Light and the photoreceptors of the eye

The visual sensory system is comprised of image-forming photoreceptor cells of rods (for low-level light) and cones (for sharpness detail and colour vision). When a photon of light reaches the rods or cones it acts on light-sensitive pigments (photopigments called rhodopsins), and this generates a sequence of chemical events that create electrical impulses in the optic nerve, which converge in the visual cortex and are interpreted as vision (Castro Reference Castro, Angus and Rosengart2011). However, it is now known that there is a third type of retinal photoreceptor in mammals called intrinsically photosensitive retinal ganglion cells (ipRGCs). The ipRGCs use the photo pigment melopsin to absorb light passing through the retina and generate electrical impulses that are transmitted via neurons to the suprachiasmatic nucleus (SCN) of the anterior hypothalamus (Lazzerini Ospri Reference Lazzerini Ospri, Prusky and Hattar2017; Fernandez Reference Fernandez, Fogerson and Lazzerini Ospri2018).

Research demonstrates that ipRCGs (which are independent of the rods and cones) mediate non-visual effects of light on a range of biological processes. The ipRGCs are particularly sensitive to the absorption of short-wavelength (blue) light but transmit information about all forms of the visible light spectrum through the retinohypothalamic tract to the SCN (Lazzerini Ospri Reference Lazzerini Ospri, Prusky and Hattar2017).

The circadian pacemaker and its connections

The SCN is often referred to as the circadian pacemaker (or ‘master clock’) because of its critical role in controlling and regulating circadian rhythms in mammals (Czeisler Reference Czeisler1999; Reppert Reference Reppert and Weaver2002). The SCN contains ~20 000 neurons that express a near-24-hour rhythm of electrical activity. Although the SCN cells can generate their own synchronised oscillations, they can be entrained (reset) by photic input (via ipRCGs) and non-photic input (from other brain regions). Hastings and colleagues (Reference Hastings, Maywood and Brancaccio2019) explain that the efferent neural and humoral signals that are generated from the SCN play a critical role in coordinating the activity of circadian oscillators located in many different target organs and peripheral tissues of the body. Thus, the SCN maintains the internal temporal order underpinning physiological adaptation to the solar cycle (Hastings Reference Hastings, Reddy and Maywood2003). Importantly, SCN projections to the pineal gland modulate secretion of the hormone melatonin. Melatonin carries ‘time’ information, for example about the time of the day (a ‘clock’ function) and time of the year (a ‘calendar’ function), from the brain to all parts of the body (Benarroch Reference Benarroch2008). Melatonin secretion begins at dark and the level of melatonin in the bloodstream plays an essential role in regulating the sleep–wake cycle (Foster Reference Foster and Wulff2005). Finally, connections between ipRGCs and the limbic system, striatum and the brain-stem may affect mood state, while connections between the thalamus, SCN, retina, visual and other cortical/subcortical areas suggest that the thalamus acts as a major interface between alertness and cognition (Fernandez Reference Fernandez, Fogerson and Lazzerini Ospri2018). The effects of light on these systems and the subsequent impacts may be of specific interest to psychiatry research and practice (Scott Reference Scott, Langsrud and Vethe2019).

