Pilot Study on Dose-Dependent Effects of Transcranial Photobiomodulation on Brain Electrical Oscillations: A Potential Therapeutic Target in Alzheimer’s Disease

Abstract

Background:

Transcranial photobiomodulation (tPBM) has recently emerged as a potential cognitive enhancement technique and clinical treatment for various neuropsychiatric and neurodegenerative disorders by delivering invisible near-infrared light to the scalp and increasing energy metabolism in the brain.

Objective:

We assessed whether transcranial photobiomodulation with near-infrared light modulates cerebral electrical activity through electroencephalogram (EEG) and cerebral blood flow (CBF).

Methods:

We conducted a single-blind, sham-controlled pilot study to test the effect of continuous (c-tPBM), pulse (p-tPBM), and sham (s-tPBM) transcranial photobiomodulation on EEG oscillations and CBF using diffuse correlation spectroscopy (DCS) in a sample of ten healthy subjects [6F/4 M; mean age 28.6±12.9 years]. c-tPBM near-infrared radiation (NIR) (830 nm; 54.8 mW/cm²; 65.8 J/cm²; 2.3 kJ) and p-tPBM (830 nm; 10 Hz; 54.8 mW/cm²; 33%; 21.7 J/cm²; 0.8 kJ) were delivered concurrently to the frontal areas by four LED clusters. EEG and DCS recordings were performed weekly before, during, and after each tPBM session.

Results:

c-tPBM significantly boosted gamma (t = 3.02, df = 7, p < 0.02) and beta (t = 2.91, df = 7, p < 0.03) EEG spectral powers in eyes-open recordings and gamma power (t = 3.61, df = 6, p < 0.015) in eyes-closed recordings, with a widespread increase over frontal-central scalp regions. There was no significant effect of tPBM on CBF compared to sham.

Conclusion:

Our data suggest a dose-dependent effect of tPBM with NIR on cerebral gamma and beta neuronal activity. Altogether, our findings support the neuromodulatory effect of transcranial NIR.

Keywords

Cerebral blood flow EEG oscillations light-emitting diode photobiomodulation transcranial light therapy transcranial near-infrared light

INTRODUCTION

Transcranial photobiomodulation (tPBM) with near-infrared radiation (NIR) is a novel intervention based on the use of low-level lasers or light-emitting diodes (LEDs) that has recently emerged as a potential valuable therapy for a range of neuro-psychiatric conditions. tPBM may have therapeutic effects in subjects with stroke, traumatic brain injury, neurodegenerative disorders, and major depressive disorder [1], as well as pro-cognitive benefits in healthy populations [1 –3]. Recent findings have pointed out the beneficial effect of tPBM in the augmentation of cognitive functions, such as memory and attention, in addition to emotional functions. In a placebo-controlled randomized study in healthy volunteers, Barrett and Gonzalez-Lima [2] showed a significant improvement in performance on the frontal lobe cognitive tasks and in the emotional state after tPBM (1064 nm laser, 250 mW/cm², 60 J/cm², 13.6 cm² x 2 sites) delivered with continuous light to participants’ forehead.

The mechanisms through which tPBM with NIR or red light, delivered to the scalp of patients, may influence the adjacent cortical areas of the brain are poorly understood [4]. One promising hypothesis is that tPBM may influence brain energy metabolism through promoting the mitochondrial function. The primary source of intracellular energy, adenosine triphosphate (ATP), which is critical to sustaining neural activity, is largely produced in mitochondria through the process of oxidative phosphorylation. This process involves a respiratory chain of five enzyme complexes which, if altered, would influence ATP synthesis. It has been suggested that tPBM, by delivering photons (energy particles) to the tissue, may promote one of such complexes—cytochrome c oxidase or respiratory chain complex IV—restoring or enhancing ATP production and leading to more energy available for neuronal activity [5]. Several studies reported an upregulation effect on cytochrome c oxidase (CCO) from both the LED and laser light therapy at or close to 830 nm, which led to the neuronal increase in energy production [6, 7]. Mitochondria dysfunction has been suggested in many common neuropsychiatric disorders [8], and tPBM may enable compensation for such dysfunction, to restore cognitive capacity.

More active mitochondria would support higher oxygen/glucose consumption, which might stimulate cerebral blood flow (CBF) to deliver such nutrients. Intriguingly, a preclinical study by Uozumi et al. (2010) reported an increase of 30%of CBF after transcranial NIR laser irradiation at 808 nm wavelength for 15–45 min (at a power density of 1.6 W/cm²), not related to heating [9]. In healthy human participants and clinical patients, cerebral oxygenation and CBF were also found to increase [10 –12]. Through its effects on both CBF and brain metabolism [3], tPBM is considered a potential non-invasive therapy for cognitive impairment based on both animal and human studies [13].

More active mitochondria have also been demonstrated to influence brain activity, especially oscillations at high frequencies. Extensive research in slice cultures of hippocampus has documented the link between the mitochondrial function and fast neural oscillations in the gamma band (∼30–90 Hz) [14, 15]. Gamma frequency activity is believed to arise from the interplay between cortical inhibitory interneurons and excitatory principal neurons—the mechanism being: high rates of interneuron firing are required to synchronize function of principal neurons [14, 16]. The high metabolic demands of such interneuron activity may be linked to the activity of the mitochondrial respiratory chain in these cells. Intriguingly, the parvalbumin-positive interneurons, which are critical to gamma oscillations, show higher levels of mitochondrial cytochrome c oxidase when compared with principal neurons [17]. Thus, by stimulating this mitochondrial enzyme (CCO), tPBM may deliver targeted stimulation of the parvalbumin-positive interneurons and would be expected to enhance gamma oscillations.

Gamma oscillations provide a fundamental mechanism of complex neuronal information processing in the hippocampus and neocortex of mammals. This type of brain electrophysiological activity has been implicated in higher brain functions such as sensory perception [18], motor activity [19], and memory formation [20], and are impaired in many common neuropsychiatric disorders [21]. Gamma oscillations represent a potential therapeutic target in Alzheimer’s disease (AD) as evidenced by both animal model and human studies that correlated electroencephalography to AD specific processes. In animal models of AD, neuronal gamma oscillations are reduced before the onset of plaque formation and of cognitive decline [22, 23]. Optogenetically-driving parvalbumin-positive interneurons at gamma frequencies (40 Hz) reduced levels of amyloid-β deposition [23]. Similarly, a non-invasive 40 Hz light-flickering regime reduced levels of amyloid-β in the visual cortex, in animal models [23]. Compared to age-matched healthy control subjects, both patients with mild cognitive impairment (MCI) and patients with AD exhibit an increase in relative power of slow oscillations (electroencephalography (EEG) delta and theta rhythms), associated with a decrease in relative power of fast oscillations (EEG alpha, beta, and gamma rhythms) [24 –26]. A reduction in spontaneous gamma-synchronization was found in patients with AD [21]. In patients with MCI, global amyloid-β deposition was inversely correlated to alpha, beta, and gamma coherence, a marker of functional and effective connectivity [27]. Despite this literature pointing to gamma oscillation as a potential therapeutic target for neurodegenerative disorders, only limited evidence is currently available on the effects of tPBM on high-frequency neural oscillations in the human brain and on its ideal dosimetry.

The goal of the current study was to elucidate the physiological processes that may underlie the pro-cognitive effects of tPBM with NIR by tracking the effects of two different treatment doses on simultaneous recordings of neural activity oscillations and CBF. We delivered 830 nm tPBM by a four LED clusters device to the forehead of healthy participants and employed EEG to quantify electrophysiological brain oscillations and diffuse correlation spectroscopy (DCS) to index CBF. We anticipated that tPBM would stimulate the CBF and the electrophysiological oscillations, especially in the high frequency bands. We also expected that pulsed light (p-tPBM) would exert a stronger effect compared to continuous light therapy (c-tPBM), which would be consistent with prior reports [28 –30]. Secondary aims were to assess the safety and tolerability of NIR-light therapy delivered transcranially.

MATERIALS AND METHODS

This single-site study, Transcranial Near-Infrared Light in Healthy Subjects: a Cerebral Blood Flow Study with Diffuse Correlation Spectroscopy (NIR-flow), was approved by the Massachusetts General Hospital (MGH) institutional review board (IRB). The main sources of recruitment were email and website ads through the Partners Health Care internal portal for clinical trials. The study clinicaltrials.gov identifier was NCT03740152. Study data are available upon request. s

Inclusion and exclusion criteria

Subjects (age 18–70) eligible for study participation were healthy by Structured Clinical Interview for Diagnostic Statistical Manual-IV (SCID) criteria. Subjects were enrolled in the study after providing written informed consent. Female subjects of childbearing potential needed to consent (without any element of coercion) to use a double-barrier method for birth control if sexually active; pregnancy and lactation were exclusionary. Other exclusionary criteria included any current psychiatric disorder, substance or alcohol use disorders (prior 6 months), lifetime psychotic episodes, bipolar disorder, unstable medical or neurological illness, recent history of stroke, active suicidal and homicidal ideation. In addition, the following criteria, potentially influencing light penetration and safety, were also exclusionary: any use of light activating drugs (prior 14 days), having a forehead skin condition (such as tattoo or open wound) and having a head-implant. Out of eleven eligible subjects, ten subjects were followed for the entire five-week study period and completed a sequence of three sessions of tPBM and sham. The study involved screening (week 1), 3 tPBM sessions (weeks 2–4), and one follow-up visit (week 5).

