Bright Light Treatment: Are We Keeping Our Subjects in the Dark?

Transcription

1 , Sleep 13(3): , Raven Press, Ltd., New York 1990 Association of Professional Sleep Societies, ' Technical Note Bright Light Treatment: Are We Keeping Our Subjects in the Dark? Drew Dawson and Scott S. Campbell Institute for Circadian Physiology, Boston, Massachusetts, U.S.A. Summary: Bright light treatments for seasonal affective disorder (SAD) and phase-disturbed circadian rhythms show considerable interindividual variation in response, Many studies have explained this variation in terms oftraditional explanations of response variance, i.e., poor diagnostic criteria, poor compliance, etc. This study examined proximity and gaze behavior during illumination as a source of variance in response. Distance from the light source and gaze were varied systematically in order to quantify the discrepancy between perceived and transmitted light. The results indicate that the discrepancy between perceived and transmitted light increases rapidly as a function of distance and gaze. At best, subjects exposed to a transmitted illuminance of 10,000 lux perceived only 20% of the transmitted light, and this dropped rapidly to between I and 2% when subjects were more than 3 feet away and not gazing directly at the light source. These results suggest that light treatment protocols may be compromised by lack of attention to compliance and administration factors. These issues may account for much of the response variance observed previously. Key Words: Bright light-seasonal affective disorder-circadian phase disturbance-gaze behavior-response variance. In recent years there has been increasing interest in the use of bright light as a treatment strategy for seasonal affective disorder (SAD) and other disorders presumed to be associated with circadian phase disturbance. While nearly all studies have found a significant improvement in the clinical status for groups of SAD patients treated with bright light (1), there has been considerable interindividual variation in response to bright light treatment (2). Several factors, including variations in diagnostic criteria, variations in presenting severity, and lack of compliance measures have been put forward to explain the inconsistency in results (2,3). More recently, there has been an increasing focus on the possibility that the antidepressive response to bright light treatment is merely a sophisticated placebo or expectation response (2,4). Interestingly, Accepted for publication November Address correspondence and reprint requests to Dr. Scott S. Campbell at Institute for Circadian Physiology, 677 Beacon St., Boston, MA , U.S.A. 267

2 268 D. DAWSON AND S. S. CAMPBELL there has been little attempt to look at treatment protocols themselves as sources of response variability. Examination of published reports indicates that there is little consistency in light administration protocols across laboratories. While most centers report the same number of treatments, there is little consistency in the spatial arrangements for the administration of light or in monitoring behavior during irradiation. The degree of discrepancy between transmitted illumination and that which is actually perceived probably reflects the combined effects of a number of variables. For any specific source of illuminance the distance between the light source and the subject is of primary importance. Light intensity drops as an inverse square of the distance between the light source and observer. Perceived illumination is further influenced by the behavior of subjects during light exposure. Subjects instructed to view the lights continuously will receive considerably more lux-hours than those permitted to engage in a wide range of behaviors in a room with a single focused point source of light. While it is clear that these differences influence the actual amount of perceived light, no studies have attempted to quantify the effect of these variables in bright light treatment protocols. This study examined how two of these factors, gaze behavior and physical distance from the light source, influence light perception. By constructing illuminance maps for the area around a "bright light workstation," it was possible to determine the relationship between the subject's physical position in space, the direction of gaze, and the corresponding level of perceived illumination. METHODS The workstation was a simulated control room panel with several computer monitors. The workstation subtended an arc of on either side of a central monitor (see Fig. 1a). A commercially available bright light source, nominally rated at 10,000 lux (Ultrabright 10,000; Medic-Lite, Inc., Lake Hopatcong, NJ, U.S.A.) was used as the lighting source. Illuminance was measured using a United Detector Power meter (model 351) equipped with a photosensor (model 263), photometric filter (model 1157), and diffusing screen (model 151). (a) (b) FIG. 1. (a) Schematic representation of the workstation viewed from floor level. (b) Schematic representation of the workstation viewed from above. Distance on the grid is in feet. Sleep, Vol. ]3, No.3, 1990

3 BRIGHT LIGHT TREATMENT 269 Illuminance maps were constructed from readings taken on a 6 x 6 foot grid placed in front of the work station. To examine the effects of behavior and gaze, measures were taken at I-foot intervals in two dimensions in each of three conditions (see Fig. 1b): (a) a direct light condition (DC) in which the sensor was pointed directly at the center of th~ light source from each grid point, (b) a horizontal gaze condition (HC) in which the sensor was pointed at the center of the work station, parallel to the floor, and (c) a task condition (TC) in which the sensor was pointed down at the same angle as a subject would read a book or write on a desk. When the subject was in the position specified as "operator" in Fig. 1b, the light source was at a distance of 3 feet and an angle of 45 from the vertical so that the angle of incidence between the light source and the photosensor in the DC, HC, and TC conditions was 0, 45, and 90 respectively. RESULTS The rated output of 10,000 lux was obtained when the light meter was pointed directly at the light source, in from the center of the plastic diffusion screen. Fig. 2a-c shows a three-dimensional representation of the percentage illuminance of rated output in each of the three conditions (DC, HC, TC) at each of the 49 grid points. Percentage illuminance contours have been smoothed and interpolated from the grid points and plotted as a continuous topographic surface. (a) (b),soil 2300 """!$O;) lc 00 seo 000 (c) 50Q,')S 00 (d) FIG. 2. (a) Perceived illuminance in the DC condition as a percentage of rated output illumination in two dimensions. (b) Perceived illuminance in the HC condition as a percentage of rated output illumination in two dimensions. (c) Perceived illuminance in the TC condition as a percentage of rated output illumination in two dimensions. (d) Percentage difference in perceived illuminance between the DC and TC conditions as a function of distance in two dimensions.

