A guide to specifying the timing layer between a Tobii eye tracker and a g.tec amplifier, written from the setups we install.
What this guide covers
Your setup has three clocks. The stimulus PC, the eye tracker and the amplifier. Nothing disciplines them to each other, and a recording synchronised only on start and stop events drifts by many milliseconds inside the first few minutes.
The fix is one 8-bit TTL word sent to two destinations. Into the amplifier digital input, where the same converter that samples the EEG samples the trigger, and into the 8-bit TTL input port on the back of the Tobii Pro Spectrum. Both files then carry the same event codes, and aligning them offline becomes a fit on known events instead of an estimate.
What is in it
- The trigger architecture, with the g.TRIGbox distributing one isolated pulse to both devices
- Where Lab Streaming Layer is enough, and the constant per-device offset it does not remove
- A photodiode calibration you run once, and which two numbers belong in your methods section
- How to prove the alignment held, on every participant rather than the first
- A timing budget listing the error on each path and whether it is correctable
Read the full guide: eye tracking with EEG, and where the timing breaks
.webp)
Cognitive Neuroscience
Explore All ProductsBehavioral research
Explore All ProductsBrain-Computer Interface Research
Explore All ProductsWhat is the most accurate way to synchronize an eye tracker with EEG?
Send the same 8-bit TTL word to both. Into the amplifier digital input, where the same converter that samples the EEG samples the trigger, and into the eye tracker TTL port. Both files then carry the same event codes, and offline alignment is a fit on known events instead of an estimate.
Is Lab Streaming Layer accurate enough for ERP work?
Sub-millisecond in the developers' own tests, but that is relative precision. Each device adds a constant offset from buffering and transmission before its data reaches LSL, and over a wireless link that reaches tens of milliseconds. Measure it once, subtract it, and LSL is fine.
Do I need a photodiode if the stimulus software already sends triggers?
Yes, at least while you are validating. Software marks the moment the draw call returns, not the moment light leaves the monitor. On a 60 Hz screen the quantisation alone is 16.7 ms.
Which products make up this setup?
g.Nautilus Research or another g.tec amplifier, a g.TRIGbox for trigger distribution, a Tobii Pro Spectrum or Pro Fusion with Tobii Pro Lab, E-Prime or PsychoPy for presentation, and a Chronos response device where reaction times matter. NBT supplies and integrates all of these in Israel.
How do I prove the synchronization worked?
Import into EEGLAB with the EYE-EEG toolbox and run its synchronisation check. It cross-correlates horizontal gaze against the EOG channels, which record the same eye movements through separate hardware. A peak at zero lag and sub-millisecond mean error is the evidence a reviewer will accept.
Still have questions?
Request for Quote
Please fill in as much details as possible and we will take care of your request as soon as possible
Request for Quote
Please fill in as much details as possible and we will take care of your request as soon as possible