Efficacy of Transcranial Photobiomodulation on Depressive Symptoms: A Meta-Analysis

Abstract

Background:

Transcranial photobiomodulation (tPBM) is a novel, noninvasive, device-based intervention, which has been tested as a possible treatment for various neurological and psychiatric conditions. Recently, it has been investigated as an innovative treatment for major depressive disorder (MDD). There have been several animal and clinical studies that evaluated the underlying mechanism and the efficacy of its antidepressant effects, but results have been conflicting.

Objective:

Thus, we conducted the first meta-analysis on effects of tPBM on depressive symptoms.

Materials and methods:

Thirty original articles on tPBM were retrieved, eight of them met criteria for inclusion to a random effects meta-analysis.

Results:

tPBM appeared effective in decreasing depressive symptom severity regardless of diagnosis (Hedges' g = 1.415, p < 0.001, k = 8), but a significant heterogeneity has been found. The meta-analysis of single-arm studies (baseline to endpoint changes) limited to participants with MDD has supported the significant effect of tPBM in reducing the depression severity, without a significant heterogeneity (Hedges' g = 1.142, 95% confidence interval = 0.780–1.504, z = 6.19, p < 0.001, k = 5). However, the meta-analysis of the few double-blind, sham-controlled studies in MDD has not supported the statistically significant superiority of tPBM over sham (Hedges' g = 0.499, p = 0.211, k = 3), although a sample size bias is likely present.

Conclusions:

Overall, this meta-analysis provides weak support for the promising role of tPBM in the treatment of depressive symptoms. Dose finding studies to determine optimal tPBM parameters followed by larger, randomized, sham-controlled studies will be needed to fully demonstrate the antidepressant efficacy of tPBM.

Introduction

Major depressive disorder (MDD) is one of the most prevalent psychiatric disorders and the third leading cause of disability worldwide.^1,2 Existing antidepressant treatments include pharmacologic interventions and psychotherapy that are proven to be safe and cost effective. However, antidepressant medications take 4–6 weeks to take a full effect and have a wide variety of side effects while having limited efficacy, and psychotherapy requires frequent sessions with specialized providers.^3–6 For patients who suffer from MDD but cannot tolerate or do not respond to these interventions, there are only limited options available such as electroconvulsive therapy (ECT), transcranial magnetic stimulation (TMS), and vagus nerve stimulation (VNS). These device-based neuromodulation methods are generally reserved only for treatment-resistant cases due to their complexity and potential risks.⁷

ECT requires multiple hospital visits and exposure to anesthesia, and a subset of patients experience transient, yet, significant adverse cognitive effects.⁸ TMS also requires multiple office visits, while VNS involves surgical device-implantation under general anesthesia. Hence, there is an unmet need for a neuromodulation modality that is safe, effective, well-tolerated, and easy to administer.

Transcranial photobiomodulation (tPBM) is a novel form of device-based neuromodulation that has suggested antidepressant effects both in animal models^9–12 and in humans.^13–19 tPBM involves applying near-infrared radiation (NIR) with wavelength of 800–1200 nm on the forehead, which has been shown to penetrate sufficiently into the cerebral cortex through the skull.^20,21 Cellular culture studies show that NIR absorbed by mitochondria causes changes in the level of expression and activity of mitochondrial enzymes, especially cytochrome c oxidase.^22,23 This has been shown to lead to increased synthesis of adenosine triphosphate^24–26 and modulate the release of reactive oxygen species and nitric oxide.^23,27 NIR also appears to reduce neuroinflammation and lesion size,²⁸ enhance cerebral blood flow,^29–31 and modulate neuronal excitability^32,33 in healthy as well as injured brains.

