e. 4.

r . .
3‘24: .1} 4
.S‘Ib’vr‘i

2L” . .
,. . x, :1,
n: {#21 .
L...
.‘

 

 

“tats

This is to certify that the

dissertation entitled

Effect of Light Exposure on Melatonin
Concentrations and Depression Scores
in Seasonal Affective Disorder

presented by

Jane Rice

has been accepted towards fulﬁllment

' of the requirements for

Ph ° D ° degree in Phys 101083)’

 

 

Date 02

MS U Lt an Afﬁrmative Action/Equal Opportunity Institution

Major professor

 

0-12771

 

l‘ll'illllllllllllll u

  
   
    

PLACE IN RETURN BOX to
TO AVOID FINES Mum on or

LIBRARY
Michigan State
University

  
 

roman this checkout from your mood.
Mot. dot. duo.

 

DATE DUE D

ATE DUE DATE DUE

‘4—

 

ewe-125‘

 

Weentﬂ

"T'TS‘

 

 

55p 1 7 19:5
/

 

 

 

 

 

 

 

 

 

L,»

‘4

:l

 

 

L,”

 

 

 

/
MSU lo M Affirmative mum/Equal Opportunity Imtltuion
WW.‘

EFFECT OF LIGHT EXPOSURE ON MELATONIN CONCENTRATIONS
AND DEPRESSION SCORES
IN SEASONAL AFFECTIVE DISORDER
By

Jane Rice

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Physiology
1993

ABSTRACT

EFFECT OF LIGHT EXPOSURE ON MELATONIN CONCENTRATIONS
AND DEPRESSION SCORES IN SEASONAL AFFECTIVE DISORDER

By

Jane Rice

Seasonal affective disorder (SAD) is a major depressive disorder in
which depressive episodes recur in autumn and winter. Daily exposure to
morning bright light is effective in reducing depressive symptoms in SAD.
The mechanism through which light exerts an antidepressant effect is
unknown, although it may involve entrainment of circadian rhythms to the
light:dark cycle.

In all experiments, subjects were diagnosed with SAD, rated for
severity of depressive symptoms, and assessed for safety during
phototherapy. Subjects entered crossover designs consisting of two week-
long trials of phototherapy separated by a l-week withdrawal period.
Phototherapy was administered daily (0600-0800 h) at an illuminance of 2500
lux.

Serial saliva samples were collected in the morning (0600-0900 h) or
evening (1800-2400 or 2000-2300 h) before and after each phototherapy trial
and assayed for melatonin to determine whether evening onset or morning
decline of nocturnal secretion changed.

Depression scores were reduced following phototherapy, with complete
remission in 40% of cases. Cool white and narrow spectrum green light
decreased depression scores as well as full spectrum light. Exposure of facial
skin to light did not increase the antidepressant effect of eye exposure alone.
Expectations of subjects for treatment did not predict outcome, but subjects’

psychological experience during phototherapy did.

Following phototherapy, salivary melatonin concentrations were
increased in the evening and decreased in the morning compared with
pretreatment values. These results are consistent with an advance in the
timing of the melatonin rhythm.

Narrow spectrum green light decreased concentrations of melatonin in
the morning similar to full spectrum light. Blended green-red light
decreased concentrations of melatonin in the morning and evening more than
full spectrum or blended blue-yellow light. However, alterations in color
vision limited the application of blended light as a treatment.

Clinical remission was not correlated with alterations in melatonin
secretion. Morning and evening concentrations of melatonin were similar in
subjects who achieved clinical remission and subjects who did not.

In conclusion, the antidepressant effect of light is not mediated
through an advance in the timing of the melatonin rhythm. The action
spectrum for the antidepressant effect may lie within the green waveband.

However, subjective experience inﬂuences treatment results.

In memory of
my sister Maureen Ann Rice Norman
(1945-1992)
who, without ever mentioning the pineal,
told me as a child that people in the olden days
had an eye in the back of the head
and ﬁrst inspired me to wonder how
the body knows without seeing
and
my aunt Geraldine Catherine McCartney
(1906-1992)
who was and still is

my hero, the wind beneath my feet

iv

ACKNOWLEDGEMENTS

I wish to express appreciation to my major professors, Drs. H. A.
Tucker and R. J. Bielski, as well as the other members of my committee, Drs.
I. Grofova, D. Jump and L. N. Kaufman, for their guidance and support
throughout this dissertation process.

I also wish to thank Mr. L. T. Chapin, Drs. N. E. Enzer, J. L. Gill, and
K. Poif, and the faculty and staff at the Psychiatry Clinic for their logistical
support.

I am especially grateful for the contributions of the 92 research
subjects who participated in this series of experiments.

I wish to thank my teamworks partners, E. A Burr and D. J. Labadie,
for their support.

I wish to acknowledge the many contributions of my extended family
for their unwavering and amazing assistance with computers, printers,
software programs, saliva collections, and childcare. They include Denise
Rice, Sr. Eileen Rice, Susan Rice, David Rice, Rose Rice, Tom Rice, Betsy
Rice, Mike Rice, Mary Rice, Pat Norman and Megan, Chris, Devin, Brendan,
Siobhan, Patrick, Caitlin, and Colin Norman, Tom Maza and Phillip Rice and
Ramona Rice Maza, Kathi Rice and Sean, Tyler, and Andrew Rice, Kate and
Charley Billings, Cathy Rice and Rebecca Rice, Mike Sprague and Deirdre,
Aidan, and Erin Sprague-Rice, Julie Rice and Jessica and Jared Rice, Philip
McCartney, and Maude Rice.

I especially wish to acknowledge the unique contributions of my

nuclear family: to Mark Hiaeshutter for standing by while I rebuilt the car

engine to teach myself a physical model of creative problem solving; to Anne
Hiaeshutter-Rice for pointing out to me that hormone-receptor binding
conveys information to the cell via sign language, i.e., shape conveys
meaning; to Daniel Hiaeshutter-Rice for teaching me the value of a picture
over the printed word.

Finally, I wish to acknowledge the ﬁnancial support of the National
Science Foundation and the College of Osteopathic Medicine Basic Science

Research Program.

vi

TABLE OF CONTENTS

List of Tables ............................................................................................. ix
List of Figures ............................................................................................. x
Introduction ...................................................................................................... 1
Review of Literature ......................................................................................... 3
A. Seasonal Changes in Human Physiology ........................................ 3
1. Seasonality of Affective Disorders ....................................... 3
2.. Animal Models for SAD ......................................................... 4
3. Human Seasonality ................................................................ 5
B. Seasonal Affective Disorder ............................................................. 8
1. Diagnostic Validity ................................................................ 8
2. Clinical and Demographic Features ..................................... 9
3. Biological Studies ................................................................... 11
C. Phototherapy of SAD ........................................................................ 14
1. Protocol ................................................................................... 14
2. Efﬁcacy ................................................................................... 15
3. Design and Methodological Issues ........................................ 17
4. Biological Effects .................................................................... 18
D. Photobiology of SAD ........................................................................ 20
1. Photobiological Effects in Humans ....................................... 20
2. Dosage .................................................................................... 21
3. Action Spectrum ..................................................................... 21
4. Quantiﬁcation of Light .......................................................... 22
E. Melatonin Hypothesis of SAD ......................................................... 23
1. Pineal Gland ........................................................................... 24
2. Synthesis and Secretion of Melatonin .................................. 25
3. Circadian Rhythm of Melatonin ............................................ 26
4. Melatonin in SAD ............ 28
F. Circadian Rhythms in SAD ............................................................. 28
1. Circadian Rhythms and Light ............................................... 28
2. Timing of Circadian Rhythms in SAD .................................. 29
3. Amplitude of Circadian Rhythms in SAD ............................ 31
Objectives .......................................................................................................... 33

vii

Chapter 1. Eﬁ'ect of Full Spectrum versus Cool White Phototherapy on

Melatonin Concentrations and Depression Scores in SAD Patients ............. 34
Introduction ........................................................................................... 35
Materials and Methods .......................................................................... 36
Results .................................................................................................... 46
Discussion .............................................................................................. 52

Chapter 2. Effect of Full Spectrum versus Narrow Spectrum Green
Light on Melatonin Concentrations and Depression Scores in SAD

Patients ............................................................................................................. 61
Introduction ........................................................................................... 62
Materials and Methods .......................................................................... 63
Results .................................................................................................... 67
Discussion .............................................................................................. 74

Chapter 3. Effect of Eye versus Head Exposure to Light on Melatonin

Concentrations and Depression Scores in SAD Patients ............................... 83
Introduction ........................................................................................... 84
Materials and Methods .......................................................................... 85
Results .................................................................................................... 86
Discussion .............................................................................................. 90

Chapter 4. Effect of White Light of Different Spectral Composition on

Melatonin Concentrations and Depression Scores 1n SAD Patients ............. 98
Introduction ........................................................................................... 99
Materials and Methods .......................................................................... 100
Results .................................................................................................... 108
Discussion .............................................................................................. 1 18

Summary and Conclusions .............................................................................. 126

Appendix A Depression Rating Scales ........................................................... 130

Appendix B. Analysis of the effect of full spectrum v. cool white

phototherapy ..................................................................................................... 135

Appendix C. Analysis of the effect of full spectrum v. narrow spectrum

green light ......................................................................................................... 138

Appendix D. Analysis of the effect of eye v. head phototherapy ................... 142

Appendix E. Analysis of variance of the effect of blended white light .......... 146

Bibliography ...................................................................................................... 151

viii

LIST OF TABLES

Table 1. 2 X 2 Latin squares for a two-stage crossover design for three
treatments. Treatments consisted of full spectrum light (FS), blue-yellow
blended white light (BY), and red-green blended white light (RG). Six
subjects completed two stages of phototherapy. Each treatment was
administered four times within the experiment ............................................. 102

Table 2. Mean (iSD) salivary melatonin concentrations in the morning
(0600-0900 h) and evening (2000-2300 h) at pretreatment (Pretrt) and the
morning following day 6 of treatment (Trt) between full spectrum light (FS),
(bIliegifed blue-yellow white light (BY), and blended red-green white light
................................................................................................................... 1 15

ix

LIST OF FIGURES

Figure 1. Experimental protocol for two stage crossover design with two
treatments ........................................................................................................... 39

Figure 2. Spectral power distribution of visible radiation of full spectrum
(Vlta-Lite F40T12) light ..................................................................................... 40

Figure 3. Spectral power distribution of visible radiation of cool white
(F40T120WSS) light ........................................................................................... 42

Figure 4. Pretreatment and treatment HRSD scores for FS and CW light.
Treatment consisted of 1 week of daily phototherapy (0600-0800 h).
Pretreatment scores were obtained in the evening before treatment began;
treatment scores were obtained in the evening of day 7 of treatment. Both
treatments decreased (p<.05) HRSD scores ...................................................... 47

Figure 5. Pretreatment and treatment melatonin proﬁles. Treatment
consisted of 1 week of daily phototherapy (0600-0800 h) with either FS or
CW light. Concentrations of melatonin in saliva were measured every 30
minutes from 1800 to 2400 h under constant dim (<50 lux) illumination.
Pretreatment melatonin proﬁles were obtained on the evening before
treatment began; treatment proﬁles were obtained on day 7 of treatment.
Treatment results were pooled for F8 and CW light. Each data point
represents the mean of 20 observations. Mean melatonin concentrations
increased (p<.03) following phototherapy compared with pretreatment ........ 50

Figure 6. Melatonin proﬁles on day 7 of treatment in subjects who

achieved clinical remission (1 1 observations) and subjects who did not
achieve clinical remission (depressed; 11 observations). Treatment

consisted of 1 week of daily phototherapy (0600-0800 h) with either FS

or CW light. Salivary samples were collected under dim (<50 lux)
illumination. Treatment results were pooled for F8 and CW light. Mean
melatonin concentrations were similar (p>.25) between the two groups
following 1 week of phototherapy ...................................................................... 53

Figure 7. Spectral power distribution of green light (Sylvania F40T12G)....65

Figure 8. SIGH-SAD, HRSD, and SAD scores at pretreatment in stage 1

(St l-Pretrt), treatment in stage 1 (Stl-Trt), pretreatment in stage 2 (St2-
Pretrt), and treatment in stage 2 (St2-Trt). Treatment consisted of 6

days of daily phototherapy (0600-0800 h) with either green or F8 light ........ 69

Figure 9. Salivary melatonin concentrations between 0600 and 0900 h at
pretreatment and treatment. Treatment consisted of 6 days of daily
phototherapy (0600-0800 h) with either green or F8 light. Treatment
results were pooled for green and FS light. Samples were collected under
indoor (<300 lux) lighting. Mean concentrations of melatonin were

reduced (p<.0001) following 1 week of phototherapy compared with
pretreatment ..................................................................................................... 72

Figure 10. Melatonin proﬁles on the morning following day 6 of
phototherapy (0600-0800 h) in subjects who achieved clinical remission

(10 observations) and subjects who did not achieve clinical remission
(depressed; 14 observations). Salivary samples were collected under dim
(<300 lux) illumination. Treatment results were pooled for F8 and green
light. Mean melatonin concentrations wer similar (p>.05) between the

two groups following 1 week of phototherapy .................................................. 75

Figure 11. HRSD, SAD, and SIGH-SAD scores at pretreatment and
treatment. Treatment consisted of 6 days of daily phototherapy

(0600-0800 h) with either eye or head exposure to light. Pretreatment

scores were obtained 2 days before the start of treatment; treatment

scores were obtained on day 6 of treatment. Treatment results were

pooled for eye and head phototherapy. Phototherapy reduced (p<.0001)
HRSD, SAD, and SIGH-SAD scores compared with pretreatment ................. 88

Figure 12. Salivary melatonin concentrations (0600-0900 h) at

pretreatment and treatment. Treatment consisted of 6 days of

phototherapy (0600-0800 h) either eye or head exposure to light.
Pretreatment proﬁles were obtained 1 day before the start of treatment;
treatment proﬁles were obtained on the morning following the last
treatment day. Treatment results (eye v. head phototherapy) were

pooled. Sample collection was performed under indoor (300 lux)

illumination levels. Phototherapy reduced (p=.06) the mean concentration

of melatonin in saliva compared with pretreatment ...................................... 91

Figure 13. Salivary melatonin concentrations (0600-0900 h) at treatment

in subjects who achieved clinical remission (10 observations) and subjects
who did not achieve clinical remission (depressed; 14 observations).
Treatment consisted of 6 days of phototherapy (0600-0800 h) with either

eye or head exposure to light. Treatment proﬁles were obtained on the

day following the last treatment day. Treatment results were pooled for

eye and head exposure. Sample collection was performed under indoor

(<300 lux) illumination levels. Mean concentrations of melatonin were
similar (p>.05) in subjects who responded positively to phototherapy
compared with subjects who did not achieve clinical remission ...................... 93

xi

' Figure 14. Spectral power distribution for BY blended white light

composed of narrow spectrum blue and narrow spectrum yellow light.
Radiant power is measured in wavebands that correspond to colors of

light; ultraviolet (<380 nm), violet (380-430 nm), blue (430-490 nm),

green (490-560 nm), yellow (560-590 nm), orange (590-630 nm), red

(630-800) ............................................................................................................ 104

Figure 15. Spectral power distribution for RG blended white light

composed of narrow spectrum red and narrow spectrum green light.

Radiant power is measured in wavebands that correspond to colors of

light (see legend for Figure 14) ........................................................................ 106

Figure 16. HRSD, SAD, and SIGH-SAD scores at pretreatment (Pre)

and on day 6 of treatment (Trt) with full spectrum (FS), blended blue-

yellow white light (BY), and blended red-green white light (RG). Pooled

for treatment, phototherapy tended to decrease HRSD (p<. 1), SAD

(p<.09), and SIGH-SAD (p<.09) scores relative to pretreatment ................... 110

Figure 17. Morning (0600-0900 h) melatonin concentrations in saliva at
pretreatment and treatment. Treatment consisted of 6 days of daily
phototherapy (0600-0800 h) with either FS, BY, or RG light. Treatment
samples were collected every 30 minutes under indoor (<300 lux)

lighting. Mean concentrations of melatonin were similar (p> .3) at
pretreatment and following 6 days of treatment. Results were pooled for

FS, BY, and RG light ........................................................................................ 113

Figure 18. Evening (2000-2300 h) melatonin concentrations in saliva at
pretreatment and treatment. Treatment consisted of 6 days of daily
phototherapy (0600-0800 h) with either FS, BY, or RG light. Treatment
samples were collected the evening following the last day of phototherapy.
Samples were collected every 30 minutes under indoor (<300 lux) lighting.
Treatment results were pooled for FS, BY, and RG light. Mean
concentrations of melatonin were similar (p>.8) at pretreatment and
treatment .......................................................................................................... 1 16

Figure 19. Morning melatonin concentrations in saliva following 6 days

of daily phototherapy (0600-0800 11) in subjects who achieved clinical
remission (4 observations) and in subjects who did not achieve clinical
remission (depressed; 8 observations). Treatment results were pooled for

FS, BY, and RG light. Mean concentrations of melatonin were similar

(p>.2) between both groups .............................................................................. 1 19

Figure 20. Evening (2000-2300 h) melatonin concentrations in saliva
following 6 days of daily phototherapy (0600-0800 h) in subjects who
achieved clinical remission (4 observations) and subjects who did not

achieve clinical remission (depressed; 8 observations). Treatment

results were pooled for FS, BY, and RG light. Mean concentrations of
melatonin were similar (p>.25) between the two groups ............................... 121

xii

Introduction

Epidemiological studies indicate that the life-time risk of major
depressive disorder in the United States is between 3 and 9 percent (Gold et
al., 1988). One type of major depressive disorder is seasonal affective
disorder, or SAD. SAD is characterized by recurrent depressive episodes
during the autumn and winter that remit spontaneously in spring. The
seasonal pattern of depressive episodes in SAD suggests that seasonal
changes in the environment, especially light, may be involved in precipitating
and remitting episodes. Many have demonstrated the efficacy of bright,
artiﬁcial light, or phototherapy, in reversing depressive symptoms in SAD
(Rosenthal et al., 1984; Lewy et al., 1987; Terman et al., 1988). However, the
mechanism through which light exerts an antidepressant effect is not clear,
although it may be associated with two photosensitive processes in humans:
secretion of the pineal hormone melatonin and entrainment of circadian
rhythms to the lightzdark cycle.

The speciﬁc focus of this thesis was to determine the waveband of light
responsible for the antidepressant effect and whether light-induced changes
in the circadian rhythm of melatonin secretion are associated with the
antidepressant effect.

In human studies, a balance must be struck between ethical and
. technical considerations and sound scientiﬁc design and experimentation.
Research in the ﬁeld of psychiatry is especially complicated by high
intersubject variability, uncontrolled secondary variables, expectations of the

l

2
subjects for treatment effects, nonspeciﬁc placebo responses to the

therapeutic setting, safety issues, and ethical considerations of withholding
treatment. My approach has been one of observing Subjects and collecting
samples under the natural light settings in which they become depressed and
achieve remission, utilizing design and statistical controls (crossover studies
with repeated measures) to maximize information from small groups of
subjects, and interpreting data within a physiological and psychological
framework.

Research ﬁndings on phototherapy for SAD are quickly put into
clinical practice without, at this time, the oversight of the Food and Drug
Administration. Patient safety must therefore be a priority in research trials.
Furthermore, the efﬁcacy of various treatment protocols must be interpreted
in terms of clinical signiﬁcance as well as statistical signiﬁcance. For many
individuals with SAD, the use of light to treat and prevent winter depressive
episodes offers a safe, fast-acting, and effective alternative to pharmacological
treatments. For many researchers in biological psychiatry, the predictable
onset and offset of depressive episodes in SAD along with its moderate
severity and quick responsiveness to an environmental intervention may

make it a useful model for major depressive disorder in general.

Review of Literature

A. Seasonal Changes in Human Physiology
1. Seasonality of Affective Disorders

Annual variations in human physiology and behavior are
common. Annual rhythms are observed in conception, birth, and mortality
rates (Aschoff, 1981), weight, sleep, energy, and mood states (Kasper et al.,
1989), affective disorders and suicide (Aschoﬁ‘, 1981; Rosenthal et al., 1983),
hormone secretion (Weitzman et al., 1975; Broadway and Arendt, 1985;
Hans et al., 1980; Illnerova et al., 1985), metabolic rate (Yurigi et al., 1972),
and immune function (Shiﬁine et al., 1982). Annual rhythms are
classiﬁed as Circannual, social, or seasonal in origin. Circannual rhythms
are generated endogenously and persist without environmental input,
although they are normally synchronized to the seasonal cycle (Gwinner,
1981). Circannual rhythms are documented in animals (Gwinner, 1981)
and may exist in humans (Win-Justice et al., 1984). Annual rhythms that
are social in origin include physiological and behavioral aspects of calendar
events, such as weight gain associated with holiday celebrations. This
review focuses on seasonal rhythms which are annual rhythms that
depend upon periodic environmental input such as light and temperature
(Zucker, 1988). Seasonal rhythms can be distinguished from social
rhythms by their occurrence 6 months apart in the northern and southern
hemispheres and by their dependence on latitude for the magnitude of the
effect (Aschoff, 1981; Gwinner, 1981).

4

Recent reports of the inﬂuence of season on psychiatric illness might
appear novel. However, the ancient humoral theory emphasized seasonal
inﬂuences on these diseases. Aristotle (1953 translation) noted that "if it be
cold beyond due measure, it produces groundless despondency."
Hippocrates believed that the black humor, associated with autumn,
induced fear when its darkness threw a shadow over the area of thought
(Jackson, 1987). Seasonal patterns have recently been documented for
several types of affective (mood) disorders (Wehr and Rosenthal, 1989) Most
prominent of these is seasonal affective disorder (SAD), the periodic
recurrence of major depression in autumn and winter with remission in

spring and summer.

2. Animal Models for SAD

The seasonality and environmental sensitivity of SAD suggests that it
may be a disorder of systems that mediate the organism's adaptation to
changes in the physical environment. The literature on animm seasonality
is helpful here. Many animals, in their normal adaptations to changes in
the physical environment, exhibit behavioral and physiological
adjustments that are similar to those exhibited by SAD patients. These
include changes in food intake, body weight, energy level, sleep duration,
and interaction with the physical and social environment (Gwinner, 1981).
Seasonal changes in animals are often induced by changes in photoperiod
and mediated physiologically by altered hormonal secretion, especially
melatonin (Reiter, 1987). SAD easily lends itself to analogies with animal
models of seasonality.

Hibernation may be a neurobiological model for SAD based on

similar behavioral and physiological changes and the periodicity of

5

symptoms (Wirz-Justice et al., 1986). Both are characterized by increased
food intake, weight gain, decreased activity, and increased sleep.
Mrosovsky (1988), however, notes several ﬂaws with this model. For
example, hibernation is characterized by increased slow wave sleep while,
in contrast, SAD patients spend a smaller proportion of sleep in the slow
wave state (Rosenthal et al., 1984). Additionally, weight gain precedes
torpor in hibernating animals while, in SAD, increased eating and weight
gain appear coincidently with decreased activity.

Nonetheless, there are other seasonal strategies used by mammals
that may be operative in SAD. Rosenthal et al. (1987a) suggest that the
symptoms of SAD may be coordinated by neurophysiological systems that
vary seasonally in an interrelated way. A growing literature on seasonal
rhythms in human physiology and behavior, discussed below, supports this
hypothesis. While human seasonality is the norm in some physiological

variables, SAD may be an extreme on a seasonality continuum.

3. Human Seasonality

A review of the literature on seasonal variations in hypothalamic
function, thermoregulatory variables, serotonergic function, and circadian
rhythmicity follows. Some of these seasonal variations occur in
physiological functions that are problematic in SAD and may have potential
explanatory value for SAD. These are discussed in a separate section on
biological studies of SAD.

Hypothalamic functions, such as sleep, appetite, mood, motivation,
and autonomic nervous system activity, are prime candidates for seasonal
variation due to their involvement in regulating physiological responses to

environmental demands. Indeed, one quarter of the normal population

6

experiences increased appetite and sleep and decreased mood in winter
(Terman 1988; Bartko and Kasper, 1989; Rosen and Roenthal, 1991). In
addition to increased sleep duration, slow wave sleep is also increased in
winter compared with summer in normal control subjects (Lacoste and
Wirz-Justice, 1989). Plasma norepinephrine and dopamine beta
hydroxylase, reﬂecting short and long term sympathetic neuronal activity,
respectively, are increased in winter (Lacoste and Wirz-Justice, 1989). Not
surprisingly, systemic blood pressure is also elevated in winter (Lacoste
and Wirz-Justice, 1989).

Seasonal changes occur in thermoregulatory and hypothalamic-
pituitary-thyroid variables, even in subjects not exposed to extreme
temperature variations. Mean core temperature decreases in winter
compared with summer or autumn (O'Malley et al., 1984; Buguet et al.,
1989). Physiological responses to cold exposure vary seasonally with
autumn being the time of the slowest rewarming rate and the greatest
subjective sensitivity to cold (Lacoste and Wirz-Justice, 1989). Thyroid
hormones and basal thyrotropin are increased in autumn and winter
(Lacoste and Wirz-Justice, 1989). Resting metabolic rate is increased in
winter (Yurigi et al., 1972). Additionally, there is seasonal variation in the
thyrotropin response to exogenous stimulation with thyrotropin releasing
hormone (Lacoste and Wirz-Justice, 1989).

The serotonergic system appears to be uniquely susceptible to
seasonal inﬂuences as evidenced by high amplitude seasonal rhythms in
several variables (Lacoste and Wirz-Justice 1989). The concentration of
serotonin in postmortem hypothalamic tissue is highest in autumn and

lowest in winter (Carlsson et al., 1980). Cerebrospinal ﬂuid (CSF) 5-

7

' hydroxyindole-acetic acid (5-HIAA), a serotonin metabolite, is highest in
winter as is platelet uptake of serotonin (Win-Justice and Richter, 1979).

Seasonal changes occur in the timing of circadian rhythms
(Pittendrigh, 1988). The timing of peak cortisol concentrations is delayed 90
minutes in winter compared with summer (Grifﬁths et al., 1986;
Kennaway and Royles, 1986). The timing of peak core body temperature is
delayed 45 minutes in winter compared with summer (Kloeppel, 1982). The
morning decline (offset) in nocturnal melatonin secretion is delayed in,
winter compared with summer (Grifﬁths et al., 1986; Kennaway and
Royles, 1986). The volume of the human suprachiasmatice nuclei and the
number of vasopressin-immunoreactive neurons contained within it are
twice as high in autumn as in summer (Hofman and Swaab, 1992). These
physiological ﬁndings of delayed circadian rhythms parallel the delay in
morning alertness reported in winter by healthy subjects (Lacoste and
Wirz-Justice, 1989).

