Many studies have reported beneficial effects from the application of red and NIR light to the body [44–46], and one group [17, 18] reported beneficial effects applying transcranial NIR to the brain in stroke patients. We embarked on this study to see if the psychological status of patients with depression might benefit from the application of NIR light to the head. Although we recruited for patients with depression, we found that 9 of those who responded also manifest an anxiety disorder by SCID, including 3 who met criteria for both generalized anxiety disorder and PTSD. Seven of these patients had also a past history of opiate abuse, 6 treated with buprenorphine and 1 with methadone. We intended this as a small pilot study for an initial evaluation of our treatment's safety (immediately and over 4-weeks) and to look for indications that it might have some efficacy immediately after each treatment, and/or at 2 and at 4-weeks post-treatment. We had 4 treatment conditions, NIR at F3 and at F4, and "no light" with the mushroom fan on at F3 and F4, as placebo conditions. We measured also rCBF by NIRS to a depth of at least 1 cm at the left and right frontal poles of the brain to see if the NIR treatment might have a definite physiological effect, and, if so, to see if the blood flow measurements might shed any information about the treatment's mechanism of action. We measured the patients' baseline HEVs because our prior studies determined that the measurement might be useful in data analysis and in guiding treatment [39, 41, 42].

Our results showed that with one 4-minute NIR treatment on each side of the head there were marked benefits in both of our primary outcome measures, the HAM-D and the HAM-A. We observed the greatest benefit at 2-weeks post-treatment for both measures. At 4 weeks both showed a statistically significant improvement over baseline but a significant decline from 2-week levels. These results should be interpreted with caution since this was not a placebo-controlled trial.

The HAM-D and HAM-A are not suited for measuring immediate effects, but are used to measure outcomes over a longer period. The PANAS was most useful for evaluating immediate post-treatment effects, but was also used at 2 and 4 weeks. There were no statistically significant associations between treatment and PANAS scores, either at the time of treatment or during the four-week follow-up period.

The size of our sample was too small to represent the larger patient populations, but still within this population, a single, brief treatment with transcranial NIR light had effects that seem to compare well with other modalities. For example in a previous study we reported [42], using transcranial magnetic stimulation, to treat 37 refractory depressed patients over two weeks, and used the same 21-item HAM-D as in the present study, and found at 2-weeks following the completion of the treatment a mean percent decrease in HAM-D of 29.4% ± 26.1. In the present study, we found a mean percent decrease in the HAM-D of 54.3% ± 26.1 at 2-weeks post-treatment. The rTMS study had a larger population (N = 37), and the mean baseline HAM-D was 29.6 ± 5.6, which was higher than that for the population in this study (23.8 ± 8.8). Therefore, the studies cannot be directly compared, but it is unusual for to find two studies with the identical outcome measures applied at identical times. In a recent rTMS study by Stern et al [43] in which they compared outcomes using different stimulation parameters, the best group had a remission (≤ 10) rate of 40% at 2-weeks post-treatment; in the present study there was a remission rate of 60% at 2-weeks.

Gershon et al [47] reviewed the efficacy of rTMS and found a wide range of " % responders" (decrease % in HAM-D ≥ 50%) from 10 to 49% among 5 sham controlled studies. In all of these studies the active treatment was far superior to the sham, which ranged from a 0% to a 25% response rate. Loo et al [48] reported in a recent rTMS study, using twice daily left-sided high frequency rTMS, a decrease in HAM-D of 38.5% immediately after a 2-week sham controlled study, compared to 54.3% in our study 2-weeks after treatment.

Some studies have compared rTMS with electroconvulsive therapy (ECT) [49–51], and found them generally to have a similar efficacy in severe depression. For instance, Janicak et al [50] compared up to 20 rTMS treatments with 3 to 12 ECT treatments and reported at the end of treatment the rTMS group had a "remission" rate (<8) of 46% compared with 56% for the ECT group. The authors did not report 2-week post-treatment results. Our 2-week post-treatment HAM-D scores indicated that 40% had achieved "remission" (< 8).

In a recent study comparing the efficacy of 6 right-sided ECT treatments with 6 bilateral over 3 weeks, Eschweiler et al. [52] found that both groups had a 37% decrease in HAM-D at the end of treatment. Each group had 26% "responders" (≥ 50%) at the end of treatment. Again, our group at 2-weeks post-treatment had a mean decrease in HAM-D of 54.3% with 40% "responders" (≥ 50%).

In a recent study, Tadi et al [53] compared the outcomes in HAM-D from baseline to 10 weeks for 223 patients with depression randomized between 4 treatment groups: sertraline, placebo pill, cognitive-behavioral therapy (CBT), and guided self-help group (GSG). At 10-weeks, the completion of the treatments, 44% of the sertraline group responded (HAM-D % decrease ≥ 50%), compared with 19% for the placebo group, 20% for the CBT group, and 19% for the GSG group. By the second week 49% of the sertraline group, 40% of the placebo group, 39% of the CBT group, and 35% of the GSG group showed improvement defined as a decrease in HAM-D of ≥ 20%. In our study at 2-weeks post-treatment 100% showed improvement (decrease in HAM-D ≥ 20%).