4 270 D. DA WSON AND S. S. CAMPBELL The results indicate that the effective perceived level of illumination is considerably lower than the rated output. For a subject seated in front and below the light source [grid position (3,6)] in the DC the perceived illuminance is 20% ofthe rated output at 12 inches. As the subject moves away from the central position, illuminance drops rapidly to 5-10% of rated output. In the TC condition the maximum effective level of illumination was ~4% of rated output at 12 inches and dropped rapidly to between 1 and 2% of rated output. The HC (Fig. 2b) condition produced a result between the DC and TC conditions with maximum output at around 13% dropping rapidly to <6%. Fig. 2d shows the percentage difference between the DC and TC conditions as a function of distance in two dimensions. This measure is an indication of the drop in perceived illumination when a subject is performing a typically permitted activity (i.e., reading or writing at a desk) during the period of illumination. Performing a task produces % reduction in perceived illuminance when the subject is close to the light, but only 45-55% reduction when the subject is at the maximum distance. DISCUSSION These results emphasize the importance of treatment compliance in the use of bright light treatment. Furthermore, they suggest that interindividual and interstudy differences in compliance may contribute significantly to the variability of patient response. While there is no doubt that poor subject selection criteria, placebo, or expectation effects have contributed to the response variability of bright light therapy, treatment differences should also be considered as a major cause of response variability. It would appear that the output rating of commercial lights should not be taken as a guide to the effective levels of perceived illumination. In those studies where behavior and gaze have not been controlled it is reasonable to assume that effective perceived illumination'is between 5 and 15% of the reported level of illumination. When subjects are required to sit in close proximity to the light source, to ensure the required illuminance, then they should be instructed to control gaze as much as possible since the difference between TC and DC is greatest at close range. If the desired level can be achieved at a greater distance (either by multiple light sources and/or greater diffusion), then stringent gaze control is less critical. Clearly, this study carries implications for the interpretation of previous research. This study shows that many treatment protocols may produce levels of illumination little greater than room level. Many of the treatment failures in SAD patients may be due to poor compliance rather than poor diagnostic or response criteria. Research into the phase-shifting response to bright light is also likely to have been compromised. Discrepancies in the phase shifts produced by similar protocols may reflect nothing more than differences in compliance. Recently derived phase-response curves (4-6) should be interpreted with respect to the compliance inherent within their respective protocols. Future studies should attempt to control the effective level of perceived illumination as much as possible. Ideally, published reports should indicate spatial variations in illuminance produced by specific protocols, and methods used to measure or ensure compliance. Where experimental manipulation of compliance is not feasible, individual variations in compliance could be monitored with some of the recently developed ambulatory illumination recording devices. A combination of illuminance mapping, ambulatory monitoring, and behavior control will improve treatment compliance and enhance our understanding of light perception and its role in circadian physiology.

5 , BRIGHT LIGHT TREATMENT 271 Acknowledgment: This work was supported by grants from the Lighting Research Institute (No. 88:SP LREF:4), the American Health Assistance Foundation (No ), and the National Institute of Mental Health (I-ROI-MH l). The assistance of J. Peters and K. Buckley is gratefully acknowledged. ' REFERENCES l. Rosenthal NE, Sack DA, Skwerer RG, Jacobsen FM, Wehr TA. Phototherapy for seasonal affective disorder. J Bioi Rhythms 1988;3: Terman M, Terman JS, Quitkin FM, McGrath PJ, Stewart JW, Rafferty AB. Light therapy for seasonal affective disorder: a review of efficacy. Neuropsychopharmacology 1989;2: Terman M. Editorial: developing the case for efficacy. SLTBR Newsletter 1989;1: Eastman C. Is bright light treatment a "brilliant" placebo? Presented at the 3rd annual conference of Professional Sleep Societies, Washington, DC, CzeislerCA, Kronauer RE, Allan JS, Duffy JF, Jewett ME, Brown EN, RondaJM. Bright light induction of strong (type 0) resetting of the human circadian pacemaker. Science 1989;244: Honma K, Honma S, Wada T. Phase dependent shift of free-running human circadian rhythm in response to a single pulse of light. Experientia 1987;43: Dawson D, Morris MM, Lack LC. Average phase response curve in humans for a 4h exposure to evening light. Sleep Res 1989;18:413.