The limitations of seasonal studies need to be acknowledged.
Variations in age, sex, methodology, latitude, and deﬁnition of season can
inﬂuence results. Similarly, zeitgebers (light, temperature), weather
modiﬁcations, and social cues (activity, eating) can confound results
(Lacoste and Wirz-Justice, 1989). Nonetheless, the widespread capacity for
seasonal change in mammalian physiology is unlikely to be absent in
humans or unsusceptible to pathology (Mrosovsky, 1988). SAD may well be
a case in point. Pathology may arise from a supersensitivity to
environmental cues that allows seasonal rhythms to reach an extreme, or,
conversely, ﬁ'om a subsensitivity to environmental signals that precludes

the development of normal seasonal rhythms.

8

B. Seasonal Affective Disorder
1. Diagnostic Validity

While seasonal changes in mood and behavior are a normal part of
human physiology, SAD represents an extreme in seasonal rhythms of
mood and energy that interferes with daily activities. SAD was ﬁrst
described in the scientiﬁc literature in the early 1980's (Lewy et al., 1982),
although there are many references to seasonal depression throughout
medical history (W ehr and Rosenthal et al., 1989). The diagnostic validity of
a psychiatric disorder is based on six areas of evidence: symptomatology,
natural course of the disease, epidemiology, family history, laboratory tests,
and response to a speciﬁc treatment (Spitzer and Williams, 1989). A
discussion of the diagnostic validity of SAD follows.

Seasonal affective disorder is diagnosed by the presence of a
depressive or bipolar (manic-depressive) disorder in which depressive
episodes recur annually in the autumn and winter and remit
spontaneously in the spring and summer. Initial diagnostic criteria,
developed by the National Institute of Mental Health, required at least two
depressive episodes that occurred in the autumn or winter and remitted in
spring and summer (Rosenthal et al., 1984), one of which fulﬁlled Research
Diagnostic Criteria for major depressive disorder (Spitzer et al., 1978).
Depressive episodes for which an antecedent psychosocial stressor could be
identiﬁed were excluded. The current diagnostic criteria is based on the
Diagnostic and Statistical Manual of Mental Disorders, third edition,
revised (DSM-III-R; American Psychiatric Association, 1987 ). DSM-III-R
contains a provision for the speciﬁcation of a seasonal pattern to both
recurrent major depression and bipolar disorder (manic-depression).

DSM-III-R criteria stipulate that recurrent onsets of depressive episodes

9

must occur within the same 60-day window each autumn. Similarly,
remission of depressive episodes must occur within the same 60-day
window each spring. Additionally, all depressive episodes must meet
criteria for major depression, with winter depressive episodes
outnumbering other depressive episodes by a ratio of at least 3:1. There
must be at least three winter episodes, two of which occur in consecutive
years. Similar to the original criteria, diagnosis is precluded when

seasonally related psychosocial stressors are in evidence.

2. Clinical and Demographic Features

The clinical symptoms of SAD include fatigue, hypersomnia
(increased sleeping, often up to 15 h per day), increased appetite, especially
for carbohydrates, weight gain, and depressed mood (Rosenthal et al.,
1984). Hypersomnia and increased appetite with weight gain are
considered atypical depressive symptoms, in contrast to the more common
symptoms of insomnia and weight loss seen in major depressive disorders
in general. However, a sizeable proportion of SAD patients experience
insomnia, loss of appetite, and weight loss (Thase, 1989; Winton and
Checkley, 1989). Depressed SAD patients consume carbohydrate-rich foods
more often than normal control subjects and exhibit a seasonal rhythm of
carbohydrate intake with a maximum value in winter (Krauchi and Win-
Justice, 1988). The most ﬁ'equent and intense symptom is a general lack of
energy (Terman et al., 1989b). Moreover, diurnal variation in symptom
severity (mood regularly worse upon awakening or before retiring) is
present in 85% of SAD patients (Graw et al., 1990).

Little is known about the natural course of SAD. The age of onset of
SAD is typically in the third decade of life (Hellekson, 1989). The onset of

10

depressive symptoms is usually in September, October, or November
(Dilsaver et al., 1990; Mayor et al., 1991). Spontaneous remission usually
occurs in March or April, with an average episode duration of 5.2 months
(Hellekson, 1989). The summer mood state of SAD patients is normal
(euthymic) or occasionally manic. While SAD is considered a chronic
disorder with recurrent depressive episodes every winter, longitudinal
studies have not yet been performed.

The incidence of SAD appears to be dependent on latitude (Potkin et
al., 1986; Rosen et al., 1990). Rosen et al. (1990) estimated the prevalence of
SAD over a range of latitudes from Florida (1.4%) to New York (9.7%).
Depue (1989b) suggests that the overall prevalence of SAD is 3% while Thase
(1989) estimates the population prevalence to be 4-6%. In addition to many
North American and European reports, SAD is reported in Asia
(Nagayama et al., 1991) and Australia (Boyce and Parker, 1988; McIntyre et
al., 1989a). The female to male ratio of SAD is usually given as 4:1
(Hellekson, 1989). A positive family history for depression in ﬁrst degree
relatives of SAD patients is reported (Hellekson, 1989).

Phototherapy is the treatment of choice for SAD, being effective in 50-
80% of cases (Terman et al., 1989b). Clinical remission usually occurs
within 4-7 days (Rosenthal et al., 1984). Maintenance of the antidepressant
response depends upon daily phototherapy, with relapse occurring within
4-7 days after withdrawal from treatment. Other types of major depression
do not respond to phototherapy (Blehar and Lewy, 1990), suggesting that
SAD is a distinct psychiatric disorder (Stewart et al., 1990). Little is known
about the response of SAD patients to antidepressant medications. Initial
reports suggest that SAD can be treated with imipramine, a noradrenergic

11.
reuptake blocker (Thase, 1989), and tranylcypromine, a monoamine oxidase

inhibitor (Dilsaver et al., 1990).

3. Biological Studies

While the validity of SAD receives support from epidemiologic and
therapeutic studies (Blehar and Rosenthal, 1989; Rosen et al., 1990),
laboratory studies do not reveal a consistent proﬁle of biological
abnormalities (Skwerer et al., 1988). Moreover, no biologic marker
discriminates unequivocally between SAD patients and either non-seasonal
depressed patients or healthy persons. Interestingly, none of the biological
abnormalities commonly seen in major depression, such as non-
suppression of cortisol with dexamethasone, decreased thyroid stimulating
hormone response to exogenous protirelin, and early onset of REM sleep,
are present in SAD (Rosenthal et al., 1984; James et al., 1986; Skwerer et
al., 1988).

However, numerous differences are noted between SAD patients and
normal control subjects, often in physiological variables that exhibit
seasonal rhythms. For example, SAD patients spend less time in slow
wave sleep in winter than in summer (Rosenthal et al., 1984), a pattern
which is opposite to that described in normal control subjects (Lacoste and
Wirz-Justice, 1989). Plasma norepinephrine concentrations, which
normally increase in winter, are decreased in SAD patients compared with
control subjects (Rosenthal et al., 1987b). Depressed SAD patients report
feeling energized 1 h after carbohydrate ingestion, whereas normal control
subjects report sedation (Rosenthal et al., 1989). A disturbance in the '
medial hypothalamic satiety system is proposed to explain the presence of

12

these hypothalamic-mediated symptoms in SAD (Krauchi and Win-
Justice, 1988).

Thermoregulatory and hypothalamic-pituitary-thyroid axis variables
often differ between SAD patients and normal control subjects. The rate of
heat loss following an endogenous heat challenge induced by exercise is
decreased in depressed SAD patients compared with normal control
subjects (Arbisi et al., 1989). SAD patients also exhibit a higher resting
metabolic rate compared to control subjects (Gaist et al., 1990). Lacoste and
Wirz-Justice (1989) suggest that the response of the hypothalamic-pituitary-
thyroid axis to long term changes in environmental temperature may be
involved in the etiology of SAD. However, thyroid hormones, thyrotropin,
and mean core body temperature do not differ between SAD patients and
healthy control subjects (Jacobsen et al., 1987a; Skwerer et al., 1988;
Rosenthal et al., 1990).

The serotonergic system which displays seasonal variation in
normal control subjects shows some abnormalities in SAD patients. The
binding of 3H-imipramine to the serotonin transporter on platelet
membrane is decreased in SAD patients compared to normal control
subjects (Szadoczky et al., 1989;1991). Additionally, stimulation with m-
chlorophenylpiperazine, a serotonin post-synaptic agonist, increases
cortisol and prolactin secretion more in SAD patients than in normal
control subjects (J acobsen et al., 1989). Wurtman and Wurtman (1989)
suggests that a deﬁciency in serotonin, perhaps related to dietary intake of
the precursor tryptophan, may be involved in triggering winter depressive
episodes. Treatment with either tryptophan (McGrath et al., 1990) or
serotonergic drugs (O’Rourke et al., 1989) has met with initial success.

13

Circadian rhythm function may differ between SAD patients and
normal control subjects in terms of phase position (Lewy et al., 1987) and
amplitude (Czeisler et al., 1987). These ﬁndings are the basis of a current
hypothesis for the etiology of SAD and are discussed fully in a separate
section.

The retinal response to light is of interest since it is the ﬁrst step in
the regulation of circadian rhythms. However, data are controversial in
this area. Electrooculographic (EOG) ratios which reﬂect the activity of the
retinal pigmented epithelium are lower in depressed SAD patients than in
normal control subjects (Lam et al., 1991). Decreased EOG ratios may thus
reﬂect a deﬁciency in photoreceptor disk renewal and, consequently, a
subsensitivity to light (Reme et al., 1990).

On the other hand, some evidence suggests that SAD patients are
supersensitive to light (Beersma, 1990). For example, depressed SAD
patients are more sensitive to detecting very dim light than normal control
subjects (Oren et al., 1991b). Additionally, SAD patients are more sensitive
to the effect of light on melatonin suppression (McIntyre et al., 1990;
Thompson et al., 1990). For example, depressed SAD patients suppress
melatonin to a greater extent than normal control subjects (40% versus 87%
of baseline values) following exposure to bright light (2000 lux for 1 hour;
Thompson et al., 1990). Interestingly, this supersensitivity to light in terms
of melatonin suppression (Thompson et al., 1990) and dim light detection
(Terman and Terman, 1991) is not present in SAD patients in the summer,
suggesting a seasonal variation in light sensitivity.

In summary, SAD appears to be a distinct psychiatric disorder in
which depressive episodes are associated with the seasonal decrease in

photoperiod. At present, the diagnosis of SAD is based on symptomatology

14

‘ and history of seasonally occurring depressive episodes. The development
of a biological marker for SAD would increase diagnostic speciﬁcity and

have implications for future research.

C. Phototherapy of SAD

The general concensus within psychiatry is that phototherapy is an
effective treatment for SAD (Blehar and Rosenthal, 1989). Despite its
efﬁcacy, the important parameters of phototherapy such as dose, spectrum,
and timing, are incompletely characterized (Lewy et al., 1986; Rosenthal et
al., 1988). Furthermore, neither the pathophysiological mechanism
underlying SAD nor the mechanism of action of the antidepressant effect of
light are known. A discussion of the important aspects of phototherapy
follows, including protocol, efficacy, dosage, and design and methodological
issues. Biological effects of phototherapy, both positive and negative, are

also discussed.

1. PI‘O‘IOCOI

The current trend of treating SAD with light began with a case study
by Lewy and colleagues (1982). In this study, phototherapy was based on the
idea of imitating the natural light of a summer day in length, intensity, and
spectrum. The subject gazed at a bright light source for 3 h in the morning
and 3 h in the evening. A therapeutic effect occurred within 1 week, but
relapse occurred within 1 week of withdrawal from treatment. Controlled
group studies soon replicated this ﬁnding (Rosenthal et al., 1984; 1985).
Subsequent studies revealed that imitating natural light in terms of
intensity (Wirz-Justice et al., 1986) and photoperiod (Wehr et al., 1986) is
unnecessary. The current protocol of phototherapy is 2 h daily during the

15

months of winter depression. Moreover, phototherapy administered prior to
onset of symptoms can prevent subsequent episodes (Meesters et al., 1992).

2. Efficacy

The efﬁcacy of phototherapy is measured by the magnitude of the
decrease in depressive symptoms. The Hamilton Rating Scale for
Depression (HRSD; Hamilton, 1961) is commonly used to measure changes
in depressive symptoms. However, this scale is not entirely appropriate for
SAD since it underestimates severity when atypical depressive symptoms
(hypersomnia, increased appetite, weight gain) are present. The Seasonal
Affective Disorder Rating Scale was developed to measure these symptoms
and serve as a supplement to the HRSD (Rosenthal and Heffernan, 1986).
The most reliable conclusions on treatment efficacy are derived from those
studies using remission criteria incorporating both absolute (HRSD score
within the normal range, i.e., <8) and relative (50% reduction of HRSD
score) improvement. Applying these criteria to pooled results of individual
studies, phototherapy is successful in 53% of patients (Terman et al., 1989b).

Individual responsiveness to phototherapy may be predicted by
clinical symptoms. Hypersomnia is associated with a superior response to
morning light (Avery et al., 1991). Pretreatment severity of atypical
depressive symptoms (hypersomnia, increased appetite, weight gain)
correlates highly with treatment success (Nagayama et al., 1991) while
pretreament severity of the typical symptoms of depression (insomnia,
decreased appetite, weight loss) predicts a poor response to phototherapy
(Stinson and Thompson, 1990). Terman et al. (1992) suggest that the ratio of
atypical symptoms to total depressive symptoms is the best predictor of

clinical success, with a ratio above .29 associated with a positive response.

16

The dose of light in phototherapy, deﬁned as the amount of light
falling on the subject, is conventionally set at 2500 lux. This dose, termed
bright light, appears to be more effective than a dose of 300 lux light, or dim
light (Rosenthal et al., 1984; 1985; James et al., 1985; Terman et al., 1989b;
Kasper et al. 1990). However, some studies report equal efﬁcacy of bright
and dim light (Wirz-Justice et al., 1986).

The efﬁcacy of a therapeutic agent often depends upon the time at
which it is administered (Reinberg, 1983). Phototherapy is no exception.
The efficacy of a given dose of light may depend upon the time at which it is
delivered relative to a critical photosensitive period (Kripke, 1984) or to the
timing of circadian rhythms (Lewy et al., 1987; Terman et al., 1989b). In
some studies, morning light has a superior antidepressant effect to that of
evening light (Lewy et al., 1987; Avery et al. 1990b; 1991; Terman et al., 1988;
Sack et al., 1990). However, other studies fail to ﬁnd a treatment difference
between morning and evening phototherapy (James et al., 1985; Hellekson
et al., 1986; Yerevenian et al., 1986; Wirz-Justice et al., 1989) or morning
and afternoon phototherapy (Jacobsen et al., 1987b). Small sample sizes
and different methodologies, especially in the deﬁnition of treatment
success, may account for these conﬂicting results. The pooled results of
individual morning versus evening crossover studies support the
superiority of morning phototherapy, with 37% of subjects responding
exclusively to morning light, 25% responding at both times of day, 3%
responding exclusively to evening light, and 35% to neither (Terman et al.,
1989b).

In addition to influencing treatment efﬁcacy, the timing of
phototherapy may inﬂuence the dose of light required to produce an
antidepressant response. For example, dim light (300 lux) is effective when

17

administered just before and during awakening (Terman et al., 1989a),
even though it is ineffective when administered after awakening (Rosenthal
et al., 1984). Furthermore, twilight transitions may be more effective than
abrupt dark-light transitions. Phototherapy delivered with graded
increases of light intensity just before and during awakening (dawn
simulation) is an effective antidepressant treatment (Terman et al., 1989a;

Terman and Schlager, 1990; Avery et al., 19903).

3. Design and Methodological Issues

Design and methodologic issues play an important role in the
interpretation of results from SAD studies. The widespread publicity about
phototherapy generates high patient expectations for treatment success
which is often associated with nonspeciﬁc placebo responses to the
therapeutic setting. The lack of a convincing placebo control further
exacerbates the problem of distinguishing a true response between various
treatment protocols.

A placebo control, indistinguishable from the active treatment, is
technically difﬁcult to develop for light since subjects are not blind to the
treatment. A low intensity light (dim light) is often used as a placebo
control, but with questionable credibility (Eastman, 1990). Even with an
adequate placebo control, the magnitude of placebo response to
phototherapy needs to be estimated. Dim light controls achieve only 11%
remissions, much lower than the typical placebo response in studies of
other depressive disorders (Terman et al., 1989b).

Experimental design issues are likewise controversial. A crossover
design with subjects serving as their own controls is often used as a means

to decrease the variability of subject response and the number of subjects

18
required. However, crossover designs often give rise to carryover and order
effects (Terman et al., 1990). While parallel designs avoid these
disadvantages, they require more subjects and have increased inter-subject
variability. Both types of designs often yield equivocal results due to the low
power to detect treatment differences (type II error) resulting from small
sample sizes.

Methodological differences among studies may also lead to disparate
results. Differences in entry criteria, pretreatment severity, and remission
criteria may account for some of the conﬂicting results. Incidental
variables, such as the changing photoperiod over the course of a study,
variability in the amount of naturally occurring sunlight, and changes in
the duration and timing of sleep, may confound results and obscure

treatment effects.

4. Biological Effects

Phototherapy induces multiple physiological effects, but it is unclear
which of these are mechanistically important or mere epiphenomena of
treatment. Moreover, it is unclear whether there are negative effects of
phototherapy in addition to the therapeutic antidepressant effect. For
example, while phototherapy is safe for short term use (Rosenthal et al.,
1984), it is unknown whether there are hazards of chronic use.

Numerous physiological effects of clinical importance are observed
with phototherapy. Phototherapy increases slow wave sleep (Skwerer et
al., 1988), decreases sleep duration (Rosenthal et al., 1984) and carbohydrate
intake (Krauchi and Wirz-Justice, 1988) and normalizes mood (Terman et
al., 1989b). Phototherapy increases plasma norepinephrine concentrations
(Skwerer et al., 1988) and theodensity of 3H-imipramine binding sites on

19

platelet membranes (Szadoczky et al., 1989; 1991). Phototherapy reduces
resting metabolic rate (Gaist et al., 1990). Additionally, phototherapy
normalizes temperature recovery from an endogenous heat challenge
(Arbisi et al., 1989).

Phototherapy also suppresses nocturnal melatonin secretion (Winton
et al., 1989) and affects circadian rhythms (Lewy et al., 1987). These
physiological effects are the basis of two current hypotheses of the
therapeutic effect of light in SAD and are discussed in following sections.

Safety with phototherapy focuses on the type of light source and the
health status of the individual (Weale, 1988; Terman et al., 1990b).
Potential harmful effects of ultraviolet B (UVB), ultraviolet A (UVA), and
visible radiation are a concern. A person staring directly at a phototherapy
unit for 2 h per day for 90 days receives an annual UVB dosage of 1.7 J/cm2
(Oren et al., 1990) which is higher than the dosage associated with an
increased risk (5- 10%) of developing cortical cataracts. UVA radiation may
present a similar, yet less serious risk. Visible radiation, especially within
the violet and blue wavebands, may have deleterious effects on the retina
(Paramore and King, 1989). While emissions from phototherapy units do
not exceed phototoxic levels by any current national standard (Sliney and
Wolbarscht, 1980 ), there are no standards outlining safe exposure limits for
long term phototherapy.

In some situations, even short term exposure may be associated with
adverse effects. For example, subjects with compromised ocular status
may be at risk. Aphakia (loss of the ocular lens), glaucoma (increased
intraocular pressure), retinal diseases, inﬂammatory diseases of the eye,
and optic nerve disease may all be exacerbated by exposure to bright light
(Terman et al., 1990b). SAD patients taking photosensitizing drugs (tricyclic

Z)

antidepressants, phenothiazines, psoralens, porphyrins, tetracycline)
during phototherapy may be at risk for erythema (sunburn; Terman et al.,
1990b). Side effects have been noted in unmedicated patients with normal
ocular status undergoing phototherapy, including headache, eye irritation,
agitation, and mania (Schwitzer et al., 1990).

In summary, phototherapy is considered the treatment of choice for
SAD with clinical remission achieved by over 50% of patients within 1 week.
Well-controlled studies of phototherapy are difﬁcult to achieve due to the
lack of a suitable placebo control for light. The development of a placebo
control depends, in part, on an understanding of the interaction of light and
human physiology and psychology.

D. Photoblology of SAD
1. Photobiologlcal Effects in Humans

Phototherapy falls within the realm of photobiology and therefore
requires an understanding of the terms and principles of photobiology
(Parrish, 1981). The science of photobiology is concerned primarily with
absorption of radiation by photosensitive compounds within the human
body. The energy of the absorbed photon may be dissipated as heat or
harnessed in a photochemical reaction which induces biochemical and
cellular changes and, ultimately, an observable biological effect.

Photobiologic effects are initiated in the skin and eye. Dermally-
mediated photobiologic effects are common due to the large surface area
and blood volume of the skin which allow the irradiation of many
photosensitive compounds (Anderson and Parrish, 1981). For example,
absorption of visible radiation by circulating bilirubin is the basis of skin

phototherapy for neonatal hyperbilirubinemia (Sisson, 1976).

21

Retinal photobiologic effects, whether visual or non-visual, are
mediated by the photosensitive rods and cones. Phototherapy is believed to
exert its antidepressant effect through eye exposure (Wehr et al., 1987 ).
Two retinal, yet non-visual, photobiologic effects, the suppression of
melatonin secretion (Lewy et al., 1980) and the entrainment of circadian
rhythms (Wever, 1986), are the focus of much SAD research and are

discussed in the following sections.

2. Dosage

Photobiologic effects depend upon three variables of light exposure:
intensity, duration, and spectrum (Thorington, 1985). Intensity, or dose,
refers to the amount of light being delivered per second. Dosage is the total
amount of light delivered (intensity x duration). Photobiologic effects are
dosage dependent. The standard dose of phototherapy (2500 lux) given for 2
h yields a radiometric dosage of 4.86 Joule/cmz. While a complete dosage-
response curve has not been elucidated, preliminary data are available.
With a constant dose of 2500 lux, .5 h of phototherapy is less effective than
1.0 or 2.0 h (Hellekson et al., 1986; Wirz-Justice et al., 1987 ; Terman et al.,
1989c). The reciprocal relationship between dose and duration suggests
that an increase in dose (10,000 lux) may allow a proportional decrease in
duration (.5 h). Preliminary results support this prediction (Terman et al.,
1990a). Unfortunately, potential risks of phototherapy increase with higher
intensities of light.

3. Action Spectrum
In addition to intensity and duration (dosage), photobiological effects
are wavelength dependent (Anderson and Parrish, 1981). The wavelength
range that produces a speciﬁc biologic effect is termed its action spectrum

22

and is determined by the photosensitive compound underlying the effect.
Preliminary data are inconclusive regarding the action spectrum for the
antidepressant effect of light. Brainard et a1. (1990) notes that neither blue
(430-465nm) nor red (615-655nm) light is as effective as broad spectrum
white light in treating SAD. Green light (505-555nm) is superior to red light
(Oren et al., 1991a), but inferior to that of white light (Stewart et al., 1991).
Methodological differences, primarily in the quantiﬁcation of light, may
account for the lack of progress in determining the optimal action spectrum

for the antidepressant effect of light.

4. Quantification of Light

Since both dosage and spectrum inﬂuence the antidepressant effect of
light, a quantity that combines both parameters is needed in phototherapy
trials. However, this has not been the case in most previous studies of SAD.
Photometric quantities are used extensively in phototherapy. Photometric
units are based on the spectral sensitivity of the human eye and describe
light in terms of its ability to evoke a visual response in humans. The
amount of light falling on a subject is termed illuminance in the
photometric system. Illuminance, measured in lux, is the most commonly
reported quantity in phototherapy trials (Rosenthal et al., 1984; 1985;
Terman et al., 1989b).

The use of photometric quantities in phototherapy implies a
dependence on visual perception which may be unfounded given the
positive effect of phototherapy administered during sleep (Avery et al.,
1990a). Radiometric quantities, which measure the absolute amount of
radiant energy without regard for its ability to evoke a visual sensation,

may be a more appropriate quantity to measure (Indyck, 1976). In the

23

radiometric system, the radiant power falling on a unit area of irradiated
surface (i.e., human subject) is termed irradiance. Irradiance, measured
in Watts per square centimeter (W/cm2), is the basis of dose in photobiology
(Yerevanian et al., 1986).

A more speciﬁc and precise measure of dose is spectral irradiance
since it encompasses both dose and action spectrum (W/cm2 within a
speciﬁc waveband). Photon density is a similar measure of dose within the
action spectrum, but it is based on the number of photons delivered
(photons/cmz), rather than the total power delivered. Photon density may
be the best quantity for dose calculations of phototherapy because
photochemical reactions ultimately depend upon the number of photons
absorbed (Brainard et al., 1990).

In summary, phototherapy of SAD may achieve its effect according to
established principles of photobiology. However, the critical variables of
photobiologic effects, dosage and action spectrum, are unknown at this
time. Both dosage and action spectrum are involved in the quantiﬁcation of
light.

E. Melatonin Hypothesis of SAD

Two main hypotheses have been proposed to explain the
antidepressant effect of light, both of which are based upon physiological
processes that are photosensitive in humans. The ﬁrst hypothesis which is
discussed in this section is based upon the suppressant effect of light on
melatonin secretion (Rosenthal et al., 1984). The second hypothesis, which
is discussed in the following section, is based upon the effect of light on

circadian rhythms (Lewy et al., 1987).

24

According to the melatonin hypothesis, SAD resembles the
behavioral and physiological changes of some animals during winter
which are induced by the seasonal change in photoperiod and mediated
physiologically through alterations in melatonin secretion (Gwinner, 1981).
Therefore, decreased daylength may induce increased melatonin secretion
in humans which, in turn, may precipitate the seasonal change in mood
and behavior in SAD patients (Rosenthal et al., 1985). Since bright light can
suppress melatonin secretion (Lewy et al., 1980), extending the winter
photoperiod with bright artiﬁcial light may alleviate depressive symptoms
(Rosenthal et al., 1985). Critical analysis of the melatonin hypothesis
requires an understanding of the physiological regulation of melatonin

secretion.
1 . Pineal Gland

Melatonin is produced by the pineal gland, a small structure located
at the base of the pineal recess of the third ventricle, superior to the tectum,
and covered by the occipital lobe of the cerebrum. The pineal gland has
been the subject of inquiry since ancient times. Three thousand years ago,
yoga practitioners described the vertex of the skull, corresponding
anatomically with the pineal, as a distribution center through which solar
radiation energized the body (Leskowitz, 1990). Ancient Greeks considered
the pineal to be the seat of the soul, a concept extended by Descartes in the
seventeenth century. The pineal was often thought of as a vestigial parietal
eye, giving rise to folklore about an eye in the back of the head. Modern
inquiry into pineal gland function began with the discovery of the

biologically active humane melatonin (Lerner, 1958). The pineal has direct

%

photosensory capabilities in lower vertebrates. In mammals the pineal is

neurosecretory in function.