Katz et al [54] reported HAM-D outcomes for 70 depressed patients (baseline HAM-D = 23.5) randomly divided between 3 treatment groups: desipramine, paroxetine, and placebo. At 2-weeks the desipramine group had a mean % decrease in HAM-D of 45%, the paroxetine group, 24%, and the placebo group, 36%.

Bech et al [55] performed a meta-analysis of 16 US trials involving depressed patients comparing fluoxetine with either tricyclic antidepressants or with placebo in trials of at least 6-weeks. The authors reported that among the 1914 patients intended to treat with fluoxetine, 38.5% were responders (HAM-D reduction ≥ 50%), while among the 686 TCA treated patients this measure was 35.5%, and among the 847 placebo treated patients the measure was 24.2%.

In regard to anxiety, Leichsenring et al [56] found that 29 patients with a generalized anxiety disorder (GAD) treated with CBT for 30 weeks achieved a 50.7% reduction on the HAM-A at the end of treatment, and that 28 patients treated with short-term psychodynamic psychotherapy over the same time period achieved a 42.8% reduction. In our study at 2-weeks post-treatment, our patients achieved a reduction in HAM-A of 63.1%. Moreover, Montgomery et al [57] reported pooled data from 6 double-blinded, placebo-controlled, 4 to 6-week trials for patients with GAD treated with a benzodiazepine (either alprazolam or lorazepam), with pregabalin (PGB), or with placebo. The benzodiazepine group had a mean decrease in HAM-A from baseline to the end of treatment of -11.0 points, the PGB group had a decrease of -11.2, and the placebo group, -8.3. In our study, we found a decrease in HAM-A from baseline to 2-weeks post-treatment of 14.9 points.

Even though our 2-week results compare well with the other reported treatments cited above, the 2 and 4-week outcomes were unblinded, and did not have a placebo control. Further, comparisons between treatments need to be made with a single study and those results replicated.

Some of our secondary experiments showed results in support of our initial hypotheses. For example, there was greater rCBF during NIR on versus off, although this difference did not achieve statistical significance. NIR on was more successful relative to NIR off when treatment was applied to a hemisphere with more positive HEV.

An increase in rCBF with NIR is consistent with an effect of NIR treatment on the brain. This effect on the brain (whatever its complex nature) likely relates to the alterations in affect. Together with our result that immediate psychological benefit of infrared treatment was associated with positive HEV, that the 2-week HAM-A outcomes related to the HEV value and left - right rCBF is consistent with the hypothesis we presented at length in a previous publication [39], stating that the right hemisphere is often associated (unexpectedly) with a positive HEV and that knowing a patient's HEV can enlighten data reduction and possibly guide treatment. We did not use HEV to guide therapy in this study, but we think that future studies should consider this possibility. In two other previous, independent publications we reported that HEV by LVFS could predict positive responses to left-sided rTMS [41, 42]. While promising, the result of an association with 2-week changes in HAM-A should be interpreted with caution since it was based on a post-hoc analysis, was sensitive to our quantification of change in HAM-A (percent vs. points change), and did not hold for HAM-D.

Because this is the first trial applying NIR to the brain, we wanted to be extremely vigilant for negative side effects. We found none, during or after the procedure. During the treatments we turned off the fluorescent lights to prevent interference with our NIRS data and the fan created a drone and light breeze, all of which seemed to relax the patients, although we did not formally measure this. Certainly, no patient complained of headaches or any other physical discomfort. No patient dropped out of the study and all continued through the 4-week follow-up. Six patients spontaneously reported feeling much improved at 2-weeks and attributed that improvement (rightly or wrongly) to the treatment. The other 4 patients did not feel any effect, positive or negative, from the treatment, including one man who had an 85% improvement on his HAM-D and a 68% improvement on his HAM-A at 2-weeks. Thus, we observed the treatment to be comfortable, pleasant, easy to apply, and safe.

The mechanism by which NIR-PBM has improved mood is not understood. PBM is known to improve blood flow in skin (as measured by laser Doppler) [58]. The fact that HEV may play a role in the response suggests that positive neural circuits might somehow be stimulated by NIR light or negative neural circuits may be inhibited. NIR is known to increase mitochondrial ATP and nerve growth factors. We feel that our outcome findings must be replicated in double blind, randomized, placebo-controlled prospective outcome studies with large numbers and various populations. The method of treatment should also be studied to attempt to optimize the results. Some possibilities are to use pulsed light, try different anatomical locations, different treatment schedules (daily, weekly, biweekly, etc) as well as different light wavelengths and total energy densities. We would like to study also whether HEV should guide treatment. Lastly, this treatment might benefit from a possible synergy with other treatments such as psychotherapy and psychotropic medications. All the subjects in this study remained on their usual treatment and no one altered their usual treatment during this study. If further study confirms our results or improves upon them, then an intense search for the mechanism of action will be highly desirable and might lead to greater knowledge of mind-brain interactions, the psychophysiology of mental states, including the effects of trauma, and of treatment benefits.