Abstract
Near-infrared (NIR) spectroscopy, also known as functional NIR spectroscopy (fNIRS), is a tool for measuring the hemodynamic response of the prefrontal cortex (PFC) by using NIR light, enabling a noninvasive indirect neural activity assessment. The application of fNIRS in emotion recognition or the differential diagnosis of psychiatric disorders of depressive patients, including major depressive disorder, bipolar disorder and schizophrenia, has previously been reported. Although the use of fNIRS has gradually expanded in cognitive neuroscience studies, few researchers have focused on the effects of light exposure in fNIRS studies. In addition, to the best of our knowledge, there are no scoping reviews of fNIRS studies on light exposure. Because light is an important topic in cognitive neuroscience and psychiatry, we evaluated fNIRS studies on light exposure in humans. We reviewed 10 papers in their entirety. Bright light (BR) modulates fear, and the color differences showed no significance in 1 study, whereas other studies delved extensively into the effects of colored light, finding some individual hemodynamic responses. In our study, we highlighted that the effects of natural light have not been studied using fNIRS. Light is becoming a critical topic in cognitive neuroscience and psychiatry, and fNIRS is critical for improving public health and managing psychiatric disorders.
Key words: prefrontal cortex, emotion recognition, light, NIR spectroscopy
Introduction
Functional near-infrared spectroscopy (fNIRS) is a tool for measuring cerebral hemodynamic responses, enabling a noninvasive assessment of indirect neural activity. This technique has developed rapidly over the 30 years since its first application.1 The fNIRS tolerates moderate movements, which is an advantage over other tools, such as functional magnetic resonance imaging (fMRI).
The first demonstration of measuring cerebral oxygen level was performed by Jöbsis in 1977,2 while the first commercial near-infrared (NIR) spectroscopy with a single channel was developed by Hamamatsu Photonics K.K. (Hamamatsu, Japan) in 1989. The original fNIRS system was expensive and cumbersome, and the many wires required rendered subjects immobile, making its application in clinical settings challenging. In 1998, a 10-channel system developed by Hitachi (Tokyo, Japan) was used on intractable epilepsy patients in a clinical setting.3 Recently, researchers have elaborated on developing more miniaturized, wireless and low-cost fNIRS systems.4, 5, 6
The principles of the mainstream continuous wave (CW) fNIRS technique are outlined in Figure 1. When NIR light travels through cerebral capillaries, the light is absorbed, scattered and transmitted. This light absorption can be calculated using the Beer–Lambert law. However, it is inappropriate to apply the original Beer–Lambert law. First, scattering light should be considered because the human brain includes blood vessels and other highly scattering tissues, such as the skin and skull. Second, the trajectory of light is not linear but banana-shaped (Figure 1) when applying NIRS to the human brain. Therefore, the law needed to be modified.7 When measuring light absorption of 2 different wavelengths, the common factor of scattering is negligible. Because the molar extinction coefficient for oxygenated hemoglobin (oxy-Hb) and deoxygenated hemoglobin (H-Hb) can be obtained experimentally, the relative concentrations of both oxy-Hb and H-Hb result from solving the 2 equations simultaneously.
Functional near-infrared spectroscopy can be a useful tool for measuring brain activity in various settings besides limited laboratory situations. However, there are critical disadvantages of fNIRS to consider. Although fNIRS measures hemodynamic response noninvasively,8, 9, 10 the spatial resolution is not as high compared to an electroencephalogram. In addition, the deep brain of adults cannot be measured since light penetration does not reach it. Therefore, fNIRS can only target the lateral surface of the brain hemispheres, mainly the cortex.11 The forehead is easy to approach regions of interest in fNIRS. These regions includes motor functionality (the primary sensorimotor cortex and the premotor cortex12) and higher-order functionality (the frontopolar cortex and the prefrontal cortex (PFC)13).
The PFC contributes to both fear acquisition and extinction, as PFC has the ability to learn from a precedent fear.14 Recent studies have suggested that the medial PFC is impaired by the basolateral complex of the amygdala (BLA), which is activated by the locus coeruleus (LC). In addition, the posterior ventromedial PFC (BA11) activates powerfully during late fear conditioning.15 In contrast, the LC activates the medial PFC, and the medial PFC inhibits the BLA.16 Therefore, the LC, medial PFC and BLA are vital for noradrenergic fear modulation, but the mechanisms differ between fear acquisition and fear extinction. In functional brain imaging studies, the fear and anxiety regions of the frontal cortex are not only in the medial prefrontal but also the dorsal, ventrolateral and prefrontal orbitofrontal cortex.17 In patients with post-traumatic stress disorder (PTSD), impairment of recalling fear extinction from the hypoactivity of the hippocampus was reported, whereas patients were conditioned to both fear acquisition and extinction normally.14 Also, the hippocampus projection to the PFC is aberrant in patients with PTSD, depression and schizophrenia,18 suggesting that hippocampal function affects fear, anxiety and cognition.