As such, tPBM has been investigated as a potential treatment targeting various neuropsychiatric disorders including stroke,^34,35 Alzheimer's disease,^36,37 traumatic brain injury (TBI),^14,15 and MDD.^13–19 Although the studies on antidepressant effects of tPBM had relatively small sample sizes and short duration and used different combinations of tPBM parameters, they have largely shown significant improvement of depressive symptoms in patients with MDD^{13,16,18,19,38} or MDD comorbid with other psychiatric disorders.^14,15,17 Encouraged by these findings, a few review articles^39–41 have been published recently but no meta-analysis on this burgeoning field has been performed yet.

The present study aims to fill this gap by performing a meta-analysis on the clinical studies of tPBM in human subjects with MDD and discussing the current evidence for tPBM as an effective and easy-to-administer device-based antidepressant modality.

Methods

This research was conducted in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.⁴²

Data sources and search strategy

The PubMed, EMBASE, and PsycNET search engines were examined with the Boolean processors for “photobiomodulation [Title/Abstract] AND depression [Title/Abstract],” “NIR [Title/Abstract] AND depression [Title/Abstract],” and “infrared [Title/Abstract] AND depression [Title/Abstract]” separately. During the search “species” and “article type” filters were activated to limit the results to “human” and “clinical trial” studies.

Inclusion and exclusion criteria

All available studies from the earliest date to August 2022, whose participants' depressive symptom severity was measured, were included in this meta-analysis. After reviewing the abstracts of listed articles, full texts of the selected articles were retrieved to extract changes in depressive symptom severity. This meta-analysis did not exclude the participants with concomitant treatment. Case-control studies and case series were also included in the meta-analysis. PRISMA flowchart of the article selection process is presented in Fig. 1.

FIG. 1.

PRISMA flowchart of the study selection. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Extraction of data

A database was prepared for the selected articles, recording the first author's name(s), the year, sample size, sex, concomitant medical and psychotropic medications, mean age, and paired mean differences and standard deviations of paired differences. If the original study did not report the raw values, the effect size was calculated from statistical test values (i.e., paired t-test). Consistent with the open science framework, the extracted dataset used in this meta-analysis can be found in Table 1.

Table 1.