Light, the SCN and the sleep–wake cycle

Although the sleep–wake cycle is the most obvious example of a circadian rhythm in humans, many other physiological processes and behaviours show rhythmicity, including body temperature, release of hormones, metabolism and gene expression (Hastings Reference Hastings, Maywood and Brancaccio2019). Several environmental stimuli can entrain the SCN (and the sleep–wake cycle), but light is the most critical factor or zeitgeber (Foster Reference Foster and Wulff2005). The role of light in resetting the phase of the SCN is partly a function of the timing, duration, intensity and wavelength of the exposure (which can range from regular light–dark cycles to seasonal ‘photo-periods’) (Czeisler Reference Czeisler1999; Reppert Reference Reppert and Weaver2002; Khalsa Reference Khalsa, Jewett and Cajochen2003). Bonmati-Carrion and colleagues (Reference Bonmati-Carrion, Arguelles-Prieto and Martinez-Madrid2014) comment that, in high-income countries, nights are excessively illuminated whereas, in contrast, more daytime hours are spent indoors. So, daytime light exposure equates to a much lower lux than under natural conditions, but there is excess light exposure during the evening and night-time, which leads to a reduced zeitgeber strength and a disorganisation of the circadian system (Box 3). Even a small change in the timing of the light exposure (e.g. 20 min of low-level artificial light at night) can cause perturbations in melatonin production and lead to a shift in the phase of circadian rhythms (Castro Reference Castro, Angus and Rosengart2011). Furthermore, the impact will be observable in the next circadian cycle (with changes in sleep–wake patterns detectable in the following 24–36 h), whereas the effects on circadian physiology (such as temperature or melatonin levels) can be measured objectively immediately after the aberrant light exposure or even during the actual exposure (Wulff Reference Wulff, Gatti and Wettstein2010).

Many biological processes show that maximal sensitivity to short-wavelength light and exposure to blue light during the daytime is important for maintaining well-being, alertness and cognitive performance during the day (Wahl Reference Wahl, Engelhardt and Schaupp2019). In contrast, chronic exposure to blue light (even as part of traditional ‘white’ light) directly before bedtime may affect sleep quality and circadian phase and/or cycle duration (Wahl Reference Wahl, Engelhardt and Schaupp2019). Interestingly, research indicates that the absence of blue light (so-called virtual darkness) is equivalent to actual darkness from the perspective of the human brain and synchronisation of circadian rhythms (Castro Reference Castro, Angus and Rosengart2011; Scott Reference Scott, Langsrud and Vethe2019).

Brief comments about sleep–wake cycle disruptions and acute illness episodes

Craig & Mathieu (Reference Craig and Mathieu2018) highlight that any acute medical illness disrupts circadian rhythms, and evidence demonstrates associations with subsequent increases in morbidity and mortality. However, sleep–wake cycle disruptions are equally, if not more, apparent in mental disorders (Foster Reference Foster and Wulff2005). For instance, sleep–wake cycle disruptions are criterion symptoms for diagnosing unipolar and bipolar disorders; sleep abnormalities may be prodromal symptoms heralding the onset of mood, psychotic and/or other major mental disorders; and emerging data identify that circadian dysrhythmias may exacerbate e.g. suicidal behaviours (Rumble Reference Rumble, Dickson and McCall2018; Scott Reference Scott, Langsrud and Vethe2019). Also, research by our own group demonstrates that day-to-day variability in sleep–wake cycles is associated with frequency of aggression or violent acts in transdiagnostic populations admitted to psychiatric intensive care units (Langsrud Reference Langsrud, Vaaler and Kallestad2016, Reference Langsrud, Kallestad and Vaaler2018).

Overall, experimental and clinical research has demonstrated a reciprocal relationship between stabilisation of sleep–wake/circadian disruptions and improvement in mental state across a range of mental disorders. These and other observations have promoted research into a range of chronotherapies that act on the circadian rhythm and/or regulation of the sleep–wake cycle in order to achieve a therapeutic effect on the mental state. As noted in Box 4, studies have examined the utility of melatonin agonists as an adjunctive pharmacological treatment in several mental disorders, while other research has explored, for example, in-patient provision of sleep deprivation (also called wake therapy). Comprehensive reviews are available on these topics (e.g. Parry Reference Parry and Maurer2003; Coogan Reference Coogan and Thome2011; Gottlieb Reference Gottlieb, Benedetti and Geoffroy2019), so we do not repeat the findings here, but we now focus specifically on the influence of light exposure on physical and mental disorders. We then explore the clinical utility of interventions that modify exposure to light (and in turn regulate sleep–wake and circadian rhythms) with particular reference to in-patient settings.