Study design, blinding, and assessment schedule

The study included three sequential sessions, separated by at least one week, which entailed three different modes of operation of the tPBM device (Fig. 1). Participants received one session of continuous light treatment (c-tPBM), followed by a session with sham treatment (s-tPBM), which was followed by a session with pulse light treatment (p-tPBM). While the sequence (c-tPBM, s-tPBM, p-tPBM) was the same for all subjects, the subjects were blind to the specific mode of device operation used in each session. Simultaneous recordings of EEG and DCS were conducted as subjects rested (with no cognitive task) before, during, and after each tPBM session. EEG was also recorded while participants performed a working memory (2-back) task: once at baseline, before the c-tPBM session, and after each tPBM session. NIR light was administered by a four LED clusters device (LiteCure^® TPBM-1000) which had the flexibility to deliver either continuous light, pulse light, or sham. In both c-tPBM and p-tPBM mode the TPBM-1000 device delivered therapeutic NIR energy. In sham (s-tPBM), the device did not deliver any light energy. The apparent behavior (i.e., the performance/output of all visible and audible indicators) of the device in any of the three programmed treatment modalities was identical. Because of the low average irradiances delivered in both c-tPBM and p-tPBM modes and because of heat sinks incorporated in the device, the subjects did not experience any skin warming from NIR. These features of the design ensured that the study was single-blind. The device was designed to shut off automatically if skin warming above 41°C was detected. Tolerability was assessed through clinician inquiry and through a self-report scale (SAFTEE-SI), which were administered before the first session (c-tPBM), as a baseline measure, and one week after each of the three sessions to examine treatment-emergent side effects.

Fig. 1

Flowchart of the study procedures

Intervention parameters (tPBM)

NIR light at 830 nm was delivered with the TPBM-1000 device (Fig. 2) to four EEG sites (Fp1, Fp2, F3, F4) targeting frontal poles and dorsolateral prefrontal cortex, covering a total surface active treatment area of 35.8 cm² [(11.52 cm²x2; Fp1, Fp2) + (6.38 cm²x2; F3, F4)] with an irradiance of 54.8 mW/cm² (average in c-tPBM and peak in p-tPBM), an average fluence of 65.8 J/cm² in c-tPBM and of 21.7 J/cm² in p-tPBM, and a total energy of 2.3 kJ in c-tPBM and 0.8 kJ in p-tPBM (Table 1). The additional parameters used for pulsed light were the following: frequency 10 Hz, duty cycle 33%. The duration of irradiation was 20 min (the 4 sites were irradiated concurrently). The entire sessions lasted about 1.5 h. The additional time was needed to have the subject complete a urine toxicology test and self-report forms, to prepare the subject, to place the necessary protections (e.g., goggles), to inspect the subject’s skin, to complete diffuse correlation spectroscopy and EEG (before, during, and after tPBM), to set the tPBM devices, and to give the subject time to rest after the irradiation. tPBM was administered by licensed physicians (i.e., MDs) who were on study staff and trained to for the use of TPBM-1000.

Fig. 2

Set up of TPBM-1000, DCS, and EEG devices, mounted on study subject’s head^*. ^*Of note that the TPBP-1000 (tPBM device) was worn by study subjects, on top of EEG set-up and in close proximity of the DCS sensors, only during the actual delivery of NIR or sham intervention.

Table 1

Treatment parameters for tPBM delivered as continuous wave (c-tPBM) and as pulse wave (p-tPBM)

	c-tPBM	p-tPBM
Wavelength	830 nm	830 nm
Irradiance (peak)	54.8 mW/cm²	54.8 mW/cm²
Pulsing rate	-	10 Hz
Duty cycle	100%	33%
Average Power	∼2W	∼0.7W
Peak Power	∼2W	∼2W
Fluence	65.8 J/cm²	21.7 J/cm²
Duration of t-PBM session	20 min	20 min
Treatment window	35.8 cm²	35.8 cm²
Cumulative dose	2.3 kJ	0.8 kJ

Electroencephalographic activity (EEG)

EEG is an imaging technique that can track electrophysiological activity of the brain with a high temporal resolution. Previously, EEG has been used to evaluate neurophysiological brain activity both during resting ‘task free’ scans, to study patterns of spontaneous dynamics of brain activity [31], and during task performance to examine how brain activity is modulated by the demands of a task [32]. Fluctuations in the EEG signal have been studied in several frequency bands, including delta (0.5–3.5 Hz), theta (4–8 Hz), alpha (8–12 Hz), beta (12–30 Hz), and gamma (>30 Hz). Such patterns of slow and fast fluctuations have been linked to different states of mental activity, and different sensory and cognitive processes [33].

The Stat^® X24 wireless EEG system (Advanced Brain Monitoring, Carlsbad, CA) was used for EEG data acquisition. The Stat X24 combines battery-powered hardware with a preconfigured sensor strip for recording twenty channel monopolar EEG. The Stat X24 provides 19 EEG channels in accordance with the International 10–20 system, including Fz, F3, F4, Cz, C3, C4, P3, P4, Pz, O1, O2, T5, T3, F7, Fp1, Fp2, F8, T4, and T6, plus adds POz. EEG (sampling rate, 256 Hz, band pass filter: 0.1 Hz high-pass, 100 Hz fifth order low-pass, referenced to linked mastoids) was recorded and analyzed during rest before and after each tPBM session and during each 2-back task session. For the 2-back task we were interested in the average power spectral density (PSD) over the entire task-performance session. Additionally, electrocardiogram was recorded from collarbone left/right locations to enable removal of cardio-artifacts. Independent component analysis (ICA) was used to remove any artifact due to eye-blinks (using a software package from OHBA: https://github.com/OHBA-analysis/osl-core). During each resting state recording, participants reclined in a comfortable chair for 5 min with eyes open (while looking at a cross shown on a computer screen) and 5 min with eyes closed. Participants were asked to relax and avoid any eyebrow movements or clenching their jaw. Because the state of brain function was expected to change from before to after tPBM not only due to the light treatment itself, but also due to relaxing in a darkened room for a considerable time during treatment [34], our primary comparison of interest was between EEG PSD in post and pre tPBM recordings, after subtracting corresponding PSD recorded on a sham treatment visit. We kept the EEG cap on during the administration of tPBM to ensure that EEG recordings were from the same identical EEG sites on the scalp, before and after applying NIR light. This spatial correspondence enhanced our ability to detect pre/post changes in the EEG signal. To prevent significant disruption from tPBM to EEG recordings, we discarded the EEG data during the tPBM administration (tPBM device “on”). Because we hypothesized that tPBM would have an effect on the brain activity in higher frequencies, we limited our analyses to the gamma, beta, and alpha bands.

Cerebral blood flow (DCS)

DCS is an optical technique that uses the temporal fluctuations of near-infrared light to measure CBF directly and non-invasively [35 –37]. DCS devices to date are not under Food and Drug Administration investigational device exemption (IDE) regulations, since they are considered non-significant risk. A homebuilt DCS system was used with a long-coherence length 852 nm laser, 4 photon counting avalanche photodiode detectors, and a custom field-programmable gate array-correlator [38]. The DCS system was placed within 1-2 mm from one of the tPBM sources, at either the left or right frontal pole. The DCS light source power at the optical probe was below 40 mW, and the beam diameter was > 1 mm (to comply with American National Standards Institute standards for skin exposure). The detectors were arranged at different distances from the source. In particular, one detector was at 5 mm separation from the source to detect superficial blood flow changes and 3 detectors were at 2.5 cm to detect blood flow in deeper tissue. The DCS measurements were done before and after tPBM and every 1 min for 10 s during tPBM, by turning off the tPBM light and turning on DCS source and detectors (to avoid detectors damage by the powerful tPBM light). Blood flow index at the short separation (BF_i) was calculated for each detector by assuming a fixed absorption and scattering across subjects. CBF_i at the three large separations was averaged together to reduce noise. Temporal changes of CBF_i (large separation), of BF_i (short separation), and the subtraction of the two were considered for the statistical analyses.