Features of the Included Studies

Article	Subjects and comorbidity	Age	Gender	Design	No. of treatments	Depression symptom assessment	Device	Area irradiated	Wave length	Pulsed or continuous	Irradiance	Fluence	Results
Iosifescu et al.38	49 MDD completers, 18 tPBM and 31 sham (SPCD design)	37.2 ± 16.5 for tPBM and 42.8 ± 15.8	14 females (77.8% for tPBM, 21 females (67.7%) for sham	Randomized, double-blind, sham-controlled, 2 phase sequential comparison design	Twice a week for 6 weeks for Phase 1, twice a week for another 6 weeks for Phase 2; sessions 20 min	HAM-D17 at baseline, week 3, week 6, week 9, week 12; QIDS weekly for 12 weeks	Transcranial PhotoBiomodulation-1000 with LED, manufactured by LiteCure LLC, DE	Frontal pole and dorsal lateral prefrontal cortex, bilateral	830 nm	Continuous	54.8 mW/cm2	65.8 J/cm2	No significant difference between tPBM and sham
Cassano et al.19	13 MDD completers	45 ± 12.8 for tPBM group, 50.7 ± 13.3 for sham group	6 females and 4 males in tPBM group, 7 females and 4 males in sham group	Randomized, double-blind, sham-controlled, pilot trial	Twice a week for 8 weeks; sessions 25–30 min	HAM-D17 at baseline, then weekly for 8 weeks	Omnilux New U, LED by Photomedex, Inc., Montgomeryville, PA	Dorsolateral prefrontal cortex, bilateral	823 nm	Continuous	36.2 mW/cm2	Up to 65.2 J/cm2	NIR increased mean change in Ham-D scores
Cassano et al.18	4 MDD completers	47 ± 14	1 female and 3 males	Pilot, proof of concept, prospective, double-blind, randomized, crossover design	Twice a week for three weeks, 4 sites, 2 min per site	HAM-D17 at 1 week before treatment, then weekly through week 8	NeuroThera, continuously emitting GaAlAs- laser, manufactured by PhotoThera Inc.	Prefrontal cortex	808 ± 10 nm	Continuous	700 mW/cm2	84 J/cm2	50% Remission of MDD at weeks 6–7 (HAMD-17 ≤ 7)
Gabel et al.51	10 Patients with low back pain, 5 among which had concurrent MDD	60.2 ± 9.2 for PBM, 58.4 ± 11.2 for control	2 females and 3 males for PBM, 3 females and 2 males for control	Case control series, retrospective, open label	5 weekly PT sessions, 3 of last sessions included PBM treatment	ARGS, OMSQ before 1st PT session, before 1st PT + PBM session, after end of treatments	Thor-UK: DDII, LED 104, THOR Photomedicine Ltd, Amersham, Buckinghamshire, UK	12 Symmetric bilateral posterior sites (thoracic, lumbar, and thighs)	850 and 660 nm	Continuous	100 mW/cm2	3 J/cm2	PBM treatment decreased depression symptoms significantly more than control group
Henderson and Morries17	39 TBI patients who completed depression questionnaires	39.0 ± 16.1 for females, 42.1 ± 17.7 for males	20 females and 19 males	Open-label single arm, retrospective, proof-of-concept study	8–34 treatments, 30 min per session	QIDS and HAM-D at baseline and post-treatment	Class IV Lasers: LT1000, (LiteCure, Newark, DE, USA), Diowave 810 (Diowave, Riviera Beach, FL, USA), or Aspen Laser (Denver, CO, USA)	Overlying forehead and temporal regions bilaterally	810/980 nm	Continuous, sweeping	8–15 W	55–81 J/cm2	92% were responders (decrease of QIDS score ≥50% from baseline), 82% were remitters (QIDS ≤5)
Morries et al.15	10 TBI who completed depression questionnaires	Unknown	6 females and 4 males	Retrospective case series	10–20 treatments, 16–30 min per session	QIDS-SR at baseline and post-treatment, BDI also at baseline and post-treatment for 7 out of 10 patients,	Class IV lasers: LT1000 (LiteCure, Newark, DE, USA) or Diowave 810 (Diowave, Riviera Beach, FL, USA)	Bilateral frontal, bilateral frontal + left temporal, bilateral frontal + bilateral temporal	810/980 nm, one pt 810 nm only	10-Hz Pulsed, scanning	Unknown	55–81 J/cm2	BDI and QIDS-SR scores decreased from moderately depressed range to non-depressed range
Schiffer et al.13	10 MDD	35 ± 7	5 females and 5 males	Pilot study (open)	One session consisting of four 4-min treatments	HAM-D, HAM-A, PANAS at baseline and 2- and 4-week follow-up's; PANAS after each treatment	LED array, Marubeni America Corp, Santa Clara, CA	Left forehead at F3 (over dorsolateral prefrontal cortex), Right forehead at F4	810 nm	Unknown	250 mW/cm2	60 J/cm2 per site	At 2 weeks post treatment, 4 out of 10 were responders (>50% reduction in HAM-D), and 4 out of 10 were in remission (HAM-D < 8)
Naeser et al.14	11 Mild TBI who completed BDI	44.3 ± 13.7	5 females and 6 males	Pilot, open-protocol study; case series	18 Sessions over 6 weeks, 20 min per session	Baseline, 1 week, 1 and 2-month follow-up's	LED Console Units (MedX Health, Model 1100, Toronto), three LED cluster heads per unit	11 Scalp placements: midline from front-to-back hairline, and bilaterally on frontal, parietal, and temporal regions	633/870 nm	Continuous	22.2 mW/cm2 per cluster head	13 J/cm2 per cluster head placement	Trend toward significance for BDI scores 1-week post treatment

ARGS, Advise Rehab Global Scale; BDI, Beck Depression Index; HAM-A, Hamilton Anxiety rating scale; HAM-D17, Hamilton Depression rating scale, 17 items; LED, light emitting diode; MDD, Major Depressive Disorder; OMSQ, Orebro Musculoskeletal Screening Questionnaire; PANAS, Positive And Negative Affect Schedule; QIDS-SR, Quick Inventory of Depressive Symptomatology—Self Report; SPCD, Sequential Parallel Comparison Design; TBI, traumatic brain injury; tPBM, transcranial photobiomodulation.

