Invisible multi-luminaire 40 Hz flicker does not entrain human electroencephalography

Introduction

There has been renewed interest in steady-state visual-evoked potentials (SSVEP) and its auditory counterpart in the 40 Hz gamma frequency range. This is due to recent animal and human studies showing that prolonged exposure to 40 Hz flickering can reduce amyloid-β plaques,^1,2 enhance glymphatic cleansing, ³ improve circadian rhythm, ² repair hippocampal structures, ⁴ and improve cognitive functioning, ⁵ even in patients with Alzheimer's disease (AD; for a review, see Sahu and Tseng ⁶ ). As a result, clinical trials have been initiated to evaluate the efficacy of at-home 40 Hz stimulation in human AD cohorts, of which phase II is complete, ⁷ and phase III is ongoing. ⁸

Importantly, to improve the usability and adaptability of these lighting devices in patients’ everyday life, studies have explored the use of “invisible spectral flicker” (ISF), which alternates between two distinct spectra of light not limited to two distinct monochromatic colors, thereby reducing the perception of flicker”.^9–11 An example of ISF on the market is the EVY light by Optoceutics, where two similar-colored lights alternate at 40 Hz speed to mask the stroboscopic discomfort. With the same purpose, others explored the use of “multi-luminaire” setups, where a weak flickering light is paired with a more intense non-flickering light such that the 40 Hz flicker is imperceivably added on top of the non-flicker light.^12,13 One example on the market of such multi-luminaire setup is Delta Electronics’ M + BrainCare light, which pairs the 40 Hz LEDs with another set of non-flickering LEDs to mask the 40 Hz flicker.

Although the efficacy of ISF light has already received support from EEG studies,^9,11 multi-luminaire light has not. If multi-luminaire light stimulation can evoke a 40 Hz neural response with minimal discomfort, this advance may add further evidence to use of 40 Hz light stimulation with reduced perception of flicker for in-home, noninvasive, and non-pharmaceutical intervention for neurodegenerative diseases and dementia. However, it remains to be demonstrated that multi-luminaire light stimulation can evoke a 40 Hz neural response. Similarly, the claims from the original human case study paper ¹³ that multi-luminaire stimulation “reduces the perceivable flickering” have yet to be corroborated. For any new entrainment technique, it is essential to assess: (1) whether it can evoke neural oscillations in sensory cortices, and (2) whether the evoked signal propagates to other regions of interest, such as the frontal cortex. These questions can be answered by measuring the 40 Hz SSVEP both at occipital (for visual) and temporal (for auditory) cortices, as well as the frontal cortex. Indeed, this was done when transcranial alternating current stimulation claimed to be able to achieve frequency-specific entrainment effects, ¹⁴ when stroboscopic flicker gamma entrainment was first introduced, ¹⁵ and when ISF was introduced.^9,11 Here, we investigate the ability of the 40 Hz multi-luminaire light source to evoke 40 Hz neural responses as measured by EEG.

Methods

We recruited 10 participants (4 females, 6 males; age range: 18–37 years; mean age: 24.60 years) for this experiment, though SSVEP has been demonstrated to be robustly measurable at just single-subject level. ¹⁶ Inclusion criteria required participants to be between 18 and 60 years old with normal or corrected-to-normal vision. Exclusion criteria included a history of psychiatric or neurological conditions, use of prescription medication, and personal or family history of epilepsy. All participants provided informed consent prior to participation, and all procedures were approved by the Research Ethics Center of National Taiwan University (202312HM050). The human research was conducted in accordance with the Declaration of Helsinki.

Participants were seated comfortably in a dimly lit room and underwent three experimental conditions: (1) “multi-luminaire direct”: Directly viewing the multi-luminaire light (M + BrainCare Light, Delta Electronics, Taiwan), (2) “multi-luminaire indirect”: Staring at a black cross at the center of a white piece of paper that is illuminated by multi-luminaire light at, and (3) “stroboscopic control”: Directly viewing a stroboscopic light panel (13.2 cm in diameter) that flickers at 40 Hz (50% duty cycle) and whose color temperature matches the multi-luminaire light (4000 K). This serves as the positive control, as 40 Hz stroboscopic flicker is known to evoke a strong neural response. ¹⁷ Task-free EEG was recorded for 2 min under each condition for every participant, with the order of the three conditions counterbalanced across participants.

EEG data were collected using a BrainAmp system and a 32-channel Brain Products ActiCap electrode system (Brain Products GmbH, Gilching, Germany). Signals were digitized at a 1000 Hz sampling rate. Electrodes were also placed around the right eye and on the canthi of both eyes to monitor vertical and horizontal eye movements. Data preprocessing and analysis were performed using EEGLAB version 2024.0 and custom MATLAB (R2023b; MathWorks Inc., Portola Valley, CA, USA) scripts. Continuous EEG data were re-referenced offline to the average of electrodes at the left and right mastoids (M1 and M2). A digital bandpass filter (0.3–100 Hz) was applied, and ocular and muscle artifacts were removed using independent component analysis. The Fourier transform of 1000-ms segments from the 2-min visual stimulation epochs was computed, and the spectra were averaged.