2. Synthesis and Secretion of Melatonin

The pinealocyte is the primary site of melatonin synthesis containing
all the necessary enzymatic machinery for hormone biosynthesis.
Melatonin, or N-acetyl-5-methoxytryptamine, is synthesized from the
amino acid tryptophan. The ﬁrst two steps in the biosynthetic pathway
produce serotonin, or 5-hydroxytryptamine (5-HT), through hydroxylation
and decarboxylation (Axelrod, 1974). Subsequently, 5-HT is acetylated by N -
acetyltransferase (NAT). NAT is the rate limiting enzyme for the
conversion of 5-HT to melatonin (Klein, 1970). The ﬁnal step in the
synthesis of melatonin is catalyzed by the enzyme hydroxyindole-O-
methyltransferase (HIOMT).

Melatonin synthesis is under sympathetic control ﬁ'om the superior
cervical ganglia (Klein, 1979). Synthesis is increased at night due to a large
and sustained increase in norepinephrine release from prejunctional
sympathetic nerve endings (Axelrod, 197 4). Binding of norepinephrine to
I betal adrenergic receptors on the pinealocytes activates the cAMP second
messenger system (Axelrod, 1974; Cowen et al., 1983), resulting in
increased 5-HT uptake and HIOMT activity. Newly synthesized melatonin
diffuses out of pinealocytes into the great cerebral vein and CSF (Reppert,
1980).

The systemic circulation is the major route for delivery of melatonin
to target sites (Rollag et al., 1978). Approximately 70% of circulating
melatonin is bound to plasma albumin (Cardinali, 1972). Melatonin can be

‘ detected in saliva in amounts equal to one third the plasma concentration

%

(Miles et al., 1985). The half-life of melatonin in blood is estimated at 20 to
40 minutes (Cardinali et al., 1983; Terman and Schlager, 1990). Systemic
melatonin is degraded by hepatic 6-hydroxylation to 6-hydroxymelatonin
which is subsequently conjugated with sulfate or glucuronate for urinary
excretion (Cardinali, 1981; Kopin et al., 1961). Binding sites for melatonin
are found in the human brain and ovary (Cohen et al., 1978). Brain binding
sites include the hippocampus and frontal cortex (Brown et al., 1981) as
well as the suprachiasmatic nuclei (Reppert et al., 1988).

Melatonin secretion decreases with age in humans. Mean 24-h
melatonin secretion in adults over age 65 is about half of that of young
adults (19-25 years; Sharma et al., 1989). Similarly, peak melatonin
concentrations in older adults is 50% that of young adults (Sharma et al.,
1989). Melatonin concentrations may rise during the luteal phase of the
menstrual cycle (Webley and Leidenberger, 1986; Brun et al., 1987),
although there is some disagreement on this point (Brzezinski et al., 1988).
The circadian rhythm of melatonin secretion is abolished by destruction of
aﬂ'erent sympathetic ﬁbers to the pineal gland (Li et al., 1989), pinealectomy
(N euwelt and Lewy, 1983), or peripheral beta adrenergic blockade by
pharmacologic agents such as atenolol (Cowen et al., 1983) and propanolol

(Hanssen, 1977).

3. Circadian Rhythm of Melatonin
Melatonin secretion exhibits one of the most stable and high
amplitude circadian rhythms in humans. Negligible concentrations are
present in the blood during the day' with a sharp rise observed after the
onset of darkness (Waldhauser and Dietzel, 1985). Maximum values (100-
200 pg/ml) are reached between 0000 and 0600 h, with peak concentrations

27

generally reached at 0200 h (Waldhauser and Dietzel, 1985). The
concentration of melatonin at 0200 h correlates highly with total 24-h
melatonin secretion (Arendt et al., 1982; r=0.97). The melatonin rhythm is
stable within an individual, but between individuals, there is often great
variability in mean and peak concentrations as well as total amount
secreted (W aldhauser and Dietzel, 1985). However, duration of nocturnal
secretion is stable between individuals at a given time (Waldhauser and
Dietzel, 1985). Duration of nocturnal melatonin secretion varies with
season, being longer in winter than in summer (Illnerova et al., 1985;
Kauppila et al., 1987). A delay in the morning decline (offset) of nocturnal
melatonin secretion may be responsible for the increased duration in
winter (Broadway and Arendt, 1985).

Melatonin secretion is photosensitive (Wurtman, 1963; Lewy et al.
1980). Light has a suppressant effect on melatonin secretion. For example,
exposure to bright light (2500 lux) in the night decreases plasma melatonin
concentrations by 50 to 70% (Lewy et al., 1980; Boyce and Kennaway, 1987).
The effect of light on melatonin secretion follows the principles of
photobiology. First, suppression is intensity dependent (McIntyre et al.,
198%). Second, the degree of suppression depends upon the wavelength of
light. Maximal suppression occurs with light at 509 nm (Brainard et al.,
1985). Photic information is transmitted from the eye to the pineal through
a multisynaptic pathway beginning with the retinohypothalamic tract to
the suprachiasmatic nuclei (SCN; Sadun et al., 1984). Fibers of the SCN
project to the paraventricular nuclei which send axons through the medial
forebrain bundle to preganglionic cells in the intermediolateral cell column
of the spinal cord. Preganglionic ﬁbers project from the superior cervical

ganglia to the pineal. Exposure to light ultimately decreases

%

norepinephrine release from sympathetic terminals and subsequently

reduces melatonin production (Wurtman et al., 1963).

4. Melatonin in SAD

The melatonin hypothesis of SAD is based on lengthening the winter
photoperiod in order to induce a summer pattern of melatonin secretion.
However, subsequent studies show that this hypothesis is inadequate to
explain the therapeutic response of light (Mrosovsky, 1988). For example,
extending the winter photoperiod with 6 h of light is unnecessary for a
positive therapeutic effect (Wehr et al., 1986). Inhibition of melatonin
secretion with a beta adrenergic antagonist rather than light is not effective
in treating SAD (Rosenthal et al., 1988a). Conversely, administration of
exogenous melatonin does not cause a relapse in remitted patients
undergoing phototherapy (Rosenthal et al., 1986). Furthermore, SAD
patients do not secrete more melatonin in winter than normal control
subjects (Rosenthal et al., 1987b; Thompson et al., 1990). Moreover,
suppression of nocturnal melatonin secretion is unnecessary for an
antidepressant effect (Wehr et al., 1986; Rosenthal et al., 1988b; Winton et
al., 1989). Therefore, inducing a summer pattern of melatonin secretion by
the acute suppression of nocturnal secretion is not associated with the
antidepressant effect of light. However, other aspects of melatonin

physiology, discussed in the next section, may be involved in SAD.

F. Circadian Rhythms In SAD
1. Circadian Rhythms and Light
The second hypothesis for the antidepressant eﬁ‘ect of light is based
upon a putative delay in the timing of circadian rhythms in SAD patients
and the ability of light to inﬂuence circadian timing (Lewy et al., 1987).

Q

Human physiology is characterized by daily rhythms. Many of these
rhythms, termed circadian, are innate and have a period (cycle length) of
about 24 h. For example, the daily rhythm of melatonin secretion is
maintained even in constant low level illumination (Wever, 1986).
However, under normal conditions, circadian rhythms are entrained, or
synchronized, to the environmental 24-h day by various time cues, or
zeitgebers (Wever, 1986). A circadian rhythm can be characterized by its
period, amplitude (peak to mean), and phase position (timing relative to
clock time). While human circadian rhythms appear to be remarkably
stable (Wirz-Justice and Wehr, 1983), disturbances in circadian rhythm
function may be involved in some depressive disorders, including SAD.
Recent evidence increasingly points to light as a chronoactive agent,
one which can affect circadian rhythms (Wever, 1986; Czeisler et al., 1989).
For example, the environmental light-dark cycle synchronizes the
endogenous rhythm of melatonin to the 24-h day. Additionally, repeated
exposure to bright light for several days shifts the timing of circadian
rhythms (Czeisler et al., 1989). The direction of the shift depends upon the
timing of light exposure. For example, morning light exposure shifts the
timing of circadian rhythms to an earlier, or advanced, phase position.
Evening light exposure shifts the timing of circadian rhythms to a later, or
delayed, phase position. The human phase-response curve expresses this
relationship between the timing of light exposure and the subsequent

response in terms of phase shifting (Czeisler et al., 1989).

2. Timing of Circadian Rhythms in SAD
Depressed SAD patients exhibit evidence of delayed timing of
circadian rhythms. The melatonin rhythm is delayed about 90 minutes in

3)

depressed SAD patients compared with normal control subjects as assessed
by the onset time of nocturnal secretion (Lewy et al., 1987; Terman et al.,
1988; Sack et al., 1990). The core temperature rhythm may be also be phase
delayed (Avery et al., 1987 ; 1990b). For example, the timing of peak core
temperature occurs about 90 minutes later in SAD patients than control 1
subjects (Avery et al., 1987). However, some studies do not show phase
delayed rhythms in SAD patients (Rosenthal et al., 1987b; 1990). This
discrepancy may be accounted for by the presence of both phase-delayed and
phase-advanced rhythms within the SAD population (Rosenthal et al.,
1990).

The phase shifting effect of light is evident after several days of
repeated timed light exposure. For example, following 1 week of morning
phototherapy, the onset time of evening melatonin secretion is advanced by
60-90 minutes compared with pretreatment values (Lewy et al., 1987;
Terman et al., 1988, 1989a; Sack et al., 1990). The timing of the circadian
rhythm of core temperature is also advanced following 1 week of morning
phototherapy (Avery et al., 1987 ; Rosenthal et al., 1990). Following 1 week of
evening phototherapy, the onset time of nocturnal melatonin secretion is
delayed. (Lewy et al., 1987). The light-induced advance in the timing of the
melatonin rhythm and core temperature rhythm correlates with an
antidepressant effect (Avery et al., 1987 ; Lewy et al., 1987; Sack et al., 1990;
Rosenthal et al., 1990). Based upon the differential effect of morning versus
evening light and the delay in circadian rhythms in SAD, Lewy et a1. (1987)
propose that light exerts an antidepressant effect in SAD because of its
ability to correct an underlying circadian rhythm disturbance.

31

3. Amplitude of Circadian Rhythms in SAD

In addition to shifting the timing of circadian rhythms, light also
enhances the amplitude of circadian rhythms (Czeisler et al., 1989). The
amplitude-enhancing effect of light is more pronounced in circadian
rhythms characterized by a low amplitude. Low amplitude circadian
rhythms are believed to be relatively unstable and susceptible to
manipulation by chronoactive agents (Wever, 1986). Low amplitude
circadian rhythms may reﬂect an abnormality in the responsiveness of the
human biological clock (Czeisler et al., 1987).

SAD patients may display low amplitude or unstable circadian
rhythms (Czeisler et al., 1987 ; Rosenthal et al., 1990). For example, SAD
patients advance the timing of the melatonin rhythm to a greater extent
than healthy subjects in response to a controlled light-dark cycle or to bright
light exposure in the morning (Lewy et al., 1987; Sack et al., 1990).
Additionally, combined morning and evening light schedule shifts both the
melatonin rhythm and the core temperature rhythm in some SAD patients,
but not in normal control subjects (Rosenthal et al., 1987b). Czeisler et a1.
(1987 ) propose that phototherapy may exert its antidepressant effect by
increasing the amplitude of circadian rhythms (Czeisler et al., 1987), thus
rendering them less susceptible to manipulation by chronoactive agents
(Mrosovsky, 1988). The amplitude of the core temperature rhythm in SAD
patients is increased following phototherapy (Rosenthal et al., 1990).

In summary, both the phase position and amplitude of circadian
rhythms, most notably melatonin, differ between depressed SAD patients
and normal control subjects. Bright light exposure in the morning
increases amplitude and advances circadian phase position. Bright light

exposure in the morning is also associated with an antidepressant effect in

32

SAD patients. It is unknown whether these changes in circadian rhythms
are essential for the therapeutic effect of light in SAD.

33

Objectives

The overall objective of the following series of experiments was to
study the effect of different types of light exposure on depressive
symptomatology and melatonin concentrations in patients with SAD. The
same general crossover design was used in all experiments. The design
consisted of 1 week of daily phototherapy followed by a 1-week withdrawal
period which was then followed by the crossover to an alternate type of
phototherapy for 1 week. Measurements were obtained at pretreatment and
the end of treatment in both the ﬁrst trial of phototherapy and in the second
trial.

Chapter 1. Effect of Full Spectrum versus Cool White Phototherapy
on Melatonin Concentrations and Depression Scores in
SAD Patients

35

introduction

Morning (0600-0800 h) phototherapy utilizing a light source with a
spectral distribution similar to sunlight (full spectrum) reduces depression
scores, sleep duration, appetite, and fatigue in SAD patients (Rosenthal et
al., '1984; Avery et al., 1987; Lewy et al., 1987; Terman et al., 1988).
Moreover, morning phototherapy has both an acute and a cumulative effect
on melatonin concentrations in blood. The acute effect is a suppression of
melatonin secretion. Melatonin concentrations in blood decrease toward
daytime concentrations within 1 h of morning bright light exposure (W ehr
et al., 1986; Terman and Schlager, 1990). However, acute suppression of
melatonin is not necessary for the antidepressant effect of light (W ehr et al.,
1986; Winton et al., 1989). The cumulative effect of phototherapy on
melatonin secretion is observed in changes in the timing of the circadian
rhythm. One week of morning phototherapy advances the onset of
nocturnal melatonin secretion to an earlier time of day (Lewy et al., 1987 ).
Clinical improvement in SAD patients undergoing phototherapy is
correlated with an advance in the timing of nocturnal melatonin secretion
(Lewy et al., 1987 ; Terman et al., 1988; Lewy and Sack, 1989: Sack et al.,
1990). Whether an advance in the timing of the melatonin rhythm is
necessary or sufﬁcient for the therapeutic effect is unknown.

Measurement of nocturnal melatonin secretion is accomplished with
serial blood sampling in the evening under dim (<50 lux) illumination in
order to avoid the acute suppressant effect of light (Lewy et al., 1987).
Onset time of nocturnal melatonin secretion is deﬁned as the time at which
melatonin concentrations rise above a set criterion (Lewy et al., 1987 ; Sack
et al., 1990). Melatonin can also be detected in saliva (Miles et al., 1985b).
Salivary melatonin exhibits a circadian rhythm which mirrors that 'of

35
plasma (Miles et al., 1985a). Whether serial saliva samples can be used to
detect light-induced changes in the onset of nocturnal melatonin secretion
is unknown.

At the present time, it is unclear whether improvement in
depression scores and modulation of the melatonin rhythm with
phototherapy depends upon full spectrum sunlight-simulating light.
Whether other types of white light, such as cool white, produce similar
results in depression scores and melatonin secretion is unknown.

Therefore, an experiment was conducted to quantify changes in
depressive symptoms and salivary melatonin concentrations in SAD
patients with full spectrum and cool white phototherapy. The purpose of
the experiment was threefold: 1) to determine whether saliva samples can
be used to measure light-induced changes in the onset of nocturnal
melatonin secretion; 2) to determine whether full spectrum light is
necessary for the antidepressant effect; and 3) to determine whether an
advance in the onset time of nocturnal melatonin secretion is suﬁcient to

induce clinical remission.
Materials and Methods

Subjects, Human subjects, aged 18-59, were recruited through
newspaper articles and referrals from local clinicians. The admission
criteria for the study included 1) a DSM-III-R diagnosis of recurrent major
depression, seasonal pattern (American Psychiatric Association, 1987), 2) a
history of depressive episodes that occurred in the fall and winter and
remitted by the following spring for the two previous years, and 3) a score of
at least 18 on the 17 -item Hamilton Rating Scale for Depression (HRSD;
Hamilton, 1967). Exclusion criteria included pregnancy, concomitant

treatment with antidepressant medications or beta adrenergic antagonists,

37
and suicidal behavior. Subjects were free of psychopharmacologic
. medications for at least 12 days prior to entering the study and remained so
for the duration of the study.

Subjects who met the admission criteria underwent medical
assessment to ensure safety during phototherapy. A medical history was
obtained with emphasis on ophthalmic and dermal status. Current health
status was determined for use as a baseline in monitoring the development
of adverse effects. A sleep history was obtained to assess hypersomnia and
the stability of the sleep-wake cycle. Subjects entered the study at least 5
weeks before spontaneous remission was expected based on recall of
previous episodes. A written informed consent was obtained from each
subject. The experiment was performed with the approval of the University
Committee for Research Involving Human Subjects (UCRIHS; IRB#88-
349).

W The design of the experiment was a 2 x 2 (2
treatments in 2 stages) crossover design consisting of two stages of
phototherapy separated by a 1-week withdrawal period (Jones and
Kenward, 1989). Subjects were assigned by a balanced procedure to either
full spectrum or cool white phototherapy in stage 1 and then crossed over to
the alternate treatment in stage 2 of the crossover. Both treatment periods
were 7 days. The duration of the study for an individual subject was 3
weeks, but subjects did not all begin on the same day.

Stage 1 of the crossover consisted of pretreatment and treatment
measures associated with the ﬁrst treatment assignment. Pretreatment
measures of depression and melatonin were obtained on the evening before
day 1 of the treatment week. Treatment measures of depression and

melatonin were obtained on the evening of day 7 of the treatment week.

38
Stage 2 likewise consisted of pretreatment and treatment measures
obtained at the same time relative to the treatment week (Figure 1).

W The two treatment conditions were full
spectrum (FS; Vita-Lite F40T12; Duro-Test Corporation) and cool white
(CW; Sylvania F40T12CWSS) phototherapy. Phototherapy in each treatment
consisted of daily exposure (0600-0800 h) to a bank of eight ﬂuorescent tubes
at a distance of 90 cm (Rosenthal et al., 1984). An acrylic diffusing lens
(Lithonia) with an absorption cut-off at 325 nm was used to prevent
transmission of UVB radiation.

Subjects were seated in front of the light source, but at a right angle
to the light source so that the direction of their gaze was 90° to the plane of
light. Subjects were instructed to stay awake and glance at the light source
for a minimum of 1 second every minute. Subjects were monitored for
development of adverse effects and to ensure compliance with the
phototherapy protocol. Subject reports were collected daily on sleep timing
and duration, mood. and energy states, and health status.

Illuminance (Gossen Panlux photometer) for F8 light was 2690 lux
and 3013 lux for CW light. Irradiance (Kettering; 250-33000 nm range) was
675 uW/cmZ for F8 light and 600 uW/cm2 for cw light, yielding a calculated
dose of 4.86 J/cm2 and 4.32 J/cm2 for F8 and CW light, respectively. The
spectral power distribution of visible radiation (ISCO spectrophotometer) for
F8 and CW light is shown in Figures 2 and 3, respectively.

W The response to phototherapy was measured
' with the HRSD and the SAD Supplementary Items, scale (Rosenthal and
Heffeman, 1986; Appendix A). Depression ratings were obtained on four
occasions: pretreatment of stage 1, last day of treatment in stage 1,

pretreatment of stage 2, and last day of treatment in stage 2. Pretreatment

Crossover Design

Stage Stage 1 Stage 2

 

 

Period Pretreatment Treatment Pretreatment Treatment

Figure 1. Experimental protocol for two stage crossover design with two

treatments.

Figure 2. Spectral power distribution of visible radiation of full spectrum
(Vita-Lite F40T12) light.

i
lo

I

g
(wU/szISIl

 

 

I I
('0 N

9M”) HEMOd

748

496 544 592 640 688
WAVELENGTH (nm)

448

400

Figure 2

Figure 3. Spectral power distribution of visible radiation of cool white
(F40T12CWSS) light. ‘

12—

I
C)
,.

I I I
00 «D St

(wUIszISIieMn) aamoa

748

496 544 592 640 688
WAVELENGTH (n m)

448

400

Figure 3

44

of stage 2 corresponded to the last day of the 1-week withdrawal period.
Ratings were performed by a nurse trained in psychiatric evaluation and
blind to treatment assignments. Ratings were obtained between 1800 and
2000 h to eliminate the effect of diurnal mood variation on depression
scores. Subjects were required to have a pretreatment HRSD score of at
least 18 at stage 1 and at least 16 in stage 2. Subjects who did not have a
HRSD score of at least 16 at the time of the crossover were rated weekly and
were only allowed to crossover to the second stage when they met this
criterion.

Sahmﬁamplgmﬂsm Serial saliva samples were collected on
four occasions that coincided with the timing of depression ratings:
pretreatment of stage 1, last day of treatment in stage 1, pretreatment of
stage 2, and last day of treatment in stage 2. Saliva samples (2 ml) were
collected without stimulation of salivary ﬂow every 30 minutes from 1800 to
2400 h. Subjects were kept in dim light (<50 lux) during the procedure and
refrained from eating, drinking, and smoking. Water was allowed
immediately after each sample collection, but not closer than 15 minutes
prior to the next collection. Samples were stored at -20° C until the time of
assay.

Saliva samples met the classiﬁcation criteria for a biosafety level 2
hazard and were assayed in accordance with the Universal Blood and
Bodily Fluids regulations of the Center for Disease Control (USPHS, 1988).
Immediately before assay, samples were thawed, refrozen, and thawed for
24 h to reduce viscosity. Samples were then centrifuged at 2500 x g for 30
minutes to remove food particles. Saliva was assayed for melatonin by

radioimmunoassay (Fraser et al., 1983) with a double antibody separation

45
method (W ebley et al., 1986) using tritiated melatonin tracer (New England

Nuclear) and antiserum G/S/7 04-6483 (Guildhay Antisera).

Wm The effect of phototherapy on HRSD scores was
modeled with a split plot crossover design (Gill, 1978). Two-way analysis of
variance with treatment (FS v. CW) and stage (1 v. 2) as between subjects
factors and period (pretreatment v. treatment) as the repeated factor was
used. Comparisons of means within periods were tested with the
Bonferroni t-test. The effect of phototherapy on SAD scores was analyzed
similarly. Clinical remission was deﬁned as at least a 50% reduction in
HRSD score to a value less than eight, adapted from the recommendation
for the 21-item HRSD (Terman et al., 1989). The frequency of clinical
remission between treatments was analyzed with a X2 test. All data are
reported as meaniSD.

The effect of phototherapy on melatonin concentrations was modeled
with a double split plot crossover design (Gill, 1978). Analysis of variance
with treatment and stage as main plot effects, period as the sub-plot effect,
and sampling time (1800-2400 h) as the sub-sub plot effect was used.
Differences in melatonin concentrations between 2000 and 2200 h (the usual
time of onset of nocturnal secretion) were compared between periods with
the Bonferroni t-test. Hormone values were then averaged over each 6-h
sampling period to obtain a representative mean. Bonferroni t-test was used
to test for differences between means of treatments within periods. Subjects
were divided (post-hoe) into two groups on the basis of their clinical status
on day 7 of phototherapy: responders were those who met clincial
remission criterion and non-responders were those who did not meet

clinical remission criterion. Mean concentration of melatonin at

46
pretreatment and at treatment was compared between responders and non-

responders with Student's t-test.
Results

One hundred and three potential subjects were initally interviewed
for eligibility and 30 of these underwent medical assessment at the
Psychiatry Clinic. Twenty-seven (90%) of these subjects met the diagnostic
criteria for SAD. Twelve subjects (nine women and three men) met the
admission criteria for the study and agreed to participate. Six subjects
received FS light ﬁrst and six received CW light ﬁrst. The two light
treatments were indistinguishable to subjects. Eleven subjects completed
the crossover (one male in the FS-CW sequence group failed to relapse
during the withdrawal period). The mean age of the subjects was 36.61123
years (range: 20-59), with a mean age of SAD onset of 23.4:103 years. All
subjects reported hypersomnia, hyperphagia, and carbohydrate craving.

Analysis of variance of the effect of phototherapy on HRSD and SAD
scores and melatonin concentrations is in Appendix B. In order to balance
the design, data ﬁ'om one subject were randomly discarded from the CW-FS
sequence group to match the subject in the FS-CW sequence group who
failed to complete the crossover. HRSD scores were similar (p>.25) before
FS and CW phototherapy (23.5:49 v. 22.6146). HRSD scores were also
similar (p>.25) after FS and CW phototherapy (9.53:6.8 v. 8915.1). Thus,
treatment results were pooled within periods. HRSD scores were reduced
(p<.0001) on day 7 of treatment compared with pretreatment (Figure 4).
Stage of the crossover did not inﬂuence (p>.2) HRSD scores. SAD scores
were similar (p>.1) before FS and CW light (17.2129 v. 15.3439). SAD
scores were also similar (p>.25) after 1 week of FS and CW phototherapy
(5.03:3.5 v. 6115.1) Thus, treatment results were pooled within periods.

47

Figure 4. Pretreatment and treatment HRSD scores for F8 and CW light.
Treatment consisted of 1 week of daily phototherapy (0600-0800).
Pretreatment scores were obtained in the evening before treatment began;
treatment scores were obtained in the evening of day 7 of treatment. Both

treatments decreased (p<.05) HRSD scores.

m._._I>> 4000
_

S

 

 

wCOEwmmﬁ. s “CO—twﬂmtmhm I

c ouowﬁm

5:550me .33....
_

 

 

 

 

 

0
38008 CISHH

I!)
'-

0N

mm

49

’ SAD scores were reduced (p<.0001) after 1 week of phototherapy compared
with pretreatment. Stage of the crossover did not inﬂuence (p>.8) SAD
scores, but there was an interaction between the effect of stage and period
(p<.06) such that SAD scores decreased more in stage 1 than in stage 2.

Both treatments were similarly effective in inducing clinical
remission (50% for F8 vs. 40% for CW light; X2=.18, p=.67, df=1).
Individually, two subjects responded to both FS and CW light, three subjects
responded only to FS light, two subjects responded only to CW light, and
three subjects responded to neither light. Subjects (except one) were unable
to maintain the antidepressant effect of light during the withdrawal week
and relapsed into a depressive state. One subject maintained a partial
antidepressant effect for 3 weeks before relapsing and entering stage 2 of
the crossover. FS phototherapy triggered a hypomanic episode in one
patient and CW phototherapy was associated with transient agitation (n=1),
headache (n=2) and eye irritation (n=1).

Melatonin concentrations varied between subjects (p<.03) and
sampling times (p<.0001). High melatonin concentrations (range: 40 to 125
pg/ml) were observed at 1800 and 1830 h on each of the four sampling days.
Melatonin concentrations then plummeted in the next hour before
beginning a steady rise over the remainder of the evening. Mean melatonin
concentrations were similar (p>.2) before FS and CW phototherapy
(26.61175 v. 34.31267 pg/ml). Mean melatonin concentrations were
similar (p>.25) between the two light treatments on day 7 of treatment (FS:
37 5120.5; CW: 41.21240 pg/ml). Thus, treatment results were pooled
within periods. Stage of the crossover did not inﬂuence (p>.9) melatonin
concentrations. Mean concentrations of melatonin in saliva increased

(p<.03) between pretreatment and day 7 of treatment (Figure 5). There was

Figure 5. Pretreatment and treatment melatonin proﬁles. Treatment
consisted of 1 week of daily phototherapy (0600-0800 h) with either FS or CW
light. Concentrations of melatonin in saliva were measured every 30
minutes from 1800 to 2400 h under constant dim (<50 lux) illumination.
Pretreatment melatonin proﬁles were obtained on the evening before
treatment began; treatment proﬁles were obtained on day 7 of treatment.
Treatment results were pooled for F8 and CW light. Each data point
represents the mean of 20 observations. Mean melatonin concentrations

increased (p<.03) following phototherapy compared with pretreatment.