Cognitive Task (2-back)

In the 2-back task, participants were presented with a sequence of alphabet letters, and were asked to press a button when the current letter matched the one from 2 steps earlier in the sequence. Participants saw 125 letters and 20 %of letters were targets. Each letter was presented for 1 s with 1 s ISI or interstimulus interval. Such 2-back task is believed to capture the engagement of working memory, one of the cognitive functions supported by the dorsolateral prefrontal cortex. To perform this task, it is not enough to simply keep a representation of recently presented items in short-term memory; the working memory buffer also needs to be updated continuously to keep track of what the current stimulus must be compared to. In other words, the participant needs to both maintain and manipulate information in working memory [39]. Our primary interest was in whether we could detect differences in the task performance after any of the active tPBM sessions (c-tPBM and p-tPBM), relative to the performance after sham (s-tPBM). Accordingly, participants were asked to perform 2-back task after each active/sham session. To reduce burden on the participants, we administered the 2-back task before tPBM only at the first visit (before c-tPBM). We decided to present only results comparing n-backs pre- and post-continuous tPBM, based on the DCS and EEG results across the three t-PBM sessions; since only continuous tPBM produced neuromodulation (see Results section).

To examine how well each participant performed in the 2-back task, we computed both the accuracy in detecting the target letters (2-back hits and the number of false positive responses) and reaction times.

Statistical analysis

Nonparametric Friedman’s test was performed to compare the effect of different light treatments on EEG with EEG channels treated as repeated measures. Similar to the parametric repeated measures ANOVA, this test is used to detect differences in treatments across multiple measurements [40, 41]. Because differences were widespread across the scalp, we also used parametric paired t-tests applied to EEG averaged across channels for comparison. Paired t-tests were also used to analyze DCS recordings and 2-back responses. Due to the pilot nature of this study, since we recruited a small number of participants, all statistics are reported including those at 0.05 level of significance (no correction for multiple comparisons). Nonetheless, we note that a number of tests would remain significant with Bonferroni correction.

RESULTS

Table 2 describes the demographic characteristics of the sample. The ten subjects (six females/four males) had a mean age of 28.6±12.9 (SD). The tPBM sessions were well tolerated without any serious adverse event. As measured by the SAFTEE scale, administered one week after each session, three subjects developed one or more adverse events: one subject (#1) experienced ‘feeling drowsy / sleepy’ and ‘weakness / fatigue’, one subject (#5) experienced ‘trouble concentrating’ and one subject (#8) reported ‘blurred vision’ and ‘nausea/vomiting’. Of note, our study included multimodal integration of several technologies and techniques, including EEG and DCS recording, tPBM delivery, 2-back cognitive, and self-report assessments, which presented several logistical challenges. This complex design, while offering new, multimodal insights on the neurophysiology of tPBM, led to the loss of data in some recording sessions, due to insufficient quality or to participants’ time constraints. We report the number of included datapoints for each comparison below. Because our data samples for EEG are small, we report results from a nonparametric Friedman’s test for repeated measures [40, 41]. We show scalp distributions of observed EEG effects for illustration, but note that in this small sample of participants, differences among EEG channels did not reach significance. Additionally, because EEG data tend to be normally distributed, we also report paired t-test results for comparison.

Table 2

Demographic characteristics on healthy subjects enrolled in the study

Subject	Age	Gender	Ethnicity	Race	Skin color
#1	56	F	Hispanic	White	5 out of 10
#2	28	F	Not Hispanic	White	2 out of 10
#3	49	M	Not Hispanic	White	1 out of 10
#4	23	F	Not Hispanic	Asian	4 out of 10
#5	23	F	Not Hispanic	White	2 out of 10
#6	24	M	Not Hispanic	White	1 out of 10
#7	20	F	Not Hispanic	Asian	2 out of 10
#8	21	F	Hispanic	More than one race	4 out of 10
#9	22	M	Not Hispanic	Asian	2 out of 10
#10	20	M	Hispanic	White	1 out of 10

Electroencephalographic activity (EEG)

Open eyes EEG recordings at rest

We observed that continuous light treatment (c-tPBM), but not pulsed light treatment (p-tPBM), significantly influenced EEG PSD. Because state changes (e.g., vigilance, drowsiness, anxiety) during rest influence PSD even within a 5 min long recording [42, 43], and because our participants remained in resting state for over 30 min directly before post-light-treatment assessment, we evaluated all resting state data collected on light-treatment days after subtracting values obtained on sham-treatment days to control for state change dynamics during the lengthy imaging/treatment sessions. We used resting state data recorded before light/sham treatment on each day as a control for any changes in EEG PSD between session days [44, 45].

Figure 3 shows the results for resting state recordings with eyes open. Results for c-tPBM treatment (n = 8) are shown in the left panel. We computed pre-light-treatment (pre-tPBM) PSD by subtracting pre-treatment recording at the sham session from pre-treatment recording at the c-tPBM session. Similarly, we computed post-light-treatment (post-tPBM) PSD by subtracting post-treatment recording at the sham session from post-treatment recording at the c-tPBM session. These PSD metrics gave us the change value that was likely due to c-tPBM treatment, while controlling for the sham effect of merely resting in a dark room with a light-treatment device applied but not active. The scalp maps on top of Fig. 3 are shown to illustrate scalp topography of PSD data averaged across participants for alpha (8–12 Hz), beta (12–30 Hz), and lower gamma (30–55 Hz) EEG recordings, with pre-tPBM PSD data in the bottom row and post-tPBM data in the top row. A widespread increase in the activity PSD is apparent particularly over frontal-central scalp regions post-tPBM, relative to pre-tPBM, especially in the gamma band. The spectrogram for these data, averaged across all EEG channels, is shown below Fig. 3, the solid lines show group average and the shaded areas shows the standard error of the mean across participants. The PSD increase in the post-tPBM relative to pre-tPBM data appears larger for higher frequencies. Statistical analyses revealed that this effect reached significance for gamma (mean change from γ-PSD _{PRE (chboxtPBMSham)} 0.05±0.38 SD to γ-PSD _{[POST (chboxtPBMSham)]} 0.29±0.31 SD; nonparametric Friedman’s test between pre- and post- with individual EEG channels as repeated measurements: Chi-sq=21.65, p < 3.27e^–06; paired t-test with data averaged across EEG channels: t = 3.02, df = 7, p < 0.02) and beta (mean change in β-PSD _{[PRE (chboxtPBMSham)]} 0.02±0.28 SD to β-PSD _{POST (chboxtPBMSham)} 0.17±0.25 SD; Chi-sq = 15.80, p < 7.03e^–05; t = 2.91, df = 7, p < 0.03) bands, and showed some significance in the alpha band (mean change from α-PSD _{PRE (chboxtPBMSham)} –0.03±0.30 SD to α-PSD _{POST (chboxtPBMSham)} 0.08±0.26 SD; Chi-sq = 7.59, p < 0.006; t = 1.40, df = 7, p > 0.2). The PSD effect was broadly distributed, and the difference between the EEG channels did not reach significance in any frequency band (Friedman’s test, p-values > 0.05). Results for pulsed-tPBM treatment (n = 9) are shown in the right panel. Again, we computed pre-tPBM PSD and post-tPBM PSD by subtracting pre- and post-treatment recordings at the sham session from pre- and post- tPBM recordings at the p-tPBM session, respectively. Even though there appears to be a frontal increase in PSD due to p-tPBM treatment, the result did not reach significance.

Fig. 3

Eyes-open resting state EEG (n = 8): spectrogram differences before (pre-LT) and after tPBM (post-LT) with continuous (CW) and pulsed light (PW). Both the effects of CW and of PW on the spectrogram were estimated after subtracting the effect of sham.

