Reducing noise in recordings is vital to EEG research.
Noise is any unwanted activity found in an EEG recording – be it muscle activity, line noise from electrical devices, or even brain activity that is unrelated to the current research task (Luck, 2014).
Understanding how to reduce noise is particularly important if the difference between the signal and the noise is small. That is, if the so-called signal-to-noise ratio is low.
Even if a researcher is studying an effect that is real, they may not find significance if their recording is too noisy. For scientists, this is critical:
high noise levels = no statistical significance = lower chances of publication
As such, researchers tend to take cumbersome measures to avoid this problem, and maximize their chances of finding significance. They might increase the number of trials they use, recruit more participants, or spend even more time setting up their EEG equipment.
There is another way, however: steady-state visually evoked potentials, or SSVEPs.
In this post, I will describe what SSVEPs are, how you can interpret them, and consider ways you can use SSVEPs to improve your SNR.
To be clear, in this article we are not aiming to reduce noise per se, but increase the signal’s visibility in relation to the noise.
What are SSVEPs?
A short introduction to SSVEPs was given in another one of our blog posts. Essentially, SSVEPs are brain signals that occur in response to a visual stimulus flickering at a fixed frequency.
Regan (1966) was one of the first neuroscientists to differentiate between a transient visually evoked potential (VEP), which is the brain’s response to a single visual stimulus, and a “steady-state” VEP, which is the brain’s response to periodically flashing stimuli.
This steady-state VEP forms a stable EEG wave whose frequency is the same as the frequency of the stimulus.
Once discovered, SSVEPs quickly piqued the community’s interest and were soon demonstrated in cats (Rager & Singer, 1998). These animal experiments, amongst others, confirmed that SSVEPs mainly originate in the primary visual cortex (V1) before spreading to other cortical areas.
The past decade has seen several reviews highlight the utility of SSVEPs. Especially in studying the processes that underlie rhythmic brain activity.
SSVEPs can be used in brain-computer interfaces (BCIs), cognitive and clinical neuroscience (Vialatte et al., 2010), as well as vision research (Norcia et al., 2015).
Here we consider how you can use SSVEPs as an alternative to cumbersome noise reduction techniques.
Two ways to interpret SSVEPs
There are two main ways to apply and interpret SSVEPs.
Treat SSVEPs as event-related potentials
The first option is to average over the SSVEPs, and consider them event-related potentials (ERP).
Every time a visual stimulus flickers, it elicits a characteristic waveform that is strongest in the visual cortex, where it is first processed.
We can analyze each stimulus response as if it were an ERP. These ERPs will then be time-locked to the flicker of the stimulus. In this way, every flicker can be treated as a single trial.
At a flickering frequency of 20 Hz, that’s 6000 trials in 5 minutes!
You can greatly improve your SNR by averaging across trials. Because the SSVEP is the only signal that is phase-locked to the stimulus, it emerges when you take an average across all trials.
This increases statistical power and reduces recording time.
Treat SSVEPs as waveforms
Alternatively, it is possible to look at an SSVEP in terms of its waveform. The frequency at which the stimulus flickers (and its harmonics) is the frequency at which the brain responds to the stimulus.
Scientists believe this happens because brain activity is entrained by the flickering stimulus (Notbohm et al., 2016).
During experiments, researchers tend to look out for certain neuronal frequencies; the ones they assume will activate during their experimental task.
If the SSVEP amplitudes at these frequencies are highest, the researcher can be confident that they have modulated the neural oscillations they were hoping for.
This is a great way to study neural oscillations. It is more effective than measuring EEG alone, and is not as invasive as, for example, transcranial alternating current stimulation.
To illustrate the benefits of SSVEPs, here are two studies we are currently performing that use SSVEPs.
Both studies are being carried out in collaboration with Dr. James Dowsett (LMU Munich, Experimental Psychology), an expert on mobile EEG and SSVEPs.
Two SSVEP use cases
Increasing SNR in EEG during movement
Recording EEG while a participant is moving can be tricky because of noisy movement artifacts. Fortunately, SSVEPs can mitigate this problem.
In this study, we elicit SSVEPs using LCD shutter glasses, which are more commonly used to view 3D television (Dowsett et al., 2020).
We control the shutter glasses using an Arduino board, which allows us to define the frequency and duration of the flickering.
Participants can then wear the shutter glasses, which elicit SSVEPs, as they move around and perceive their environment. This method enables us to study neural oscillations while participants walk to a target location, providing high flexibility and high SNR.
The goal of this study is to verify the signal quality of Mentalab Explore when stationary and ambulatory using both dry and gel electrodes.
If we had only recorded EEG, it would have been difficult to differentiate the brain signals from noise. However, because we used SSVEPs, which provide a characteristic EEG signature, we could objectively compare the data in each condition.
For example, we know what the wave form should look like, and it should be visible across trials.