51

m chomam

Acv>zﬁ_unvmgsh

oovw comm . CONN 00 _.N ooom oom _. oow _.
o

— _ _ _ _ _ b _ _ _ r _

 

 

2953: .1 , ,

 

oow

wCOEwQObOEQ .I. VONF

(Iw/ﬁd) NINOlV‘IEiW

52

no increase in melatonin concentrations between 2000 and 2200 h in
pretreatment proﬁles. However, treatment proﬁles showed a rise (p<.05) in
melatonin concentrations between 2000 and 2200 h.

On the basis of clinical status on day 7 of phototherapy (remission: a
50% decrease in HRSD score to a value less than eight v. depressed: HRSD
score greater than or equal to eight), there was no difference (p>.25) in
mean melatonin concentrations at pretreatment between subjects who
eventually achieved clinical remission (9 observations) and subjects who
did not (11 observations). Additionally, there was no difference (p>.25) in

mean melatonin concentrations on day 7 between the two groups (Figure 6).
Discussion

The symptomatology, demographic characteristics, and response to
phototherapy in this sample of SAD patients is consistent with previous
reports (Wirz-Justice et al., 1986; Rosenthal et al., 1988). The decrease in
HRSD and SAD scores following 1 week of morning phototherapy conﬁrms
previous observations using similar protocols (Rosenthal et al., 1984).
Pretreatment HRSD scores were comparable before both stages of the
crossover, indicating that the 1 week withdrawal period was sufﬁcient to
prevent carryover effects of light on depression scores. This observation is
consistent with previous reports (Rosenthal et al., 1984; Wehr et al., 1986).
Other studies which failed to use a 1-week withdrawal period have reported
carryover effects of light (Salinas et al., 1992). The equal efﬁcacy of FS and
CW light to decrease HRSD and SAD scores indicates that a spectral
distribution similar to natural light is not critical for an antidepressant
response. Furthermore, the positive response to phototherapy in the
absence of UVB radiation indicates that the action spectrum for the

antidepressant response is not within this waveband (290-320 nm).

Figure 6. Melatonin proﬁles on day 7 of treatment in subjects who achieved
clinical remission (9 observations) and subjects who did not (depressed; 11
observations). Treaatment consisted of 1 week of daily phototherapy (0600-
0800 h) with either FS or CW light. Salivary samples were collected under
dim (<50 lux) illumination. Treatment results were pooled for F8 and CW
light. Mean melatonin concentrations were similar (p>.25) between the two

groups following 1 week of phototherapy.

oovm comm

— _ .

:o_mm_Emm +
83230 .1

CONN

3 >3 ”.0 m2:

ooPN

OOON

o snowem

com _.

002
o

 

 

(Iw/Bd) NiNOiV'lSW

55

Subjects found saliva sampling relatively easy to perform and
preferable to blood sampling. Melatonin concentrations in saliva showed
high interindividual and low intraindividual variability, a ﬁnding that is
consistent with reports based on plasma melatonin (Waldhauser et al.,
1985; Laasko et al., 1990). In the present experiment, melatonin
concentrations increased steadily between 2000 and 2400 h, reﬂecting
thenocturnal rise in pineal uptake of serotonin (Axelrod, 1974), pineal NAT
activity (Klein and Weller, 1970), and plasma melatonin concentrations
(Waldhauser et al., 1985). The data in the present experiment are
consistent with previous reports of circadian rhythmicity in salivary
melatonin (Miles et al., 1985a, 1985b, 1989; McIntyre et al., 1987). The light-
induced increase in salivary melatonin concentrations in this experiment
parallels that reported for plasma melatonin following a similar
phototherapy protocol (Hansen et al., 1987). Melatonin concentrations in
saliva are generally 30% of plasma concentrations (Miles et al., 1985a).
Approximately 70% of circulating melatonin is bound to plasma albumin
(Cardinali et al., 1972) and cannot diffuse into acinar cells of the salivary
glands. Thus, salivary melatonin concentrations may reflect the
biologically active free fraction of circulating melatonin (Miles et al., 1985b).

The ﬁnding of very high melatonin concentrations in saliva early in
evening is not readily explainable. Others have also noted it (Miles et al.,
1989; Nickelson et al., 1991). One possibility is that salivary ﬂow rate may
have varied throughout the 6-h proﬁle. Salivary ﬂow rate exhibits a
circadian rhythm with ﬂow rate higher early in the evening (1800 h)
compared with later in the evening (2400 h; Dawes, 197 2). However, Miles
et a1. (1989) reported that changes in salivary ﬂow rate do not affect the

concentration of melatonin in saliva. Another possibility is that entry into

56

' the dim light environment removed the suppressant effect of light on
melatonin. However, this possibility appears unlikely since melatonin
concentrations at this time were often twice as high as midnight
concentrations. Furthermore, this effect has not been observed in plasma
samples obtained under similar protocols (Lewy et al., 1987 ). More likely
the effect is related to the salivary assay. In the present experiment as well
as in the two previous reports (Miles et al., 1989; Nickelson et al., 1991),
high concentrations of melatonin in saliva were observed at sampling times
closely associated with food ingestion. Miles at al. (1989) suggests that the
presence of certain kinds of food in saliva may contaminate the
radioimmunoassay system. Thus, high concentrations of melatonin in
saliva early in the evening are most likely an artefact of the assay rather
than a measure of circulating melatonin.

The similar response in melatonin concentrations to F8 and CW
light indicates that a spectral distribution similar to sunlight is not
necessary to modulate melatonin secretion. This observation is consistent
with other photobiologic effects in which the full spectral range of light is
unnecessary to elicit a response (Anderson and Parrish, 1981). The
observation of similar melatonin concentrations before both stages of the
crossover indicates that there is no carryover eﬁ‘ect of light on melatonin
concentrations following the 1-week withdrawal period. Others have
reported carryover effects of light persisting beyond 1 week (Rosenthal et al.,
1990). This inconsistency may be due to differences in the duration of daily
phototherapy between experiments (2 vs. 6 h).

Sleep deprivation does not account for the antidepressant effect of
morning phototherapy. In a study of hospitalized SAD patients, Rosenthal
et al. (1985) reported that awakening patients at 0500 h and administering 3

57

h of dim (<5 or <300 lux) light was not effective in reducing depression
scores. I

Studies of light-induced changes in the circadian rhythm of
melatonin usually employ 24-h proﬁles. However, for practical reasons
(subject compliance and availability of subjects), studies of SAD patients
frequently employ shorter time periods such as overnight (Terman and
Schlager, 1990) or evening (Lewy et al., 1987; Sack et al., 1990) proﬁles.
Determination of onset time of nocturnal melatonin secretion within these
proﬁles is often based upon visual inspection of the raw data (Terman and
Schlager, 1990) or the time at which melatonin concentrations rise above a
set criterion, such as 15 or 20 pg/ml (Lewy et al., 1987; Sack et al., 1990). In
the present experiment, high variability between subjects and high
concentrations of melatonin early in the evening precluded the use of a set
criterion in determining onset time. Visual inspection was deemed
unreliable. Another approach is to determine whether a signiﬁcant
increase in melatonin concentration occurs within the usual timeframe
(2000-2200 h) of onset of nocturnal secretion (Lewy et al., 1987). In the
present experiment, melatonin concentrations remained low between 2000
and 2200 h before treatment, suggesting that the onset of nocturnal
secretion had not yet occurred. In contrast, on day 7 of treatment,
melatonin concentrations increased signiﬁcantly between 2000 and 2200 h,
suggesting that the onset of nocturnal secretion had occurred. These
observations are consistent with previous reports of advanced onset of
nocturnal melatonin secretion in SAD patients following 1 week of
phototherapy compared with pretreatment (Lewy et al., 1987; Terman et al.,
1988; Sack et al., 1990).

58

Others have used mean hormone concentrations within a speciﬁed
time period to monitor light-induced changes in the timing of circadian
rhythms (Hansen et al., 1987; Wirz-Justice et al., 1990). For example, an
increase, relative to pretreatment, in the mean concentration of melatonin
between 2200 and 2400 h following 7 days of phototherapy may be indicative
. of a phase advance of the rhythm such that more melatonin is secreted
earlier in the evening (Hansen et al., 1987 ). In the present experiment,
comparison of mean melatonin concentrations before and after
phototherapy revealed an increase in evening melatonin secretion following
phototherapy. This ﬁnding may be due to an advance in the timing of the
melatonin rhythm or to a general increase in 24-h melatonin secretion.
Skwerer et al. (1988) compared 24-h proﬁles of melatonin secretion in SAD
patients before and after 1 week of phototherapy and found no evidence of an
overall increase in 24-h melatonin secretion. The result of the Skwerer
study suggests that an increase in evening melatonin concentrations
following phototherapy as occurred in the present experiment may be
speciﬁc to that timeframe and may reﬂect an advance in the timing of the
circadian rhythm. Assessment of the morning decline (offset) of nocturnal
melatonin secretion will strengthen this interpretation if morning
melatonin concentrations are decreased following phototherapy.

In the present experiment, the increase in evening melatonin
concentrations and possible advance in the melatonin rhythm following 1
week of phototherapy was associated with clinical remission. Others have
reported the same association of clinical remission and a circadian
advance in the melatonin rhythm (Lewy et al., 1987; Terman et al., 1988;
Sack et al., 1990). However, these studies have utilized a morning versus

evening crossover design which leads to circadian phase shifts in opposing

59

directions with each week of phototherapy (advance with morning light
and delay with evening light). Since a therapeutic effect occurs only with
morning light, the circadian phase advance is confounded with the
therapeutic result. Nonetheless, several investigators have proposed that
the antidepressant effect of light is mediated by the phase advance in the
melatonin rhythm (Lewy et al., 1987 ; Sack et al., 1990). In the present
study, only morning light was used and a possible circadian advance was
observed in the onset of nocturnal melatonin secretion. However, when
subjects were divided on the basis of clinical status on day 7 of phototherapy
(subjects who achieved clinical remission versus subjects who remained
depressed), it was observed that both groups showed a similar increase in
evening melatonin concentrations. This observation suggests that there is
a dissociation between an advance in nocturnal melatonin secretion and
clinical remission. Others have noted the same dissociation in individual
case studies (Terman et al., 1990; Allen et al., 1992). While both clinical
remission and an advance in the onset of nocturnal melatonin secretion are
the result of cumulative exposure to morning bright light, the two eﬁ'ects
are not directly related. However, it cannot be ruled out that a circadian
phase advance in the melatonin rhythm may be necessary for an
antidepressant effect, although it is not sufﬁcient in itself to cause clinical
remission.

In conclusion, measurement of melatonin in saliva offers a non-
invasive technique for assessing the effect of phototherapy in SAD patients.
FS light is not essential to decrease depression scores or modulate the
circadian rhythm of melatonin. Both FS and CW light contain the
necessary waveband of light with a sufﬁcient intensity to induce clinical

remission and increase nocturnal melatonin concentrations. The results

so
of this study indicate that there is a dissociation between the effect of light

on depression scores and evening melatonin concentrations following 1

week of morning phototherapy in SAD.

Chapter 2. Effect of Full Spectrum versus Narrow Spectrum Green
Light on Melatonin Concentrations and Depression

Scores in SAD Patients

62

Introduction

Daily phototherapy for 2 h with full spectrum light decreases
depression scores (Rosenthal et al., 1984) and advances the timing of the
nocturnal onset of melatonin secretion (Lewy et al., 1987 ). The eﬁ‘ect of
narrow spectrum light on depression scores and melatonin secretion is not
fully known. In other areas of photomedicine, narrow spectrum light is
often as effective in producing the desired result as the full spectrum of
visible light. For example, neonatal hyperbilirubinemia is effectively treated
with narrow spectrum blue or green light (Ennever, 1990).

Narrow spectrum blue and narrow spectrum red light are not as
effective as full spectrum light in reducing depression scores in SAD
patients (Brainard et al., 1990). Whether the loss of efﬁcacy with blue and
red light is due to the absence of white light or the lack of a critical
waveband of light is unknown. In addition, it is not known whether white
light is required to modulate the circadian rhythm of melatonin. The
timing of the circadian rhythm of melatonin can be assessed from the time
of onset of the nocturnal rise, the time of peak concentration, or the time of
the morning decline (offset) of nocturnal secretion. In Chapter 1, it was
shown that increasing evening melatonin concentrations (consistent with
advancing the timing of nocturnal secretion) was insufﬁcient to induce an
antidepressant response. However, it is not known whether decreasing
melatonin concentrations in the morning (also consistent with advancing
the timing of nocturnal secretion) may be important in inducing clinical
remission in SAD patients. Therefore, an experiment was conducted to
determine the effect of white light versus green light on morning melatonin
concentratiOns and depression scores in SAD patients. The purpose of the

experiment was threefold: to determine 1) whether green light can

63
decrease depression scores, 2) whether green light can modulate the
circadian rhythm of melatonin, and 3) whether clinical remission is

associated with a decrease in melatonin concentrations in the morning.

Materials and Methods

Subjgcts. Subjects were recruited through an advertisement in the
Lansing State Journal newspaper as well as through the methods
described in Chapter 1. The admission criteria for the study were identical
to that described in Chapter 1, with the exception that the criterion for the
HRSD score was decreased to 14. Exclusion criteria were expanded to
include previous photosensitizing reactions and cataract surgery based
upon the advice of a dermatologist and optometrist. Lactation was also
added to the exclusion criteria due to the possibility of stimulating vitamin
D synthesis (Adams et al., 1982). The medical assessment was identical to
that described in chapter 1. Subjects did not maintain sleep logs during
this experiment. The experiment was performed with the approval of
UCRIHS (IRB#89-506).

W The experimental design was that described
in Chapter 1 with the following modiﬁcations. The treatment period lasted
6 days. Day 7 of the treatment week was reserved for morning saliva
sampling. In order to differentiate the circadian effects of light from the
direct suppressing eﬁ‘ect on melatonin secretion, phototherapy was not
administered on the morning of saliva collection (Lewy et al., 1980).
Pretreatment and treatment values were obtained for both stage 1 and stage
2 of the crossover. Pretreatment scores for depression were obtained 2 days
before the start of treatment. Pretreatment melatonin proﬁles were

obtained 1 day before the start of treatment. Treatment depression scores

64

were obtained on day 6 of the treatment week and treatment melatonin
proﬁles were obtained on the following morning (day 7).

WW. Phototherapy was administered daily at the
Psychiatry Clinic between 0600-0800 h. The two treatments were full
spectrum light (FS; Vita-Lite F40T12, Duro-Test Corp.) and narrow
spectrum green light (Sylvania F40T12G) with a half-peak bandwidth
between 505 and 555 nm. Illuminance for F8 phototherapy was 2690 lux
and 4500 lux for green phototherapy. Irradiance was 675 uW/cm2 for F8
light and 600 uW/cm2 for green light. These measured irradiances yielded
a calculated dose of 4.86 and 4.32 Joules/cm2 for F8 and green light,
respectively. The spectral power distribution for F8 and green light is
shown in Figures 2 and 7, respectively.

WW5. Depression ratings (HRSD and SAD) were
obtained between 0900-1200 h on four occasions: pretreatment and
treatment in stage 1 of the crossover and pretreatment and treatment in
stage 2 of the crossover. Additionally, a SIGH-SAD (Structured Interview
Guide for the Hamilton - Seasonal Affective Disorder) score, which is a
composite of the HRSD score and SAD score, was calculated on each of the
four occasions (Williams et al., 1988). Subjects were required to have a
HRSD score of at least 13 to crossover to stage 2.

Wm Serial saliva samples were collected on
four occasions: pretreatment and treatment in stage 1 of the crossover and
pretreatment and treatment in stage 2 of the crossover. Samples were
collected every 30 minutes ﬁ-om 0600-0900 h. Subjects were allowed to eat
and drink brieﬂy between sample collections, but not within 15 minutes of

the next sample collection. Sample collection was performed at each

Figure 7. Spectral power distribution of green light (Sylvania F40T12G).

25—

I
C)
C“

I ' I
8 .9
(WU/ZwOISiieMn) aaMOd

J
U)

748

496 544 592 640 688
WAVELENGTH (nm)

448

400

Figure 7

67
subject's home under the natural and artiﬁcial lighting (<300 lux) to which
they were normally exposed (below the threshold for melatonin
suppression; Lewy et al., 1980). Samples were stored in home freezers
(approximately -10 to -20° C until returned to the clinic and then stored at
-20°C until time of assay. Saliva samples were assayed for melatonin as
described in the previous chapter.

WW5. As described in Chapter 1, a split plot crossover
design was used to model the effect of phototherapy on depression scores.
In brief, treatment (FS v. green) and stage (1 v. 2) were the main plot
(between subjects) effects and period (pretreatment v. treatment) was the
repeated factor of the sub-plot. Separate analyses of variance were
performed on HRSD, SAD, and SIGH-SAD scores. Bonferroni t-test was
used to determine whether pretreatment scores were comparable before
light treatments (or stages) and whether treatment scores were comparable
after light treatments (or stages). Change in depression score
(pretreatment score minus treatment score) within treatments (or stages)
was calculated and tested with Bonferroni t-test. The criterion for clinical
remission was a 50% decrease in HRSD score to a value less than eight.
The effect of phototherapy on melatonin concentration was modeled as a
double split plot as described in Chapter 1 with sampling time (0600-0900 h)
as a sub-sub-plot effect. Hormone values were averaged over each 3-h
sampling period to obtain a representative mean. Bonferroni t-test was
used to test for differences between means of treatments (or stages) within

periods. All data are reported as mean1SD.
Resuns

One hundred fourteen potential subjects were assessed to determine

study eligibility over the course of two winter seasons. Forty-ﬁve

68

underwent medical assessment at the Psychiatry Clinic and forty-one (91%)
met the diagnostic criteria for SAD. Twenty-two subjects (17 women, 5
men) met the study criteria and agreed to participate. Eighteen subjects
completed the crossover. One female and one male subject failed to relapse
during the withdrawal period from F8 light, one female subject developed
hypomania with FS light, and one female subject attempted suicide during
the withdrawal period from green light. Data from these subjects were not '
included in the analyses. The mean age of study subjects was 43818.5
(rangez28-60) with an average age of onset of SAD of 30.71102 (range:10-51).
Subjects had not previously participated in a phototherapy study.

Analysis of variance for the effect of phototherapy on HRSD, SAD,
and SIGH-SAD scores and melatonin concentrations is in Appendix C.
There was no difference in HRSD (p>.1), SAD (p>.1), or SIGH-SAD (p>.05)
scores between green and FS light. Thus, results were pooled for green and
FS light. Phototherapy reduced HRSD scores (19.3132 v. 10.3168; p<.0001),
SAD scores (12.7149 v. 7.115.5; p<.0001), and SIGH-SAD scores (32.0156 v.
17 .3111.2; p<.0001). However, there was an interaction (p<.05) between the
effect of stage (1 v. 2) and period (pretreatment v. treatment) on depression
scores on all three rating scales (Figure 8). Depression scores were
comparable at pretreatment between the stages, but treatment scores were
lower (p<.005) in stage 1 of the crossover than in stage 2. The change in
depression scores (pretreatment score - treatment score) was greater in
stage 1 of the crossover than in stage 2 for all depression rating scales
(HRSD change score: 13.8 v. 4.3, p<.005; SAD change score: 8.8 v. 2.5, p<.01;
SIGH-SAD change score: 22.6 v. 6.8, p<.005).

Based on clinical remission criterion, one subject responded to both

green and FS light, seven responded only to F8 light, ﬁve responded only- to

Figure 8. SIGH-SAD, HRSD, and SAD scores at pretreatment in stage 1
(Stl-Pretrt), treatment in stage 1 (Stl-Pretrt), pretreatment in stage 2 (St2-
Pretrt), and in stage 2 (St2-Trt). Treatment consisted of 6 days of daily
phototherapy (0600-0800 h) with either green or F8 light.

w shaman

wwl.<0m 20_wmm¢n_m_n_
O<m DmmI O<m-_.._0_w

 

 

 

ooooo
.....

.....

----------

cccccccccc
oooooooooo
.....

 

.....

.....

CD
,_

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

G

3..

”In” [G

i a

..... \ a. .s:

\ s

\\ m

on S

o.

O

H

mm a
Edam I

Eccdém S on

E- :m E
Edi- :m I

 

71
green light, and ﬁve did not respond to either light treatment. The
frequency of clinical remission did not differ between light treatments
(X2=.28, p>.9). However, the frequency of clinical remission was higher
(X2=7.14; p<.01) in stage 1 (12 out of 18 subjects) than in stage 2 (2 out of 18
subjects). All subjects relapsed during the withdrawal week and
immediately entered stage 2 of the crossover. ‘

A substantial proportion of subjects reported difﬁculty with
phototherapy during stage 2. Six out of nine subjects who received green
light in stage 2 and seven out of nine subjects who received FS light in stage
2 reported a strong dislike for the light and discomfort (nausea, headache)
during the 2-h sessions. Three of these subjects experienced a worsening of
depressive symptoms (insomnia, loss of appetite, psychomotor retardation)
during stage 2. One subject experienced discomfort in both stages of the
crossover. Following 1 week of phototherapy, HRSD scores for subjects (15
observations) who reported a dislike for and discomfort with phototherapy
averaged 16214.8 compared with 8814.3 for subjects (8 observations) who
reported no preference for one light over the other and 4.4132 for subjects
(13 observations) who reported a positive experience with phototherapy.

The mean concentration of melatonin in saliva was similar between
FS and green light at pretreatment (36617.9 v. 39.6166 pg/ml; p>.1) and
treatment (29.9168 v. 28.115 pg/ml; p>.1). Thus, results were pooled for
green and FS light. Melatonin concentrations were reduced between
pretreatment and treatment (38.2185 v. 29.0162 pg/ml; p<.001; Figure 9).
Pretreatment concentrations of melatonin were 35.7158 and 40.7110.4
pg/ml (p>.1) in stage 1 and 2, respectively. Concentrations of melatonin
following 6 days of phototherapy were 31.516 and 26515.8 pg/ml (p>.1) in

stage 1 and 2 of the crossover, respectively. Melatonin concentrations

Figure 9. Salivary melatonin concentrations between 0600 and 0900 h at
pretreatment and treatment. Treatment consisted of 6 days of daily
phototherapy (0600-0800 h) with either green or F8 light. Treatment results
were pooled for green and FS light. Samples were collected under indoor
lighting (<300 lux). Mean concentrations of melatonin were reduced

(p<.001) following 1 week of phototherapy compared with pretreatment.

oomo

 

wCOEuMOﬁ. .1.
“£02.30ngth ll:

oomo

Ev ><o do us:

oono
_ _

m ouowam

coco
_ _

 

 

8 $2
(Iw/Bd) NINOiV'BW

C)
0)

C)
V.

cm

74
decreased (p<.005) more in stage 2 of the crossover than in stage 1. The
concentration of melatonin was similar in subjects who achieved clinical
remission (14 observations) and subjects who did not (22 observations) at

pretreatment (p>.2) and following 6 days of phototherapy (p>.05; Figure 10).

Discussion

The monochromatic green light used in this study was intensely
colored light. Thus, subjects were not blind to differences in the
treatments. However, subjects did not know which treatment was
considered the active treatment. Upon initial exposure, some subjects
perceived green light as pleasant, while others perceived it as unpleasant.
However, a person’s ﬁrst impression was transient and did not affect
compliance with the phototherapy protocol or treatment results.

The ﬁnding that there was no difference between FS and green light
in decreasing depression scores and inducing clinical remission would
seem to indicate that the action spectrum for the antidepressant effect of
light lies within the green waveband. This observation is inconsistent with
a previous study (Stewart et al., 1991) which found green light not as
effective as FS light in reducing HRSD scores. The decreased efﬁcacy of
green light in the study of Stewart et al. (1991) may have been due to the
lower intensity (2367 lux) compared with the intensity (4500 lux) used in the
present study. The results of the present experiment are consistent with a
previous study (Oren et al., 1991a) which found green light capable of
inducing clinical remission in SAD patients and more effective than red
light in reducing HRSD scores. The intensity of green light (4680 lux) in
the study of Oren et al. (1991a) was similar to that used in the present study.
Therefore, green light may decrease depression scores and induce clinical

remission when administered at a sufﬁcient dose.

75

Figure 10. Melatonin proﬁles on the morning following 6 days of
phototherapy (0600-0800 h) in subjects who achieved clinical remission (14
observations) and subjects who did not (depressed; 22 observations).
Salivary samples were collected under indoor (<300 lux) lighting.
Treatment results were pooled for F8 and green light. Mean melatonin

concentrations were similar (p>.05) between the two groups.

76

come some

. ummmoﬁoo +.
co_mm_Emm t.

3 :5 do us:
echo

_ _

CH ouowwm

oomo

 

a e
(Iw/ﬁd) NINOiV'lEiW

CD
00

CD
V

cm

77

However, the pronounced stage effect in the present study
complicates the conclusion that the action spectrum for the antidepressant
effect of light lies within the green waveband. Green light was as effective
as FS light in inducing clinical remission in stage 1 of the crossover (5 out
of 9 subjects with green light and 7 out of 9 with FS light). In contrast,
green light and FS light were similarly ineffective in inducing clinical
remission in stage 2 (1 out of 9 subjects both green and FS light). The stage
effect on depression scores is not readily explainable, although it has been
noted previously (Rafferty et al., 1990). Others have reported that
expectations for treatment often predict treatment response (Evans, 1985;
Wehr et al., 1987 ). For example, positive expectations may affect a subject’s
response by producing an emotional state which heightens the
physiological effect (Uhlenhuth et al., 1959; Eastmen, 1990). Various
physiological mechanisms, such as secretion of endorphins and steroids,
have been proposed to explain the effect of positive expectations on
treatment results (Ross and Olson, 1981; Grunbaum et al., 1986). Terman
et al. (1990) suggest that positive expectations for treatment may be higher
in the ﬁrst stage than in subsequent stages of a crossover study and thus
may give rise to a stage effect. The ﬁnding of the present experiment is
consistent with this explanation. However, in the experiment described in
Chapter 1, treatment results were similar in both stages of the crossover,
indicating that initial expectations for treatment do not always cause a
stage effect.