We also examined the EEG PSD changes over time while participants remained at rest during the sham visit (e.g., due to changes in vigilance, anxiety, drowsiness), and how such changes relate to the light-treatment sessions. We computed a pre- versus post- sham PSD effect by subtracting pre-treatment recording at the sham session from post-treatment recording at the sham session. Similarly, we computed pre- versus post- PSD effects for the c-tPBM and p-tPBM sessions by subtracting pre-treatment recording from post-treatment recording at each session. We observed differences among these pre/post PSD effects for gamma (nonparametric Friedman’s test among sham, c-tPBM, and p-tPBM treatments with EEG channels as repeated measurements: Chi-sq = 14.28, p < 0.0008) and beta (Chi-sq 10.82, p < 0.005) but not alpha (Chi-sq 1.68, p > 0.43) bands. The follow-up comparisons showed that EEG PSD decreased more from pre- to post- treatment during the sham session than during the c-tPBM session for gamma (mean change in γ-PSD _{[POST (chboxtPBM) –PRE (chboxtPBM)]} –0.02±0.32 SD and mean change in γ-PSD _{[POST (Sham) –PRE (Sham)]} –0.26±0.24 SD; nonparametric Friedman’s test between c-tPBM and sham sessions with EEG channels as repeated measurements: Chi-sq = 25.6, p < 4.20e^–07; t = 3.60, df = 7, p < 0.009) and beta (mean change in β-PSD _{[POST (chboxtPBM) –PRE (chboxtPBM)]} –0.08±0.18 SD and mean change in β-PSD _{[POST (Sham) –PRE (Sham)]} –0.23±0.17 SD; Chi-sq = 27.00, p < 2.05e^–07; t = 3.60, df = 7, p < 0.009) bands, with some significance in the alpha band (mean change in α-PSD _{[POST (chboxtPBM) –PRE (chboxtPBM)]} –0.04±0.13 SD and mean change in α-PSD _{[POST (Sham) –PRE (Sham)]} –0.15±0.19 SD; Chi-sq=8.63, p < 0.004; t = 2.12, df = 7; p < 0.08) bands. This result indicates that the EEG PSD changed during the 30 + min period when participants remained in resting state due to factors unrelated to the light treatment (e.g., change in vigilance, anxiety, drowsiness), but that c-tPBM counteracted this effect. No significant differences were observed between EEG PSD changes during the sham session and the p-tPBM session. None of the differences between the EEG channels reached significance in any frequency band (Friedman’s test, p-values > 0.05).

Closed eyes EEG recordings at rest

Figure 4 shows the results for resting state recordings with closed eyes, which were obtained according to the same procedure as with open eyes. Results for c-tPBM treatment (n = 7) are shown in the left panel. Again, there was a widespread increase in the activity power in post-tPBM relative to pre-tPBM scans for gamma and beta bands. In the alpha band, activity power was high both pre-tPBM and post-tPBM, which would be expected as alpha is known to increase in power when participants close their eyes. The global effect is illustrated in the lower panel of the Fig. 4. Similarly to Open eyes EEG recordings at rest, the effect on PSD appears progressively larger with higher EEG frequencies. The effect reached significance in the gamma band (mean change from γ-PSD _{PRE (chboxtPBMSham)} –0.11±0.31 SD to γ-PSD _{[POST (chboxtPBMSham)]} 0.24±0.22 SD; nonparametric Friedman’s test between pre- and post- with individual EEG channels as repeated measurements: Chi-sq = 31.11, p < 2.43e^–08; t = 3.61, df = 6, p < 0.015), showed some significance in the beta band (mean change in β-PSD _{[PRE (chboxtPBMSham)]} 0.01±0.26 SD to β-PSD _{POST (chboxtPBMSham)} 0.17±0.13 SD, Chi-sq=8.82, p < 0.003; t = 1.77, df = 6, p > 0.1), but was not significant in the alpha band (mean change from α-PSD _{PRE (chboxtPBMSham)} 0.20±0.37 SD to α-PSD _{POST (chboxtPBMSham)} 0.20±0.33 SD, Chi-sq = 0.31, p > 0.57; t = 0.026, df = 6, p > 0.8). Results for pulsed-tPBM treatment (n = 8) are shown in right panel. The effect showed some significance in the gamma band (mean change from γ-PSD _{PRE (phboxtPBMSham)} –0.10±0.40 SD to γ-PSD _{[POST (phboxtPBMSham)]} 0.08±0.27 SD; nonparametric Friedman’s test between pre- and post- with individual EEG channels as repeated measurements: Chi-sq = 9.35, p < 0.003; t = 1.27, df = 7, p > 0.2) and the beta band (mean change in β-PSD _{[PRE (phboxtPBMSham)]} –0.08±0.33 SD to β-PSD _{POST (phboxtPBMSham)} 0.10±0.16 SD, Chi-sq = 4.39, p < 0.04; t = 1.5, df = 7, p > 0.1). None of the differences between the EEG channels reached significance in any frequency band (Friedman’s test, p-values > 0.05).

Fig. 4

Eyes-closed resting state EEG (n = 7): spectrogram differences before (pre-LT) and after tPBM (post-LT) with continuous (CW) and pulsed light (PW). Both the effects of CW and of PW on the spectrogram were estimated after subtracting the effect of sham.

In the analysis of differences in the over-time changes across each session among the three treatment conditions, we observed significant effects between pre/post PSD effects for gamma (nonparametric Friedman’s test among sham, c-tPBM, and p-tPBM treatments with EEG channels as repeated measurements: Chi-sq = 21.49, p < 2.16e^–05) but not beta (Chi-sq 5.00, p < 0.09) or alpha (Chi-sq 4.00, p > 0.1) bands. Additional comparisons showed differences between c-tPBM and sham sessions for the gamma band (mean change in γ-PSD _{[POST (chboxtPBM) –PRE (chboxtPBM)]} 0.09±0.28 SD and mean change in γ-PSD _{[POST (Sham) –PRE (Sham)]} –0.26±0.34 SD; nonparametric Friedman’s test between c-tPBM and sham treatments with EEG channels as repeated measurements: Chi-sq = 33.31, p < 7.87e^–09; t = 3.27, df = 6, p < 0.02), and some significance in the beta band (mean change in β-PSD _{[POST (chboxtPBM) –PRE (chboxtPBM)]} 0.03±0.22 SD and mean change in β-PSD _{[POST (Sham) –PRE (Sham)]} –0.13±0.23 SD; Chi-sq = 7.45, p < 0.007; t = 1.87, p > 0.1), but not in the alpha band (mean change in α-PSD _{[POST (chboxtPBM) –PRE (chboxtPBM)]} 0.21±0.34 SD and mean change in α-PSD _{[POST (Sham) –PRE (Sham)]} 0.21±0.31 SD; Chi-sq = 0.56, p < 0.45; t = 0.03, p > 0.98). This result again confirmed that the EEG PSD changed during the 30 + min period when participants remained in resting state due to factors unrelated to the light treatment, and that c-tPBM counteracted this effect. Some significance was also observed for differences in EEG PSD changes between the sham session and the p-tPBM session in the gamma band (mean change in γ-PSD _{[POST (phboxtPBM) –PRE (phboxtPBM)]} –0.03±0.25 SD and mean change in γ-PSD _{[POST (Sham) –PRE (Sham)]} -0.22±0.39 SD; nonparametric Friedman’s test between c-tPBM and sham treatments with EEG channels as repeated measurements: Chi-sq = 7.47, p < 0.007; t = 1.4, p > 0.2). None of the differences between the EEG channels reached significance in any frequency band (Friedman’s test, p-values > 0.05).

EEG recordings during the cognitive task

Light-induced changes in EEG, recorded while participants performed the 2-back task, were similar to the changes described for the resting scans: a significant increase in power density of fast oscillations was confirmed with c-tPBM. We used a two-step approach to the analysis of EEG during the 2-back task. First (step 1), we directly compared EEG PSD obtained after each active treatment session to EEG PSD obtained after the sham session (always recorded during the 2-back task). In Fig. 5, these differences (n = 5) are illustrated in the scalp maps on the top row, showing subtractions of post- c-tPBM PSD minus post-sham PSD in the left panel and subtractions of post-p-tPBM PSD minus post-sham PSD in the right panel. During 2-back performance, an increase in the post-c-tPBM PSD relative to post-sham PSD was most evident over the frontal-temporal scalp sites. In the omnibus analysis with all three treatment types, we observed differences in gamma (nonparametric Friedman’s test among sham, c-tPBM, and p-tPBM treatments with EEG channels as repeated measurements: Chi-sq=8.25, p < 0.02) and beta (Chi-sq = 6.63, p < 0.04) bands. The overall increase in post-c-tPBM PSD, relative to post-sham, was significant in the gamma band (mean change in γ-PSD _{[POST (chboxtPBMSham)]} 0.1±0.1 SD; nonparametric Friedman’s test between c-tPBM and sham sessions with EEG channels as repeated measurements: Chi-sq = 13.82, p < 0.0002; t = 3.79, df = 4, p < 0.02). There was some significance in the beta band (mean change β-PSD _{[POST (chboxtPBMSham)]} 0.1±0.1 SD; Chi-sq = 10.98, p < 0.0009; t = 2.42, df = 4, p < 0.08). The effect was not significant for the post-p-tPBM comparison. Second (step 2), we examined if this difference between c-tPBM and sham visits was over and above any difference in the general state of the brain, which might have been session-dependent. We showed the post-c-tPBM versus post-sham effect (recorded during the 2-back task) was larger than the pre-treatment difference in the PSD during the baseline resting scans with eyes open between these sessions (shown in bottom-row scalp maps in Fig. 5). The overall increase (n = 4 –in the bottom left panel) was larger in the post-c-tPBM versus post-sham (during the 2-back task) relative to the pre-c-tPBM versus pre-sham rest subtraction in the gamma (mean change in γ-PSD _{[2hboxBack (chboxtPBMSham) –PRE (chboxtPBMSham)]} 0.3±0.2 SD; nonparametric Friedman’s test between post-2-back and pre-rest with EEG channels as repeated measurements: Chi-sq = 23.41, p < 1.31e^–06, t = 3.20, df = 3, p < 0.05) and beta (mean change in β -PSD _{[2hboxBack (chboxtPBMSham) –PRE (chboxtPBMSham)]} 0.2±0.1 SD; Chi-sq = 27.6, p < 1.49e^–07; t = 4.30, df = 3, p < 0.03) bands. This difference was not observed for the pulsed-tPBM condition. Despite these results are obtained in a small sample of participants, there is consistency in the continuous-tPBM treatment response across the subjects, which is reflected in statistical significance. Even though the n-back performance of participants suggests some practice effect as performance slightly improves from session 1 (c-tPBM) to session 2 (sham) to session 3 (p-tPBM), the increase in the brain activity (EEG) in the post-c-tPBM relative to post-sham condition occurred at session 1 relative to session 2; no improvements were noticed in the post-p-tPBM (session 3) relative to post sham condition (session 2). None of the differences between the EEG channels reached significance in any frequency band (Friedman’s test, p-values > 0.05).