The fNIRS application in the differential diagnosis of depressive patients, including major depressive disorder, bipolar disorder and schizophrenia, has been reported.19 In Japan, fNIRS, as an optical topography test, has been covered by the healthcare insurance system for the differential diagnosis of the depressive state. However, clinical applications in neuroimaging are still controversial. In 2018, the American Psychiatric Association (APA) concluded that neuroimaging should remain a research tool only. There may be 2 main reasons for this. Firstly, most studies have a small number of participants and identify a relatively tiny effect size. Moreover, although neuroimaging is noninvasive, it is challenging to draw an identical assessment between clinicians, which makes it difficult to support its application for clinical diagnosis.
The use of fNIRS has expanded gradually in cognitive neuroscience studies. Cognitive neuroscience is a broad multidisciplinary field focusing mainly on the neural mechanisms behind cognitive function by linking together disciplines such as physiology, anatomy, psychology, psychiatry, and computational biology.20 Table 1 shows the main subjects of cognitive neuroscience with reported topics for fNIRS application. It should be noted that the mentioned studies were almost but not entirely separate, and each field will overlap with other fields. For example, dyslexia is in the reading field but can also be categorized in the decision or language fields. In terms of the attention field, attention deficit hyperactivity disorder (ADHD) during the developmental ages is a critical topic in fNIRS studies.21 Studies regarding traumatic brain injury,22, 23 autism spectrum disorders,24, 25 drowsy driving,26 and meditation27 are also recent clinical applications of fNIRS. Studies on consciousness, anesthesia management,28 disorders of consciousness,29 and locked-in syndrome30 have been reported. Decision-making is one of the essential applications of cognitive neuroscience. The application of the brain–computer interface31 to examine group decision,32 purchase decision33, 34 and risk-taking35, 36 yielded some pieces of evidence. In terms of executive control, obesity,37 exercise38 and older adults39 have been researched. In the intelligence field, problem-solving40 and work performance41 have been studied. The language field42, 43 is the most extensively studied topic in cognitive neuroscience, and the verbal fluency task (VFT) has recently demonstrated its ability to measure depressive disorder.44 We perceive almost all stimuli under any circumstances; therefore, it is challenging to aggregate perception studies. To the best of our knowledge, perception of depth,45 emotion,46, 47, 48 face,49 reward,50, 51 self-agency,52 and time53 have been reported. Finally, reading studies include dyslexia54 and overt reading55 (reading aloud).
Objectives
Light is essential not only for visualization but also for sleep, mood and cognition. Meta-analyses and systematic reviews support the impact of light on alertness.56 In addition, light therapy is an effective treatment for depression in Parkinson’s disease57 and premenstrual dysphoric disorder,58 and the effects of light therapy on sleep disturbance have also been reported.57, 59, 60, 61 In cognitive neuroscience and psychiatry, light is an important topic. However, few researchers have focused on the effects of light exposure measured using fNIRS. In addition, to the best of our knowledge, there is no scoping review of fNIRS studies on light exposure. Therefore, in the present review, we evaluate fNIRS studies on light exposure in humans.
Review protocol
We delineated 4 inclusion criteria. First, studies must be performed on humans. Second, studies should focus on light exposure. Therefore, papers that only mentioned light exposure for the technical principles of fNIRS or light effect on fNIRS experiments itself62 (ambient light) were excluded from our review. Third, not only direct but also indirect light exposure measurements were included, as was smartphone use63 (effect of blue light (BL)). However, studies were excluded if there were other non-negligible effects, such as the natural light effect on garden therapy64 with green plants. Fourth, since the first fNIRS study was published in 1991, studies published only after that year were included in the analysis.1 Following an initial Internet search, we concluded the literature on light exposure encompassed various fields such as architecture, education, religion, and complementary medicine. Therefore, from the standpoint of scientific rigor, we required papers to have a PubMed Identifier (PMID) in PubMed and did not take into account information available from other websites, dissertations and books. Our review presents the whole PubMed-indexed research on the topic.