The extent of entrainment induced by the three conditions—ISF-direct, ISF-indirect, and stroboscopic-control—was quantified using the signal-to-noise ratio (SNR). SNR measures the prominence of the 40 Hz amplitude response relative to the surrounding frequency neighborhood (baseline: 38–42 Hz, excluding 40 Hz) during the 2-min visual stimulation epochs. Topographic maps of 40 Hz SNR values were generated using EEGLAB's topoplot function. These maps represent the prominence of the 40 Hz amplitude response in comparison to the baseline frequencies. As the evoked response to visual stimulation is typically strongest in the occipital region (i.e., O1, Oz, and O2), ¹⁸ analyses focused on this region. Statistical analysis involved paired t-tests to compare the 40 Hz amplitude response to the surrounding frequency neighborhood (38–42 Hz, excluding 40 Hz) in the occipital region across the three conditions.

Results

We did not observe any significant 40 Hz EEG oscillations in any multi-luminaire conditions. This null finding was true for both multi-luminaire direct (t(9) = −1.310, p = 0.223, Cohen's d = −0.414) and multi-luminaire indirect (t(9) = 0.910, p = 0.387, Cohen's d = 0.288) condition. Considering that the multi-luminaire technology is designed as an indirect lighting device (e.g., reading lamp), the lack of evoked potentials, even in the most direct viewing condition, is concerning.

In contrast, we observed stronger 40 Hz EEG oscillations in response to the single-luminaire stroboscopic light (t(9) = 5.153, p < 0.001, Cohen's d = 1.630). In fact, all 10 out of 10 participants exhibited a 40 Hz entrainment effect in the occipital region in the stroboscopic-control condition. Propagation to the frontal region is also visible in Figure 1.

OPEN IN VIEWER

Figure 1.

Evoked EEG response to multi-luminaire direct, multi-luminaire indirect, and stroboscopic control conditions.

Discussion

Our results suggest that multi-luminaire light stimulation, where 40 Hz luminance flickering is paired with a brighter non-flickering light source, does not evoke a 40 Hz neural response detectable by human EEG. Specifically, no increase in 40 Hz power was observed in the occipital region, nor was propagation to the frontal region observed (Figure 1).

How can we reconcile these null results with two prior studies that reported positive outcomes using the same light? The first study investigated 40 Hz multi-luminaire light in an in-vitro setup, where SH-SY5Y cells were exposed to the light stimulation for durations of 15, 30, 45, and 60 min, and found a reduction in tau protein levels. ¹² However, the direct exposure of light to cell cultures is fundamentally different from light interacting with the human retina. For instance, Lee et al. demonstrated that red and white light produce stronger neural entrainment than green and blue light, a pattern that aligns with the distribution of photoreceptors on the retina. ¹⁵ These sensory limitations (whether retinal, auditory, or central nervous system-related) are not mechanisms in and of themselves but impose constraints on the practical application of 40 Hz stimulation in real-world contexts. This leaves significant uncertainty about how in-vitro findings translate to sensory stimulation in humans.

The second study by the same group involved a non-randomized design with AD patients and reported stable cognitive functioning, reduced psychiatric symptoms, and decreased caregiver burden in the multi-luminaire group. ¹³ However, this study presents several methodological issues. First, the non-randomized design meant participants were assigned to either the multi-luminaire or control group based on their daycare center, introducing numerous potential confounding variables. Second, at baseline, the multi-luminaire group already exhibited significantly worse cognitive function (e.g., Mini-Mental State Examination and Cognitive Abilities Screening Instrument scores) and higher caregiver burden compared to the control group, leaving more room for improvement in the multi-luminaire group and biasing the results. Third, there was no placebo or sham group; the control group simply engaged in normal daily activities without the use of a non-40 Hz light, while the multi-luminaire group participated in beneficial activities (e.g., singing, painting) under multi-luminaire lighting. Furthermore, key outcome measures such as neuropsychiatric symptom severity and caregiver burden were subjective and could easily have been influenced by participant or caregiver expectations. Without a non-flickering light control, it is difficult to disentangle these effects. Lastly, the study had multiple endpoints but did not correct for multiple comparisons, thereby increasing the likelihood of a false positive outcome. Based on our findings in this study, it seems unlikely that the positive outcomes described in Li et al. ¹³ were the result of evoking 40 Hz neural responses with multi-luminaire light stimulation. Future investigations into cognitive effects of 40 Hz multi-luminaire light stimulation should include EEG validation to confirm that a 40 Hz neural response can indeed be evoked.

In conclusion, our SSVEP results indicate that 40 Hz multi-luminaire light stimulation, specifically the Delta M + BrainCare light, does not evoke a 40 Hz neural response, neither in the occipital nor the frontal cortex. Therefore, it is highly unlikely that multi-luminaire light, with its current parameters and implementations, can be of use as an in-clinic or at-home intervention against AD progression. It is worth noting, however, that perhaps adjusting the relative intensity between the flickering and non-flicker lights may reach a balance between 40 Hz EEG response and imperceptibility. Indeed, similar efforts have been made using ISF ⁹ and heterochromatic flicker, ¹⁹ where ISF showed smaller effect of evoked 40 Hz response (3.03 dB) compared to stroboscopic stimulation (12.04 dB) but achieved better comfort. ⁹ Therefore, strategy can be adopted here for multi-luminaire lights, though this would entail fine-tuning the intensity ratio between flickering and non-flickering lights, which requires further research and EEG validation. As of now, we see no evidence of 40 Hz EEG response to multi-luminaire light. Future research should prioritize EEG validation when designing devices and parameters in 40 Hz sensory stimulation.