Possible Effects of Bright Light Therapy on Electroencephalogram-Vigilance in the Treatment of Depression in Adolescents: A Pilot Study

Background: Chronotherapeutic treatments for depression, such as bright light therapy (BLT), are non-invasive and produce almost no side effects. However, study evidence for reliable neurobiological changes associated with treatment response is still rare. Several studies using EEG-vigilance indicate higher arousal and a later decline during resting state in adult depressive patients compared to healthy controls. To our knowledge, there are no study reports on EEG-vigilance in depressive youth to date.

Methods: A total of 11 adolescents with depression receiving BLT were compared to 11 age and gender-matched patients with depression receiving treatment as usual (TAU). The BLT was administered in the morning for 2 weeks on five consecutive days per week. The depressive symptomatology was assessed using the Beck Depression Inventory (BDI-II) and the resting state electroencephalogram (EEG) of 20 min was recorded. EEG and BDI-II were assessed before and after 10 days of treatment. Vigilance level and vigilance decline were estimated using the VIGALL toolbox.

Results: Brain arousal increased after 10 days of bright light therapy in adolescents with depression. Severe depressive symptoms were associated with higher brain arousal levels; the BDI-II sum score correlated negatively with the amount of drowsiness.

Limitations: The sample size was small and participants’ brain arousal at baseline was not matched and differed between BLT and TAU groups.

Conclusion: The BLT might have an additional effect on brain arousal. EEG-vigilance seems to be a reliable and valid marker for neurobiological changes that are probably associated with depression and its treatment and, therefore, might be of clinical relevance.

Introduction

Depression is a major health problem that arises during adolescence with increasing incidence rates (1, 2). Prevalence rates of depression in youth range between 5 and 6%, with a lifetime prevalence rate of 20% by late adolescence (3). The risk of suicide might be increased by 30 times (4). Despite extensive research efforts on cognitive-behavioral approaches and medication, treatment outcomes for depressive youth remain unsatisfying (5, 6). Hence, better treatment is needed. Chronotherapeutic treatments, such as bright light therapy (BLT), are non-invasive and produce almost no side effects (7). Preliminary evidence shows its feasibility not only in adults [e.g., (8)], but also in adolescents with depression [e.g., (9–12)].

However, responses to treatment mostly rely on self-reports, which underlie possible bias and lack of elucidated neurobiological changes, and are perhaps associated with treatment response. Accordingly, there has been a call for more objective markers, such as EEG biomarkers, to estimate treatment responses. However, it is unclear which EEG biomarkers should be used when depression treatment outcome is of interest (13).

Hereby, the concept of vigilance may be of help: several studies indicate higher arousal and a later decline during a resting state EEG-session in adult depressive patients compared to healthy controls (14) as well as a more stable vigilance regulation pattern in adult depressive patients compared to healthy controls (15). Hegerl and Hensch (16) discussed the avoidance of stimulation as a reaction to a tonic increased vigilance in depression in adults. To our knowledge, there are no study reports on EEG-vigilance in depressive youth to date.

The BLT in the morning stabilizes the circadian rhythm and therefore probably also changes brain arousal. BLT can decrease daytime sleepiness and induces a forward shift of sleep onset when it is carried out about 7.5 up to 9.5 h after the secretion of melatonin, referred to as dim light melatonin onset (DLMO) (17).

To summarize, vigilance can be seen as directly related to the negatively biased and increased default mode network (DMN) activity in depression, which expresses itself as rumination, resulting in increased brain arousal, perhaps, changed by BLT. Up to date, no study investigated EEG vigilance in depressed adolescents. The study aimed to assess vigilance before and after BLT for depressive adolescent inpatients, compared to usual treatment for depression [treatment as usual (TAU)].

Methods

Procedure

All patients admitted to the LWL university hospital Hamm between 2013 and 2014, with a clinically diagnosed depression based on the International Statistical Classification of Diseases and Related Health Problems (ICD-10) and a BDI-II (18), indicating moderate to severe depression, were screened for in- and exclusion criteria of a larger study described elsewhere (COWALI) (19). A positive vote from the ethics committee of the Ruhr-University Bochum, Germany (Registry number: 4,418–12) was obtained. Patients and their legal guards/parents signed informed consent for taking part in the study and being investigated with an EEG. A pre-test BDI-II and an EEG have been recorded non-more than 3 days before treatment (details see below). The recording started with controlled eye movements, the Berger test, and a mental calculation. Then, a resting EEG was recorded for 20 min. On the last day of 10 days of BLT or TAU, a post-EEG assessment and a second assessment via BDI-II were conducted similar to pre-test conditions.

