These authors contributed equally to this work.
We analyzed theta-band phase synchrony (TBPS) under reduced and ordinary flicker lighting to determine the effect of light flickers on neurocognitive processes.
Nineteen healthy participants (mean age, 30.4±4.5 years; male, 63.2%) performed the Sternberg working memory tasks with event-related potential recording under reduced and control flicker conditions, respectively. We measured the P300 amplitude during memory retrieval, and for TBPS analysis, we calculated the weighted phase lag index within the P300 time window. Furthermore, we used standardized low-resolution brain electromagnetic tomography (sLORETA) to determine differences in functional cortical source connectivity between the two flicker conditions.
The hit rate (F1,18=0.862,
Reduced flicker lighting enhanced TBPS during the working memory task compared with control flicker lighting. Reduced flicker light may improve cognitive functioning by facilitating information transfer within the brain network. Flicker conditions should be considered when optimizing lighting, especially in environments demanding high-level cognitive performance.
Flicker, also known as temporal light modulation, refers to a rapid variation in the brightness of light over time [
Light sources based on light-emitting diodes (LEDs) are becoming increasingly common for general lighting because of their high energy efficiency, long operational lifetime, and high durability compared with conventional light sources [
Cognitive functions are largely dependent on appropriate large-scale integration between different brain regions [
In this study, we used electroencephalography (EEG) to measure the neural activities of healthy participants while performing the Sternberg working memory task. Working memory refers to a cognitive system that temporarily stores and manipulates information to perform complex cognitive tasks [
We prospectively enrolled 30 healthy participants (mean age, 30.4±4.4 years; male, 66.7%) via advertisements from April to June 2016. The exclusion criteria were as follows: 1) history of neurological and psychiatric disorders, traumatic brain injury, and previous brain surgery and 2) use of medications affecting the central nervous system, including sleeping pills and antidepressants, within the last 2 weeks. Following eligibility screening, all participants provided written informed consent before enrollment in the study. Participants were then subjected to the Sternberg working memory tasks with ERP recordings under both reduced and control flicker lighting conditions. Among the 30 participants who completed the experiments, 11 were excluded from the analysis because their averaged ERP waveforms did not show typical ERP-like patterns, including the positive peak corresponding to the P300 component. Finally, the data from 19 participants (mean age, 30.4±4.5 years; male, 63.2%) were analyzed. This study was approved by the Institutional Review Board of Seoul National University Hospital (H-1601-043-733) and was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines.
We used two different light flicker conditions in this study: reduced flicker lighting (percent flicker, 1.78%; flicker index, 0.28%) and control flicker lighting (percent flicker, 40.27%; flicker index, 13.37%) (
Before starting the experiment, the participants rested in the dark for 3 min. Following this, LED light was presented for a total of 22 min in the following sequence (
The Sternberg working memory task was performed using Presentation® (Neurobehavioral Systems Inc., Berkeley, CA, USA) while recording EEG [
During the Sternberg working memory task, scalp EEG was recorded with 60 cap-based electrodes (Quick-cap, Charlotte, NC, USA), which were placed according to the international 10–20 system. We used the average reference and set a bandpass filter from 0.1 to 70 Hz. The sampling rate was 400 Hz. The electrode impedances were kept below 5 kΩ.
Preprocessing and analysis of the ERP data were performed using MATLAB (MathWorks, Natick, MA, USA). For each trial, ERP recordings were epoched into 1,400-ms segments (from -200 to 1,200 ms after the stimulus), which corresponded to the retrieval phase of the working memory task (
For cortical source analysis, standardized low-resolution brain electromagnetic tomography (sLORETA) was used to solve the inverse problem because sLORETA has the advantage of zero error localization based on images of standardized current density [
We estimated the TBPS by measuring wPLI within the P300 time window (400–600 ms). EEG connectivity with zero phase lags contributes to spurious functional connectivity, which results from volume conduction to the scalp electrodes. The phase lag index (PLI) was originally developed to measure non-zero phase lags between two time series data by averaging the sign of the imaginary part of the cross-spectrum. As a modified version of PLI, wPLI is weighted by the magnitude of the imaginary component of the cross-spectrum to minimize the sensitivity to volume conduction and noise while increasing the capacity to detect changes in phase synchronization [
To extract the characteristic of phase perturbation at the theta band from EEG signals, a band-pass filter was adapted with a second-order Butterworth filter at a 4–6 Hz frequency band. Cross-spectrum X was computed for electrodes i and j using the extracted complex-valued Hilbert spectra vector Z,
Therefore, we estimated the wPLI between the two signals, i and j, which is defined as:
N is the total number of epochs, and θ (Xi,j) is the difference in the phase of cross-spectrum X between signals i and j. The wPLI ranges between 0 and 1, where 0 indicates no coupling, and 1 indicates perfect phase locking between the two signals. Baseline corrections were applied by subtracting the wPLIs at the baseline interval (-200–0 ms). Finally, we averaged the normalized wPLIs over the P300 time window (400–600 ms).