Statistics

The exact bias-corrected Hedges' g values and standard errors were calculated as the effect size measure. The number of participants (N) and studies (k) were reported for each arm. The main random-effects meta-analysis and subgroup analyses were performed using the Hunter & Schmidt method as the tau² estimator. Heterogeneity (I²) between studies was tested by Cochrane's Q test, and the I² value was inspected. Publication bias across the studies was explored with funnel plots. Funnel plot asymmetry was evaluated via Egger's⁴³ and Begg's tests.⁴⁴ Linear nonparametric Trim & Fill method was used to crosscheck and re-evaluate the effect sizes against publication bias.⁴⁵

Effect size calculations were done using “MOTE”⁴⁶ and “esc” packages⁴⁷ in R. All analyses were performed using the “meta, version 4.18-0” package⁴⁸ in the R software.⁴⁹ R codes used in the analyses can be found in Supplementary Data S1. Forest plots, including prediction intervals, were used to summarize effect sizes.

Results

A pool of 102 participants, 59 in single-arm studies (k = 5) and 43 in sham-controlled studies (k = 3), were obtained from 8 published studies. Five studies included participants with MDD whereas three studies measured depressive symptom severity in subjects with other diagnoses. Figure 2 presents the summarized study characteristics and effect sizes. More details and characteristics of the included studies are presented in Table 1.

FIG. 2.

Forest plot of reduction in depressive symptom severity following tPBM (change from baseline to endpoint, single arm). tPBM, transcranial photobiomodulation.

Main meta-analysis

As seen in Fig. 2, tPBM significantly reduced depressive symptom severity from baseline to endpoint in our single-arm analysis pooled across all studies (k = 8) regardless of diagnosis [Hedges' g = 1.415, 95% confidence interval (CI) = 0.870–1.960, z = 5.09, p < 0.001]. However, there was a significant heterogeneity across the studies (I² = 76%, p < 0.01). To decrease heterogeneity, we then limited the analysis to studies only conducted on the participants with MDD (k = 5). Again, we found a significant effect of tPBM in decreasing the depressive symptom severity (Hedges' g = 1.142, 95% CI = 0.780–1.504, z = 6.19, p < 0.001; Fig. 3). No significant heterogeneity was observed in this subgroup (I² = 3%, p = 0.39). tPBM resulted in a greater reduction in depression symptoms compared to sham in randomized controlled studies (RCT's; k = 3).

FIG. 3.

Forest plot of reduction in depressive symptom severity following tPBM in studies with MDD subjects only (change from baseline to endpoint, single arm). MDD, major depressive disorder.

However, the standardized difference between tPBM and sham was not significantly different from 0 (Hedges' g = 0.499, 95% CI = −0.283 to 1.281, z = 1.25, p = 0.211; Fig. 4). A significant overall between-study heterogeneity was observed in RCT's (I² = 71%, p = 0.03).

FIG. 4.

Forest plot of comparative reduction in depressive symptom severity in randomized controlled studies of tPBM versus sham in MDD.

Publication bias

A visual inspection of the adjusted funnel plot has suggested a possible publication bias (Fig. 5); however, the statistical evaluation of the funnel plot returned a nonsignificant publication bias in the Begg's test (Kendall tau = 12, z = 1.36, p = 0.174) and in the Egger's test (β = 1.77, z = 1.02, p = 0.307). A nonparametric, linear trim-and-fill test did not suggest that imputation or correction on the observed effect size would be required.