Participants

Twenty-two female patients aged between 13 and 18 years were included in the study. The two treatment groups were recruited consecutively, after the BLT group recruitment (n = 11) age-matched patients for the TAU group (n = 11) were recruited. Participants were not matched for EEG parameters. Exclusion criteria comprised psychotic symptoms, suicidality, organic brain damage, anorexia nervosa, intake of neuroleptics or beta-blockers, and an IQ of < 85. Antidepressant medication had to be stable for at least 4 weeks. Two patients in the BLT group and one patient in the TAU group dropped out for similar reasons, while 9 patients in the BLT group and 10 patients in the TAU group were finally included in the analysis.

Treatment

Patients received BLT starting 14.56 days after admission (SD = 6.10 days). The BLT comprised ten sessions of 45 min with 10,000 Lux. The BLT was administered in the morning for 2 weeks on five consecutive days per week. BLT started 8.5 h after melatonin-onset, which was calculated with the help of the Morningness-Eveningness-Questionnaire (MEQ) (20). The MEQ is a self-report questionnaire, which can determine the individual chronotype. During weekends, no BLT was administered.

The TAU is based on a multimodal therapy approach including individual and group psychotherapy, as well as all kinds of care and therapy, in-hospital schooling, and family-focused interventions are usually provided.

Electroencephalogram Measurement and Preprocessing

The resting-state electroencephalogram (EEG) of 20 min was recorded according to the 10/20 system (Fp1, Fp2, F3, F4, C3, C4, P3, P4, O1, O2, F7, F8, CP1, CP2, CP5, CP6, Fz, Cz, Pz, FC5, FC6, T7, T8, FC1, FC2, EOG, and ECG). The sampling rate was set to 200 Hz. Offline, EEG was preprocessed with BrainVision Analyzer (BrainProducts, Gilching, Germany) according to the manual of the VIGALL toolbox,¹ which included the following steps: Filtering with Butterworth’s zero-phase filter (0.5–70 Hz, notch at 50 Hz), creating 1-s segments, rough artifact screening by visual inspection, execution of independent component analysis (ICA) and exclusion of ICA components reflecting continuous artifacts like blinks, eye movements or cardioballistic artifacts, and marking of remaining artifacts. EEG data were screened for sleep-indicating graphoelements (sleep-spindles, K-complexes), which did not occur.

Vigilance

To keep data comparable to prior EEG vigilance research (21, 22), we used similar methods for arousal analysis. Consecutive segments of 1-s length were classified into 6 different EEG-vigilance stages: 0, A1, A2, A3, B, and B2/3 (C was not observed), ranging from wakefulness to drowsiness by using the add-on VIGALL 2.1 for BrainVision Analyzer. To do so, the VIGALL algorithm uses source localization in different frequency bands with LORETA. Further information about VIGALL, which is licensed under GPL3 is available at https://github.com/danielboettger/VIGALL or see text FOOTNOTE 1. The arousal stability index was calculated based on 1-min intervals (interval 1, segments 1–60; interval 2, segments 2–61; etc.) (For scoring criteria see Table 1). Furthermore, we calculated the mean arousal level from the whole EEG acquisition period and the relative EEG vigilance stages occurrence (number of segments of one stage/number of all artifact-free segments). Vigilance stage C of sleep onset was not observed.

TABLE 1

Table 1. Scoring criteria of the arousal stability index.

Statistical Analysis

Of all vigilance classification stages, we only used A1, B23, and mean vigilance as dependent variables, because A1 and B23 were the most classified stages and can, therefore, be treated as most relevant (Figure 1). The effects of time on A1, B23, and mean vigilance were normally distributed and were tested with paired t-tests. For the same dependent variables, repeated measures ANOVAs were additionally modeled with the grouping factor, the inner subject factor time (pre and post), and the interaction of both. Baseline measures of vigilance were compared between the groups with a test of equivalence, the minimal detectable effect size for the two-sample t-test was used as the smallest effect size of interest (SESOI, d = 1.366, N1 = 9, N2 = 10, power = 0.8, and alpha error probability = 0.05). Baseline variables were seen as equivalent if confidence interval boundaries of effect sizes were smaller than the SESOI. Effects of the group-factor on the distribution of pre-post difference of arousal stability indices were analyzed with crosstabs and χ²-tests; the effect of time was tested with the Wilcoxon test, separately for the BLT and TAU group. We used Pearson’s correlation index to analyze the association between the BDI-II sum score and vigilance classification B23 separately for pre and post-measures. Additionally, we tested with Pearson’s correlation index whether changes in vigilance level were associated with baseline scores or with changes in the BDI-II sum score.

