To investigate brain oscillatory characteristics according to brightness and color temperature of light emitting diode (LED) light in young and elderly subjects.
We analyzed 22 young (age, 29.0±5.2 years) and 23 elderly (age, 64.8±4.5 years) healthy subjects. A LED light source was used with a combination of two color temperature (6,500 K vs. 3,000 K) and two brightness (700 lx vs. 300 lx) conditions. Participants were exposed to each light condition in relaxed wakefulness. Then, we analyzed power spectral density and functional connectivity from eye-open electroencephalography.
A main effect of brightness on delta (
The present study demonstrated that spectral power and functional connectivity as well as subjective feelings are affected by the brightness and color temperature of LED light. These results might help us to understand the neurophysiological effects of light and identify the optimal indoor lighting conditions for an individual’s environment.
Light has profound effects on human health [
Although there has been substantial research into the effects of lighting conditions on mood and behavior, previous investigations were mostly based on subjective measures such as self-reported questionnaires. Accordingly, there has been relatively little evidence objectively showing that brain activities are altered by different lighting conditions. Electroencephalography (EEG) is an electrophysiological method to record the brain activity and its change according to different lighting conditions [
We prospectively recruited healthy young (n=30, age range 20–39 years) and old (n=30, age range 60–76 years) volunteers through advertisement from February to June 2016. Exclusion criteria were history of any neurological or psychiatric disorders, significant head trauma, and brain surgery. Those who took medication within the last two weeks, which can affect EEG recordings, such as antiepileptic drugs, sleeping pills, and antidepressants, were also excluded. All participants visited the Department of Neurology of the Seoul National University Hospital and were screened for the selection criteria. All subjects gave written informed consent before enrollment into the study. Among 60 subjects enrolled, 8 young and 7 old subjects were excluded from the analysis because of poor EEG data quality. Finally, data from 22 young (mean age, 29.0± 5.2 years; men 50%) and 23 elderly (mean age, 64.8±4.5 years; men 82.6%) subjects were analyzed in this study. This study was approved by the Institutional Review Board of the Seoul National University Hospital (H-1601-043-733) and was conducted in compliance with the Declaration of Helsinki and the Good Clinical Practice guidelines.
We used four different light conditions as a combination of color temperature (6,500 K vs. 3,000 K) and brightness (700 lx vs. 300 lx): high-bright (6,500 K, 700 lx); high-dark (6,500 K, 300 lx); low-bright (3,000 K, 700 lx); and low-dark (3,000 K, 300 lx) light. These light parameters were chosen within the range of indoor lighting conditions which are commonly used in the real world. The experimental space (width 2 m, length 2.5 m, height 2.7 m) was shielded from ambient light. Light emitting diode (LED) light panels (Mimilighting, In-cheon, Korea) were suspended from the ceiling. We used a controller to regulate color temperature and brightness of LED light per session. The lighting condition was checked at desk level (0.7 m).
For each light condition, subjects first took a rest in a lightless condition for 5 min. Then, LED light of specific parameters was presented for 5 min when subjects opened and closed their eyes alternately every 1 min (
EEG was recorded using a 64-channel recording system (Grass Technologies, Quincy, MA, USA) with 60 cap-based electrodes (Quick-Cap, Charlotte, NC, USA) that were located according to the international 10–20 system. Impedances were kept below 10 kΩ in all electrodes. EEG recordings were sampled at 400 Hz and referenced to an average reference. A bandpass filter was set between 0.5–70 Hz. We thoroughly inspected EEG recordings and selected 30 artifact-free epochs of 2-s duration from eye-open resting-state EEG for each light condition.
Selected EEG data were pre-processed and further analyzed for power spectral density and functional connectivity using MATLAB (MathWorks, Natick, MA, USA). We applied 60-Hz notch filter and interpolated bad EEG channels. Then, we performed independent component analysis decomposition and removed components containing ocular movement, heartbeat, and muscle artifacts. The current source density transformation was employed to enhance EEG spatial resolution and minimize the volume conduction effect. In addition, EEG data were demeaned and detrended to remove nonstationarity of the time series mean.