I speculate that expectations for treatment inﬂuence treatment
response only if subjects can distinguish treatments and one treatment is
positively or negatively experienced during the course of the trial. In the
preceding study, two indistinguishable white light sources (FS and CW)

78

were used, subjects reported similar experiences during the course of each
treatment, and no stage effect was observed. In contrast, in the present
experiment, subjects were exposed to physically different types of light.
Since many subjects (12 out of 18) responded positively to light treatment in
stage 1 (whether the treatment was FS or green light), treatment with a
different type of light in stage 2 may have been perceived as withdrawal
from the active treatment. While expectations for treatment were not
measured prior to beginning the study, subjectS’ experience during
phototherapy was recorded. A positive experience during phototherapy
enhanced the effect of light. Depression scores were 50% lower in subjects
who reported a positive experience during phototherapy compared with
subjects who reported a neutral experience.

On the other hand, a negative experience during treatment may
inhibit a positive response to treatment. Physiological mechanisms may
also mediate the effect of a negative psychological experience (Brier, 1989).
For example, a negative experience of phototherapy in stage 2 may have
induced a stress response in some subjects which ultimately inﬂuenced
both depression scores and melatonin concentrations. HRSD scores were '
twice as high in subjects who reported a negative experience compared
with subjects who reported a neutral experience. Stress is associated with
increased melatonin secretion in some animal species (Khan et al., 1990).
Human melatonin secretion appears to be unaffected by physical stress
(Vaughan et al., 1979), yet may be inﬂuenced by psychological stress
associated with increased plasma concentrations of cortisol. Many reports
indicate that increased cortisol secretion is associated with decreased
melatonin concentrations (Claustrat et al., 1984; Brismar et al., 1985). In

the present experiment, melatonin concentrations decreased the most .in

79
subjects who disliked treatment. Moreover, some of the adverse effects
reported by subjects during stage 2, such as insomnia and loss of appetite,
are commonly seen in other types of major depression in association with
hypercortisolemia (Gold et al., 1988). However, since cortisol was not
measured in the present experiment, it is impossible to draw a ﬁrm
conclusion regarding the relationship of cortisol and melatonin.

If a person's physiological and psychological experience during
treatment explains the stage effect, two possibilities exist. One possibility is
that since green light parallels the effect of FS light in both stages, the
antidepressant effect of light lies within the green waveband. However, the
effect of green light (or white light containing the green waveband) can be
modulated by subjective experience during treatment. The second
possibility is that subjective experience during treatment determines the
antidepressant effect, regardless of treatment. Analysis of treatment
results from subjects who reported a neutral experience with both
treatments may provide insight into the true effect of green versus FS light
on depression scores. In these subjects, phototherapy decreased depression
scores equally with both types of light, lending further credence to the
hypothesis that the action spectrum for the antidepressant effect lies within
the green waveband.

Unexpectedly high melatonin concentrations were observed only
intermittently in saliva samples as opposed to the consistent pattern
observed in Chapter 1. This difference is most likely due to a greater
variability in food intake in subjects in the present study as compared to the
previous study. In the present study, some subjects did not eat before or
during the 3-h sample collection while in the previous study, all subjects ate
prior to the 6-h sample collection.

so

In the present experiment, phototherapy was not administered
during sample collection so the acute suppressant effect of bright light on
melatonin secretion was avoided. Subjects collected samples under similar
lighting conditions at pretreatment and treatment. Thus, changes in
morning melatonin concentrations before and after phototherapy reﬂect the
cumulative effect of 6 days of early morning light exposure. Others have
reported changes in the melatonin rhythm induced by cumulative exposure
to morning light (Lewy and Sack, 1989; Terman et al., 1988; Czeisler and
Shanahan, 1991). The decrease observed in melatonin concentrations
between pretreatment and treatment is consistent with a phase advance in
the morning decline in nocturnal melatonin secretion. In Chapter 1 it was
shown that melatonin concentrations increased between 1800-2400 h
following 1 week of morning phototherapy compared with pretreatment.
Together, these results suggest that a phase advance in nocturnal
melatonin secretion occurs with morning phototherapy such that more
melatonin is secreted in the evening (1800-2400 h) and less in the morning
(0600-0900 h).

However, in the present experiment, concentrations of melatonin
were similar following phototherapy in subjects who responded positively to
phototherapy and those who did not. This ﬁnding is consistent with the
results of Chapter 1 in which evening melatonin concentrations were
similar in subjects who achieved clinical remission and those who did not.
Taken together, these results indicate that changes in the timing of the
melatonin rhythm as measured by evening onset and morning decline of
nocturnal secretion are not associated with clinical outcome. Moreover,
concentrations of melatonin were similar at pretreatment between subjects

who eventually achieved clinical remission and those who did not. This

81
ﬁnding is also consistent with the results of Chapter 1. Therefore, mean
melatonin concentrations measured in the morning or evening are not
good predictors of clinical response.

It may be important that mean melatonin concentrations in the
morning were reduced by morning phototherapy, regardless of any effect
on the timing of the melatonin rhythm. High concentrations of melatonin
are associated with drowsiness and decreased performance (Arendt et al.,
1984; Wright et al., 1986). Rosenthal et al. (1984) proposed that high
circulating melatonin may be implicated in the hypersomnia and daytime
drowsiness observed in SAD patients. It follows that a decrease in
circulating melatonin may lead to remission of these symptoms. However,
concentrations of melatonin were similar following treatment in subjects
who responded to phototherapy and those who did not, indicating that
merely decreasing melatonin concentration in the morning is not sufﬁcient
in itself to induce clinical remission.

The ﬁnding of the present study would seem to indicate that green
light is able to modulate secretion of melatonin by reducing morning
concentrations, at least in the second stage of the crossover. Others have
reported (Brainard et al., 1985;-Horne et al., 1991) that green light can
acutely suppress melatonin secretion. Together, these ﬁndings suggest
that the light-induced suppression of melatonin and the photic regulation of
the circadian rhythm may be initiated by the same class of photoreceptor in
the retina, namely, that which has a spectral sensitivity to the green
waveband of light (Nathans et al., 1986). To date, no other studies have been
performed on the action spectrum of the photic regulation of the circadian
rhythm of melatonin.

82

In summary, morning concentrations of melatonin and depression
scores are decreased following 1 week of phototherapy compared with
pretreatment. The decrease in morning melatonin concentrations
following phototherapy is consistent with a phase advance in the timing of
the melatonin rhythm. However, decreases in melatonin concentrations
were observed in subjects who achieved clinical remission and in subjects
who remained depressed. It appears likely that advancing the timing of the
decline of nocturnal melatonin secretion is not sufﬁcient to induce clinical
remission in SAD patients. Green light has an efﬁcacy comparable to that
of FS light in reducing depression scores and modulating melatonin
secretion. The action spectrum for the antidepresant effect of light may lie
within the green waveband. However, treatment response for both FS and
green light can be strongly inﬂuenced by subjective experience during

treatment.

Chapter 3. Effect of Eye versus Head Exposure to Light on
Melatonin Concentrations and Depression Scores in
SAD Patients

84
Introduction

The antidepressant effect of light is believed to be dependent upon eye
exposure (Rosenthal et al., 1984). Retinal absorption of light is the ﬁrst step
in the photic regulation of circadian rhythms which are believed to be
involved in the antidepressant effect of light (Lewy et al., 1980; Czeisler et
al., 1989). Skin exposure to light in the absence of retinal exposure is not as
effective as eye exposure in reducing depressive symptoms in SAD patients
(Wehr et al., 1987 ). However, it is not known whether skin exposure may
supplement the antidepressant effect of eye exposure. For example,
depressed patients, including SAD patients, frequently exhibit decreased
immune function (Calabrese et al., 1987; Kasper et al., 1991). UVA
radiation is known to affect immune function through dermal absorption
(Morison, 1985). SAD patients treated with phototherapy exhibit an
increase in immune function (Kasper et al., 1991).

Expectations for treatment often predict outcome (Evans, 1985; Wehr
et al., 1987 ). Additionally, physiological and psychological experience
during treatment may inﬂuence treatment response as shown in Chapter
2. Whether expectations for treatment measured prior to undergoing
phototherapy are directly related to subjective experience during treatment
response is unknown.

Therefore, an experiment was conducted to determine the effect of
eye versus full head exposure to light on depression scores and melatonin
concentrations in SAD patients. A second objective of the experiment was
to determine whether expectations for treatment predict subjective

experience during treatment and treatment response.

85
Materials and Methods

Sgbjggts, Subject recruitment and admission into the study were as
described in Chapter 2. Based upon consultation with an oncologist,
melanoma was added to the exclusion criteria. Patients were assessed for
color blindness by the Ishihara Test, but were not excluded if they tested
positive. At the initial assessment, subjects had the treatment conditions
explained and treatment apparatus shown to them. Subjects were asked to
evaluate their expectations for each treatment condition by means of a scale
(-4 to +4) on which -4 indicated an expectation of very much worsening, 0
indicated an expectation of no improvement, and +4 indicated an
expectation of very much improvement. The experiment was performed
with the approval of UCRIHS (IRB#90-498).

W The experimental design was a 2 x 2 crossover
design as described in Chapter 1. In brief, pretreatment measures of
melatonin were obtained 1 day before the start of treatment and
pretreatment measures of depression 2 days before. Treatment measures
of depression were obtained on day 6 of the treatment week and treatment
measures of melatonin on the following day.

Ehgmhgmpy. Phototherapy was administered in each subject's
home for 6 days between 0600 and 0800 h. The light source for phototherapy
consisted of four full spectrum (FS; Vita-Lite F40; Duro-Test) ﬂuorescent
tubes. Subjects sat 60 cm away from the light source which resulted in an
illuminance of 2500 lux at eye level. The two treatments were head
phototherapy and eye phototherapy. Head phototherapy consisted of light
exposure to the eyes and the skin of the face, head, neck, and hands. In eye
phototherapy, only the eyes were exposed to light. Subjects were cotton

gloves, long-sleeve shirts, and a black double-knit face mask which

86
completely covered the head and neck with the exception of eyes, nose, and

mouth. Illuminance on the skin under the face mask was 25 lux.

W. Depression ratings (HRSD, SAD, SIGH-SAD)
were obtained according to the protocol described in Chapters 1 and 2. In
brief, depression ratings were obtained on four occassions: pretreatment
and treatment in stage 1 of the crossover and pretreatment and treatment
in stage 2 of the crossover.

Wm. Subjects followed the protocol and
schedule for collection of morning (0600-0900 h) saliva samples as described
in Chapter 2. Samples were obtained at pretreatment and treatment in
stage 1 of the crossover and pretreatment and treatment in stage 2.
Samples were stored at -20° C until the time of assay for melatonin as
described in Chapter 1.

W. As described in Chapter 1, a split plot crossover
design was used to model the effect of phototherapy on HRSD, SAD, and
SIGH-SAD scores with treatment (eye v. head) and stage (1 v. 2) as main
plot (between subjects) effects and period (pretreatment v. treatment) as the
repeated factor in the sub-plot. The effect of phototherapy on melatonin
concentration was modeled similarly with sampling time (0600-0900) added
as the sub-sub plot. Comparisons of means within treatments (or stages)
was tested with the Bonferroni t-test. Clinical remission criteria was a 50%

decrease in HRSD score to a value less than eight.
Resuns

Potential subjects (n=105) were interviewed to determine study
eligibility. Of these, 39 underwent medical assessment at the Psychiatry
Clinic and 36 (92%) met the diagnostic criteria for SAD. Thirteen subjects

entered the study and 12 (7 women, 5 men) completed the crossover (one

87
female subject developed severe insomnia with eye phototherapy). The

mean (18D) age of subjects was 40217.3 (range: 24-49). The mean age of
onset of SAD in subjects was 27219.2 (range: 15-43). One male subject was
red-green color blind. One subject had previously participated in a study of
phototherapy. Subjects reported no adverse effects during phototherapy
with either treatment, other than one subject who reported ocular pain and
pressure during eye phototherapy.

Analysis of variance for the effect of phototherapy on HRSD, SAD,
and SIGH-SAD scores melatonin concentrations is in Appendix D. There
was no difference in the effects of eye versus head phototherapy on HRSD
scores (p=.5), SAD scores (p=.3), or SIGH-SAD scores (p=.4). There was no
interaction between treatment and period. Thus, treatment results were
pooled for eye and head exposure. Depression scores did not differ between
stages 1 and 2 of the crossover. Phototherapy reduced HRSD, SAD, and
SIGH-SAD scores (p<.0001; Figure 11).

Subjects' expectations for treatment did not predict response. Six
subjects expected no improvement from either treatment, two subjects
expected both treatments to yield improvement, and four subjects expected
eye+skin phototherapy to be superior to eye phototherapy. Only two out of
the twelve subjects correctly predicted their response to phototherapy based
upon the treatment evaluation of the nurse. Expectations of subjects did not
enhance treatment effect. Subjects (10 observations) who had positive
expectations for treatment averaged 21.4 on the HRSD scale before
treatment and 9.1 following treatment. Subjects (14 observations) who had
no expectations for treatment (0 on the expectation scale) averaged 22.0

before phototherapy and 10.1 following phototherapy on the HRSD scale.

Figure 11. HRSD, SAD, and SIGH-SAD scores at pretreatment and
treatment. Treatment consisted of 6 days of daily phototherapy (0600-0800 h)
with either eye or head exposure to light. Pretreatment scores were
obtained 2 days before the start of treatment; treatment scores were obtained
on day 6 of treatment. Treatment results were pooled for eye and head
phototherapy. Phototherapy reduced (p<.0001) HRSD, SAD, and SIGH-SAD

scores compared with pretreatment.

I Pretreatment
3 Treatment

 

|_
C)
v

I
CD
03

or
38008 NOISSSHdBCI

I
C)
1-

    
 

 

 

 

 

 

HRSD SAD

DEPRESSION SCALES

SlGH-SAD

Figure 11

90
Subjects reported similar experiences during both stages of the

crossover and with both light treatments. N 0 subjects reported intense
dislike for or discomfort with either light condition.The frequency of clinical
remission was similar (X2=.4, p>.9) between head phototherapy (6
observations) and eye phototherapy (4 observations). Individually, three
subjects responded to both light treatments, three responded only to head
phototherapy, one responded to eye phototherapy, and ﬁve did not‘respond to
either light treatment. All subjects relapsed during the withdrawal week
and immediately crossed over to stage 2.

Pretreatment melatonin concentrations in saliva averaged 30.71202
and 2681161 pg/ml in eye- and head-exposed subjects, respectively (p>.25).
Melatonin concentrations following treatment averaged 22.2111.1 and
2281118 pg/ml in saliva in eye- and head-exposed subjects, respectively
(p>.25). Concentrations of melatonin in saliva tended (p=.06) to decrease
between pretreatment and treatment (Figure 12). Stage of the crossover did
not affect (p=.8) melatonin concentration. Melatonin concentrations were
lower (p<.0005) at pretreatment in subjects (10 observations) who
subsequently failed to achieve clinical remission than in subjects (14
observations) who subsequently achieved clinical remission. However,
following the week of phototherapy, concentrations of melatonin were
similar in both groups due to a decrease (p<.0005) in subjects who achieved

clinical remission (Figure 13).
Discussion

In agreement with previous studies (Rosenthal et al., 1984; Wehr et
al., 1987), depression scores were reduced in SAD patients following 1 week
of eye exposure to light. Exposure of facial skin to light did not augment the
antidepressant response observed with eye exposure alone. This ﬁnding is

91

Figure 12. Salivary melatonin concentrations (0600-0900 h) at
pretreatment and treatment. Treatment consisted of 6 days of phototherapy
(0600-0800 h) with either eye or head exposure to light. Pretreatment
proﬁles were obtained 1 day before the start of treatment; treatment proﬁles
were obtained on the morning following the last treatment day. Treatment
results (eye v. head phototherapy) were pooled. Sample collection was
performed under indoor (<300 lux) illumination levels. Phototherapy
reduced (p=.06) the mean concentration of melatonin in saliva compared

with pretreatment.

NH enemas

EV ><o do us:

come oomo oono oooo
_ IL _ _ E _ _

 

“£02.56th .1.

wCOEummbuOEQ .I.

 

0..

O
N
(Iw/ﬁd) NINOiV‘lEW

om

mos

Figure 13. Salivary melatonin concentrations (0600-0900 h) at treatment in
subjects who achieved clinical remission (10 observations) and subjects who
did not achieve clinical remission (depressed; 14 observations). Treatment
consisted of 6 days of phototherapy (0600-0800 h) with either eye or head
exposure to light. Treatment proﬁles were obtained on the day following the
last treatment day. Treatment results were pooled for eye and head
exposure. Sample collection was performed under indoor (<300 lux)
illumination levels. Mean concentrations of melatonin were similar (p>.05)
in subjects who responded positively to phototherapy compared with

subjects who did not achieve clinical remission.

mH unawam

3 :3 do m2:
come coco case came

P _ — r — _ _

 

co_mm_Eom 1..
252550 +

59
58’

lm"

til
ngii

Hill".

or

CD
cu
(Iw/ﬁd) NiNOlV'iBlN

cm

as

9‘5
consistent with thereport of Wehr et al. (1987) that eye phototherapy (eyes

open to light, no skin exposure) is more effective than skin phototherapy
(eyes covered by goggles, maximal skin exposure) in reducing depression
scores in SAD patients. Since partial skin exposure in this experiment did
not increase treatment response, it appears likely that the antidepressant
effect is dependent upon eye exposure only. The antidepressant effect of eye
exposure to light is evident even during sleep (Avery et al., 1990) with closed
eyelids (Moseley et al., 1988).

In previous studies of SAD, subject’s expectations for treatment have
been measured before subjects began phototherapy, but actual experience
during treatment has not been assessed. In the present study, both
expectations for treatment and subjective experience during phototherapy
were recorded during each treatment week. Expectations for treatment did
not predict treatment outcome. Subjects for the most part expected little
therapeutic beneﬁt from phototherapy. Only two subjects correctly predicted
their response to treatment. This ﬁnding is in contrast to the report of
Wehr et al. (1987 ) at the National Institute for Mental Health in which
subjects expected positive results from treatment and treatment results
satisﬁed those expectations. The difference between the present study and
the Wehr study may be due to expectations for treatment being higher at a
prestigious research institution (Eastman, 1990). Another possibility is that
apparent differences between treatments within each study inﬂuenced
expectations for treatment. In the Wehr study, treatments varied
considerably (subjects were goggles and swim suits in the skin condition
versus total skin coverage in the eye condition) and subjects expected eye
phototherapy to be superior. In the present study, treatments were
relatively similar in that both allowed eye exposure to white light. Since

%
both treatment conditions were similar with regard to eye exposure,

subjects may have assumed the treatments would be equally efﬁcacious.

In contrast to the ﬁndings with green and FS light in Chapter 2,
subjective experience during phototherapy in this experiment was neutral
in both stages of the crossover and with both light conditions. Subjects
reported neither a dislike for nor discomfort with either treatment. There
was no stage effect observed in either depression scores or melatonin
concentrations. HRSD scores following phototherapy averaged just above
the cut-off for complete clinical remission. This observation conﬁrms the
results of Chapter 1 in which subjects reported a neutral experience during
phototherapy and mean depression scores were near the cut-off point for
complete clinical remission. This ﬁnding is also consistent with the results
from the sub-group of subjects in Chapter 2 who reported a neutral
experience with phototherapy. '

Melatonin concentrations were similar following eye exposure and
full head exposure to light. Full head exposure to light allowed for dermal
absorption of light (Anderson and Parrish, 1981) and possibly cranial
transmission of light (Lucey and Hewitt, 1976; Wan et al., 1981; Svaasand
and Ellingsen, 1983). However, any supplemental absorption of light in the
full head exposure condition did not affect melatonin secretion. This is not
unexpected since the retinal-hypothalamic-pineal- pathway is the only
known anatomical route for photic regulation of melatonin secretion.

Melatonin concentrations (0600-0900 h) were decreased following 1
week of phototherapy compared with pretreatment. This ﬁnding conﬁrms
the results reported in Chapter 2 for the cumulative effect of 6 days of
morning phototherapy on salivary melatonin concentrations. The

decrease in melatonin concentrations between pretreatment and the

97
morning following day 6 of treatment may reﬂect a phase advance in the

morning decline (offset) of nocturnal melatonin secretion (Terman and
Schlager, 1990). .

Melatonin concentrations were similar following treatment in
subjects who achieved clinical remission and in those who did not. This
ﬁnding conﬁrms the results of the previous study. However, in contrast to
the results of Chapters 1 and 2, melatonin concentrations at pretreatment
varied between subjects who subsequently achieved clinical remission and
those who did not. Melatonin concentrations were already low before
treatment in subjects who failed to respond positively to phototherapy and
no further decrease was noted following treatment. This result is not
readily explanable since the two groups were similar in age, sex, and
depressive symptomatology.

In summary, skin exposure to light during phototherapy does not
increase the magnitude of the antidepressant effect of light. Expectations of
subjects for treatment effects did not predict treatment outcome. On the
whole, depression scores and melatonin concentrations decreased
following 1 week of morning phototherapy compared with pretreatment. In
the subgroup of patients who did not achieve clinical remission, the
concentration of melatonin at pretreatment was low and remained stable
following phototherapy in contrast to the light-induced decrease in

melatonin observed in subjects who did achieve clinical remission.

Chapter 4. Effect of White Light of Different Spectral Composition
on Melatonin Concentrations and Depression Scores In SAD

Patients

99
introduction

Depression scores in SAD patients are decreased by daily (0600-0800
h) exposure to full spectrum white light (Rosenthal et al., 1984;Winton et
al., 1989). Concentrations of melatonin in plasma are increased in the
evening (1800-2300 h) and exhibit an earlier nocturnal rise relative to
pretreatment following 1 week of morning phototherapy (Lewy et al., 1987;
Sack et al., 1990). It was shown in Chapters 2 and 3 that concentrations of
melatonin in saliva in the morning are decreased following 1 week of
morning phototherapy. It is unknown whether these light-induced
changes in evening and morning melatonin concentrations are
synchronized to each other. Furthermore, it is unknown whether
synchronized shifts in the melatonin rhythms are functionally related to
the antidepressant effect. If the antidepressant effect and circadian phase
shifting effect of light are responses to the same waveband of light, it would
suggest they may be related. There is no information available on the
action spectra for advancing the timing of the melatonin rhythm, although
green light is more effective than red or blue light in suppressing
melatonin secretion (Brainard et al., 1985). Red and blue light are not as
effective as full spectrum light in reducing depression scores in SAD
patients (Brainard et al., 1990).

Difﬁculties arise in using narrow spectrum light to determine the
action spectrum for the antidepressant effect of light. Narrow spectrum
light is strongly colored light; thus, subjects are not blind to treatment
differences. Humans often have a psychological response to speciﬁc colors.
Since patients' expectations for treatment response often predict outcome
(W ehr et al., 1987), narrow spectrum light that elicits favorable expectations

may strongly inﬂuence treatment response. Conversely, light that is

100
strongly disliked by subjects may lead to poor compliance with phototherapy

protocol (e.g., decreased eye exposure) and, consequently, may negatively
inﬂuence treatment response (Stewart et al., 1991). Narrow spectrum light
can also produce undesirable effects such as headache, vertigo, and nausea
(Sisson and Kendall, 1973) which may further inﬂuence treatment
response.

However, it is possible to blend complementary pairs of narrow
spectrum light to produce white light. For example, blending blue and
yellow light produces white light without the green and red wavebands.
Thus, subjects can be exposed to narrow bands of light, but without
perceiving color differences between treatments. This methodology can be
used to investigate the action spectrum of the antidepressant effect of light
without having expectations for treatment responsible for a large
component of treatment response. An experiment was performed to
measure the cumulative effect of daily exposure to blended white light on

depression scores and melatonin secretion in SAD patients.

Materials and Methods

Subjm Subjects were recruited through an advertisement in the
Lansing State Journal newspaper and through referrals from local
clinicians. The admission criteria for the study remained the same as
described in Chapters 13 Potential subjects underwent medical
assessment at the Psychiatry Clinic as described in Chapter 3. The
experiment was performed with the approval of UCRIHS (IRB#90-498).

W The design of the study was a 3 x 2 (3
treatments within 2 stages) crossover study. Treatment assignment was by

a 2 X 2 Latin square in which subjects represented the two rows of the

101
square and stages represented the two columns. Three squares were

performed in order that each treatment occurred twice in each stage and
each subject received two treatments (Table 1).

Pretreatment and treatment measures of depression and melatonin
were obtained in stages 1 and 2 of the crossover as described in Chapter 2.
In brief, pretreatment depression scores were obtained 2 days before the

start of treatment; pretreatment melatonin proﬁles were obtained the day
1 before the start of treatment. Treatment depression scores were obtained on
the last of phototherapy (day 6 in the treatment week) and treatment
melatonin proﬁles were obtained on the following day on which no
phototherapy was administered (day 7).

Ehmmhempy. Phototherapy was administered from 0600 to 0800 h at
the Psychiatry Clinic for 6 days. The three treatments were full spectrum
(FS) light, blended white light composed of blue and yellow wavebands (BY),
and blended white light composed of red and green wavebands (RG). The
FS treatment, described in Chapter 3, produced an illumination of 2500 lux.
The goal of the two blended white light sources was to create a convincing
match to white full spectrum (2500 lux) light with subjects blind to
wavelength, and thus color, composition. BY blended white light was
produced with narrow spectrum blue ﬂuorescent lamps (Sylvania F40T 12B)
and narrow spectrum yellow ﬂuorescent lamps (Sylvania F40T12GO). The
half-peak bandwidth (bandwidth with at least 50% of peak emission) of
narrow spectrum blue light was between 430 and 470 nm. The half-peak
bandwidth of narrow spectrum yellow light was between 550 and 650 nm.
RG blended white light was produced with narrow spectrum red
ﬂuorescent lamps (Sylvania F40T12B) and narrow spectrum green

ﬂuorescent lamps (Sylvania F40T12G). The half-peak bandwidth for red

102

Table 1. 2 X 2 Latin squares for a two-stage crossover design for three
treatments. Treatments consisted of full spectrum light (FS), blue-yellow
blended white light (BY), and red-green blended white light (RG). Six
subjects (subj) completed two stages of phototherapy. Each treatment was
administered four times within the experiment.

Staseofﬁmssoxer
1 L 1 2 1 L

 

 

Subj Subj SUbj
1: FS BY 3: FS RG 5: BY RG
2: BY FS 4: RG FS 6: RG BY

103
light was between 620 and 680 nm and for green light, between 510 and 555

nm. Cumulative spectral power distributions for BY and RG blended light
are given in Figures 14 and 15, respectively. For the BY treatment, three
blue lamps and three yellow lamps were used which yielded an
illuminance of 2500 lux. For the RG treatment, ten red and four green
lamps were used which yielded an illuminance of 2800 lux.

A light box was built to produce the blended white light treatments.
The light box was 48" tall and trapezoid in shape (15" wide in the front, 24"
wide in the back, and 30" wide on the sides. A 24" x 48" white matte board
(reﬂectance >.9) served as the back wall of the light box. The diagonal sides
of the trapezoid were made from a series of ﬂuorescent light ﬁxtures. The
metal frames of the light ﬁxtures formed the outside walls of the box and
the bare ﬂuorescent lamps formed the inner walls. The front wall of the
box consisted of a 15" x 48" clear acrylic diffuser. The light box rested on the
ﬂoor and was covered on the top with a piece of white matte board. Light
from each side of the box fell on the white matte board. The two
complementary colors of light blended at the surface of the matte board to
create white light. The white light was reﬂected off the matte board and
transmitted through the clear diffuser. The subject sat outside the light box
at a distance of 30 cm in front of the diffuser.