Fig. 5

Eyes-open EEG during n-back task (n = 5): spectrogram differences before light therapy (pre-LT at rest) and after light therapy (post-LT during n-back), with continuous (CW) and pulsed light (PW). Both the effects of CW and of PW on the spectrogram were estimated after subtracting the effect of sham.

Cerebral blood flow (DCS)

In our participants’ sample (n = 10), there were no statistically significant differences in the changes of DCS signals at each session (c-tPBM, s-tPBM, p-tPBM), for 2.5 cm and 5 mm separation and when subtracting the short separation to discount the contribution of superficial tissues to the measure of blood flow.

Cognitive tests (n-Back)

We analyzed performance quality during the 2-back task. Figure 6 summarizes the levels of performance on each 2-back task at session 1 (n = 6, c-tPBM). The performance was not significantly changed from pre- to post- c-tPBM treatment both in terms of accuracy in detecting targets (mean change –4.49±20.84 %, t = –0.53, df = 6, p > 0.5) or reaction times [from 711.4±116.2 SD to 733.9±61.1 SD (sec), t = 0.45, df = 5, p > 0.5].

Fig. 6

Performance on 2-back task (n = 6): accuracy in target detection (%) and number of false positives (#) before and after tPBM with continuous light.

DISCUSSION

Our study presents two important findings, worthy of being discussed: 1) tPBM potentiated brain fast oscillations, and 2) tPBM-induced enhancement of brain fast oscillation was dose-dependent. These findings have broad implications for the field of neuromodulation with tPBM.

tPBM potentiation of brain fast oscillations

Delivered by a four LED clusters device in a single irradiation of 20 min to the forebrain, c-tPBM led to a significant increase in the high frequency neural activity, in the gamma and beta bands, believed to support higher-order cognition. During the resting state scans with eyes either open or closed, this increase was over broad frontal-temporal regions of the scalp, showing scalp topography common of the neural activity in these bands [46]. During the 2-back working memory task, the increase also reached significance in the gamma and beta bands and was primarily over the frontal sites, likely because this task engages the prefrontal cortex [47].

Our results are consistent with several previous studies of tPBM effects on the brain function. A recent investigation in healthy older adults at risk of cognitive decline reported an increase in the gamma EEG power, and a smaller increase in the beta power, over bilateral temporal scalp regions, during laser tPBM (continuous wave, 1064nm, 250 mW/cm², 137.5 J/cm², 13.6 cm² x 1 site) [48]. In the gamma band, this power enhancement persisted after tPBM. Another recent study with slightly different parameters of laser tPBM (continuous wave, 1064 nm, 9.72 J/cm² per minute, and 106.94 J/cm² over 11 min), applied to the right forehead of healthy participants, showed an enhancement effect on the neural activity in high frequency bands (alpha and beta) [49]. Whereas both these and the present study suggest a potential neuromodulatory effect, their differences in the analysis approach can explain some discrepancies in the findings. Our study which observed the tPBM effect on gamma and beta, not only compared the active tPBM to sham (similarly to prior studies), but also measured the change from pre- to post- tPBM (within session) thus controlling for any additional differences between sessions (e.g., in the quality of EEG setup). The study by Vargas and colleagues took a within-session comparison approach without sham-control, which cannot partial out if participants were growing drowsier as they rested during tPBM. The study by Wang and colleagues did compare tPBM and sham EEG measurements; however, measurements were collected on different days without accounting for potential between-session differences, such as in the EEG set-up, which gamma might be sensitive to. Similarly, a recent randomized, sham-controlled, double-blinded study conducted by Zomorrodi et al. demonstrated a significant increase in alpha, beta, and gamma power and a reduction in delta and theta power after a single session of pulsed (40 Hz) tPBM at 810 nm wavelength during resting state. In this latter study both sham and active mode caused a power increase in all frequency bands comparing post to pre-stimulation, however the active tPBM facilitated a power increase in alpha, beta, and gamma oscillations compared to sham [50]. Our study both replicates and adds to the findings of Zomorrodi et al., as it extends the effects of transcranial NIR on brain oscillations to the CW mode (instead of PW at 40 Hz), to an irradiance of ∼55 mW/cm² (lower than 100 mW/cm²) and to wavelength of 830 nm (instead of 810 nm).

Ability of tPBM to influence gamma band neural activity in the human brain may have important clinical significance. This type of brain activity has been linked to performance on complex and attention-demanding tasks [51], and was implicated in support of diverse sensory and cognitive processes, including perceptual processing, object representations, visual awareness, and language [52 –59]. Furthermore, it was found that the presentation of a previously learned stimulus evokes a stronger neural response in the gamma band than that of a new stimulus, suggesting a core role gamma activity may play in the mechanism of memory [57]. Remarkably, clinical states, such as major depressive disorder, mild cognitive impairment, dementia due to AD and Down syndrome may alter gamma band brain activity [60 –62].

tPBM dose-dependent effect on fast brain oscillations

Most of the effects of tPBM on high frequency brain oscillations disappeared when pulsed light was used (p-tPBM). We interpreted this change as a dose-dependent effect, either due to the pulsed nature of tPBM (10 Hz, 33%duty cycle), i.e., the on-off cycling of the light delivered affected the response, or due to the overall lower total energy delivered—0.8 kJ in pulsed-tPBM compared to 2.3 KJ in continuous-tPBM. A dose-dependent effect of laser tPBM with continuous light on behavior, accompanied by changes in the electrocorticogram spectra, was previously demonstrated in an animal model of depression [63]. After pharmacological induction of depression, rats presented with both reduced survival behaviors on a forced swim test and reduced gamma-beta power at intracranial EEG. While laser tPBM led to normalization of behavior and electrocorticogram, only its lowest dose was effective; the middle dose produced no behavioral changes and the highest worsening of depressive behaviors. Observing a dose-dependent effect of tPBM on the neural activity in humans has great scientific significance as this suggests a causal link. It will be important in future research to further establish if different doses of tPBM influence the degree of the evoked neural changes. Characterizing the dose-response curve also has clinical significance, enabling the delivery of effective tPBM treatments. Our data suggests tPBM thresholds of insufficient (≤0.8 kJ) and likely effective (≥2.3 kJ) total energy for a single session with low irradiance in healthy subjects.

Lack of tPBM effect on cognition and CBF

Large enhancement on the resting state gamma and beta power, which was consistent across participants, suggests promise of tPBM for treatment of cognitive deficits. Intriguingly, the resting state gamma power has previously been linked to language and cognition during early development and in clinical conditions such as schizophrenia. Prior studies have also showed the effects of a single session of tPBM on cognition. For instance, Barrett and Gonzalez-Lima (2013) found that measures of attention (psychomotor vigilance task, PVT) and memory (delayed match to sample, DMS) improved in response to tPBM. In our study, we employed the n-back working memory task, which like PVT and DMS depends on the function of the prefrontal cortex. Unfortunately, in this pilot study the sample size was small, which is a likely explanation for why we failed to observe improvement on this task due to tPBM. Our very low accuracy rate at 2-back (67–71%), in a sample of mostly high-functioning and healthy subjects, also suggests that subjects might have underperformed, therefore jeopardizing the reliability of the task. Finally, it is possible that tPBM with low irradiance might require multiple sessions and longer follow-up to demonstrate an effect on cognition.