FIG. 5.

Funnel plot of the pooled studies.

Discussion

tPBM is an innovative, device-based experimental treatment that involves illuminating the brain with near infrared light. Several studies have highlighted tPBM's underlying cellular mechanisms as well as its antidepressant effects; however, the results have not been consistent. This article describes the first meta-analysis on antidepressant effects of tPBM in an attempt to form consensus in the field.

The meta-analysis of single-arm studies, either regardless of the diagnosis (TBI and MDD) or limited to participants with MDD only, suggested that tPBM is effective in decreasing the depressive symptom severity. The effect of tPBM in single-arm MDD studies was robust and not associated with significant heterogeneity, which increases the relevance of these results. On the contrary, tPBM was not statistically superior to sham in the few sham-controlled studies limited to participants with MDD. This may be due to several reasons: high heterogeneity between studies, overall small number of studies available, and possible observer bias in the single-arm studies. The source of between-study heterogeneity may originate from differences in the tPBM parameters used (as specific tPBM wavelengths and intensities,⁵⁰ and areas of brain stimulation may result in different biological and clinical effects) and characteristics of study populations (e.g., participants' sex, comorbid diagnoses, and concomitant medications).

Since tPBM is a new method, the standards of its antidepressant treatment have not been established yet. For example, a consensus has not yet been formed on the optimal treatment parameters, such as the wavelength, dose (i.e., irradiance and fluence) or time interval of administration, the ideal target brain location to be exposed to NIR light, and continuous versus pulsatile laser stimulation. Thus, a wide range of these parameters were included in this meta-analysis as long as their studies met the inclusion criteria.

There was no significant publication bias or heterogeneity in the antidepressant effect size of tPBM when we limited the meta-analysis to single-arm studies only, including patients with MDD, which supports the reliability of the results. Overall, our meta-analysis provides weak evidence for tPBM's antidepressant efficacy. However, we cannot exclude the observer bias in single-arm studies and readers should avoid definitive interpretations. As such, there are only a small number of clinical studies available, all with small or medium sample size, and even fewer studies with sham control. Particularly, among the three sham-controlled studies, there is a bias in terms of the number of participants in each study (19 and 10 vs. 47). This could have contributed to the aforementioned, statistically insignificant effect of tPBM compared to sham.

We could not attempt to perform meta-regression on possible confounders such as medication use and duration, as the total number of available studies was significantly smaller than what was needed (n > 10).

Overall, tPBM would be an exciting addition to the current antidepressant treatments, as it could potentially be administered at home easily and safely. Clinical studies published so far generally support its antidepressant effects, however, with significant limitations. Larger, well-designed RCT's, testing tPBM doses previously proven in dose-finding studies, will be needed to fully demonstrate the efficacy and limitations of this treatment.

Authors' Contributions

Y.C.: conceptualization, data curation, writing—original draft preparation, writing—reviewing and editing; U.T.: data curation, formal analysis, writing—original draft preparation, and writing—reviewing and editing; D.V.I.: conceptualization, resources, and writing—reviewing and editing.

Author Disclosure Statement

Y.C. and U.T. report no conflict of interest. In the last 10 years, D.V.I. has received consulting honoraria from Alkermes, Allergan, Angelini, Axsome, Biogen, Boehringer Ingelheim, Centers for Psychiatric Excellence, Clexio, Global Medical Education, Jazz, Lundbeck, Neumora, Otsuka, Precision Neuroscience, Relmada, Sage, and Sunovion and research support (through his academic institutions) from Alkermes, AstraZeneca, Brainsway, LiteCure, NeoSync, Otsuka, Roche, and Shire.

Funding Information

This work is supported by NIH R61/R33 MH122647 (D.V.I.).

Supplementary Material

Supplementary Data S1