W Depression ratings were obtained by a
psychiatric nurse who was blind to treatment assignments. Depression
ratings (HRSD, SAD, and SIGH-SAD) were obtained on four occasions:
pretreatment and treatment in stage 1 of the crossover and pretreatment
and treatment in stage 2 of the crossover. The criterion for clinical

remission was a 50% decrease in HRSD score to a value less than eight.

104

Figure 14. Spectral power distribution for BY blended white light composed
of narrow spectrum blue and narrow spectrum yellow light. Radiant
power is measured in wavebands that correspond to colors of light:
Ultraviolet (<380 nm), violet (380-430 nm), blue (430-490 nm), green (490-560
nm), yellow (560-590 nm), orange (590-630 nm), red (630-800 nm).

105

«a unawam

E5 Isozmhm><>>
oowéoo 806% 8903 08-8». onions 93.08 owav

     

 

 

 

 

 

 

 

 

 

 

a

 

 

 

2.6 I
;o__d>§

 

F

(O N
(sue/VI) aaMOd 1NVICIVH

106

Figure 15. Spectral power distribution for RG blended white light composed
of narrow spectrum red and narrow spectrum green light. Radiant power
is measured in wavebands that correspond to colors of light (see legend for

Figure 14).

107

ms enemas

AEE Eczmd><>>
08.08 80.03 condom condos 02709 09.68 onov

S

 

 

I...‘.~I

 

 

 

:35 I
Com 3

impel
ll pl
gendfi

 

ID V’ (O N 1-
(sum) aaMOd .LNVICIVH

(O

108
W. Serial saliva samples were collected on

four occasions: pretreatment and treatment in stage 1 of the crossover and
pretreatment and treatment in stage 2 of the crossover. On each occasion,
samples were collected every 30 minutes for a 3-h interval in the morning
(0600-0900) and a 3-h interval in the evening (2000-2300). The protocol for
sample collection and storage was as described in Chapters 2 and 3.
Samples were assayed for melatonin as previously described in Chapter 1.
W As described in previous chapters, a split plot
crossover design was used to model the effect of phototherapy on HRSD,
SAD and SIGH-SAD scores and a double split plot crossover for melatonin.
In brief, hormone concentrations were averaged within sample periods to
get a representative mean. A pretreatment to treatment change score was
calculated for depression scores and mean melatonin concentrations
within each treatment and tested with the Student's t-test (Gill, 1978).
Separate analyses of variance were performed on morning (0600-0900 h)

and evening (2000-2300 h) melatonin concentrations.
Results

Potential candidates (n=35) were interviewed to determine eligibility
and interest in the study. Of these, 22 underwent medical assessment at
the Psychiatry Clinic and 21 (95%) met the criteria for SAD. Six subjects
(all female) entered and completed the study (four subjects per treatment).
The mean age of subjects was 4116 (rangez22-58). The mean age of onset of
SAD was 29.71143 (range: 1049).

Subjects could not determine the spectral (color) composition of BY
light. Some subjects initially perceived RG light as having a greenish tint.
However, after a few minutes of exposure, these subjects reported that the

light appeared white. Subjects who underwent RG phototherapy reported

109
that white objects appeared pink immediately following the completion of

phototherapy (duration < 1 h). One subject experienced headache, vertigo,
and nausea with BY and RG light. Two subjects experienced an increase
in depressive symptoms during phototherapy (one with RG light and one
with FS light). One subject experienced prolonged (2 days) alteration in
color vision after BY light in which red objects appeared intensely bright.

Analysis of variance of HRSD, SAD, and SIGH-SAD scores and
melatonin concentrations is in Appendix E. There was no difference (p>.6)
in HRSD scores between light treatments. Thus, treatment results were
pooled for FS, BY, and RG light. There was a trend (p<.1) for HRSD scores
to decrease between pretreatment and treatment (22813.6 v. 11818.5,
respectively; Figure 16). Stage of the crossover did not affect HRSD scores
(p>.8). There was no difference (p>.8) in SAD scores between light
treatments. Pooled for treatments, there was a trend (p<.09) for SAD scores
to decrease between pretreatment and treatment (14.4142 v. 6616.2,
respectively). Stage of the crossover did not affect SAD score. There was no
difference (p>.7) in SIGH-SAD scores between light treatments. Pooled for
treatment, there was a trend (p<.09) for SIGH-SAD scores to decrease
between pretreatment and treatment (36.816 v. 18.41138, respectively).
Stage of the crossover did not affect (p>.9) SIGH-SAD score.

In terms of clinical signiﬁcance, one out of four subjects achieved
clinical remission (50% decrease in HRSD score to a value less than eight)
with FS light, one out of four with RG light, and two out of four with BY
light. The frequency of clinical remission was similar (X2=.51; p>.9)
between treatments. All subjects who achieved clinical remission in stage

1 relapsed during the withdrawal week. Three subjects reported a neutral

110

Figure 16. HRSD, SAD, and SIGH-SAD scores at pretreatment (Pre) and on
day 6 of treatment (Trt) with full spectrum (FS), blended blue-yellow white
light (BY), and blended red-green white light (RG). Pooled for treatment,
phototherapy tended to decrease HRSD (p<.1), SAD (p<.09), and SIGH-SAD

(p<.09) scores relative to pretreatment.

111

Vooooeeeeeo.

O

 

mm._<0m szwwmn—mo

 

cs spawns

D<w-10_m

 

 

BHOOS NOISSEHdEO

 

#:0536055. @
wCOEwmmzmhn— I

 

‘112
experience during both phototherapy trials. One subject reported intense

dislike and discomfort during both trials of phototherapy (BY and RG). One
subject reported intense dislike for FS light and a positive experience with
RG light. Another subject reported intense dislike for RG light and a
positive experience with BY light. Subjective experience during
phototherapy predicted treatment response in terms of HRSD score.
Subjects (2 observations) who reported a positive experience during
phototherapy averaged the lowest HRSD scores following phototherapy
(3.5121). Subjects (4 observations) who reported a negative experience
averaged the highest post-treatment HRSD scores (21.517). Subjects (6
observations) who reported a neutral experience scored intermediate on the
HRSD scale (8.213).

Melatonin concentrations declined steadily (p=.0001) from 0600 to 0900
h (32.81208. v. 14.51167 pg/ml). Melatonin concentrations were similar
(p>.8) between all three treatments. Pooled for treatment, melatonin
concentrations were not different (p>.3) between pretreatment and
treatment (23.4199 v. 16918.7 pg/ml, respectively; Figure 17). Stage of the
crossover did not affect (p>.7) melatonin concentration at pretreatment or
treatment. However, within the RG light condition, melatonin
concentrations decreased between pretreatment and treatment (29.21149 v.
16519 pg/ml; p<.025; Table 2).

Evening (2000-2300 h) melatonin concentrations were similar (p>.3)
between treatments. Melatonin concentrations (pooled for treatments) were
similar (p>.8) between pretreatment and treatment (20.61128 v. 19.91122
pg/ml, respectively; Figure 18). Stage of the crossover did not inﬂuence

(p>.9) melatonin concentration at pretreatment or treatment. However,

113

Figure 17. Morning (0600-0900 h) melatonin concentrations in saliva at
pretreatment and treatment. Treatment consisted of 6 days of daily
phototherapy (0600-0800 h) with either FS, BY, or RG light. Treatment
samples were collected on the morning following day 6 of phototherapy.
Samples were collected every 30 minutes under indoor (<300 lux) lighting.
Mean concentrations of melatonin were similar (p>.3) at pretreatment and
following 6 days of treatment. Results were pooled for FS, BY, and RG
light.

0000

114

29:89... .1

“COEHMOSOEQ .I.

come

EV >5 .6 wsE

oono
_

 

NH ouowﬁm

oooo
_

 

or

00

CV

(Iw/Bd) NINOlV‘lSW

115

' Table 2. Mean (+SD) salivary melatonin concentrations in the morning
(0600-0900 h) and evening (2000-2300 h) at pretreatment and the morning
following day 6 of treatment (treatment) between full spectrum light (FS),
blended blue-yellow white light (BY), and blended red-green white light

 

 

 

 

(RG).
Treatment
FS BY Bi}

Morning Pretrt 21.8152 19. 119.4 2921149a
Melatonin

(pg/ml) Trt 17.51123 1681104 165193
Evening Pretrt 16.6152 13.519 318113.61?
Melatonin

(pg/ml) Trt 18.5183 16.91158 24.51148b

aChange in pretreatment to treatment; p<.025

bChange in pretreatment to treatment; p<.025

116

Figure 18. Evening (2000-2300 h) melatonin concentrations in saliva at
pretreatment and treatment. Treatment consisted of 6 days of daily
phototherapy (0600-0800 h) with either FS, BY, or RG light. Treatment
samples were collected the evening following the last day of phototherapy.
Samples were collected every 30 minutes under indoor (<300 lux) lighting.
Treatment results were pooled for FS, BY, and RG light. Mean
concentrations of melatonin were similar (p>.8) at pretreatment and

treatment.

117

oomw

CONN

ALV>RRHunvmgsh

ooPN
_ _

 

 

ma shaman

OOON
_

 

29:39... .1

wCOEuMOh—Ohm .l.

 

C)
v-

8 .‘2
(Iw/ﬁd) NINOIV'Iaw

mm

on

118
within the RG light condition, melatonin concentrations decreased (p<.025)

between pretreatment and treatment. There was no effect of food
consumption on melatonin concentrations.

7 Morning melatonin concentrations were similar between subjects
who achieved clinical remission (four observations) and subjects who did
not (8 observations) at pretreatment (p>.25) and at treatment (p>.2; Figure
19). Evening melatonin concentrations were similar between subjects who
achieved clinical remission and subjects who did not at pretreatment

(p>.25) and at treatment (p>.25; Figure 20).

Discussion

The results of this experiment indicate that complementary pairs of
narrow spectrum light can be blended convincingly to appear as white
light. This ﬁnding raises the possibility of a placebo control for
phototherapy. Upon elucidation of the action spectrum for the
antidepressant effect of light, blended white light that lacks the action
spectrum can be used as a placebo control. Blended white light can be
produced to match the illuminance (2500 lux) of bright light phototherapy,
adding to its credibility. However, while subjects cannot perceive the
spectral composition of blended white light, alterations in color vision occur
which can inﬂuence the experience of phototherapy.

The trend for scores on all three depression scales to decrease
between pretreatment and day 6 of treatment is consistent with the results
of the previous three chapters. Treatment results were similar with all

three light conditions.

119

Figure 19. Morning (0600-0900 h) melatonin concentrations in saliva
following 6 days of daily phototherapy (0600-0800 h) in subjects who achieved
clinical remission (4 observations) and subjects who did not achieve clinical
remission (depressed; 8 observations). Treatment results were pooled for
FS, BY, and RG light. Mean concentrations of melatonin were similar

(p>.2) at pretreatment and treatment.

as madman

30 :5 no msE

 

0000 0000 00h0 0000
_I _ - _ _ _ _
. """
<1 9,1,,

[III I‘ll]
1” ’I.’l
IIII \\\\‘IIIIIII./
4- A x
[I //
Ill/I

,o,,,

4",, le
,4

:o_wm_Emm.9

commence .1

 

O O O
(O N 1-

(IUJ/5d) NINOlV'Iaw

C
v

on

121

Figure 20. Evening (2000-2300 h) melatonin concentrations in saliva
following 6 days of daily phototherapy (0600-0800 h) in subjects who achieved
clinical remission (4 observations) and subjects who did not achieve clinical
remission (remission; 8 observations). Treatment results were pooled for
FS, BY, and RG light. Mean conceentrations of melatonin were similar

(p>.25) at pretreatment and treatment.

5 >5 no 92:.
comm ocmm ooam

— _ _ _ _
869.com .9

ummmoﬁoo .1

om ouomﬁm

Doom

 

2 $2
(Iw/ﬁd) NINOiV'IEiW

O
N

mm

123
The small number of subjects in the study was a result of the

ﬁ'equency and severity of adverse effects experienced by subjects. Short-
term alterations (<1 h) in color vision were reported by all four subjects who
underwent RG phototherapy. This was not unexpected since green light in
RG phototherapy was of sufﬁcient intensity to bleach the photopigment in
the retinal cone cells that selectively absorb green light. Upon later
exposure to full spectrum light, these cone cells could not be further
stimulated by the green waveband and, consequently, the complementary
color within white light (red) is perceived. One subject experienced a
prolonged (2 days) alteration in color vision following BY light in which red
objects appeared extremely bright. This was most likely caused by a
supersensitivity in red-sensitive cone cells resulting from a lack of
stimulation during BY phototherapy. Side effects such as these need to be
addressed before further studies with blended white light are performed.

Subjects who reported a diser for and discomfort with phototherapy
actually worsened in depressive symptomatology. This observation
conﬁrms the ﬁnding in Chapter 2 in which a negative experience during
phototherapy predicted a poor response to light. Differences in spectral
sensitivity between individuals may account for the discomfort experienced
with blended white light (Curcio, et al., 1989).

The ﬁnding that morning (0600-0900 h) melatonin concentrations did
not decrease between pretreatment and treatment is at odds with results
reported in Chapters 2 and 3. However, in the two previous experiments,
changes in melatonin concentrations between pretreatment and treatment
did not vary with different types of light or light exposure. In the present
experiment, the absence of a decrease in melatonin concentration between

pretreatment and treatment is the result of combining treatments that have

124
different effects. Within the RG treatment, morning melatonin

concentrations decreased 43% between pretreatment and treatment, an
observation consistent with the effect of green light in stage 1 in Chapter 2.
This ﬁnding suggests that the action spectrum for decreasing morning
melatonin concentrations and perhaps phase advancing the circadian
rhythm of melatonin is within the green waveband of light.

Evening melatonin concentrations did not change between
pretreatment and treatment. This ﬁnding is at odds with the 30% increase
in melatonin concentration following 1 week of phototherapy observed in
Chapter 1. This discrepancy is most likely due to differences in the length of
the melatonin proﬁle (3 h in the present experiment and 6 h in Chapter 1).
The evening proﬁle of melatonin in the present experiment excluded
sampling times 2330 and 2400 h which exhibited the highest melatonin
concentrations in Chapter 1. There was no effect of food consumption on
melatonin concentrations as observed in Chapter 1. This was most likely
due to a longer time interval (2 h) between the evening meal and the
beginning of saliva sampling in the current experiment.

Within the RG light condition, evening melatonin concentrations
actually decreased. This ﬁnding is inconsistent with a phase advance in
the onset of nocturnal melatonin secretion. This ﬁnding suggests that
alterations in evening onset and morning decline of nocturnal melatonin
secretion may be mediated by different wavelengths of light.

Between individuals, morning and evening melatonin proﬁles
showed large variation in response to light. On three occasions melatonin
concentrations increased in both the morning and evening proﬁles
following phototherapy. On three other occasions melatonin concentrations

decreased in both the morning and evening proﬁles following phototherapy.

125
On the remaining six occasions, melatonin concentrations showed

inconsistent changes. These results indicate that the evening onset and
morning offset of nocturnal melatonin secretion are not necessarily coupled
in terms of photic responsiveness.

The observation that melatonin concentrations were similar between
subjects who responded to phototherapy and those who did not is consistent
with the ﬁndings reported in Chapters 1-3. Melatonin concentrations did
not differentiate between those who achieved clinical remission frOm those
who did not at pretreatment. This ﬁnding indicates that pretreatment
melatonin concentrations in saliva at the time of evening onset or morning
offset of nocturnal melatonin secretion are not predictors of response to
phototherapy. Moreover, melatonin concentrations at the time of evening
onset or morning offset of nocturnal melatonin secretion following 1 week of
I phototherapy were not correlated with clinical improvement.

In summary, blended white light composed of complementary pairs
of narrow spectrum light can be used as a convincing match for
phototherapy. All three types of white light (FS, blended BY, and blended
RG) tended to decrease depression scores following 6 days of phototherapy.
Blended white light composed of narrow spectrum green and narrow
spectrum red light decreased melatonin concentrations at the time of
evening onset and morning offset of nocturnal melatonin secretion.
However, clinical remission was not correlated with changes in evening or
morning melatonin concentrations, indicating that the time of onset and
offset of nocturnal melatonin secretion may not be associated with clinical

improvement.

126

Summary and Conclusions

Phototherapy is a fast-acting, effective treatment for SAD. In the 49
subjects who underwent two week-long trials of phototherapy in this series
of experiments, complete clinical remission was achieved 40% of the time.
However, relapse occurred within 1 week of withdrawal from treatment.
Minor side effects, such as headache and eye irritation, occurred in 7% of
trials; severe side effects occurred in 4% of trials, but only with narrow
spectrum phototherapy.

Expectations of subjects for treatment effects measured prior to
phototherapy do not predict treatment outcome. However, subjective
experience during phototherapy can inﬂuence results. For example,
subjects who reported a negative experience during phototherapy in terms
of physiological and psychological discomfort showed little, if any,
improvement in depressive symptoms following 1 week of treatment.
However, this effect was limited primarily to the second stage of crossover
trials. Conversely, subjects who reported a positive experience during
phototherapy received a greater antidepressant effect than subjects who
reported a neutral experience. This effect was also limited by stage of the
crossover occurring primarily in stage one.

Phototherapy with either FS or CW (2500 lux) light is effective in
reducing depressive symptoms in SAD patients. The antidepressant effect
appears to be dependent solely upon retinal exposure to light. Skin
exposure to light does not supplement the antidepressant effect.
Phototherapy with narrow spectrum green light is equally effective as FS

light in reducing depressive symptoms as long as subjects have a neutral

127
experience during the trial. Intense discomfort during phototherapy with

the green waveband inhibits the antidepressant effect.

Blended white light composed of complementary pairs of narrow
spectrum light can be used as a convincing match for full spectrum
phototherapy. Blended white light has an advantage over monochromatic
light in blinding subjects to color composition and thus avoiding subjective
color preferences. However, while the human brain adapts well to
exposure to blended light, subjects may experience both short- (< 1 h) and
long-term (>2 d) alterations in color vision. These alterations in spectral
sensitivity may be associated with a negative experience during
phototherapy.

Serial saliva samples can be used to measure melatonin and to
monitor light-induced changes in its circadian rhythm. Saliva sampling
offers a safe, non-invasive alternative to blood sampling for collecting data
on human circadian rhythmicity.

One week of phototherapy in the early morning (0600-0800 h)
decreases the concentration of melatonin in morning samples (0600-0900 h).
Moreover, phototherapy in the early morning increases the concentration of
melatonin in evening (1800-2400 h) saliva samples. The light-induced
changes in morning and evening melatonin concentrations may reflect a
phase advance of the melatonin rhythm. Morning and evening melatonin
concentrations revert back to the initial pretreatment concentrations
following a 1-week withdrawal from phototherapy. The effects of light on
melatonin secretion are dependent upon the retinal-hypothalamic-pineal
pathway and are not modulated by supplemental skin exposure. Green
light, whether administered monochromatically or in conjunction with red
light to produce blended white light, appears to have equal potency as ,FS

128
light in reducing morning melatonin concentrations. However, a subject’s

psychological and physiological experience of phototherapy may inﬂuence
melatonin secretion. For example, a negative experience during
phototherapy is associated with a larger decrease in morning melatonin
concentrations compared with a neutral experience.

Clinical remission and decreased concentrations of melatonin in the
morning occur following 1 week of phototherapy. Moreover, clinical
remission and increased concentrations of melatonin in the evening occur
following 1 week of phototherapy. However, these light-induced changes in
melatonin secretion are not directly related to the antidepressant response.
For example, melatonin concentrations are similar following phototherapy
in subjects who achieve clinical remission and those who are still
depressed. Similarly, evening melatonin concentrations following
phototherapy are similar in subjects who achieve clinical remission and
subjects who do not. It appears that an advance in the timing of nocturnal
melatonin secretion is not sufﬁcient to induce clinical remission in SAD
patients. However, light-induced alterations in other aspects of the
melatonin circadian rhythm, such as peak concentration, amplitude, and
total 24-h secretion, cannot be ruled out in playing a role in clinical
remission. .

It is concluded that early morning exposure to bright light decreases
depressive symptoms in SAD patients. Furthermore, daily exposure to
bright light in the morning increases evening melatonin concentrations
and decreases morning melatonin concentrations in SAD patients.
However, decreases in depressive symptoms do not appear to be mediated
through light-induced changes in melatonin secretion. Retinal exposure to

light alone accounts for the antidepressant effect and the changes in

129
' melatonin secretion. Subjects’ expectations for treatment do not inﬂuence

treatment results, but subjective experience during phototherapy, both
positive and negative, affects clinical outcomes. The mechanism through
which phototherapy decreases depression scores appears not to be
dependent upon modulation of evening onset or morning offset of nocturnal
melatonin secretion. Light within the green waveband appears to be at
least partially responsible for the decreasing depression scores and
modulating the melatonin rhythm. However, the antidepressant effect of
light may ultimately depend upon the physiological effects of light and the
psychological experience of undergoing treatment.

APPENDICES

APPENDIX A

Depression Rating Scales

Hamilton Rating Scale for Depression

1. Depressed Mood

0 Absent

1 These feeling states indicated only on questioning

2 These feeling states spontaneously reported verbally

3 Communicates feeling states non-verbally -- i.e., through
facial expression, posture, voice, and tendency to weep

4 Patient reports virtually only these feeling states in
his spontaneous verbal and non-verbal communication

2. Feelings of Guilt
0 Absent
1 Self-reproach, feels she or he had let people down
3 Present illness is a punishment. Delusions of guilt
4 Hears accusatory or denunciatory voices and/or
experiences threatening visual hallucinations

3. Suicide
0 Absent
1 Feels life is not worth living
2 Wishes she or he were dead or any thoughts of possible death
to self
3 Suicide ideas or gesture
4 Attempts at suicide

4. Insomnia (Early)
0 No difﬁculty falling asleep
1 Complains of occasional difﬁculty falling asleep - i.e., more than
1/2 hour
2 Obvious and severe; more than 1/2 hour usually

5. Insomnia (Middle)
0 No difﬁculty
1 Complains of being restless and disturbed during the night

2 Waking during the night - any getting out of bed rates 2
(except for purposes of voiding)

130

131

6. Insomnia (Late)
0 No difﬁculty
1 Waking in early hours of the morning but goes back to sleep

2 Unable to fall asleep again if she or he gets out of bed

7. Work Activities

0 N o difﬁculty

1 Thoughts and feelings of incapacity, fatigue or weakness
related to activities, work or hobbies

2 Loss of interest in activities, hobbies, or work --
either directly reported by patient, or indicated in listlessness,
indecision and vacillation (feels she or he has to push self to work or
activities) , ’

3 Decrease in actual time spent in activities or decrease in
productivity .

4 Stopped working because of present illness

8. Retardation
0 Normal speech and thought
1 Slight retardation at interview
2 Obvious retardation at interview
3 Interview diﬁicult
4 Complete stupor

9. Agitation
0 None
1 Fidgetiness
2 Playing with hands, hair, etc.
3 Moving about, can't sit still
4 Hand wringing, nail biting, hair pulling, biting of lips

10. Anxiety (Psychic)
0 N o difﬁculty
1 Subjective tension and irritability
2 Worry about minor matters
3 Apprehensive attitude apparent in face or speech
4 Fears expressed without questioning

11. Anxiety (Somatic)

Physiological concomitants of anxiety, such as:
Gastrointestinal - dry mouth, wind, indigestion, diarrhea, cramps
Cardiovascular - palpitations, headaches
Respiratory - hyperventilation, sighing
Sweating

0 Absent

1 Mild

2 Moderate

3 Severe

4 Incapacitating

132

12. Somatic Symptoms (Gastrointestinal)
0 None
1 Loss of appetite but eating without encouragement. Heavy feeling in
abdomen
2 Difﬁculty eating without urging. Requests or requires laxatives or
medication for bowels or medication for GI symptoms

13. Somatic Symptoms (General)
0 None
1 Heaviness in limbs, back or head. Backaches, headache, muscle
aches. Loss of energy and fatigability
2 Any clear-cut symptom rates 2

14. Genital Symptoms
symptoms such as loss of libido, menstrual disturbance
0 Absent
1 Mild
2 Severe

15. Hypochondriasis
0 Not present
1 Self-absorption (bodily)
2 Preoccupation with health
3 Frequent complaints, requests for help, etc.
4 Hypochondriacal delusions

16. Loss of Weight
0 N 0 weight loss
1 Probable wieght loss associated with present illness
2 Deﬁnite (according to patient) weight loss

17 . Insight .
0 Acknowledges being depressed and ill
1 Acknowledges illness but attibutes cause to bad food, climate,

overwork, virus, need for rest, etc.
2 Denies being ill at all

133
APPENDIX A

Depression Rating Scales
Seasonal Affective Disorder Rating Scale

. Fatigability (or low energy level, or feelings of being heavy, laden,
weighed down) .