It was surprising that our study found no significant changes in the CBF, as indexed by DCS recorded at 5 mm and at 2.5 cm source-detector separation in response to both active tPBM sessions, compared to the sham mode. In fact, to this date, one of the most validated and replicated neurophysiological findings in humans treated with tPBM is the increase in CBF [11], although, prior studies, demonstrating an effect of tPBM on CBF, typically used devices with higher irradiance in the order of 75–250 mW/cm² [64 –66]. Furthermore, DCS, the technology used for CBF detection, is a validated technique with numerous studies supporting its use in humans. We believe that our measure of CBF was not sensitive enough and consequently inaccurate. There are several reasons for our current interpretation of our DCS findings: 1) Cerebrovascular coupling is expected whenever cortical brain areas are activated, no matter the mechanism of neuromodulation. In other terms, even if minimal photochemical reactions occurred in cortical mitochondria, still neuronal activity would consequently evoke vascular changes. 2) While every effort was made to place the DCS detectors in close proximity to the tPBM NIR source (1-2 mm), the light was distributed over four sources, of which only one was chosen for adjacent placement of DCS (therefore reducing the NIR dose in proximity of the detector). Moreover, the complexity of our assembly—including four rigid NIR light sources, a full EEG cap, a protective eyewear and DCS sensors, as well as DCS fibers—is likely to have affected our measure by either shielding the NIR light close to the DCS sensor or by distancing the DCS sensor. 3) The DCS sensor measures both short (5 mm) and long (2.5 cm) separation effects, which are expected to reflect blood flow changes respectively within the superficial tissues (skin) and within superficial and deep tissues combined (including brain). Strikingly, in our measures no changes in the short-separation DCS signal were recorded. While we were not interested in the effects on skin vasculature, tPBM NIR has intrinsic vasomotor effects which should have resulted in a detectable DCS signal. 4) Other groups have reported on tPBM effects on both brain oscillations and CBF, when using a total energy of NIR equal or less than one quarter of our study dose [65]. Of note, in our study, two of the light sources (F3 and F4) were frequently placed above hair and this might have resulted in an overestimate of the light energy deposition. 5) Finally, it is unlikely that the small sample size might explain the lack of tPBM effect on CBF, since this effect size was negligible.

Strengths and limitations

Several limitations should be acknowledged for our study: 1) This pilot study is based on a small sample of healthy participants; therefore, despite the strong effect on brain oscillations, the current results can only be considered preliminary. It is unproven whether the effects on brain oscillation would generalize to a wider population, including patients suffering brain disorders, such as AD. 2) Similarly, due to the small study sample, we cannot judge conclusively the potential effects of low dose p-tPBM on brain oscillations; it is striking that large effects were observed for continuous light tPBM, suggesting that, at the very least, effect sizes of pulsed and continuous light dosages were quite different. 3) In each session of our pilot study we recorded EEG during eyes open resting state, eyes closed resting state and during the cognitive task with no sequence randomization. 4) It is also noteworthy that we only used a 20-channel EEG recording; it may be necessary to record with higher EEG sensor density to adequately characterize tPBM effects on brain oscillations (e.g., the effect of the p-tPBM over the anterior scalp regions). 5) Because the dose of tPBM was much lower in pulse mode, this prevented from testing the impact of the pulsing of the light on brain oscillations. In the future, studies might adequately compare continuous and pulsed waves effects, by matching average irradiance and total energy per session. 6) An additional limitation, also related to the study device, is the use of four clusters of LED sources mounted on a rigid frame. The rigid frame prevented the repositioning of two sources (on F3, F4) below the hairline; as already mentioned, it is therefore likely that the actual dose of NIR was less, due to the shielding effect of hair in this young cohort. 7) Our measure of the effects on cognitive function appears to be affected by the subjects’ poor effort at the n-back task, in addition to the small sample size. The strengths of our approach are: 1) In our study, tPBM delivered with an LED device was well tolerated with no serious adverse events. 2) LED tPBM has several advantages over laser, such as low or no risk of retina injury, lower cost, and potential self-administration. These features are decisive in terms of broadening of the clinical use of tPBM, as they offer a considerable progress in safety and comfort, compared to prior studies using laser sources. 3) The multimodal neurophysiological testing applied in our study offered a rare opportunity to understand the effects of tPBM in humans. 4) The single-blind, sham-controlled design, with careful match of all visible and auditory outputs of the tPBM device between active and sham sessions contributes to rigorous testing of our hypotheses. Of note, only two investigators (PC and EB) were always aware of the exact mode of tPBM applied at each session (c-tPBM, sham, p-tPBM), while investigators involved with neurophysiological testing remained mostly blind.

Considerations on tPBM dosimetry, penetration, and mechanism of action

A final consideration should be made to the postulated mechanism of action of tPBM, responsible for the potentiation of fast brain oscillations. While most human studies on tPBM do not assess the expected NIR penetration based on study parameters, on the sample characteristics, on the position of the light source and of the target area (through cadaver models or through specific simulations), the uncertainty over photon deposition at cortical level could be considered a limitation of our study. Henderson and Morries (2019) have extensively reviewed the literature on penetration of the light through human living tissues and in cadaver models and they concluded that there are scenarios when the expected deposition of photons to the brain is negligible [4]. They also postulated that in these cases the observed neurophysiological or therapeutic effects of tPBM are likely to be related to the systemic effects of NIR such as anti-inflammatory and anti-oxidant effects [4]. Our group is familiar with the therapeutic effects on brain disorders of systemic PBM [67]. As a team, we have speculated that in addition to direct effects onto the brain, associated with photon deposition to the brain, and in addition to indirect, systemic effects, likely associated with the irradiation of blood [68], there might be a third potential mechanism related to indirect but local effects. Although, we have no proof to offer, we speculated that tPBM might induce very weak skin currents which might induce changes in neurophysiology such as the ones demonstrated in our experiment, even in the absence of actual deposition of NIR onto the brain. The effects on brain oscillations of very weak skin currents applied to the forehead, and their potential therapeutic effects, are well-known [69, 70]. Some of these modalities of neuromodulation are transcranial DCS, transcranial alternating current stimulation, and cranial electrotherapy stimulation [71]. While we expected in our experiment a direct effect of NIR light onto brain function, associated with photon deposition onto the brain, we acknowledge that this might not be the case and either an indirect systemic and/or an indirect local effect might be solely responsible for the observed changes in fast brain oscillations. In favor of a direct effect is that our team members previously reported that a fluence of 0.3 J/cm² (810 nm) at the target tissue was indeed effective in modulating neuronal metabolism (e.g., ATP production) and mitochondrial function (e.g., mitochondrial membrane potential) [72]. Also, our group has shown that light deposition with the current study parameters (NIR in CW mode) is expected to achieve at least 0.3 J/cm² at the cortical level [73]. We also demonstrated in a follow up study that penetration of transcranial NIR light is significantly greater in young adults, compared to middle-aged and older adults; of note, in our current study cohort, all but two study subjects were in their twenties [74]. According to these data, our CW parameters of tPBM could have been barely sufficient to produce neurophysiological effects based on the actual penetration of the NIR light to the brain surface. However, a case could be made for the opposite, in line with the work of Henderson and Morries (2019), penetration of NIR light through scalp and skull can be as low as 1-2%of incident light and can even be none, depending on selected parameters [4]. Interestingly, based on the work of Lychagov and colleagues (2006) and of Tedford and colleagues (2015), even when using laser sources and high power of NIR light there is remarkable interindividual variability in light penetration, which might render direct effects onto the brain questionable at times [75, 76]. Lastly, if we consider 0.9 J/cm² as the threshold of photon deposition at target tissue, necessary to achieve a direct effect [76], this would be above the average expected deposition in our study; therefore, solely indirect effects should be postulated for the NIR mechanism of action. One additional level of complexity is that there is extensive variability of the expected deposition of NIR light within any target area of the brain, such a Brodmann’s area of the cortex. For instance, the upper quartile of a given area (based on light deposition), potentially receives 3-fold higher fluence than the overall average for the same brain area [73]. Depending on the proportion of a given brain area needed to be modulated in order to trigger macroscopic neurophysiological changes, the upper quartile fluence at target tissue rather than the average fluence could be considered to predict direct effects of NIR on brain function.

CONCLUSIONS

We observed a significant and large enhancement of the power spectral density of neural activity in the gamma-band in response to c-tPBM with NIR. This result is consistent with our hypothesis that tPBM influences high-frequency synchronized brain activity in the gamma band. By modulating the brain gamma activity, linked to higher-order cognition, tPBM might have a promise as a procognitive therapy, and more specifically as an intervention for AD. It is encouraging that a large and significant increase in EEG metrics was found at rest with eyes open and closed and during cognitive challenge, despite the small sample of participants.

Footnotes

ACKNOWLEDGMENTS

The authors acknowledge the help of Arianna Riccio for drawing the depiction of our Transcranial Photobiomodulation device (TPBM-1000), diffuse correlation spectroscopy, and electroencephalogram systems ().