FIGURE 1

Figure 1. Vigilance classification distribution. BLT, Bright Light Therapy; shown are the mean percentage of vigilance level classifications. The 20 min resting-state EEG was second-wise classified by the VIGALL algorithm.

For the paired t-tests and the repeated measures ANOVA we calculated the minimum of the detectable effect size. For the two-tailed paired t-tests with total sample size = 18 (BLT) and 20 (TAU), alpha error probability = 0.05, and power = 0.8 resulting in minimum Cohen’s d = 0.7 (BLT) and d = 0.66 (BLT) (medium to large effect). Repeated measures ANOVA, with total sample size = 36, alpha error probability = 0.05, power = 0.8, and correlation among replication measures = 0.5, resulting in minimum Cohen’s d = 0.295 (small effect) for within factor, d = 0.961 (large effect) for between factor, and d = 0.48 (medium effect) for between and within factor interaction.

Results

Baseline Characteristics

The groups neither differed regarding age [t(17) = 0.021; p = 0.983; M = 15.61 years SD = 1.19] nor BDI-II sum score at admission [t(17) = 0.378; p = 0.710; M = 31.53 SD = 10.98]. Three patients under BLT got antidepressant medication (quetiapine or SSRIs). One TAU-patient got an SSRI medication. All patients in the TAU group had a (recurrent) depressive disorder according to ICD-10 (six in the BLT group). Two patients under BLT were diagnosed with a combined disorder of conduct and emotions and one patient had a post-traumatic stress disorder. All pre-EEGs were recorded 15.42 (SD = 9.75) days after admission [t (17) = –1.787; p = 0.092].

Baseline measures of vigilance classification stages A1, B23, and mean vigilance were not equivalent between the groups. Although t-tests did not show a significant difference between the groups [A1: T(17) = 1.983;p = 0.057; B23: T(17) = –1.639; p = 0.12; mean vigilance: T(17) = 1.727; p = 0.102], the confidence intervals of the effect sizes laid not inside the boundaries of the SESOI, even though the SESOI was determined by the small sample size and was quite large [CI of Cohen’s d, A1(-0.29 1.88); B23(-1.68.19); and mean vigilance (-0.16 1.72)]. Therefore, we decided to focus on the effect of time and time by group interaction effects but not on the group effect.

Level of Vigilance

Of all paired t-tests undertaken for the different vigilance levels, only the test on B23-classification revealed an effect of time for the BLT group but not the TAU group. The percentage of B23 classification was lower (pre = 43%, post = 34%), meaning that patients who received BLT got more aroused. In a repeated-measures ANOVA, the group*time interaction was not significant (see Table 1). Looking at Figure 1, one recognizes a higher percentage of A1 classification after treatment for both groups (decreased vigilance in the post-intervention measure, see Table 2 and Figure 1).

TABLE 2

Table 2. Effect of time on EEG vigilance level.

The same picture was observed for the mean vigilance classification: BLT showed a significant effect in paired t-tests [T (8) = 2,944; p = 0.019] indicating higher mean vigilance classification after BLT (pre = 4.04, post = 4.53) but there was no group*time interaction in the repeated measures ANOVA [F(1, 16) = 0.41, p = 0.53].

Vigilance Stability

There was no effect of time (BLT: Z = –0.535; p = 0.593; r = 0.12; TAU: Z = –1.131; p = 0.258; r = 0.26) and no effect of group [χ²(4) = 4.958, p = 0.292, w = 0.54] on pre-post difference in vigilance stability.

Electroencephalogram Vigilance and Beck Depression Inventory-II Association

For the BDI-II sum score, we found a negative correlation with vigilance stage B23 (r = –0.557; p = 0.016) and a positive correlation with mean vigilance (r = 0.528; p = 0.024) only at follow up, indicating that severe depressive symptoms are associated with higher brain arousal level. At baseline, there was no correlation between vigilance measures and the BDI-II sum score. Additionally, we found no correlation between changes in vigilance and changes in BD II score, nor with baseline BDI-II sum score.

Discussion

In this study, we observed a statistically significant increase in vigilance in depressive adolescent patients after 10 days of BLT. Some increase in EEG-vigilance was observed for the TAU group as well, although to a lesser extent and not reaching statistical significance. However, this vigilance increase after BLT was neither associated with a change in the BDI-II sum score nor with the BDI-II sum score at baseline. This finding indicates that BLT might have an additional and immediate effect on vigilance, which is probably not related to upregulated brain arousal in depression with symptoms, such as inner tension or prolonged sleep latency (14). Instead, these results are more in line with the improvement of sustained attention and sleepiness reported for sleep-deprived young adults after BLT (23). The normalization effect of the antidepressant intervention on an increased level of brain arousal and EEG-vigilance in adult depressive patients (14) was not observed, which might be due to the short period of only 2 weeks between baseline and follow-up. Add-on effects of BLT probably emerge later in the treatment timeline. Nevertheless, we found a positive correlation of the BDI-II sum score with the B23 vigilance level, indicating that severe depressive symptoms are associated with higher brain arousal levels, as was described for adults (22).