We started the literature search on November 15, 2022, and finished it on January 20, 2023. The process of selecting literature is shown in Figure 2. We used Google Scholar, Scopus and PubMed databases to identify relevant studies, and carefully selected queries in order to identify fNIRS research on light exposure. The queries that were used are listed in Table 2. Google Scholar facilitates literature identification,65 but many irrelevant results appear because of its low precision.66 In addition, search results are not reproducible.67 Therefore, using Google Scholar solely is challenging in identifying literature68; thus, Scopus and PubMed were used to improve coverage and maintain search consistency. On January 20, 2023, 2 authors (RN and TF) independently searched the databases and found that all results were consistent. In addition, we confirmed that the results had not been updated since the previous search. After literature identification, we deduplicated the references using Zotero v. 6.0.18 (https://www.zotero.org/). After deduplication, articles with unavailable full-text articles were excluded, abstracts were screened and 10 papers remained for full-text review. The main highlight of the fNIRS studies was the oxy-Hb concentration.
The bias risk assessment was performed based on the characteristics and number of participants and study protocols. Moreover, the study protocol was assessed. We used robvis tool to create risk-of-bias plots.69
Results
A brief overview of the 10 reviewed studies is presented in Table 3. In summary, bright light (BR) modulates fear. Regarding the color differences, no significance was indicated in 1 study, whereas other studies suggested that colored light produced some individual hemodynamic responses.
Bias risk assessment
Figure 3 summarizes the risk of bias. All studies included both male and female participants. Although the number of participants ranged from 10 to 34 in most studies, 1 study had 757 participants. Due to the inhomogeneous signal processing methods, comparing oxy-Hb concentrations between studies would be inappropriate. In 4 studies,70, 71, 72, 73 the reported number, male/female proportion, standard deviation (SD), mean (M), and age ranges of participants were identical. In 1 study, a machine learning technique was used.71 However, although the machine learning technique is overused in various fields, artificial intelligence (AI) does not apply to small biased datasets.74 Therefore, even if machine learning methods successfully find plausible clustering, these models are not always suitable for all datasets, and sometimes the clustering is a result of sheer luck.
Detailed description
of the reviewed literature
Hori et al. stimulated 757 subjects with various colored lights and waterfall sounds using different frequencies to study the emotional effects on the PFC.75 Their feedback system enabled them to alter the stimuli in response to increased oxy-Hb concentration in the PFC, resulting from pleasantness or unpleasantness. Pleasantness or unpleasantness was determined according to a previous study.11 After the stimulation, 298 participants answered a questionnaire to assess their comfort with the stimuli, ranging from −5 (unpleasant) to +5 (pleasant). Results indicated that sound frequency rather than light color affects how pleasant or unpleasant the feelings of the participants were. High-frequency sounds accounted for approx. 43% of both pleasantness and unpleasantness responses.75
Yoshiike et al. performed an fNIRS study to show the effect of BR on PFC activity during human fear conditioning and fear extinction. They conducted a single-blinded study by exposing 29 participants to either bright (9000 lx) or regular (<500 lx) light, and presented visual conditioned stimuli (CS) randomly, followed by unconditioned stimuli (US) using an unpleasant but not painful electric shock. For CS, each sign represented a learning process using triangles, squares and circles to represent fear extinction, fear acquisition and a safety state, respectively. In addition to the PFC activity on fNIRS, skin conductance was recorded with each response to stimuli.76 The results were intriguing. Bright light decreases skin conductance during the recall session of any learning process. However, fNIRS showed that BR modulates fear immediately and differently, promoting fear extinction and safe learning while inhibiting fear acquisition. Moreover, BR modulates fear in any type of learning, which generally lasts at least 24 h.
Wolf’s research group extensively focused on the color of light and conducted 8 fNIRS studies in the years 2011–2022.70, 71, 72, 73, 77, 78, 82, 83 In 2011, they first investigated the effect of red and blue light on oxygen consumption in the brain and muscles of 10 participants.77 Then, they exposed the 12 volunteers to these 2-colored lights sequentially to examine the hypothesis stating that red light enhances the effects of BL.78 The source of light was not a light-emitting diode (LED) but a white light bulb with colored filters. The constant hemoglobin concentration and high tissue oxygen saturation indicated that BL, but not red light, exposure decreases oxygen demand in the brain.77, 78 However, this contradicts the activation of the PFC in both long (30 min) and short (1 min) BL exposures.79, 80, 81 Interestingly, blood oxygenation was independent of the sequence, and the oxygenation in both sequences was equivalent to a single red light exposure, suggesting that red light is preponderant to BL.78
Metz et al. in 2017 proposed physiology-augmented fNIRS (SPA-fNIRS) and conducted the first 2 SPA-fNIRS studies. As the name suggests, SPA-fNIRS associates fNIRS with physiological parameters, including skin conductance, end-tidal carbon dioxide (PETCO2) and heart rate (HR).82, 83 Both studies exposed the participants to colored light, but the major difference was whether the exposure was continuous82 or intermittent.83
Metz et al. continuously exposed the 12 participants to blue, green, red, and yellow LED lights for 10 min while recording their physiological and neurological activities using SPA-fNIRS. Skin conductance increased for all colors except green, and only yellow light significantly elevated oxy-Hb concentration in the left but not the right PFC. Additionally, as in their previous study, oxygen demand decreases were only observed when exposed to blue, but not red light. Except for the green light, HR increased while the partial pressure of CO2 (PaCO2, equivalent to PETCO284) decreased. The HR elevation and PETCO2 reduction indicate that participants hyperventilated during fNIRS. They described that this hyperventilation occurred due to weak pain in the recovery phase, but not from the light itself.