None of the funding sources had any involvement neither in the study design, in the collection, analysis, interpretation of study data [but in the writing of the report, Luis DeTaboada offered his comments to the final manuscript], nor in the decision to submit the article for publication.

LiteCure LLC provided all funding to support this study. There is no associated grant number to provide.

Authors’ disclosures available online ().

References

Cassano

, Petrie

, Mischoulon

, Cusin

, Katnani

, Yeung

, De Taboada

, Archibald

, Bui

, Baer

, Chang

, Chen

, Pedrelli

, Fisher

, Farabaugh

, Hamblin

, Alpert

, Fava

, Iosifescu

(2018) Transcranial photobiomodulation for the treatment of major depressive disorder. The ELATED-2 Pilot Trial. Photomed Laser Surg 36, 634–646.

Barrett

, Gonzalez-Lima

(2013) Transcranial infrared laser stimulation produces beneficial cognitive and emotional effects in humans. Neuroscience 230, 13–23.

Blanco

, Maddox

, Gonzalez-Lima

(2017) Improving executive function using transcranial infrared laser stimulation. J Neuropsychol 11, 14–25.

Henderson

, Morries

(2019) Near-infrared photonic energy penetration–principles and practice. In Photo biomodulation in the Brain, Hamblin M, Huang Y-Y, eds. Academic Press, pp. 67–88.

Hennessy

, Hamblin

(2017) Photobiomodulation and the brain: A new paradigm. J Opt 19, 013003.

Mochizuki-Oda

, Kataoka

, Cui

, Yamada

, Heya

, Awazu

(2002) Effects of near-infra-red laser irradiation on adenosine triphosphate and adenosine diphosphate contents of rat brain tissue. Neurosci Lett 323, 207–210.

Wong-Riley

, Liang

, Eells

, Chance

, Henry

, Buchmann

, Kane

, Whelan

(2005) Photobiomodulation directly benefits primary neurons functionally inactivated by toxins: Role of cytochrome c oxidase. J Biol Chem 280, 4761–4771.

Marazziti

, Baroni

, Picchetti

, Landi

, Silvestri

, Vatteroni

, Catena Dell’Osso

(2012) Psychiatric disorders and mitochondrial dysfunctions. Eur Rev Med Pharmacol Sci 16, 270–275.

Uozumi

, Nawashiro

, Sato

, Kawauchi

, Shima

, Kikuchi

(2010) Targeted increase in cerebral blood flow by transcranial near-infrared laser irradiation. Lasers Surg Med 42, 566–576.

10.

Tian

, Hase

, Gonzalez-Lima

, Liu

(2016) Transcranial laser stimulation improves human cerebral oxygenation. Lasers Surg Med 48, 343–349.

11.

Salgado

, Zangaro

, Parreira

, Kerppers II (2015) The effects of transcranial LED therapy (TCLT) on cerebral blood flow in the elderly women. Lasers Med Sci 30, 339–346.

12.

Wang

, Tian

, Reddy

, Nalawade

, Barrett

, Gonzalez-Lima

, Liu

(2017) Up-regulation of cerebral cytochrome-c-oxidase and hemodynamics by transcranial infrared laser stimulation: A broadband near-infrared spectroscopy study. J Cereb Blood Flow Metab 37, 3789–3802.

13.

de la Torre

(2016) Cerebral perfusion enhancing interventions: A new strategy for the prevention of Alzheimer dementia. Brain Pathol 26, 618–631.

14.

Galow

, Schneider

, Lewen

, Ta

T-T

, Papageorgiou

, Kann

(2014) Energy substrates that fuel fast neuronal network oscillations. Front Neurosci 8, 398–398.

15.

Whittaker

, Turnbull

, Whittington

, Cunningham

(2011) Impaired mitochondrial function abolishes gamma oscillations in the hippocampus through an effect on fast-spiking interneurons. Brain 134, e180–e181.

16.

Buzsaki

, Wang

(2012) Mechanisms of gamma oscillations. Annu Rev Neurosci 35, 203–225.

17.

Gulyas

, Buzsaki

, Freund

, Hirase

(2006) Populations of hippocampal inhibitory neurons express different levels of cytochrome c. Eur J Neurosci 23, 2581–2594.

18.

Gross

, Schnitzler

, Timmermann

, Ploner

(2007) Gamma oscillations in human primary somatosensory cortex reflect pain perception. PLoS Biol 5, e133.

19.

Nowak

, Zich

, Stagg

(2018) Motor cortical gamma oscillations: What have we learnt and where are we headed? Curr Behav Neurosci Rep 5, 136–142.

20.

Sederberg

, Schulze-Bonhage

, Madsen

, Bromfield

, McCarthy

, Brandt

, Tully

, Kahana

(2007) Hippocampal and neocortical gamma oscillations predict memory formation in humans. Cereb Cortex 17, 1190–1196.

21.

Herrmann

, Demiralp

(2005) Human EEG gamma oscillations in neuropsychiatric disorders. Clin Neurophysiol 116, 2719–2733.

22.

Nakazono

, Jun

, Blurton-Jones

, Green

, Igarashi

(2018) Gamma oscillations in the entorhinal-hippocampal circuit underlying memory and dementia. Neurosci Res 129, 40–46.

23.

Iaccarino

, Singer

, Martorell

, Rudenko

, Gao

, Gillingham

, Mathys

, Seo

, Kritskiy

, Abdurrob

, Adaikkan

, Canter

, Rueda

, Brown

, Boyden

, Tsai

(2016) Gamma frequency entrainment attenuates amyloid load and modifies microglia. Nature 540, 230–235.

24.

van der Hiele

, Vein

, van der Welle

, van der Grond

, Westendorp

, Bollen

, van Buchem

, van Dijk

, Middelkoop

(2007) EEG and MRI correlates of mild cognitive impairment and Alzheimer’s disease. Neurobiol Aging 28, 1322–1329.

25.

Czigler

, Csikos

, Hidasi

, Anna Gaal

, Csibri

, Kiss

, Salacz

, Molnar

(2008) Quantitative EEG in early Alzheimer’s disease patients - power spectrum and complexity features. Int J Psychophysiol 68, 75–80.

26.

Moretti

, Pievani

, Geroldi

, Binetti

, Zanetti

, Rossini

, Frisoni

(2010) EEG markers discriminate among different subgroup of patients with mild cognitive impairment. Am J Alzheimers Dis Other Demen 25, 58–73.

27.

Michels

, Muthuraman

, Anwar

, Kollias

, Leh

, Riese

, Unschuld

, Siniatchkin

, Gietl

, Hock

(2017) Changes of functional and directed resting-state connectivity are associated with neuronal oscillations, ApoE genotype and amyloid deposition in mild cognitive impairment. Front Aging Neurosci 9, 304.

28.

Lapchak

, De Taboada

(2010) Transcranial near infrared laser treatment (NILT) increases cortical adenosine-5’-triphosphate (ATP) content following embolic strokes in rabbits. Brain Res 1306, 100–105.

29.

Brondon

, Stadler

, Lanzafame

(2009) Pulsing influences photoradiation outcomes in cell culture. Lasers Surg Med 41, 222–226.

30.

Lapchak

, Salgado

, Chao

, Zivin

(2007) Transcranial near-infrared light therapy improves motor function following embolic strokes in rabbits: An extended therapeutic window study using continuous and pulse frequency delivery modes. Neuroscience 148, 907–914.

31.

Rosazza

, Minati

(2011) Resting-state brain networks: Literature review and clinical applications. Neurol Sci 32, 773–785.

32.

Bridwell

, Cavanagh

, Collins

AGE

, Nunez

, Srinivasan

, Stober

, Calhoun

(2018) Moving beyond ERP components: A selective review of approaches to integrate EEG and behavior. Front Hum Neurosci 12, 106.

33.

Herrmann

, Struber

, Helfrich

, Engel

(2016) EEG oscillations: From correlation to causality. Int J Psychophysiol 103, 12–21.

34.

Ward

, McLaren

, Schultz

, Chhatwal

, Boot

, Hedden

, Sperling

(2013) Daytime sleepiness is associated with decreased default mode network connectivity in both young and cognitively intact elderly subjects. Sleep 36, 1609–1615.

35.

Mesquita

, Durduran

, Yu

, Buckley

, Kim

, Zhou

, Choe

, Sunar

, Yodh

(2011) Direct measurement of tissue blood flow and metabolism with diffuse optics. Philos Trans A Math Phys Eng Sci 369, 4390–4406.

36.

Durduran

, Yodh

(2014) Diffuse correlation spectroscopy for non-invasive, micro-vascular cerebral blood flow measurement. Neuroimage 85(Pt 1), 51–63.