As a limitation, it should be mentioned that treatment effects may depend on the baseline level of vigilance. In this study, patients were not assigned as matches in vigilance, which produced somewhat higher arousal in the BLT group before treatment. Further limitations are the lack of treatment response data in terms of follow-up measures of symptoms and EEG vigilance.

Nevertheless, the chosen markers proved to be reliable and valid indicators for neurobiological changes probably associated with depression and its treatment. Future studies with larger samples and longer follow-up times are necessary to replicate the presented results.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics Statement

The studies involving human participants were reviewed and approved by Ethics Committee of the Ruhr University Bochum. Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.

Author Contributions

CB analyzed the data and drafted the manuscript for intellectual content. SG acquired the data. SG, AD, LJ, JS, and MK revised the manuscript for intellecutal content. FM revised the manuscript for intellectual content and drafted procedure and study sample description of methods part. OR helped in conceptualizing the study and revised the manuscript for intellectual content. TL and MH designed the study, drafted main part of the introduction, and revised the manuscript for intellectual content. All authors contributed to the article and approved the submitted version.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s Note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

1. ^ https://research.uni-leipzig.de/vigall/

References

1. Clarke G, Harvey AG. The complex role of sleep in adolescent depression. Child Adolesc Psychiatr Clin N Am. (2012) 21:385–400. doi: 10.1016/j.chc.2012.01.006

PubMed Abstract | CrossRef Full Text | Google Scholar

2. Skurtveit S, Bramness JG, Hjellvik V, Hartz I, Nesvag R, Hauge LJ, et al. Increase in diagnosis of depressive disorders contributes to the increase in antidepressant use in adolescents. Acta Psychiatr Scand. (2018) 137:413–21. doi: 10.1111/acps.12877

PubMed Abstract | CrossRef Full Text | Google Scholar

3. Thapar A, Collishaw S, Pine DS, Thapar AK. Depression in adolescence. Lancet. (2012) 379:1056–67. doi: 10.1016/S0140-6736(11)60871-4

CrossRef Full Text | Google Scholar

4. Windfuhr K, While D, Hunt I, Turnbull P, Lowe R, Burns J, et al. Suicide in juveniles and adolescents in the United Kingdom. J Child Psychol Psychiatry. (2008) 49:1155–65. doi: 10.1111/j.1469-7610.2008.01938.x

PubMed Abstract | CrossRef Full Text | Google Scholar

5. Emslie GJ, Kennard BD, Mayes TL, Nakonezny PA, Zhu L, Tao R, et al. Insomnia moderates outcome of serotonin-selective reuptake inhibitor treatment in depressed youth. J Child Adolesc Psychopharmacol. (2012) 22:21–8. doi: 10.1089/cap.2011.0096

PubMed Abstract | CrossRef Full Text | Google Scholar

6. Kennard B, Silva S, Vitiello B, Curry J, Kratochvil C, Simons A, et al. Remission and residual symptoms after short-term treatment in the treatment of adolescents with depression study (TADS). J Am Acad Child Adolesc Psychiatry. (2006) 45:1404–11. doi: 10.1097/01.chi.0000242228.75516.21

CrossRef Full Text | Google Scholar

7. Moscovici L, Kotler M. A multistage chronobiologic intervention for the treatment of depression: a pilot study. J Affect Disord. (2009) 116:201–7. doi: 10.1016/j.jad.2009.01.015

PubMed Abstract | CrossRef Full Text | Google Scholar

8. Wirz-Justice A, Benedetti F, Terman M. Chronotherapeutics for Affective Disorders, A Clinician’s Manual for Light and Wake Therapy. 2nd ed. New York, NY: Cambridge University Press (2013).