0 Does not feel more fatigued than usual

1 Feels more fatigued than usual but this has not impaired
function signiﬁcantly

2 More fatigued than usual; at least 1 hour a day; three days a
week

3 Fatigued much of the time, most days

4 Fatigued almost all the time

. Social withdrawal

0 Interacts with other people as usual

1 Less interested in socializing with others but continues to do so
2 Interacting less with other people in social (optional) situations
3 Interacting less with other people in work or family situations
4 Marked withdrawal from others in family or work situations

. Appetite increase

0 N 0 increase in appetite

1 Wants to eat a little more than usual

2 Wants to eat somewhat more than usual
3 Wants to eat much more than usual

. Increased eating

0 Is not eating more than usual

1 Is eating a little more than usual

2 I s eating somewhat more than usual
3 Is eating much more than usual

. Carbohydrate craving

0 No change in food preference

1 Eating more carbohydrates (starches or sugars) than before
2 Eating much more carbohydrates than before

3 Irrestible craving for sweets or starches

134

6. Weight gain
0 No weight gain
1 Probable weight gain associated with present illness

2 Deﬁnite weight gain

7. Hypersomnia
0 No increase in sleep length
1 At least 1 hour increase in sleep length
2 2+ hour increase in sleep length
3 3+ hour increase in sleep length
4 4+hour increase in sleep length

135

APPENDIX B

Analysis of variance of the effect of full spectrum v. cool white
phototherapy

 

HRSD Scores
WmL—JL—Mus 1’
Subject 9 3.09 .06
Treatment 1 .44 .5
Stage 1 1.41 .3
Error 8
Period 1 93.12 .0001
Period*Subject 9 2.06 .2
Period*Treatment 1 .01 .9
Period*Stage 1 .88 .4

Error(Period) 3

136
APPENDIX B

Analysis of variance of the effect of full spectrum v. cool white
phototherapy

 

 

SAD Scores
Wanna df F value P
Subject 9 1.01 .5
Treatment 1 .08 .8
Stage 1 .02 .9
Error 8
Period 9 168.37 .0001
Period*Subject 1 1.75 .2
Period*Treatment 1 3.31 .1
Period*Stage 1 5.31 .05

Error(Period) 3

137
APPENDIX B

Analysis of variance of the effect of full spectrum v. cool white
phototherapy

Melatonin Concentrations

 

w 1
Subject 9 4.28 .03
Treatment 1 .95 .4
Stage 1 1.36 .3
Error 8

Period 1 8.53 .02
Period*Subject 9 1.16 .4
Period*Treatment 1 .43 .5
Period*Stage 1 .00 .9
Error(Period) 8

Sample 12 32.74 .0001
Sample*Subject 108 5.3 .0001
Sample*Treatment 12 .62 .8
Sample*Stage 12 .53 .9
Error(Sample) %

Period*Sample 12 2.06 .03
Period*Sample*Subject 108 2.34 .0001
Period*Sample*Treatment 12 .54 .9
Period*Sample*Stage 12 .56 .9

Error(Period*Sample)

138
APPENDIX C

Analysis of variance of the effect of full spectrum v. narrow spectrum
green light

 

HRSD Scores
WL—JL—Mm P
Subject 17 .96 .5
Treatment 1 2.14 .2
Stage 1 23.9 .0002
Error 16
Period 1 74.75 .0001
Period*Subject 17 1.2 .4
Period*Treatment l ‘ 2.81 .1
Period*Stage 1 20.69 .0003

Error(Period) 16

139
APPENDIX C

Analysis of variance of the effect of full spectrum v. narrow spectrum
green light

 

 

SAD Scores
We df F m. p
Subject 17 4.06 .004
Treatment 1 2.61 .13
Stage 1 12.45 .003
Error 16
Period 1 37.8 .0001
Period*Subject 17 .7 .8
Period*Treatment 1 1.86 .2
Period*Stage . 1 11.71 .004

Error(Period) 16

140
APPENDIX C

Analysis of variance of the effect of full spectrum v. narrow spectrum
green light ‘

SIGH-SAD Scores

 

 

Wee df F value P
Subject 17 1.94 .1
Treatment 1 3.94 .06
Stage 1 29.78 .0001
Error 16

Period 1 70.42 .0001
Period*Subject 17 .82 .7
Period*Treatment l 2.95 .1
Period*Stage 1 20.37 .004

Error(Period) 16

141
APPENDIX C

Analysis of variance of the effect of full spectrum v. narrow spectrum
green light

Melatonin Concentrations

 

W P
Subject 17 8.38 .0001
Treatment 1 .02 .9
Stage 1 .00 1
Error 16

Period 1 16.25 .001
Period*Subject 17 2.25 .06
Period*Treatment 1 1.09 .31
Period*Stage 1 4.84 .04
Error(Period) 16

Sample 6 7.8 .0001
Sample*Subject 102 1.14 .26
Sample*Treatment 6 1.74 .12,
Sample*Stage 6 2.16 .05
Error(Sample)‘ %

Period*Sample 6 1.43 .2
Period*Sample*Subject 102 1.19 .2
Period*Sample*Treatment 6 1.74 .12
Period*Sample*Stage 6 1.5 .19

Error(Period*Sample) %

142
APPENDIX D

Analysis of variance of the effect of eye v. head phototherapy

 

 

HRSD Scores

We df F value P
Subject 11 1.42 .3
Treatment 1 .42 .5
Stage 1 .01 .9
Error 10

Period 1 55.08 .0001
Period*Subject 11 1.46 .3
Period*Treatment 1 2.05 .2
Period*Stage 1 .44 .5

Error(Period) 10

143
APPENDIX D

Analysis of variance of the effect of eye v. head phototherapy

 

SAD Scores

Spurs: gf yan'angg {If E: mine P
Subject 11 1.55 .2
Treatment 1 1.23 .3
Stage 1 .05 .8
Error 10

Period 1 50.56 .0001
Period*Subject 11 1.21 .4
Period*Treatment 1 1.4 .3
Period*Stage 1 1.7 .2

Error(Period) 10

144
APPENDIX D

Analysis of variance of the effect of eye v. head phototherapy
SIGH-SAD Scores

 

Sgurgg of Vaganm dﬁ E yams P
Subject 11 1.27 .4
Treatment 1 .84 .4
Stage 1 .02 .9
Error 10

Period 1 64.56 .0001
Period*Subject 11 1.47 .3
Period*Treatment 1 2.12 .2
Period*Stage 1 1.13 .3

Error(Period) 10

145
APPENDIX D

Analysis of variance of the effect of eye v. head phototherapy

Melatonin Concentrations

 

 

Snumflariance df F value 1’
Subject 11 4.81 .01
Treatment 1 .37 .6
Stage 1 .05 .8
Error 10

Period 1 4.64 .06
Period*Subject 11 2.82 .06
Period*Treatment 1 .6 .5
Period*Stage 1 3.66 .08
Error(Period) 10

Sample 6 2.63 .02
Sample*Subject 66 1.6 .03
Sample*Treatment 6 1.18 .3
Sample*Stage 6 1.41 .2
Error(Sample) 6)

Period*Sample 6 1.51 .2
Period*Sample*Subject 66 1.7 .02
Period*Sample*Treatment 6 5 .8
Period*Sample*Stage 6 .97 .5
Error(Period*Sample) 6)

146

APPENDIX E

Analysis of variance of the eﬂ'ect of blended white light

 

 

HRSD Scores

We df FvalusL P
Subject 5 .48 .8
Treatment 2 .21 .8 '
Stage 1 .13 .7
Error 3

Period 1 5.97 .09
Period*Subject 5 35 .9
Period*Treatment 2 .45 .7
Period*Stage 1 .04 .9

Error(Period) 3

147

APPENDIX E

Analysis of variance of the effect of blended white light

 

 

SAD Scores

Saumuflanance df anlus P
Subject 5 4.11 .1
Treatment 2 1.38 .4
Stage 1 19.97 .02
Error 3

Period 1 6.11 .09
Period*Subject 5 1 1.0
Period*Treatment 2 .19 .8
Period*Stage 1 0 1.0

Error(Period) 3

148

APPENDIX E

Analysis of variance of the effect of blended white light

Sanctum—M

Subject
Treatment

Stage

Error

Period
Period*Subject
Period*Treatment
Period*Stage

Error(Period)

SIGH

5
2

-SAD Scores

.95

.12

2.06

6.35
.21

.01

 

ioéo'oz'U

.09

149
APPENDIX E

Analysis of variance of the effect of blended white light
Melatonin Concentrations (0600-0900 h)

 

 

Wanna df Maine P
Subject 5 4.95 .1
Treatment 2 2.41 .2
Stage 1 8.12 .07
Error 3

Period 1 1.26 .3
Period*Subject 5 2 .9
Period*Treatment 2 .2 .8
Period*Stage 1 .11 .8
Error(Period) 3

Sample 6 12.9 .0001
Sample*Subject 30 1.56 .2
Sample*Treatment 12 1.08 .4
Sample*Stage 6 .97 5
Error(Sample) 18

Period*Sample 6 1.39 .3
Period*Sample*Subject 30 1.4 .2
Period*Sample*Treatment 12 1.49 .2
Period*Sample*Stage 6 3.21 .03
Error(Period*Sample) 18

150
APPENDIX E

Analysis of variance of the effect of blended white light

Melatonin Concentrations (2000-2300 h)

 

 

Snumflariance df Fm P
Subject 5 1.73 .3
Treatment 2 1.75 .3
Stage 1 1.78 .3
Error 3

Period 1 .05 .8
Period*Subject 5 .58 .7
Period*Treatment 2 1.32 .4
Period*Stage 1 0 1.0
Error(Period) 3

Sample 6 1.72 .2
Sample*Subject 30 2.69 .02
Sample*Treatment 12 1.79 .1
Sample*Stage 6 .69 .7
Error(Sample) 18

Period*Sample 6 .61 .7
Period*Sample*Subject 30 .39 1.0
Period*Sample*Treatment 12 .31 1.0
Period*Sample*Stage 6 1.11 .4
Error(Period‘Sample) 18

BIBLIOGRAPHY

Bibliography

Adams, J. S., Clemens, T. L., Parrish, J. A, Holick, M. I. 1982. Vitamin D
synthesis and metabolism after ultraviolet radiation of normal and
vitamin D deﬁcient subjects. N. Engl. J. Med. 306:722. '

Allen, N., D. Kerr, P. Smythe, N. Martin, K. Osola, C. Thompson. 1992.
Insulin sensitivity after phototherapy for seasonal affective disorder.
Lancet 339:1065.

American Psychiatric Association. 1987. Diagnostic and Statistical Manual of
Mental Disorders, 3rd ed. revised. American Psychiatric Association,
Washington.

Anderson, R. R. and J. A Parrish. 1981. The optics of human skin. J. Invest.
Dermatol. 77:13.

Arbisi, P. A, R. A. Depue, M. B. Spoont, A. Leon, B. Ainsworth. 1989.
Thermoregulatory response to thermal challenge in seasonal affective
disorder. Psych. Res. 28:323.

Arendt, J ., A A. Borbely, C. Franey, J. Wright. 1984. The effect of chronic,
small doses of melatonin given in the late afternoon on fatigue in man:
a preliminary study. Neurosci. Lett. 45:317.

Arendt, J ., S. Hampton, J. English, P. Kwasowski, V. Marks. 1982. 24-hour
proﬁles of melatonin, cortisol, insulin, C-peptide and GIP following a
meal and subsequent fasting. Clin. Endo. 16:89.

Aristotle. 1953. Problemata, book xxx, problem I. In: W. D. Ross (Ed.) The
Works of Aristotle Translated into English, vol VII. p. 954. Oxford
University Press, Oxford.

Aschoff, J. 1981. Annual rhythms in man. In: J. Aschoff (Ed.) Handbook of
Behavioral Neurobiology. Biological Rhythms vol 4. pp. 475-487.
Plenum Press, New York.

Avery D., M. A. Bolte, S. Cohen, D. Dunner. 1990a. Is dawn simulation
effective in treating winter depression? Bull. Soc. Light Treatment
Biol. Rhythms Abst. 2:16.

151

152

Avery D., A. Khan, S. Dager, S. Cohen, G. Cox, D. Dunner. 1990b. Is morning
light superior to evening light in treating seasonal affective disorder?
Psychopharmacol. Bull. 26:521.

Avery D. H., A Kahn, S. R. Dager, S. Cohen, G. B. Cox, D. L. Dunner. 1991.
Morning or evening bright light treatment of winter depression? The
signiﬁcance of hypersomnia. Biol. Psych. 29:117.

Avery D.H., A. Khan, S. Dager, D. Dunner. 1987. Temperature rhythm
phase-typing of seasonal affective disorder and response to AM and PM
bright light. Sleep Res. 16:595.

Axelrod, J. 1974. The pineal gland: a neurochemical transducer. Science
184:1341.

Bartko, J. J. and S. Kasper. 1989. Seasonal changes in mood and behavior: a
cluster analytic approach. Psych. Res. 28:227.

Beersma, D. G. M. 1990. Do winter depressives experience summer nights in
winter? Arch. Gen. Psych. 47:879.

Blehar, M. C. and A. J. Lewy. 1990. Seasonal mood disorders: Concensus and
controversy. Psychopharmacol. Bull. 26:465.

Blehar, M. C. and N. E. Rosenthal. 1989. Seasonal affective disorders and
phototherapy. Arch. Gen. Psych. 46:469.

Boyce, P., G. Parker. 1988. Seasonal affective disorder in the southern
hemisphere. Am. J. Psych. 145:96.

Boyce, P., D. J. Kennaway. 1987 . Effects of light on melatonin production.
Biol. Psych. 22:473.

Brainard, G. C., A. J. Lewy, M. Menaker, R. H. Fredrickson, L. S. Miller, R.
G. Weleber, V. Cassone, D. Hudson. 1985. Effect of light wavelength on
the suppression of nocturnal plasma melatonin in normal volunteers.
Ann. N. Y. Acad. Sci. 4532376

Brainard, G., A Lewy, M. Menaker, R. Fredrickson, L. Miller, R. Weleber, V.
Cassone, D. Hudson. 1988. Dose-response relationship between light
irradiance and the suppression of plasma melatonin in human
volunteers. Brain Res. 454:212.

Brainard, G. C., D. Sherry, R. G. Skwerer, M. Waxler, K. Kelly, N. E.
Rosenthal. 1990. Effects of different wavelengths in seasonal affective
disorder. J. Affective Disord. 20:209.

153
Brewerton, T., W. H. Berrettini, J. I. Numberger, M. Linnoila. 1988. An
analysis of seasonal ﬂuctuations of CSF monoamine metabolites and
neuropeptides in normal controls: Findings with 5-HIAA and HVA
Psych. Res. 23:257.

Brier, A. 1989. Experimental approaches to human stress research:
Assessment of neurobiological mechanisms of stress in volunteers and
psychiatric patients. Biol. Psych. 26:438.

Brismar, K., S. Werner, M. Thoren, L. Wetterberg. 1985. Metyrapone: an
agent for melatonin as well as ACTH and cortisol secretion. J.
Endocrinol. Invest. 8:91.

Broadway, J. W. and J. Arendt. 1985. 24-hour rhythm of human 6-
hydroxymelatonin sulphate in Antarctica. J. Endocrinol. 106:96.

Brun, J ., B. Claustrat, M. David. 1987. Urinary melatonin, LH, oestradiol,
progesterone excretion during the menstrual cycle or in women taking
oral contraceptives. Acta Endocrinol. (Copenh.) 116:145.

Brzezinski, A, H. J. Lynch, M. M. Seibel, M. H. Deng,,T. M. Nader, R. J.
Wurtman. 1988. The circadian rhythm of plasma melatonin during the
normal menstrual Cycle and in amenorrheic women. J. Clin.
Endocrinol. Metab. 66:891.

Buguet, A, R. Gati, G. Soubiran, J. P. Straboni, A. M. Hanniquet, G.
Livecchi-Gonnet, J. Bittel. 1988. Seasonal changes in circadian
rhythms of body temperatures in humans living in a dry tropical
climate. Eur. J. Appl. Physiol. 58:334.

Calabrese, J. R. M. A. Kling, P. W. Gold. 198'7. Alterations in
immunocompetence during stress, bereavement, and depression. Focus
on neuroendocrine regulation. Am. J. Psych. 144: 1123.

Cardinali, D. P., H. J. Lynch, R. J. Wurtman. 1972. Binding of melatonin to
human and rat plasma proteins. Endocrinol. 91:1213.

Cardinali, D. P., M. I. Vacas, M. I. Keller Sarmiento, E. Morguerstern. 1983.
Melatonin action: sites and possible mechanisms in brain. In: J.
Axelrod, F. Fraschini, G.P. Velo (Ed.) The Pineal Gland and its
Endocrine Role. pp. 277-299. Plenum Press, New York.

Carlsson, A, L. Svennerholm, B. Winblad. 1980. Seasonal and circadian
monomaine variations in human brains examined postmortem. Acta
Psychiatr. Scand. 61(supp1280):75.

154

Claustrat, B., D. Chazot, J. Brun, D. Jordan, G. Sassolas. 1984. A
chronobiological study of melatonin and cortisol secretion in depressed
subjects: Plasma melatonin, a biochemical marker in major depression.
Biol. Psych. 19:1215.

CDhen M., D. Roselle, T. J. Schmidt, M. Lippman. 1978. Evidence for a
cytoplasmic melatonin receptor. Nature 274:894.

Cowen, P. J ., S. Fraser, R. Sammons, A R. Green. 1983. Atenolol reduces
plasma melatonin concentration in man. Br. J. Clin. Pharmacol.

15:579.

Curcio, C. A, K. R. Sloan, R. E. Kalina, A F. Hendrickson. 1989. Human
photoreceptor topography. J. Comp. Neurol. 292:497.

Czeisler, C. A, R. E. Kronauer, J. S. Allan, J. F. Duffy, M. E. Jewett, E. N.
Brown, J. M. Ronda. 1989. Bright light induction of strong (type 0)
resetting of the human circadian pacemaker. Science 244: 1328.

Czeisler, C. A, R. E. Kronauer, J. J. Mooney, J. L. Anderson, J. S. Allan.
1987. Biologic rhythm disorders, depression, and phototherapy. In: M.
K. Erman (Ed.) Psychiatric Clinics of North America, vol 10, no. 4.
pp.687-709. W. B. Saunders Co., Philadelphia.

Dawes, C. 197 2. Research in human salivary ﬂow rate and composition. J.
Physiol. (Lond) 220:529.

DeLuca, H. F. 1984. The metabolism, physiology and function of vitamin D.
In: Kumar (Ed) Vitamin D, Basic and Clinical Aspects. pp. 1-68.
Martinus Nijhogg Publishing, Boston.

Depue, R. A, P. Arbisi, S. Krauss, W. G. Iacono, A. Leon, R. Muir, J. Allen.
1990. Seasonal independence of low prolactin concentration and high
spontaneous eye blink rates in unipolar and bipolar II seasonal
affective disorder. Arch. Gen. Psych. 47:356.

Depue, R. A, P. Arbisi, M. R. Spoont, S. Krauss, A. Leon, B. Ainsworth.
1989a. Seasonal and mood independence of low basal prolactin
secretion in premenopausal women with seasonal affective disorder.
Am. J. Psych. 146:989.

Depue, R. A, P. Arbisi, M. R. Spoont, A. Leon, B. Ainsworth. 1989b.
Dopamine functioning in the behavioral facilitation system and
seasonal variation in behavior: Normal population and clinical studies.
In: N. E. Rosenthal, M. C. Blehar (Ed.) Seasonal Affective Disorder
and Phototherapy. pp. 230-259. Guilford Press, New York.

Dilsaver, S. C. 1990. Onset of winter depression earlier than generally
thought? J. Clin. Psych. 51:258.

155

” Dilsaver, S. C. and R. S. Jaeckle. 1990. Winter depression responds to an
open trial of tranylcypromine. J. Clin. Psych. 51:326.

Eastman, C. I. 1990. What the placebo literature can tell us about
phototherapy in SAD. Psychopharmacol. Bull. 26:495.

Ennever, J. 1990. Blue light, green light, white light, more light: Treatment
of neonatal jaundice, Clin. Perinatal. 17:467.

Evans, F. J. 1985. Expectancy, therapeutic instructions, and the placebo
response. In: L. White (Ed.) Placebo Theory, Research, and
Mechanisms. pp. 215-228. Guilford Press, New York.

Fraser, S., P. Cowen, M. Franklin, C. Franey, J. Arendt. 1983. Direct
radioimmunoassay for melatonin in plasma. Clin. Chem. 29:396.

Gaist, P. A, E. Obarzanek, R. G. Skwerer, C. C. Duncan, P. M. Shultz, N. E.
Rosenthal. 1990. Effects of bright light on resting metabolic rate in
patients with seasonal affective disorder and control subjects. Biol.
Psych. 28:989.

Gill, J. 1978. Design and Analysis of Experiments in the Animal and Medical
Sciences. Vol 1. pp. 159- 185. Iowa State University Press, Ames, Iowa.

Gill, J. 1978. Design and Analysis of Experiments in the Animal and Medical
Science. Vol 2. pp. 169-259. Iowa State University Press, Ames, Iowa.

Gillin, J. C., N. Sitaram, T. A Wehr, W. Duncan, R. Post, D. L. Murphy, W.
B. Mendelson, R. J. Wyatt, W. E. Bunney. 1984. Sleep and affective
illness. In: R. M. Post and J. C. Ballenger (Ed.) Neurobiology of Mood
Disorders. pp. 175-180. Williams and Wilkins, Baltimore.

Gold, P. W., F. K Goodwin, G. P. Chrousos. 1988. Clinical and biochemical
manifestations of depression (parts I and II). New Engl. J. Med.
319:348 and 413.

Gould, S. J. 1987 . Time's Cycle and Time's Arrow; Myth and Metaphor in
the Discovery of Geological Time. Harvard University Press,
Cambridge.

Graw, P., K. Krauchi, A. Wirz-Justice, W. Poldinger. 1991. Diurnal variation
of symptoms in seasonal affective disorder. Psych. Res. 37:105.

Grifﬁths, P. A, S. Folkard, C. Bojkowski, J. English, J. Arendt. 1986.
Persistent 24-h variations of urinary 6-hydroxy melatonin sulphate
and cortisol in Antarctica. Experientia 42:430.

156

Grota, L. J ., B. I. Yerevanian, K. Gupta, J. Kruse, L. Zborowski. 1989.
Phototherapy for seasonal major depressive disorder: Effectiveness of
bright light of high or low intensity. Psych. Res. 29:29.

Grunbaum, A. 1986. The placebo concept in medicine and psychiatry.
Psychol. Med. 16:19.

Gwinner, E. 1981. Annual rhythms: Perspective. In: J. Aschoff (Ed.)
Handbook of Behavioral Neurobiology. Biological Rhythms vol 4. pp.
381-389. Plenum Press, New York.

Hamilton, M. 1967. Development of a rating scale for primary depressive
illness. Br. J. Soc. Clin. Psych. 6:278.

Hansen, T., T. Bratlid, O. Lin-arde, T. Brenn. 1987 . Midwinter insomnia in
the subarctic region: Evening levels of serum melatonin and cortisol
before and after treatment with bright artiﬁcial light. Acta Psychiatr.
Scand. 75:428.

Hanssen, T., T. Heyden, I. Sunberg, L. Wetterberg. 1977. Effect of propanolol
on serum melatonin. Lancet 309.

Haus, E., D. J. Lakatua, F. Halberg, G. Comelissen, L. L. Sacket, H. G. Berg,
T. Kawasaki, M. Ueno, K. Uezone, M. Matsuoka, T. Omae. 1980.
Chronobiological studies of plasma prolactin in women in Kyushu,
Japan and Minnesota, USA J. Clin. Endocrinol. Metab. 51:632.

Hellekson, C. 1989. Phenomenology of seasonal affective disorder: An
Alaskan perspective. In: N. E. Rosenthal and M. Blehar (Ed.) Seasonal
Affective Disorders and Phototherapy. pp. 33-45. Guilford Press, New

ork.

Hellekson, C. J ., J. A Kline, N. E. Rosenthal. 1986. Phototherapy for seasonal
affective disorder in Alaska. Am. J. Psych. 143:1035.

Hoﬁnan, M. A, D. F. Swaab. (1992). Seasonal changes in the suprachiasmatic
nuclei of man. Neurosci. Lett. 139:257.

Horne, J. A, J. Donlon, J. Arendt. 1991. Green light attenuates melatonin
output and sleepiness during sleep deprivation. Sleep 14:233.

Illnerova, H., P. Zvolsky, J. Vanacek. 1985. The circadian rhythm in plasma
melatonin concentration of the urbanized man: Effect of summer and
winter time. Brain Res. 328:186.

Indyck, L. 1976. The physics of phototherapy. In: A. K. Brown and J.
Showacre (Ed.) Phototherapy for neonatal hyperbilirubinemia. pp. 207-
218. DHEW pub. No. (NIH) 76-1075.

157

Jackson, S. W. 1987. Melancholia and Depression from Hippocratic Times to
Modern Times. p. 42. Yale University Press, New Haven.

Jacobsen, F. M., E. A Mullerr, S. L. B. Rogers, N. E. Rosenthal. 1989. The
'role of serotonin in seasonal affective disorders and the antidepressant
response to phototherapy. In: N. E. Rosenthal and M. C. Blehar (Ed.)
Seasonal Affective Disorders and Phototherapy. pp. 333-341. Guilford
Press, New York.

Jacobsen, F. M., D. A Sack, T. A Wehr, S. Rogers, N. E. Rosenthal. 1987a.
N euroendocrine response to 5-hydroxytryptophan in seasonal affective
disorder. Arch. Gen. Psych. 44:1086. .

J acobsen, F. M., T. A. Wehr, R. A Skwerer, D. Sack, N. E. Rosenthal. 1987b. '
Morning versus midday phototherapy of seasonal affective disorder.
Am. J. Psych 144:1301.

James, S. P., T. A. Wehr, D. A Sack, B. L. Parry, S. Rogers, N. E. Rosenthal.
1986. The dexamethasone suppression test in seasonal affective
disorder. Compr. Psych. 27:224.

James, S.P., T. A. Wehr, D. A Sack, B. L. Parry, N. E. Rosenthal. 1985.
Treatment of SAD with light in the evening. Br. J. Psych. 147:424.

Jones, R, M. Kenwar. 1989. Design and Analysis of Cross-Over Trials. pp. 16-
88. Chapman and Hall, London.

Kasper, S., N. E. Rosenthal, S. Barberi, A Williams, L. Tamarkin, S. L. B.
Rogers, S. R. Pillemar. 1991. Immunological correlates of seasonal
ﬂuctuations in mood and behavior and their relationship to
phototherapy. Psych. Res. 36:253-264.

Kasper, S., L. B. S. Rogers, P. A. Madden, J. R. Joseph-Vanderpool, N. E.
Rosenthal. 1990. The eﬂ'ects of phototherapy in the general population.
J. Affective Disord. 18:211.

Kasper, S., T. A Wehr, J. J. Bartko, P. A. Gaist, N. E. Rosenthal. 1989.
Epidemiological ﬁndings of seasonal changes in mood and behavior: A
telephone survey of Montgomery County, Maryland. Arch. Gen. Psych.
46:823.

Kauppila, A, A. Kivela, A. Pakarinen, O. Vakkuri. 1987 . Inverse seasonal
relationship between melatonin and ovarian activity in humans in a
region with a strong seasonal contrast in luminosity. J. Clin.
Endocrinol. Metab. 65:823.

Kennaway, D. J. and P. Royles. 1986. Circadian rhythms of 6-sulphatoxy
melatonin, cortisol and electrolyte excretion at the summer and winter
solstices in normal men and women. Acta Endocrinol. 113:450. '

158

Khan, R., S Daya, B. Potgieter. 1990. Evidence for a modulation of the stress
response by the pineal gland. Experientia 46:860.

Klein, D. C. 1979. Circadian rhythms in the pineal gland. In: D. T. Krieger
(Ed.) Endocrine Rhythms. pp. 203-223. Raven Press, New York.

Klein, D. C. and J. L. Weller. 1970. Indole metabolism in the pineal gland, a
circadian rhythm in N-acetyltransferase. Science 169: 1093.

Kloeppel, H. B. 1982. Circannual changes of the circadian phase position in
man. In: A. Hildebrandt and H. Hankel (Ed.) Biological Adaptation:
International Symposium at the University of Larburg/Lahn. pp. 137-
145. Georg Thieme, Stuttgart.

Kopin, I. J., C. M. B. Pare, J. Axelrod, H. Weissbach. 1961. The fate of
melatonin in mammals. J. Biol. Chem. 236:3072.

Krauchi, K. and A. Wirz-Justice. 1988. The four seasons: Food intake
frequency in seasonal affective disorder in the course of a year. Psych.
Res. 25:323.