Spectral power analysis was performed using a fast Fourier transformation with the Welch’s method (1-s hamming window, 50% overlap)[
To measure EEG functional connectivity, we computed the weighted phase lag index (wPLI)[
For power spectral density, we conducted repeated-measures analysis of variance (ANOVA). Within-subject variables included brightness (two levels: bright and dark), color temperature (two levels: high and low), and brain regions (three levels: frontal, central and parietal), while a between-subject variable was age groups (two levels: old and young). Thereafter, we conducted repeated-measures ANOVA per each age group. For functional connectivity and subjective evaluation data, within-subject variables included brightness (two levels: bright and dark) and color temperature (two levels: high and low), and between-subject variables were age groups (two levels: old and young). Both power spectral density and functional connectivity data were analyzed for each frequency band. When sphericity assumptions were violated, we applied the Greenhouse-Geisser correction to control for type 1 error rates. Multiple comparisons in post hoc analysis were corrected by the Bonferroni method. A two-tailed
A significant interaction between brightness and age group was noted in theta power (F1,43=4.704,
A main effect of brightness on delta-band power was significant in the old age group (F1,22=4.545,
Similarly, theta-band power were significantly higher in bright light than in dark light (F1,22=4.882,
By contrast, power spectral density of alpha, beta bands were not affected by brightness or color temperature. In addition, there was no significant interaction between brightness and color temperature at these frequencies. Taken together, power spectral density of elderly subjects was influenced by brightness rather than color temperature. Bright light enhanced the delta- and theta-band power particularly in the frontal region.
We observed a marginally significant effect of color temperature on beta power (F1,21=3.839,
At delta, theta, and alpha frequencies, however, there was no significant effect of different light conditions on power spectral density. In summary, power spectral density of young subjects was mainly affected by color temperature; high color temperature increased beta-band power in the central regions. Further, these effects of color temperature were particularly significant under the dark light condition.
A significant main effect of color temperature on functional connectivity was observed in delta (F1,43=8.428,
There was a significant effect of color temperature on comfort (F1,43=7.302,
The present study demonstrated that brain oscillations were influenced by brightness and color temperature of the LED light. Subjective feelings were also affected by the different lighting conditions, which was in agreement with previous studies [
Brain oscillatory activity changes with age, which is characterized by general slowing of background EEG activity together with decrement in amplitude and power of alpha rhythms [
Another significant finding in this study was that color temperature conditions modulated resting-state functional connectivity. High color temperature light increased overall connectivity strength particularly in delta and beta frequencies. Physiological effects of color temperature on brain activity have been evaluated by recording EEG under different light conditions [
There are several limitations in this study. Brain oscillatory characteristics shown in this study represented short-term responses to the light conditions with exposure time of five min. Therefore, long-term effects of the light conditions cannot be concluded from this study. Another limitation is that subjective feelings except for sleepiness were not measured by validated scales. Moreover, the interaction effects between light conditions and age groups appeared to be modest with marginal significance levels. The subtle differences between the different age groups might be attributed to relatively short exposure to the light. Another possible explanation is that the resting state was not the optimal condition to detect brain oscillatory responses to the different light. In addition, we detected the changes in resting-state brain oscillations without neuropsychological tests, so that we could not determine that the altered EEG activities have functional implications particularly in neurocognitive performances. Further research such as cognitive event-related potentials under different light conditions will be required to address these concerns.
The online-only Data Supplement is available with this article at
The authors would like to thank Byeong Uk Lee for his contribution to EEG recording and Min Hee Jeong for her contribution to EEG analysis. This work was supported by a research grant from Mimi Lighting Inc. (grant No. 06-2016-0560).
Experimental design. (A) Schematic illustration of exposure to each lighting condition. (B) Schematic illustration of a counterbalanced design for the order of lighting conditions. Each subject is exposed to all 4 lighting conditions in a different order.