Sunny versus sunless wards

Although the negative impact of some types of light exposure has been well-publicised, a smaller body of research explores the potential health benefits of optimising light exposure in community and clinical settings. For example, a review by Park et al (Reference Park, Chai and Lee2018) suggested that full-spectrum light exposure in hospital in-patients can be associated with increased levels of patient satisfaction, decreased levels of distress and improved physical capacity (by decreasing heart rate and systolic blood pressure, and increasing oxygen uptake). Such findings support evidence from earlier observational or cohort studies of medical patients, with several assessments of the impact of exposure to natural light on the clinical outcomes of patients admitted to intensive care units (ICU). In one study, Beauchemin & Hays (Reference Beauchemin and Hays1998) noticed that beds in a cardiac intensive care unit (CICU) in Alberta, Canada, were oriented in such a way that four beds were north facing and virtually sunless (light exposure at 09.45 h registering 200–400 lux) and four were south facing and mainly bright (light exposure at 09.45 h registering 1200–1300 lux). They reviewed CICU admissions over 4 years and identified that individuals who experienced an acute myocardial infarct whose hospital beds were in bright rooms had better outcomes in terms of both shorter length of stay (LOS) and significantly lower mortality rates than those in dark rooms (8 v. 13%).

Undertaking randomised controlled trials (RCTs) with very acutely ill patients is problematic (not least, the assessment of capacity and ethics of obtaining informed consent), but some RCTs do exist. For example, an RCT of 89 patients admitted for cervical or lumbar spinal surgery found that the participants allocated to a brightly lit ward reported less distress or pain than those allocated to a dimly lit ward, and the former group needed 22% less analgesic medication per hour, which contributed to a significant reduction in the cost of care and treatment (Walch Reference Walch, Rabin and Day2005). Nearly all the medical in-patient studies comment that the impact of light on the circadian system was an important mediator of the observed clinical outcomes (Scott Reference Scott, Langsrud and Vethe2019).

In psychiatry, theories regarding the origin of seasonal affective disorders, such as the circadian phase-shift hypothesis of winter depression, played an important role in stimulating research into light therapies (Lewy Reference Lewy, Rough and Songer2007). Like general medicine, the initial studies mainly comprised careful clinical observations of cohorts of patients admitted to psychiatric wards with hospital rooms that had different orientations and light exposure (e.g. they were east or west facing). For example, a retrospective study of hospital records by Beauchemin & Hays (Reference Beauchemin and Hays1996) found that the LOS of patients with depression in sunny rooms was reduced by about 2.6 days (15%) compared with depressed patients in sunless rooms (P < 0.05). Open studies also indicated the benefits of bright light in unmedicated in-patients with depression (Wirz-Justice Reference Wirz-Justice, Graw and Roosli1999). However, observational and open studies have largely been overtaken by prospective cohort studies, and by controlled trials and RCTs that actively manage the timing and duration of light exposure and (gradually) have established the minimum intensity of light (lux) required to achieve significant reductions in depressive symptoms.

Timing, duration and type of light

One of the first studies that alluded to the importance of the timing of light exposure was undertaken by Lewy et al (Reference Lewy, Bauer and Cutler1998). Researchers noted that being exposed to bright daylight led to a significant decrease in symptoms within 2 weeks in in-patients with seasonal depressive disorders, whereas it took 4–6 weeks to show the equivalent benefit from antidepressants. However, the rapid improvement in symptoms was mainly associated with exposure to morning light, whereas afternoon or evening/night light had limited effects. Reviews and meta-analyses have identified that treatment effects in depression persist for longer with light illuminations >2000 lux (compared with artificial lighting or exposure <300 lux) and that morning BLT for about 2 h duration is more effective than BLT (of the same lux) for about 30 min (Parry Reference Parry and Maurer2003).

