3.1. Results of the Questionnaire Study

Figure 1 demonstrates that most of winter seasonals reported annual variations in such bipolar neurovegetative symptoms as under/overweightiness, under/oversleeping and hyper/hypoactivity. Only one symptom, post/presleep insomnia that could be

Figure 4. Diurnal patterns of melatonin in subjects of different diurnal types. Solid lines show patterns before light treatment, thin lines show patterns after light treatment.

Figure 5. Evening-morning differences for melatonin and cortisol. The significant differences between pre- and posttreatment values are indicated by p-values (Student's paired t-test).

Table 3. Evening-morning difference for serum melatonin before/after a.m. or p.m. light treatment

Table 3. Evening-morning difference for serum melatonin before/after a.m. or p.m. light treatment


All treatments

A.M. treatment

P.M. treatment

Controls, n = Mean (SD)




-84.4 /-30.2



(177.3) /(174.8)

(132.4) /(123.9)

(217.8) /(221.3)

Patients, n = Mean (SD)




Paired t-test


-49.1 /63.5


(128.2) /(132.3)

(135.4) / (156.6)

(125.3) / (83.3)

t = 2.67, p < 0.05

t = 2.44, p < 0.05

Nonresponders, n = Mean (SD)




Paired t-test



-124.3 /-20.7




t = 3.60, p < 0.05

Responders, n = Mean (SD)




Paired t-test

-54.0 /15.6


14.1 / 5.2

(134.9) /(88.9)

(125.2) /(95.5)

(140.5) / (83.4)

t = 2.01, p < 0.1

t = 3.17, p < 0.05

Note. t-values (Student paired t-test) are not shown if the level of significance for a difference between pre- and post-treatment values was found to be 0.1 orhigher.

Note. t-values (Student paired t-test) are not shown if the level of significance for a difference between pre- and post-treatment values was found to be 0.1 orhigher.

related to delaying/advancing of sleep timing, was much less common in comparison with other neurovegetative symptoms. The annual patterns were very similar in N-SAD, S-SAD and SAD. Only the amplitude parameter was considerably different in these diagnostic groups.

3.1.1. Delays of theAnnual Rhythms of Depressive Symptoms. Although the seasonal variations in depressive symptoms closely follow the variations in day length, most of the symptoms lag behind the seasonal changes of daylight and darkness. Figure 2 shows that only two phases of one of the neurovegetative symptoms, namely the phase of fall downcrossing of the mean level (c) and the phase of winter minimum (m) for the symptom of post/presleep insomnia (=early/late waking), correspond directly with the seasonal decrease of daylight. All other phases appear to be delayed one to six weeks relative to the changes in day length. Besides, the differences between similar phases of different symptoms as well as the differences between different phases of the same symptom often reach a statistically significant level. These results suggest that, although the photoperiod appears to be the main seasonal timer for depressive symptoms, the modulating influence of some other environmental factors may account for the observed specific of the annual patterns of these symptoms. Winter cold may produce the increase in sleep duration and, due to that, the winter minimum of the corresponding symptom may be shifted towards JanuaryFebruary (the months of the lowest temperatures). The specific effects of summer heat could include the decrease of appetite leading to the delay of summer minimum of the eating symptoms on July/August (the months of highest temperatures). Besides, especially in the subsam-ples from the southernmost location (Turkmenia, #1), both summer heat and winter cold + winter photoperiod could affect several symptoms in a similar way, i.e. they both may cause such problems as late waking, daytime drowsiness and bad feeling. When the impacts of summer and winter seasons were similar, the annual patterns of the symptoms of post/presleep insomnia, hyper/hypoactivity and good/bad mood were bimodal and, as a result, the maximal crosscorrelations between the annual curves of symptom and scotoperiod were lower than +0.82 (due to that the data on these 3 symptoms in Turkmenian subsamples are not shown in Figure 2). In contrast, the effects of winter and summer environment on the symptoms related to sleep length and metabolism were always opposite and, therefore, all maximal crosscorrelations between the annual rhythms of these symptoms and seasonal changes in day length were highly significant (+0.99 in the whole sample).

In general, the findings on seasonality of neurovegetative depressive symptoms raise a question about multi-component nature of physiological dysfunction in SAD. We suggest that, depending on the particular underlying dysregulation, the MLT rhythm may be of more or less importance for control of seasonal variations in a certain symptom. Such chronobiological mechanisms as phase resetting and daytime measurement could be primarily responsible for the symptoms closely related to circadian and metabolic dysfunctions in winter depressives, while the disturbances in sleep home-ostasis and arousal may be associated with both chronobiological and non-chronobio-logical mechanisms including thermoregulation.

3.1.2. DelayoftheSleep-Wake Cycle. Significant differences on scores of evening lateness were found neither between nonseasonals and seasonals in the questionnaire sample, nor between healthy controls and patients with winter depression in the LT-study. By contrast, the scores of morning lateness were significantly higher in season-als compared to nonseasonals (Student's unpaired t-test, p < 0.0001 for both studies). Table 1 and Figure 3 indicate that high current level of depression and, especially, high degree of seasonality could account for the elevation of M-scores in SAD. The symptom of late awakening appears to be caused by both an increase in sleep need and a delay of circadian phase. The lack of significant association between E-score and seasonal depression may be explained by contradictions between the actions of two physiological factors on sleep onset: the increased sleep need produces an advance of bed time, while the delay of circadian phase leads to a delay of bed time.