Google Scholar

9. Bogen S, Legenbauer T, Gest S, Holtmann M. Lighting the mood of depressed youth: Feasibility and efficacy of a 2 week-placebo controlled bright light treatment for juvenile inpatients. J Affect Disord. (2016) 190:450–6. doi: 10.1016/j.jad.2015.09.026

PubMed Abstract | CrossRef Full Text | Google Scholar

10. Holtmann M, Mokros L, Kirschbaum-Lesch I, Kolch M, Plener PL, Ruckes C, et al. Adolescent depression: study protocol for a randomized, controlled, double-blind multicenter parallel group trial of bright light therapy in a naturalistic inpatient setting (DeLight). Trials. (2018) 19:568. doi: 10.1186/s13063-018-2949-0

PubMed Abstract | CrossRef Full Text | Google Scholar

11. Kirschbaum-Lesch I, Gest S, Legenbauer T, Holtmann M. Feasibility and efficacy of bright light therapy in depressed adolescent inpatients. Z Kinder Jugendpsychiatr Psychother. (2018) 46:423–9. doi: 10.1024/1422-4917/a000603

PubMed Abstract | CrossRef Full Text | Google Scholar

12. Kirschbaum I, Straub J, Gest S, Holtmann M, Legenbauer T. Short-term effects of wake– and bright light therapy on sleep in depressed youth. Chronobiol Int. (2018) 35:101–10. doi: 10.1080/07420528.2017.1388251

PubMed Abstract | CrossRef Full Text | Google Scholar

13. Widge AS, Bilge MT, Montana R, Chang W, Rodriguez CI, Deckersbach T, et al. Electroencephalographic biomarkers for treatment response prediction in major depressive illness: a meta-analysis. Am J Psychiatry. (2019) 176:44–56. doi: 10.1176/appi.ajp.2018.17121358

PubMed Abstract | CrossRef Full Text | Google Scholar

14. Hegerl U, Wilk K, Olbrich S, Schoenknecht P, Sander C. Hyperstable regulation of vigilance in patients with major depressive disorder. World J Biol Psychiatry. (2012) 13:436–46. doi: 10.3109/15622975.2011.579164

PubMed Abstract | CrossRef Full Text | Google Scholar

15. Olbrich S, Sander C, Minkwitz J, Chittka T, Mergl R, Hegerl U, et al. EEG vigilance regulation patterns and their discriminative power to separate patients with major depression from healthy controls. Neuropsychobiology. (2012) 65:188–94. doi: 10.1159/000337000

PubMed Abstract | CrossRef Full Text | Google Scholar

16. Hegerl U, Hensch T. The vigilance regulation model of affective disorders and ADHD. Neurosci Biobehav Rev. (2014) 44:45–57. doi: 10.1016/j.neubiorev.2012.10.008

PubMed Abstract | CrossRef Full Text | Google Scholar

17. Wirz-Justice A. From the basic neuroscience of circadian clock function to light therapy for depression: on the emergence of chronotherapeutics. J Affect Disord. (2009) 116:159–60. doi: 10.1016/j.jad.2009.04.024

PubMed Abstract | CrossRef Full Text | Google Scholar

18. Hautzinger M, Keller F, Kühner C. BDI-II Beck Depressions-Inventar. Frankfurt: Harcourt Test Services (2006).

Google Scholar

19. Gest S, Holtmann M, Bogen S, Schulz C, Pniewski B, Legenbauer T. Chronotherapeutic treatments for depression in youth. Eur Child Adolesc Psychiatry. (2016) 25:151–61. doi: 10.1007/s00787-015-0720-6

PubMed Abstract | CrossRef Full Text | Google Scholar

20. Griefahn B, Künemund C, Bröde P, Mehnert P. Zur validität der deutschen übersetzung des morningness-eveningness-questionnaires von horne und östberg. Somnologie (Berl). (2001) 5:71–80. doi: 10.1046/j.1439-054x.2001.01149.x

CrossRef Full Text | Google Scholar

21. Berger C, Duck A, Perin F, Wunsch K, Buchmann J, Kolch M, et al. Brain arousal as measured by EEG-assessment differs between children and adolescents with attention-deficit/hyperactivity disorder (ADHD) and DEPRESSION. Front Psychiatry. (2021) 12:633880. doi: 10.3389/fpsyt.2021.633880

PubMed Abstract | CrossRef Full Text | Google Scholar

22. Ulke C, Tenke CE, Kayser J, Sander C, Bottger D, Wong LYX, et al. Resting EEG measures of brain arousal in a multisite study of major depression. Clin EEG Neurosci. (2019) 50:3–12. doi: 10.1177/1550059418795578

PubMed Abstract | CrossRef Full Text | Google Scholar

23. Comtet H, Geoffroy PA, Kobayashi Frisk M, Hubbard J, Robin-Choteau L, Calvel L, et al. Light therapy with boxes or glasses to counteract effects of acute sleep deprivation. Sci Rep. (2019) 9:18073. doi: 10.1038/s41598-019-54311-x

PubMed Abstract | CrossRef Full Text | Google Scholar