Scholkmann et al. used blue, green and red LEDs to beam light for 10 min, on and off every 20 s, but did not use yellow LEDs. In addition to PFC activity, visual cortex (VC) activity was measured using the SPA-fNIRS. The results indicated that BL and red light significantly increased respiratory rate (RR), whereas green light decreased it. Regarding HR and PETCO2, only BL elevated HR, and PETCO2 showed no significant alteration. In addition, only BL elicited an elevation in PFC oxy-Hb, while VC activity-related oxy-Hb elevation was observed with any color.83
In 2021, followed by another SPA-fNIRS study to investigate the colored light effect (CLE) during a VFT,73 Wolf’s research group delved into the interindividual differences of hemodynamic responses and consecutively published 4 studies in January 2021, April 2021, May 2022, and October 2022.70, 71, 72, 73 Briefly, followed by performing a VFT under CLE, they patterned or clustered hemodynamic responses using various techniques, including unsupervised machine learning,71 namely k-means.
In 2022, Honma et al. enrolled 34 healthy individuals to compare reading comprehension on a smartphone compared to paper by recording the number of sighs the subjects made while reading and bilateral PFC activity obtained using fNIRS. The 10-point comprehension test used 10 multiple-choice questions. This crossover study revealed that reading on paper achieves at least 1 point higher score than reading on a smartphone. In addition, sighs decreased by approx. 70% while reading on a smartphone, and the left PFC was simultaneously overactivated. These findings indicated that BL decreases sighs and increases left PFC overactivation, resulting in poor comprehension due to excessive cognitive loads.63
Discussion
Results show that the CLEs on verbal fluency have been extensively researched, and the hemodynamic responses may differ individually. In addition, BL overly activates the PFC. Moreover, BR modulates fear immediately and differently between learning processes, promoting fear extinction and safe learning, as well as inhibiting fear acquisition.
Although our review provides some evidence of the effects of colored light, some concerns should be mentioned. First, according to previously mentioned findings,75 light color does not affect emotions. On the other hand, Battaglia et al. suggested that both positive and negative emotions equally control our behaviors.85 Therefore, light color differences may not affect our behaviors. However, it is not surprising that numerous studies suggest that color underlies mood, quality of life (QOL) and cognition. In 2010, Carruthers et al. developed the Manchester Color Wheel to assess patients by the color of light they were exposed to using 38 colors representing positive, neutral or negative conditions that were rigorously reproduced on the participants.86 Color-using art therapy can increase purpose in the life of stroke patients and their caregivers, resulting in QOL improvements.87 In addition, as Berndt et al. indicated in 2020, color plays a vital role in language comprehension, facilitating anagram solving by presenting the referent color of the solution words.88
Second, as previously mentioned, Metz et al. reported that exposure to green light did not increase HR and decrease PaCO2,82 suggesting that green light did not induce hyperventilation (in turn caused by weak pain) during fNIRS. However, the light color should not be the only difference studied, and hyperventilation during fNIRS needs to be elucidated more complex. Although previously mentioned, it should be noted again that noninvasiveness is a core feature of fNIRS. Moreover, hyperventilation suggesting HR elevation and concurrent PETCO2 reduction were not observed with intermittent light exposure.83 Research on the effect of green light is scarce because exposure to green light in the natural environment is rare; nonetheless, the effects of green light need to be elucidated.