TBPS was calculated based on the theta-band wPLI for 1,770 connections between the scalp electrodes and 44 connections between the cortical sources. Thereafter, we compared the overall TBPS according to the two flicker conditions and two memory loads. In addition, we performed pairwise comparisons to identify significantly altered functional connections using reduced flicker light.
Repeated-measures analysis of variance (ANOVA) was used to determine the effect of flicker conditions and memory loads on working memory performance, ERP P300 amplitudes, and TBPS. Within-subjects factors included flicker conditions (reduced vs. control) and memory loads (item 2 vs. 3). Post-hoc comparisons between the two flicker conditions were performed using the Wilcoxon signed-rank test. Statistical significance was set at a two-tailed
Behavioral performance was measured by HR and RT during the Sternberg task (
We analyzed the effect of light flicker on the mean P300 amplitude during working memory tasks. Repeated-measures ANOVA showed that the flicker condition had no significant effect on the parietal P300 amplitude (F1,18=3.992,
We analyzed the effect of light flicker on TBPS from scalp EEG recordings.
Thereafter, we performed pairwise comparisons of 1,770 electrode pairs between the two flicker conditions. The reduced flicker light significantly enhanced TBPS compared with the control flicker light at both memory loads (unadjusted p<0.0025) (
EEG source imaging with sLORETA showed that the sources of theta oscillations of the ERP P300 component were mainly located in the bilateral prefrontal areas (
Next, we measured the TBPS between the cortical sources and compared them between the two flicker conditions. At memory load 3, the reduced flicker light significantly increased TBPS between the left prefrontal cortex and right hippocampus compared with the control flicker light (FDR<0.1) (
In this study, we evaluated the effect of LED flicker on brain oscillations during working memory tasks. The data from our experiments showed that behavioral performance and parietal P300 amplitude did not differ significantly when the flicker level was reduced. However, the functional connectivity data demonstrated that the reduced flicker light significantly enhanced TBPS compared with the control flicker light. Furthermore, the cortical source analysis confirmed that the thetaband connection between the left prefrontal lobe and right hippocampus significantly increased under reduced flicker light compared with the control flicker light. To the best of our knowledge, this is the first neurophysiological study to demonstrate that theta-band functional connectivity during cognitive functioning is modulated by the degree of light flicker.
EEG signals provide valuable information on oscillatory brain activity according to different frequency bands, which subserve distinct neural processes and have different functional significance [
We evaluated the P300 component and TBPS during the retrieval phase of working memory. Retrieval from working memory is implicated in several cognitive operations, such as stimulus evaluation, memory search, and decision-making [
The scalp topography of TBPS showed increased connections between the frontal and parietal regions under reduced flicker light. These results are in agreement with previous studies suggesting that theta synchronization between anterior and posterior brain regions significantly increased during the working memory processes [
A limitation of this study is that the functional connectivity changes only represented short-term responses with an exposure time of 19 min. Therefore, the long-term effects of LED flicker on functional connectivity and cognitive performance should be determined in future studies. Furthermore, the fact that data from 11 participants were excluded from the analysis is another significant limitation. Although those excluded failed to show typical ERP waveforms, this selection might have biased the results of this study. We evaluated slow theta oscillations of 4–6 Hz because the effect of the flicker condition was significantly observed only in this frequency range. Therefore, these results should be carefully interpreted.
In summary, we found that reduced flicker LED light increased theta-band functional connectivity during the working memory task compared with the control flicker light. Cortical source imaging showed enhanced TBPS between the hippocampus and prefrontal cortex under reduced flicker conditions. Therefore, reduced flicker light may facilitate working memory processes. LED lighting with reduced flicker should be considered in an environment that requires high-level cognitive performance.
The online-only Data Supplement is available with this article at
Waveforms of the light-emitting diode light. The x-axis indicates time (sec) and the y-axis indicates illuminance (lx). (A) Control flicker light. Frequency, 120 Hz; mean illuminance, 501 lx (max, 709.2 lx; min, 286.8 lx); percent flicker, 40.27%; flicker index, 13.37%. (B) Reduced flicker light. Frequency, 120 Hz; mean illuminance, 499.7 lx (max, 508.8 lx; min, 491.1 lx); percent flicker, 1.78%; flicker index, 0.28%.