37.

Buckley

, Parthasarathy

, Grant

, Yodh

, Franceschini

(2014) Diffuse correlation spectroscopy for measurement of cerebral blood flow: Future prospects. Neurophotonics 1, 011009.

38.

Selb

, Wu

, Sutin

, Lin

, Farzam

, Bechek

, Shenoy

, Patel

, Boas

, Franceschini

, Rosenthal

(2018) Prolonged monitoring of cerebral blood flow and autoregulation with diffuse correlation spectroscopy in neurocritical care patients. Neurophotonics 5, 045005.

39.

Chai

, Abd

Hamid AI

, Abdullah

(2018) Working memory from the psychological and neurosciences perspectives: A review. Front Psychol 9, 401.

40.

Hogg

(1987) Engineering Statistics. Macmillan Pub Co, New York.

41.

Hollander

, Wolfe

(1999) Nonparametric Statistical Methods, 3rd Edition. John Wiley & Sons, Inc., Hoboken, NJ.

42.

Batashvili

, Staples

, Baker

, Sheffield

(2019) Exploring the relationship between gamma-band activity and maths anxiety. Cogn Emot 33, 1616–1626.

43.

Guragain

, Rad

, Wang

, Verma

, Archer

, Wilson

, Tavakolian

(2019) EEG-based classification of microsleep by means of feature selection: An application in aviation. Annu Int Conf IEEE Eng Med Biol Soc 2019, 4060–4063.

44.

Keil

, Debener

, Gratton

, Junghofer

, Kappenman

, Luck

, Luu

, Miller

, Yee

(2014) Committee report: Publication guidelines and recommendations for studies using electroencephalography and magnetoencephalography. Psychophysiology 51, 1–21.

45.

Newson

, Thiagarajan

(2018) EEG frequency bands in psychiatric disorders: A review of resting state studies. Front Hum Neurosci 12, 521.

46.

Tewarie

, Hillebrand

, van Dijk

, Stam

, O’Neill

, Van Mieghem

, Meier

, Woolrich

, Morris

, Brookes

(2016) Integrating cross-frequency and within band functional networks in resting-state MEG: A multi-layer network approach. Neuroimage 142, 324–336.

47.

Spitzer

, Haegens

(2017) Beyond the status quo: A role for beta oscillations in endogenous content (re)activation. eNeuro 4, ENEURO.0170-17.2017.

48.

Vargas

, Barrett

, Saucedo

, Huang

, Abraham

, Tanaka

, Haley

, Gonzalez-Lima

(2017) Beneficial neurocognitive effects of transcranial laser in older adults. Lasers Med Sci 32, 1153–1162.

49.

Wang

, Dmochowski

, Zeng

, Kallioniemi

, Husain

, Gonzalez-Lima

, Liu

(2019) Transcranial photobiomodulation with infrared laser increases power of brain oscillations. bioRxiv, 535757.

50.

Zomorrodi

, Loheswaran

, Pushparaj

, Lim

(2019) Pulsed near infrared transcranial and intranasal photobiomodulation significantly modulates neural oscillations: A pilot exploratory study. Sci Rep 9, 6309.

51.

Tallon-Baudry

, Kreiter

, Bertrand

(1999) Sustained and transient oscillatory responses in the gamma and beta bands in a visual short-term memory task in humans. Vis Neurosci 16, 449–459.

52.

Senkowski

, Herrmann

(2002) Effects of task difficulty on evoked gamma activity and ERPs in a visual discrimination task. Clin Neurophysiol 113, 1742–1753.

53.

Herrmann

, Frund

, Lenz

(2010) Human gamma-band activity: A review on cognitive and behavioral correlates and network models. Neurosci Biobehav Rev 34, 981–992.

54.

Tallon-Baudry

, Bertrand

(1999) Oscillatory gamma activity in humans and its role in object representation. Trends Cogn Sci 3, 151–162.

55.

Guntekin

, Basar

(2014) A review of brain oscillations in perception of faces and emotional pictures. Neuropsychologia 58, 33–51.

56.

Busch

, Herrmann

, Muller

, Lenz

, Gruber

(2006) A cross-laboratory study of event-related gamma activity in a standard object recognition paradigm. Neuroimage 33, 1169–1177.

57.

Busch

, Debener

, Kranczioch

, Engel

, Herrmann

(2004) Size matters: Effects of stimulus size, duration and eccentricity on the visual gamma-band response. Clin Neurophysiol 115, 1810–1820.

58.

Gruber

, Tsivilis

, Montaldi

, Muller

(2004) Induced gamma band responses: An early marker of memory encoding and retrieval. Neuroreport 15, 1837–1841.

59.

Keil

, Muller

, Ray

, Gruber

, Elbert

(1999) Human gamma band activity and perception of a gestalt. J Neurosci 19, 7152–7161.

60.

Basar

, Demiralp

, Schurmann

, Basar-Eroglu

, Ademoglu

(1999) Oscillatory brain dynamics, wavelet analysis, and cognition. Brain Lang 66, 146–183.

61.

Basar

, Basar-Eroglu

, Guntekin

, Yener

(2013) Brain’s alpha, beta, gamma, delta, and theta oscillations in neuropsychiatric diseases: Proposal for biomarker strategies. Suppl Clin Neurophysiol 62, 19–54.

62.

Roh

, Park

(2016) Brain multimodality monitoring: Updated perspectives. Curr Neurol Neurosci Rep 16, 56.

63.

Mohammed

(2016) Transcranial low-level infrared laser irradiation ameliorates depression induced by reserpine in rats. Lasers Med Sci 31, 1651–1656.

64.

Pruitt

, Wang

, Wu

, Kallioniemi

, Husain

, Liu

(2020) Transcranial photobiomodulation (tPBM) with 1,064-nm laser to improve cerebral metabolism of the human brain. Lasers Surg Med 52, 807–813.

65.

Chao

(2019) Effects of home photobiomodulation treatments on cognitive and behavioral function, cerebral perfusion, and resting-state functional connectivity in patients with dementia: A pilot trial. Photobiomodul Photomed Laser Surg 37, 133–141.

66.

Holmes

, Barrett

, Saucedo

, O’Connor

, Liu

, Gonzalez-Lima

(2019) Cognitive enhancement by transcranial photobiomodulation is associated with cerebrovascular oxygenation of the prefrontal cortex. Front Neurosci 13, 1129.

67.

Caldieraro

, Cassano

(2019) Transcranial and systemic photobiomodulation for major depressive disorder: A systematic review of efficacy, tolerability and biological mechanisms. J Affect Disord 243, 262–273.

68.

Song

, Hu

, Yu

, Wang

, Zhao

, Korngold

, Zhao

(2020) Existence of circulating mitochondria in human and animal peripheral blood. Int J Mol Sci 21, 2122.

69.

Del Felice

, Castiglia

, Formaggio

, Cattelan

, Scarpa

, Manganotti

, Tenconi

, Masiero

(2019) Personalized transcranial alternating current stimulation (tACS) and physical therapy to treat motor and cognitive symptoms in Parkinson’s disease: A randomized cross-over trial. Neuroimage Clin 22, 101768.

70.

Kim

, Jang

, Roh

, Kim

(2020) A direct comparison of the electrophysiological effects of transcranial direct and alternating current stimulation in healthy subjects. Brain Res 1747, 147065.

71.

Guleyupoglu

, Schestatsky

, Edwards

, Fregni

, Bikson

(2013) Classification of methods in transcranial electrical stimulation (tES) and evolving strategy from historical approaches to contemporary innovations. J Neurosci Methods 219, 297–311.

72.

Sharma

, Kharkwal

, Sajo

, Huang

, De Taboada

, McCarthy

, Hamblin

(2011) Dose response effects of 810nm laser light on mouse primary cortical neurons. Lasers Surg Med 43, 851–859.

73.

Cassano

, Tran

, Katnani

, Bleier

, Hamblin

, Yuan

, Fang

(2019) Selective photobiomodulation for emotion regulation: Model-based dosimetry study. Neurophotonics 6, 015004.

74.

Yuan

, Cassano

, Pias

, Fang

(2020) Transcranial photobiomodulation with near-infrared light from childhood to elderliness: Simulation of dosimetry. Neurophotonics 7, 015009.

75.

Tedford

, DeLapp

, Jacques

, Anders

(2015) Quantitative analysis of transcranial and intraparenchymal light penetration in human cadaver brain tissue. Lasers Surg Med 47, 312–322.

76.

Lychagov

, Tuchin

, Vilensky

, Reznik

, Ichim

, De Taboada

(2006) Experimental study of NIR transmittance of the human skull. Proceedings Volume 6085, Complex Dynamics and Fluctuations in Biomedical Photonics III; 60850T (2006) https://doi.org/10.1117/12.650116.