Moreover, the frequency of the flickering shutter glasses (9 Hz) should be most prominent during visual stimulation.
Studying high-frequency oscillations
In 2016, Notbohm and colleagues showed that they could manipulate brain oscillations using a rhythmic, visual stimulus.
Other studies used this idea to identify higher EEG frequencies, which are otherwise difficult to capture if the brain is left passive – the lower frequencies tend to be too loud.
In particular, 40 Hz is related to many important cognitive processes, possibly even consciousness, but can be tricky to identify.
Several labs have used a simple LED set-up, which flickers at 40 Hz, to promote 40 Hz neural oscillations (Jones et al., 2019; Sahin & Figueiro, 2021).
We decided to take this idea further by studying 40 Hz oscillations during sleep.
If the 40 Hz frequency is correlated with consciousness (even at sleep; Voss et al., 2014), then it should be strongest during wakefulness, and may also differ between sleep stages.
We already know that SSVEPs can be elicited during sleep (Norton et al., 2017; Sharon & Nir, 2017). However, it is unclear if this set-up can target higher frequencies, and what that means during sleep.
In the current study, we are creating a fully mobile sleep mask, which uses flickering LEDs and Mentalab Explore. In this way, it can be taken to participants’ homes, preserving their natural sleep environments.
Because Mentalab Explore can sample at up to 1000 Hz, record multimodal data offline comfortably and portably, it was by far the best choice for this paradigm.
We are happy to discuss your ideas and provide guidance on how to use our application code. If you’d like to learn more, please contact us at contact@mentalab.com.
References
Dowsett, J., Dieterich, M., & Taylor, P. C. (2020). Mobile steady-state evoked potential recording: dissociable neural effects of real-world navigation and visual stimulation. Journal of Neuroscience Methods, 332, 108540. https://doi.org/10.1016/j.jneumeth.2019.108540
Jones, M., McDermott, B., Oliveira, B. L., O’Brien, A., Coogan, D., Lang, M., Moriarty, N., Dowd, E., Quinlan, L., Mc Ginley, B., Dunne, E., Newell, D., Porter, E., Elahi, M. A., O’ Halloran, M., & Shahzad, A. (2019). Gamma Band Light Stimulation in Human Case Studies: Groundwork for Potential Alzheimer’s Disease Treatment. Journal of Alzheimer’s disease : JAD, 70(1), 171–185. https://doi.org/10.3233/JAD-190299
Luck, S. J. (2014). An introduction to the event-related potential technique. MIT press.
Norcia, A. M., Appelbaum, L. G., Ales, J. M., Cottereau, B. R., & Rossion, B. (2015). The steady-state visual evoked potential in vision research: A review. Journal of vision, 15(6), 4. https://doi.org/10.1167/15.6.4
Norton, J. J., Umunna, S., & Bretl, T. (2017). The elicitation of steady-state visual evoked potentials during sleep. Psychophysiology, 54(4), 496–507. https://doi.org/10.1111/psyp.12807
Notbohm, A., Kurths, J., & Herrmann, C. S. (2016). Modification of Brain Oscillations via Rhythmic Light Stimulation Provides Evidence for Entrainment but Not for Superposition of Event-Related Responses. Frontiers in human neuroscience, 10, 10. https://doi.org/10.3389/fnhum.2016.00010
Rager, G., & Singer, W. (1998). The response of cat visual cortex to flicker stimuli of variable frequency. The European journal of neuroscience, 10(5), 1856–1877. https://doi.org/10.1046/j.1460-9568.1998.00197.x
Regan D. (1966). Some characteristics of average steady-state and transient responses evoked by modulated light. Electroencephalography and clinical neurophysiology, 20(3), 238–248. https://doi.org/10.1016/0013-4694(66)90088-5
Sahin, L., & Figueiro, M. G. (2020). Flickering Red-Light Stimulus for Promoting Coherent 40 Hz Neural Oscillation: A Feasibility Study. Journal of Alzheimer’s disease: JAD, 75(3), 911–921. https://doi.org/10.3233/JAD-200179
Sharon, O., & Nir, Y. (2018). Attenuated Fast Steady-State Visual Evoked Potentials During Human Sleep. Cerebral cortex (New York, N.Y. : 1991), 28(4), 1297–1311. https://doi.org/10.1093/cercor/bhx043
Vialatte, F. B., Maurice, M., Dauwels, J., & Cichocki, A. (2010). Steady-state visually evoked potentials: focus on essential paradigms and future perspectives. Progress in neurobiology, 90(4), 418–438. https://doi.org/10.1016/j.pneurobio.2009.11.005
Voss, U., Holzmann, R., Hobson, A., Paulus, W., Koppehele-Gossel, J., Klimke, A., & Nitsche, M. A. (2014). Induction of self awareness in dreams through frontal low current stimulation of gamma activity. Nature neuroscience, 17(6), 810–812. https://doi.org/10.1038/nn.3719