3.2. Results of LT- Study

LT caused clinically and statistically significant reduction of depressive scores. Time of LT was not important for antidepressant response.

3.2.1. Physiological Effects of LT. The reduction of HRSD and SIGH-SAD scores was lower than 60% in all 4 groups without expected physiological changes in sleep homeostasis, metabolism, circadian or sympatho-adrenal systems. In contrast, the groups with expected changes had mean score reductions higher than 70%. The associations between different physiological responses to LT were not strong. Only the direction of change in non-REMS homeostasis was found to be associated significantly with the direction of change in metabolic rate (Chi-square = 4.89, p < 0.05, n = 21). Additionally, an insignificant association between metabolic activation and phase advance was noted (Chi-square = 3.32, p = 0.1, n = 61).

Thus, despite rather weak positive interrelations between studied physiological effects of light, each of these effects contributed to reduction of depressive symptoms. Since any of biological responses and their sum was related positively to clinical efficacy of LT, the observed physiological effects appear to be additive.

3.2.2. Effects ofLTon MLT and Cortisol. Although the diurnal means of MLT and cortisol did not change significantly throughout the study, the daytime MLT levels

(at 12:00 and 16:00) were significantly higher in patients compared to controls in wintertime. This difference disappeared after LT and in summer (See Figures 1-4 in (12)). Moreover, a five-way ANOVA indicates that the mean MLT levels were higher in patients compared to controls (Table 2).

Both hormonal variables underwent significant diurnal variations with a pattern similar in patients and controls (see Figure 1 in (11), and Figures 1-4 in (12)). Figure 4 and Table 2 show that the circadian pattern of serum MLT reflected the sleep-wake pattern. Even among SAD patients, we found the morning type subjects (evening sleepers) with phase-advanced MLT rhythm and the short sleepers (habitually larks in the morning and owls in the evening) with low daytime MLT levels.

Evening-morning differences for MLT in patients were similar to that in controls throughout the study (Figure 5). The Evening-morning difference for cortisol was lower in patients compared to controls before (t = -2.765,p < 0.01), but not after LT (t = -0.310,p > 0.1) or in summer (t = -0.708,p > 0.1). The phases of both rhythms did not change significantly in controls, but the MLT phase in controls showed the same tendency as did the patients' phase (Figures 4 and 5; Table 3). An interaction between time of day and condition was significant for MLT both in the whole group of subjects and in the patients' group alone ((Table 2). This result indicates an advance shift of the circadian phase following LT (Figures 4 and 5). Moreover, the significant interaction between time of LT and condition was revealed by 4-way ANOVA of cortisol in patients (F = 5.72, Df = 1,19, p < 0.05).This interaction reflects the decrease of morning levels of cortisol after morning, but not after afternoon LT (assumably, due to an advance shift of the cortisol rhythm). As can be seen in Figure 5, LT and change of season affected the evening-morning differences in a similar way.

It may be generalized that no difference between patients and controls on the timing of MLT secretion was observed in our sample, but patients more likely responded to LT by advancing circadian phase.

For the patients group, the significant interaction between diurnal profile of MLT, time of LT and response was revealed by the 4-way ANOVA. Moreover, a triple interaction (response x time of LT x time of day) as well as the interaction of all four factors (condition x response x time of LT x time of day) reached statistically significant levels (Table 2).

The relative values of MLT and cortisol at 08:00 and 24:00 were similar in control and patient groups. However, compared to the whole control group, responders to morning light had significantly higher cortisol levels at 08:00. Besides, the significant positive correlation was found between HRSD and relative cortisol levels at 24:00 in patients treated in the afternoon, and, in patients treated in the morning, higher baseline HRSD and SIGH-SAD determined increase of relative value of cortisol at 08:00. In contrast to nonresponders to morning light, nonresponders to afternoon light showed significantly lower MLT levels at 24:00 and significantly higher MLT levels at 08:00. Light seems to work worse when patients with advanced MLT phase are treated in the morning, and when patients with delayed phase are treated in the afternoon (see also Figure 2 in (52)).

In general, only morning LT, and then only in patients, was able to produce a significant advance of circadian phase. Phase shifts were significant in responders and the same tendency was noted in nonresponders. Neither in patients, nor in controls were the phase changes following afternoon LT an significant. However, the significant decrease of MLT levels at 08:00, indicating advance shift of the rhythm, was observed when the nonresponder subgroup was analyzed separately. The results show that, in the majority of patients treated in the afternoon, the antidepressant response occurred despite the lack of phase shift of circadian rhythms, whereas the advance shifts were observed in both responders and nonresponders treated in the morning.

Beating Insomnia

Beating Insomnia

Discover How to Beat Insomnia Naturally & Enjoy a Great Night’s Sleep. The Secrets You Need to Know to Fall Asleep Fast, Sleep Through the Night & Awaken Feeling Rested, Refreshed and Rejuvenated.

Get My Free Ebook

Post a comment