Finally, there are some discrepancies to be noted when comparing continuous light82 and intermittent light83 exposure. Regarding the BL effect, continuous light does not induce activation,77, 82 whereas intermittent light induces an oxy-Hb concentration increase in the PFC.83 However, interindividual differences can be the main plausible cause of varying results, as mentioned.70, 71, 72
Blue light activates intrinsically photosensitive retinal ganglion cells (ipRGCs), and the signal is projected to several brain regions, including the ventral tegmental area, raphe and LC via the suprachiasmatic nucleus, to modulate circadian rhythms.89 The ipRGCs do not form a vision, suggesting that even in blind people exposure to colored light may result in circadian modulation.80, 90, 91 A LED display on smartphones is an example of BL emitter. Blue light harms photoreceptors and pigmented epithelium cells,92 and the reviewed fNIRS study suggests the negative impact of BL on reading comprehension.63
In the reviewed studies, BL increased total hemoglobin concentrations when the participants were exposed to light emitted by a color-filtered white lightbulb, but it remained constant under LED illumination. However, the difference between incandescent light and LED can be applied to this result. Recently, Farghly et al. reared rabbits under various light conditions and found that rearing animals under LED conditions resulted in the highest glucose concentrations.93 In addition, Niemierzycka et al. showed that LED light with a color temperature of 4200 K achieves higher speeds with maintained accuracy on the Kraepelin test.94 As the Kraepelin test induces PFC activity95 and glucose is primarily present in the brain, it can be inferred that LED light plays a different role than incandescent light on the PFC.
The present review indicated that there are only 10 studies on the effects of light exposure measured using fNIRS measurements in PubMed. In addition, we should carefully interpret these results. For example, since colored light effects have been studied only by a single research group, there are some non-negligible biases, such as confirmation bias. Thereby, the present review identified the sparsity of light exposure studies with fNIRS usage.
The effect of natural light has not been studied using fNIRS. As previously mentioned, garden therapy was studied,64 but the effect of green plants is not negligible. However, studying solely natural light is challenging because natural light is a component of the environment. For example, when downtown, there can be additional stimuli like noise from cars or rising temperatures caused by the heat island effect.
We hypothesize that fNIRS can yield critical findings in combination with other biomarkers. For example, because fNIRS measures the oxidation of hemoglobin in the cerebral capillaries, cerebral oxidative stress can be noninvasively measured in combination with other oxidative indicators, such as the kynurenine-3-monooxygenase, which is activated by oxygen.96, 97, 98, 99, 100, 101 Moreover, fNIRS might facilitate the understanding of fear learning, in combination with cardiac autonomic dynamics of heart rate variability.102
Another important hypothesis is that fNIRS can bring critical findings for psychiatric disorders in terms of light effects. For example, since light therapy was reported to be effective in seasonal affective disorder (SAD) in 1984,103 only the atypical pattern of rod electroretinogram has been reported as a possible biomarker.104 Due to a much stronger invasiveness of electroretinogram than fNIRS, the present review suggests the need for developing a biomarker in patients with SAD. The fNIRS can also be used in combination with inflammatory cytokines in dementia105, 106 or the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in various gut microbiota-related psychiatric illnesses.107 Although MRI is essential for diseases such as cerebral small vessel disease,108 there might be a possibility of using fNIRS to predict the critical need for an MRI, which would significantly spare the clinician’s time and effort.
As previously mentioned, fNIRS is an effective tool for cognitive neuroscience due to its non-invasiveness, tolerance of movements and high resolution. On the other hand, most fNIRS studies corroborate PFC activity by associating oxy-Hb concentration increase with brain activity.109, 110, 111, 112, 113 The idea of such association is derived from the neurovascular coupling theory. Although the mechanism by which neural tissue activation leads to increased cerebral blood flow remains unclear,114, 115 in terms of neurovascular coupling, the integration of fNIRS and functional MRI can be helpful in understanding brain hemodynamics.116
Limitations
Although this scoping review presents the whole PubMed-indexed research on fNIRS and the effects of light exposure, it has some limitations. We focused only on PubMed-indexed literature, which excludes information available from websites, dissertations and books. In addition, fNIRS-related fields, such as engineering, computer science and physics, were not incorporated because PubMed does not include biomedical journals. Finally, we did not include studies published after January 2023.
Conclusions
The fNIRS techniques enable real-time monitoring of human brain activity with tolerance to motion artifacts to some extent. Humans live with light, but its quality changes from natural to artificial light exposure. Notwithstanding that natural light is still essential, its effect on the brain’s hemodynamic response has not been thoroughly researched. Delving into light effects under commonplace environments using fNIRS is a reasonable approach. In conclusion, light will be a more critical topic in cognitive neuroscience and psychiatry, and fNIRS can be a critical tool for improving public health and managing psychiatric disorders.