Grand-averaged event-related potentials (ERPs) over the frontal (F3, Fz, and F4) and central (C3, Cz, and C4) regions. Blue lines indicate the ERP waveform of the control flicker light while red lines indicate the waveform of the reduced flicker light. There was no significant difference in the P300 amplitude between the two flicker conditions.
The authors would like to thank Byeong Uk Lee for his contribution to the EEG recordings. This work was supported by a research grant from Mimi Lighting Inc. (grant no. 06-2016-0560).
The authors have no potential conflicts of interest to disclose.
Conceptualization: Jun-Sang Sunwoo, Ki-Young Jung. Data curation: Sanghun Lee. Formal analysis: Jun-Sang Sunwoo, Sanghun Lee. Funding acquisition: Ki-Young Jung. Investigation: all authors. Methodology: JunSang Sunwoo, Ki-Young Jung. Resources: Ki-Young Jung. Visualization: Sanghun Lee. Writing—original draft: Jun-Sang Sunwoo, Sanghun Lee. Writing—review & editing: Jun-Sang Sunwoo, Ki-Young Jung.
Schematic illustration of the study. (A) Experimental paradigm. (B) Sternberg working memory task. Participants pushed button 1 when they were ready. After a fixation cross, a series of digits were displayed on the screen according to memory loads. Each digit in the memory set was presented for 1.2 s with 0.2-s intervals. After a retention phase for 2.2 s, a test digit (the probe) was presented on the screen and participants responded by pressing one of the buttons based on whether the probe was a member of the memory set. A time window of event-related potential (ERP) and functional connectivity analysis was between 400–600 ms after the probe onset (retrieval phase).
Event-related potentials (ERPs) according to the two flicker conditions. (A) Grand-averaged ERP plots over the parietal regions (P3, Pz, and P4). Blue lines indicate the ERP waveform of the control flicker light while red lines indicate the waveform of the reduced flicker light. A P300 component was defined as the waveform ranging between 400–600 ms after the stimulus onset (blue box). There was no significant difference in the parietal P300 amplitude between the two flicker conditions. (B) ERP topography of the P300 component. The maximum positivity was localized to the parietal regions. Color scale depicts the P300 amplitude (μV/m2).
TBPS from scalp EEG recordings. (A) The highest 5% of theta-band (4–6 Hz) connections were illustrated. (B) TBPS under the control (blue line) and reduced (red line) flicker conditions. TBPS within the P300 time window (400–600 ms, blue box) was compared between the two flicker conditions. *
Cortical source imaging and functional connectivity analysis. (A) Theta (4–6 Hz) cortical source activity of the P300 component (400–600 ms) measured by standardized low-resolution brain electromagnetic tomography (sLORETA). (B) Pairwise comparisons of theta-band phase synchrony (TBPS) between the two flicker conditions at memory load 3. Dots indicate regions of interest and red lines indicate significantly increased TBPS under the reduced flicker condition compared to the control flicker condition (false discovery rate<0.1). Note that TBPS between the left prefrontal lobe and right hippocampus increased under the reduced flicker light.
Coordinates for regions of interest
Brain regions | Side | Brodmann area | MNI coordinates (mm) |
||
---|---|---|---|---|---|
x | y | z | |||
Prefrontal lobe | |||||
Anterior prefrontal cortex | R | 10 | 24 | 56 | 4 |
L | 10 | -23 | 57 | 8 | |
Orbitofrontal cortex | R | 11 | 25 | 40 | -17 |
L | 11 | -23 | 37 | -19 | |
Hippocampus | |||||
Piriform cortex | R | 27 | 17 | -34 | -6 |
L | 27 | -16 | -37 | -2 | |
Ventral entorhinal cortex | R | 28 | 30 | -28 | -13 |
L | 28 | -30 | -23 | -17 | |
Perirhinal cortex | R | 35 | 15 | -40 | -5 |
L | 35 | -12 | -42 | -5 | |
Parietal lobe | |||||
Sensory association cortex | R | 5 | 13 | -44 | 62 |
L | 5 | -19 | -49 | 64 |
Location of regions of interest in MNI coordinates. MNI: Montreal Neurological Institute
Behavioral performances according to lighting flicker conditions
Variables | Control flicker | Reduced flicker |
---|---|---|
HR, % | ||
Item 2 | 99.23 (1.64) | 98.91 (1.92) |
Item 3 | 98.64 (2.12) | 98.12 (2.41) |
Mean | 98.94 (1.89) | 98.51 (2.19) |
RT, ms | ||
Item 2 | 278.09 (37.84) | 277.51 (42.49) |
Item 3 | 288.79 (38.60) | 290.48 (41.13) |
Mean | 283.44 (38.09) | 284.00 (41.77) |
Data are presented as mean (standard deviation). n=19. HR: hit rate, RT: reaction time