Kripke, D. F. 1984. Critical interval hypotheses for depression. Chronobiology
International 1:73.

Laasko, M., T. Porkka-Heiskanen, A Alila, D. Stenberg, G. Johansson. 1990.
Correlation between salivary and serum melatonin: Dependence on
serum melatonin levels. J. Pineal Res. 9:39.

Lacoste, V. and A. Wirz-Justice. 1989. Seasonal variation in normal subjects:
An update on variables current in depression research. In: N. E.
Rosenthal and M. Blehar (Ed.) Seasonal Affective Disorders and

- Phototherapy. pp. 167-229. Guilford Press, New York.

Lam, R. W., C. W. Beattie, A Buchanan, R. A Remick, A. P. Zis. 1991. Low
electrooculographic ratios in patients with seasonal affective disorder.
Am. J. Psych. 148:1526.

Lam, R. A, A. Buchanan, J. A Mador, M. R. Corral, R. A Remick. 1992. The
effects of ultraviolet-A wavelength in light therapy for seasonal
depression. J. Affective Disord. 24:237.

Lawwill, T., S. Crockett, G. Currier. 1977. Retinal damage secondary to
chronic light exposure. Doc. Ophthalmol. 44:379.

Lerner, A. B., J. D. Case, Y. Takahasld, T. H. Lee, N. Mori. 1958. Isolation of
melatonin, pineal factor that lightens melanocytes. J. Am. Chem. Soc.
80:2587. .

159

Leskowitz, E. 1990. Seasonal affective disorder and the yoga paradigm: A
reconsideration of the role of the pineal gland. Med. Hypotheses
33:155.

Lewy, A. J. and R. L. Sack. 1986. Light therapy and psychiatry. Proc. Soc.
Exp. Biol. Med. 183:11.

Lewy, A. J ., R. L. Sack. 1989. The dim light melatonin onset as a marker for
circadian phase position. Chronobiology Int. 6:93.

Lewy, A. J ., H. A. Kern, N. E. Rosenthal, T. A Wehr. 1982. Bright artiﬁcial
light treatment of a manic-depressive patient with a seasonal mood
cycle. Am. J. Psych. 139:1496.

Lewy, A. J ., R. L. Sack, S. Miller, T. M. Hoban. 1987. Antidepressant and
circadian phase shifting effects of light. Science 235:352.

Lewy, A. J ., T. A Wehr, F. K. Goodwin, D. A Newsome, S. P. Markey. 1980.
Light suppresses melatonin secretion in humans. Science 210: 1267.

Li, Y., D. H. Jiang, M. L. Wang, R. Jiao, S. F. Pang. 1989. Rhythms of serum
melatonin in patients with spinal lesions at the cervical, thoracic or
lumbar region. Clin. Endocrinol. 30:47.

Lucey, J. F., J. Hewitt. 1976. Recent observations on light and neonatal
jaundice. In A. K. Brown, J. Showacre (Eds.) Phototherapy for
Neonatal Hyperbilirubinemia. DHEW Pub. No. 76-1075. Government
Printing Center, Washington, D. C.

Mayor, J ., J. Rice, R. J. Bielski. 1991. Environmental inﬂuences on the onset
of winter depression. J. Clin. Psych. 52:480.

McGrath, R. E., B. Buckwald, E. V. Resnick. 1990. The effect of L-tryptophan
on seasonal affective disorder. J. Clin. Psych. 51:4.

McIntyre, I. M., S. M. Armstrong, T. R. Norman, G. D. Burrows. 1989a.
Treatment of seasonal affective disorder with light: Preliminary
Australian experience. Austral. and New Zea. J. Psych. 23:369.

McIntyre, I., T. Norman, G. Burrows, S. Armstrong. 1987 . Melatonin rhythm
in human plasma and saliva. J. Pineal Res. 4:177.

McIntyre, I. M., T. R. Norman, G. D. Burrows, S. M. Armstrong. 1989b.
Human melatonin suppression by light is intensity dependent. J.
Pineal Res. 6:149.

McIntyre, I. M., T. R. Norman, G. D. Burrows, S. M. Armstrong. 1990.
Melatonin supersensitivity to dim light in seasonal aﬁ‘ective disorder.
Lancet 335:488.

160

Meesters, Y., J. H. C. Jansen, D. G. M. Beersma, A. L. Bouhuys, R. H. van ’
den Hoofdakker. 1992. Early light treatment and the prevention of
winter depression. Presented at the annual meeting of the Soc. Light
Treatment Biol. Rhythms, Washington, D. C.

Miles, A, D. R. S. Philbrick, D. M. Shaw, S. F. Tidmarsh, A. J. Pugh. 1985a.
Salivary melatonin estimation in clinical research. Clin. Chem.
3122041.

Miles, A, D. Philbrick, S. Tidmarsh, D. Shaw. 1985b. Direct
radioimmunoassay of melatonin in saliva. Clin. Chem. 31:1412.

Miles, A, D. Philbrick, J. Grey. 1989. Salivary melatonin estimation in
assessment of pineal-gland function. Clin. Chem. 35:514.

Morison, W. L. 1985. Photoimmunology: Study of the effects of nonionizing
radiation of the immune system. Ann. New York Acad. Sci. 453:1105.

Moseley, M. J ., S. C. Bayliss, A R. Fielder. 1988. Light transmission through
the human eyelid: in vivo measurement. Ophthal. Physiol. Optics.
8:229.

Mrosovsky, N. 1988. Seasonal affective disorder, hibernation, and annual
cycles in animals: Chipmunks in the sky. J. Biol. Rhythms 3:189.

Nagayama, H., M. Sasaki, S. Ichii, K. Hanada, M. Okawa, T. Ohta, Y. Asano,
Y. Sugita, J. Yamazaki, M. Kohsaka, T. Kotorii, K. Maeda, N.
Okamoto, Y. Ishizuka, K. Takahashi, Y. Honda, S. Takahashi. 1991.
Atypical depressive symptoms possibly predict responsiveness to
phototherapy in seasonal aﬁ’ective disorder. J. Aﬁ'ective Disord. 23:185.

Nathans, J ., D. Thomas, D. S. Hogness. 1986. Molecular genetics of human
color vision: The genes encoding blue, green, and red pigments. Science
2322193.

' Neuwelt, E. A. and A. J. Lewy. 1983. Disappearance of plasma melatonin
after removal of a neoplastic pineal gland. New Eng. J. Med. 308:1132.

Nickelsen, T., A. Samel, H. Maass, M. Vejvoda, H. Wegmann, K. Schoming.
1991. Circadian patterns of salivary melatonin and urinary 6-
sulfatoxymelatonin before and after a 9 hour time-shift. Adv. Exp.
Med. Biol. 294:493.

O'Malley, B. P., A. Richardson, N. Cook, S. Swart, F. D. Rosenthal. 1984.
Circadian rhythms of serum thyrotropin and body temperature in
euthyroid individuals and their responses to warming. Clin. Sci.
67:433.

161

O’Rourke, D., J. J. Wurtman, R. J. Wurtman, R. Chebli, R. Gleason. 1989.
Treatment of seasonal depression with d-fenﬂuramine. J. Clin. Psych.

50:9.

Oren, D., G. Brainard, S. Johnston, J. Joseph-Vanderpool, E. Sorek, N.
Rosenthal. 1991a. Treatment of seasonal affective disorder with green
light and red light. Am. J. Psych. 148:509.

Oren, D. A, F. S. Rosenthal, N. E. Rosenthal, M. Waxler, T. A Wehr. 1990.
Exposure to ultraviolet B radiation during phototherapy. Am. J.
Psych. 147:675.

Oren, D. A, J. R. Vanderpool, N. E. Rosenthal. 1991b. Adaptation to dim
light in depressed patients with seasonal affective disorder. Psych.
Res. 36: 187.

Paramore, J. E. and V. M. King. 1989. Ophthalmic implications of seasonal
affective disorder. J. Am. Optom. Assoc. 60:508.

Pittendrigh, C. S. 1988. The photoperiodic phenomenona: Seasonal
modulation of the "day within". J. Biol. Rhythms 3:173.

Potkin, S. G., M. Zetin, V. Stamenkovic, D. Kripke, W. E. Bunney. 1986.
Seasonal affective disorder: Prevalence varies with latitude and
climate. Clin. N europharmacol. 9(suppl 4):181.

Rafferty, B., M. Terman, J. Terman, C. Reme. 1990. Does morning light
prevent evening light effect? A statistical model for morning/evening
crossover studies. (Abstr) Presented at the meeting of the Sec. Light
Treatment Biol. Rhythms, New York.

Reinberg, A. 1983. Clinical chronopharmacology: An experimental basis for
chronotherapy. In: A. Reinberg and M. H. Smolensky (Ed.) Biological
Rhythms and Medicine: Cellular, Metabolic, physiopathologic and
Phannacologic Aspects. pp. 211-263. Spring-Verlag, New York.

Reiter, R. J. 1987 . The melatonin message: Coincidence hypotheses. Life Sci.
40:2119.

Reiter, R. J. 1988. NeurOendocrinology of melatonin. In: A. Miles, D. B. S.
Philbrick, C. Thompson (Ed.) Melatonin: Clinical Perspectives. pp. 1-
42. Oxford University Press, Oxford.

Reme, C, M. Terman, A Wirz-Justice. 1990. Are deﬁcient retinal
photoreceptor renewal mechanisms involved in the pathogenesis of
winter depression? Arch. Gen. Psych. 47: 878.

Reppert, S. M., D. R. Weaver, S. A. Rivkees, E. G. Stopa. 1988. Putative
melatonin receptors in a human biological clock. Science 242:7 8.

162

Rollag, M. D., R. J. Morgan, G. D. Niswender. 1978. Route of melatonin
secretion in sheep. Endocrinology 102:1.

Rosen, L. N. and N. E. Rosenthal. 1991. Seasonal variation in mood and
behavior in the general population: A factor-analytic approach. Psych.
Res. 38:271.

Rosen, L. N., S. D. Tergum, M. Terman, M. J. Bryant, H. Hoffman, S. F.
Kasper, J. R. Hamovit, J. P. Docherty, B. Welch, N. E. Rosenthal. 1990.
Prevalence of seasonal affective disorder at four latitudes. Psych. Res.
31:131. '

“Rosenthal, N. E. and M. M. Heffernan. 1986. Bulimia, carbohydrate craving,
and depression: A central connection? In: R. J. Wurtman and J. J.
Wurtman (Ed.) Nutrition and the Brain. pp. 139-166. Raven Press,
New York.

”Rosenthal, N. E., M. J. Genhart, B. Caballero, F. M. Jacobsen, R. G. Skwerer,
R. D. Coursey, S. Rogers, B. J. Spring. 1989. Psychobiological effects of
carbohydrate- and protein-rich meals in patients ' with seasonal
affective disorder and normal controls. Biol. Psych. 25:1029.

V'Rosenthal, N. E., M. Genhart, F. M. Jacobsen, R. G. Skwerer, T. A. Wehr. '
1987 a. Disturbances of appetite and weight regulation in seasonal
affective disorder. Ann. New York Acad. Sci. 499:216.

0 Rosenthal, N. E., F. M. Jacobsen, D. A Sack, J. Arendt, S. P. James, B. L.
Parry, T. A Wehr, 1988a. Atenolol in seasonal affective disorder: A test
of the melatonin hypothesis. Am. J. Psych. 145:52.

Rosenthal, N. E., A A. Levendosky, R. G. Skwerer, J. R. J oseph-Vanderpool,
K. A. Kelly, T. Hardin, S. Kasper, P. DellaBella, T. A. Wehr. 1990.
Effects of light treatment on core body temperature in seasonal
aﬂ‘ective disorder. Biol. Psych. 27:39.

cRosenthal, N. E., D. A. Sack, C. J. Carpenter, B. L. Parry, W. B. Mendelson,
T. A. Wehr. 1985. Antidepressant effects of light in seasonal affective
disorder. Am. J. Psych. 142:163.

“Rosenthal, N. E., D. A. Sack, J. C. Gillin, A. J. Lewy, F. K. Goodwin, Y.
Davenport, P. S. Mueller, D. A Newsome, T. A. Wehr. 1984. Seasonal

affective disorder: A description of the syndrome and preliminary
ﬁndings with light. Arch. Gen. Psych. 41:72.

Rosenthal, N. E., D. A. Sack, F. M. Jacobsen, S. P. James, B. L. Parry, J.
Arendt, L. Tamarkin, T. A Wehr. 1986. Melatonin in seasonal affective
disorder and phototherapy. J. Neural Transm. (Suppl) 21:257. .

163

Rosenthal, N. E., D. A. Sack, R.G. Skwerer, F. M. Jacobsen, T. A. Wehr.
1988b. Phototherapy for seasonal affective disorder. J. Biol. Rhythms
3:101. ‘

Rosenthal, N. E., D. A. Sack, T. A. Wehr. 1983. Seasonal variations in
affective disorders. In: T. A Wehr and F. K. Goodwin (Ed.) Circadian
ghythms in Psychiatry. pp. 185-201. The Boxwood Press, Paciﬁc

rove.

Rosenthal, N. E., R. G. Skwerer, D. A Sack, C. C. Duncan, F. M. Jacobsen, L.
Tamarkin, T. A. Wehr. 1987b. Biological effects of morning-plus-
evening bright light treatment of seasonal affective disorder.
Psychopharmacol. Bull. 23:364.

Ross, M., J. M. Olson. 1981. An expectancy-attribution model of the effects of
placebos. Psychol. Rev. 88:408.

Sack, R. L., A. J. Lewy, D. M. White, C. M.‘ Singer, M. J. Fireman, R.
Vandiver. 1990. Morning versus evening light treatment for winter
depression. Arch. Gen. Psych. 47:343.

Sadun, A. A, J. D. Schaechter, L. E. H. Smith. 1984. A retinohypothalamic
pathway in humans - light mediation of circadian rhythms. Brain Res.
302:37 1.

Salinas, E., C. Hzakim-Kreis, M. Piketty, R. Dardennes, C. Musa. 1992.
Hypersecretion of melatonin following diurnal exposure to bright light
in seasonal affective disorder: Preliminary results. Biol. Psych. 32:387.

Schwitzer, J ., C. Neudorfer, H. Belcha, W. W. Fleischhacker. 1990. Mania as
a side effect of phototherapy. Biol. Psych. 28:532.

Shanahan, T. L., C. A. Czeisler. 1991. Light exposure induces equivalent
phase shifts of the endogenous circadian rhythms of circulating plasma
melatonin and core body temperature in men. J. Clin. Endocrinol.
Metab. 73:227.

Sharma, M., J. Palacios-Bois, G. Schwartz, H. Iskandar, M. Thakur, R.
Quirion, N. P. V. Nari. 1989. Circadian rhythms of melatonin and
cortisol in aging. Biol. Psych. 25:305.

Shifrine, M., A. Garsd, L. S. Rosenblatt. 1982. Seasonal variation in
immunity of humans. J. Interdiscipl. Cycle Res. 13:157.

Sisson, T. R. C. 1976. Visible light therapy of neonatal hyperbilirubinemia.
In: K. C. Smith (Ed.) Photochemical and Photobiological Revliews, vol
1. pp. 241-268. Plenum Press, New York.

164

Skwerer, R. G., F. M. Jacobsen, C. C. Duncan, K A. Kelly, D. A Sack, L.
Tamarkin, P. A Gaist, S. Kasper, N. E. Rosenthal. 1988. Neurobiology
of seasonal affective disorder and phototherapy. J. Biol. Rhythms
3:135.

Sliney, D. and M. Wolbarsht. 1980. Safety with Lasers and Other Optical
Sources. pp. 325-343. Plenum Press, New York.

Souetre, E. 1988. Twenty- four hour proﬁles of body temperature and plasma
TSH in bipolar patients during depression and during remission and 1n
normal control subjects. Am. J. Psych. 145: 1122.

Spitzer, R. L., J. Endicott, E. Robins. 1978. Research diagnostic criteria.
Arch. Gen. Psych. 35:773.

Spitzer, R. L. and J. B. W. Williams. 1989. The validity of seasonal aﬁ'ective
disorder. In: N. E. Rosenthal and M. C. Blehar (Ed.) Seasonal Affective
Disorders and Phototherapy. pp. 79-84. Guilford Press, New York.

Stewart, J. W., F. M. Quitkin, M. Terman, J. S. Terman. 1990. Is seasonal
affective disorder a variant of atypical depression? Differential
response to light therapy. Psych. Res. 33:121.

Stewart, K., J. Gaddy, B. Byme, S. Miller, G. Brainard. 1991. Effects of
green or white light for treatment of seasonal depression. Psych. Res.
38:261.

Stinson, D. and C. Thompson. 1990. Clinical experience with phototherapy. J.
Affective Disord. 18:129.

Stryd, R. P., T. J. Gilbertson, M. N. Brunden. 1979. A seasonal variation
study of 25-hydroxyvitamin D serum levels in normal humans. J. Clin.
Endocrinol. Metab. 48:771.

Svaasand, L. 0., R. Ellingsen. 1983. Optical properties of human brain.
Photochem. Photobiol. 38:293. .

Szadoczky, E., A Falus, M. Arato, A. Nemeth, G. Teszeri, E. Moussong-

Kovacs. 1989. Phototherapy increases platelet 3H-imipramine binding
in patients with winter depression. J. Affective Disord. 16:121.

Szadoczky, E., A Falus, A. Nemeth, G. Teszeri, E. Moussong-Kovacs. 1991.
Effect of phototherapy on 3H-imipramine binding sites in patients with
SAD, non-SAD and in healthy controls. J. Affective Disord. 22:179.

Terman, M. 1988. On the question of mechanism in phototherapy for seasonal

affective disorder: Considerations of efﬁcacy and epidemiology. J. Biol.
Rhythms 3:155.

“,

165

Terman, M. and D. S. Schlager. 1990. Twilight therapeutics, winter
depression, melatonin, and sleep. In: J. Montplaisir and R. Godbout
(Ed.) Sleep and Biological Rhythms. pp.113-128. Oxford University
Press, New York.

Terman, M. and J. Terman. 1991. Seasonal variation in detection of dim light
by SAD patients and normals. Presented at the annual meeting of the
Soc. Light Treatment Biol. Rhythms, Toronto.

Terman, M., C. E. Reme, B. Rafferty, P. F. Gallin, J. S. Terman. 1990b.
Bright light therapy for winter depression: Potential ocular effects and
therapeutic implications. Photochem. Photobiol. 5 1:781.

Terman, M., D. Schlager, S. Fairhurst, B. Perlman. 1989a. Dawn and dusk
simulation as a therpeutic intervention. Biol. Psych. 25:966.

Terman, M., J. S. Terman, L. Amire. 1992. Light-refractory vs.light-
responsive SAD patients: Depression scales predictors. Presented at
the annual meeting of the Soc. Light Treatment Biol. Rhythms.
Washington, D. C.

Terman, M ., J. S. Terman, B. Raﬁ'erty. 1990. Experimental design and
measure of success in the treatment of winter depression by bright
light. Psychopharmacol. Bull. 26:505.

Terman, M., J. S. Terman, F. M. Quitkin, T. B. Cooper, E. S. Lo, J. M.
Gorman, J. W. Stewart, P. J. McGrath. 1988. Response of the
melatonin cycle to phototherapy for seasonal affective disorder. J.
Neural Transm. 72:147.

Terman, M., J. S. Terman, F. M. Quitkin, P. J. McGrath, J. W. Stewart, B.
Rafferty. 1989b. Light therapy for seasonal affective disorder: A review
of efﬁcacy. N europsychopharmacology 2: 1.

Terman, M., J. S. Terman, F. M. Quitkin, J. W. Stewart, P. J. McGrath, E. V.
Nunes, S. G. Wager, E. Tricamo. 1989c. Dosing dimensions of light
therapy: Duration and time of day. In: C. Thompson and T.
Silverstone (Ed.) Seasonal Affective Disorder. pp. 187-204. CNS
Publishers, London.

Terman, J. S., M. Terman, D. Schlager, B. Rafferty, M. Rosofsky, M. J. Link,
F. M. Quitkin. 1990a. Efﬁcacy of brief, intense light exposure for
treatment of winter depression. Psychopharmacol. Bull. 26:3.

Thase, M. E. 1989. Comparison between seasonal affective and other forms
of recurrent depression. In: N. E. Rosenthal and M. C. Blehar (Ed.)
Seasonal Affective Disorders and Phototherapy. pp.64-78. Guilford
Press, New York.

166

Thompson, C., D. Stinson, A. Smith. 1990. Seasonal affective disorder and
season-dependent abnormalities of melatonin suppression by light.
Lancet 336:703.

Thorington, L. 1985. Spectral, irradiant and temporal aspects of natural and
artiﬁcial light. Ann. New York Acad. Sci. 453:28.

USPHS CDC NIH. 1988. Biosafety in microbiological and biomedical
laboratories. pp. 11-13.

Uhlenhuth, E. H., A. Canter, J. O. Neustadt, H. E. Payson. 1959. The
symptomatic relief of anxiety with meprobamate phenobarbital and
placebo. Psychiatry 115:905.

Vaughan, G. M., S. D. McDonald, R. M. Jordan, J. P. Allen, R. Bell, E. A
Stevens. 197 9. Melatonin, pituitary function and stress in humans.
Psychoneuroendocrinology 4:351.

Waldhauser, F. and M. Dietzel. 1985. Daily and annual rhythms in human
melatonin secretion. Ann. New York Acad. Sci. 453:205.

Wan, S., J. A. Parrish, R. R. Anderson, M. Madden. 1981. Transmittance of
nonionizing in human tissues. Photochem. Photobiol. 34:679.

Weale, R. 1988. Light on seasonal affective disorders? Br. Med. J. 296:359.

Webley, G. E. and F. A. Leidenberger. 1986. The circadian pattern of
melatonin and its positive relationship with progesterone in women. J.
Clin. Endocrinol. Metab. 63:323.

Webley, G., H. Mehl, K. Willey. 1985. Validation of a sensitive direct assay for
melatonin for investigation of circadian rhythms in different species. J.
Endocrinol. 106:387.

Wehr, T. A. and N. E. Rosenthal. 1989. Seasonality and affective illness. Am.
J. Psych. 146:829.

Wehr, T. A, F. M. Jacobsen, D. A. Sack, J. Arendt, L. Tamarkin, N.
Rosenthal. 1986. Phototherapy of seasonal affective disorder: Time of

day and suppression of melatonin are not critical for anti-depressant
effect. Arch. Gen. Psych. 43:870.

Wehr, T. A, R. G. Skwerer, F. M. Jacobsen, D. A Sack, N. E. Rosenthal.
1987. Eye versus skin phototherapy of seasonal affective disorder. Am.
J. Psych. 144:753.

167

Weitzman, E. D., A. S. deGraaf, J. F. Sassin, T. Hanse, O. B. Godtlibsen, M.
Perlow, L. Hellman. 1975. Seasonal patterns of sleep stages and
secretion of cortisol and growth hormone during 24 hour periods in
northern Norway. Acta Endocrinol. 78:65.

Wever, R. A. 1986. Characteristics of circadian rhythms in human functions.
J. Neural Transm. 21(suppl)323.

Williams, J ., M. Link, N. Rosenthal, M. Terman. 1988. Structured Interview
Guide for the Hamilton Depression Scale-Seasonal Affective Disorders
Version (SIGH-SAD). pp. 1-10. New York State Psychiatric Institute,
New York.

Winton, F. and S. Checkley. 1989. Clinical characteristics of patients with
seasonal affective disorder. In: C. Thompson and T. Silverstone (Ed.)
Seasonal Affective Disorder. pp. 59-68. CNS Publishers, London.

Winton, F., T. Corn, L. W. Huson, C. Franey, J. Arendt, S. A. Checkley. 1989.
Effects of light treatment upon mood and melatonin in patients with
seasonal affective disorder. Psychol. Med. 19:585.

Wirz-Justice, A. 1992. Light therapy and circadian rhythms circa 1992. Paper
presented at the annual meeting of the Soc. Light Treatment Biol.
Rhythms, Washington, D. C.

Wirz-Justice, A. and R. Richter R. 1979. Seasonality in biochemical
determinations: A source of variance and a clue to the temporal
incidence of affective illness. Psych. Res. 1:53.

Wirz-Justice, A. and T. A. Wehr. 1983. NeuropsychopharmaCology and
biological rhythms. Adv. Biol. Psych. 11:20.

Wirz-Justice, A., C. Bucheli, P. Graw, P. Kielholz, H. U. Fisch, B. Woggon.
1986. Light treatment of seasonal affective disorder. Acta Psychiatr.
Scand. 74:193.

Wirz-Justice, A., P. Graw, C. Bucheli, A. C. Schmid, B. Gisin, C. Jochum, W.
Poldinger. 1989. Seasonal affective disorder in Switzerland: A clinical
perspective. In: C. Thompson and T. Silverstone (Ed.) Seasonal
Affective Disorder. pp. 69-76. CNS Publishers, London.

Wirz-Justice, A., A. C. Schmid, P. Graw, K. Krauchi, P. Kielholz, W.
Poldinger, H. U. Fisch, C. Buddenberg. 1987. Dose relationships of
morning white light in seasonal affective disorder (SAD). Experientia
43:574.

Wirz-Justice, A., R. A. Wever, J. Aschoff. 1984. Seasonality in freerunning
circadian rhythms in man. Naturwissenschaﬁen 71:316. '

 

168

Wright, J ., M. Aldous, C. Franey, J. English, J. Arendt. 1986. The effects of
exogenous melatonin on endocrine function in man. Clin. Endocrinol.
24:375.

Wurtman, R. J ., J. Axelrod, J. E. Fischer. 1964. Melatonin synthesis in the
pineal gland: Effect of light mediated by the sympathetic nervous
system. Science 143:1328.

Wurtman, R. J ., J. J. Wurtman. 1989. Carbohydrates and depression. Sci.
Am. 253:68.

Wurtman, R. J ., J. Axelrod, L. S. Phillips. 1963. Melatonin synthesis in the
pineal gland: Control by light. Science 142:1071.

Yerevanian, B I., J. L. Anderson, L. J. Grota, M. I. Bray. 1986. Effects of
bright incandescent light on seasonal and non-seasonal major
depressive disorder. Psych. Res. 18:355.

Young, M. A., L. G. Watel, H. W. Lahmeyer, C. I. Eastman. 1991. The
temporal onset of individual symptoms in winter depression:
differentiating underlying mechanisms. J. Aﬂ'ective Disord. 22:191.

Yurugi, R., T. Sasaki, M. Yoshimura. 1972. Seasonal variation of basal
metabolism in Japanese. In: S. Itoh, K. Ogata, H. Yoshimura (Ed.)
Advances in Climatic Physiology. pp. 395-410. Igaku Shoin Ltd., Tokyo.

Zucker, I. 1988. Seasonal affective disorders: Animal models non ﬁngo. J.
Biol. Rhythms 3:209.

 

"‘ﬁllllﬂllllllliilliES