Power spectral density of the elderly subjects according to brightness of light. (A) Delta-band (0.5–4 Hz) power. (B) Theta-band (4–8 Hz) power. Error bars indicate the standard errors of the mean. n=23. *
Power spectral density of the young subjects according to color temperature of light. (A) Beta-band (12–30 Hz) power per brain region. (B) Interaction between color temperature and brightness on beta power. Error bars indicate the standard errors of the mean. n=22. *
Effect of color temperature on functional connectivity. (A) Comparisons of wPLI for each frequency band. n=45. (B) Beta-band functional connectivity for each age group. Error bars indicate the standard errors of the mean. Old group, n=23; Young group, n=22. *
Subjective feelings according to different light conditions. (A) Comfort. (B) Happiness. (C) Sleepiness. (D) Refreshment. Error bars indicate standard deviations. n=45. *
Repeated-measures analysis of variance for power spectral density in elderly subjects
Factors | Delta |
Theta |
Alpha |
Beta |
||||
---|---|---|---|---|---|---|---|---|
F | F | F | F | |||||
Color temperature | 0.639 | 0.433 | 0.166 | 0.687 | 0.168 | 0.686 | <0.001 | 0.984 |
Brightness | 4.545 | 0.044 |
4.882 | 0.038 |
0.503 | 0.486 | 0.379 | 0.545 |
Region | 2.117 | 0.151 | 6.743 | 0.003 |
33.176 | <0.001 |
0.841 | 0.417 |
Color temperature×brightness | 1.293 | 0.268 | 0.170 | 0.684 | 0.798 | 0.381 | 0.114 | 0.739 |
Region×color temperature×brightness | 1.105 | 0.326 | 0.350 | 0.707 | 0.804 | 0.420 | 0.057 | 0.906 |
Degree of freedom=1, 22.
Repeated-measures analysis of variance for power spectral density in young subjects
Factors | Delta |
Theta |
Alpha |
Beta |
||||
---|---|---|---|---|---|---|---|---|
F | F | F | F | |||||
Color temperature | 1.219 | 0.282 | 3.238 | 0.086 | 1.026 | 0.323 | 3.839 | 0.063 |
Brightness | 0.465 | 0.503 | 1.168 | 0.292 | 1.780 | 0.196 | 0.030 | 0.863 |
Region | 25.447 | <0.001 |
26.333 | <0.001 |
14.033 | <0.001 |
26.406 | <0.001 |
Color temperature×brightness | 0.015 | 0.904 | 0.158 | 0.695 | 1.652 | 0.213 | 6.484 | 0.019 |
Region×color temperature×brightness | 0.102 | 0.903 | 1.322 | 0.274 | 0.210 | 0.680 | 4.816 | 0.023 |
Degree of freedom=1, 21.
Summary of repeated-measures analysis of variance for functional connectivity
Factors | Delta |
Theta |
Alpha |
Beta |
||||
---|---|---|---|---|---|---|---|---|
F | F | F | F | |||||
Color temperature | 8.428 | 0.006 |
2.842 | 0.099 | 3.242 | 0.079 | 4.207 | 0.046 |
Color temperature×group | 1.175 | 0.284 | 1.273 | 0.265 | 0.314 | 0.578 | 4.058 | 0.050 |
Brightness | 2.031 | 0.161 | 0.023 | 0.881 | 0.508 | 0.480 | 0.272 | 0.605 |
Brightness×group | 0.137 | 0.714 | 0.018 | 0.893 | 2.013 | 0.163 | 0.245 | 0.623 |
Color temperature×brightness | 0.016 | 0.900 | 1.808 | 0.186 | 0.177 | 0.676 | 2.916 | 0.096 |
Color temperature×brightness×group | 0.015 | 0.902 | 0.096 | 0.759 | 0.515 | 0.477 | 0.407 | 0.527 |
Group | 2.258 | 0.140 | 0.281 | 0.599 | 0.225 | 0.637 | 2.171 | 0.148 |
Degree of freedom=1, 43.