Some research suggests that spectral frequencies of light are associated with differential improvements in depressive symptoms, with inconsistent evidence that green light may be more beneficial than red light. However, the current consensus is that bright white light (often 10 000 lux at eye level and a distance of about 30 cm) for 30 min per day during the early morning (for up to 6 weeks) is likely to be most effective for depressive disorders. Evidence suggests that morning sunlight and/or BLT may be particularly useful for treating bipolar depression and/or reducing hospital LOS (Benedetti Reference Benedetti, Colombo and Barbini2001), but a recent meta-analysis suggests that findings are inconsistent regarding any specific benefit for bipolar compared with unipolar depression (Takeshima Reference Takeshima, Utsumi and Aoki2020). Research is ongoing on the use of BLT for atypical depression, but there is emerging evidence of its utility in treatment-resistant disorders (Kragh Reference Kragh, Larsen and Martiny2018) and its use with adolescent in-patients (Holtmann Reference Holtmann, Mokros and Kirschbaum-Lesch2018; Kirschbaum Reference Kirschbaum, Straub and Gest2018).

Bright-light therapy beyond the mood disorders

The majority of studies of BLT focus on mood disorders, but there are suggestions that benefits may extend to other in-patient populations, such as individuals with eating disorders (although the timing of symptom response and persistence of treatment effects remain unknown) (Beauchamp Reference Beauchamp and Lundgren2016). Studies of BLT in psychotic disorders are scarce (Bellivier Reference Bellivier, Geoffroy and Etain2015), with most research focused on schizoaffective or puerperal disorders (Oren Reference Oren, Cubells and Litsch2001; van Ravesteyn Reference van Ravesteyn, Lambregtse-van den Berg and Hoogendijk2017). As evidence indicates that in-patients with schizophrenia experience significant sleep/circadian disruptions (~50% show severe circadian misalignment such as phase shifts, non-24-hour sleep–wake and melatonin cycles, etc.), it appears that intervention trials could be justified (e.g. Langsrud Reference Langsrud, Vaaler and Kallestad2016; Scott Reference Scott, Langsrud and Vethe2019). Also, this research will allow examination of whether the therapeutic effect of light in mental disorders is mediated solely through an effect on the circadian system or, as suggested by some research, whether there is a direct effect of light on mental health (Fernandez Reference Fernandez, Fogerson and Lazzerini Ospri2018). As clinicians will be aware, these mediation studies will need to consider the effects of confounding by medications prescribed for in-patients with severe mental disorders (many of which have an impact on neurotransmitters and neuronal pathways linked to the circadian system).

Dark therapies

Difficulties in delivering chronotherapy to acutely ill psychiatric patients is best exemplified by considering dark therapies. The interventions with published evidence of effectiveness are darkness therapy (spending 14 h in darkness per day) or the use of blue-blocking glasses with patients who are admitted to hospital for acute mania (Barbini Reference Barbini, Benedetti and Colombo2005; Henriksen Reference Henriksen, Skrede and Fasmer2016). However, the research protocol required active engagement by the study participants, for example they were asked to repetitively engage in specified behaviours (resting in forced darkness from about 18.00 to 08.00 h) or use particular equipment (such as blue-blocking glasses) at specified times of the day (Scott Reference Scott, Langsrud and Vethe2019). Unsurprisingly, there are concerns about the level of adherence in individuals who are acutely unwell and about imposing personal responsibility on them regarding the timing of exposure to different intensities or spectra of light. Also, given the prevalence of sleep–wake cycle disturbances in in-patients with mental disorders, it is logical to consider the possibility of delivering ‘optimised’ or circadian lighting to transdiagnostic in-patient populations.

‘Circadian lighting’ in in-patient facilities: Scandinavian research and experience

In response to all the above considerations, clinical and research groups in Denmark (e.g. the Rigshospitalet in Copenhagen, Aabenraa and Aarhus University) and Norway (St Olavs Hospital in collaboration with the Norwegian University of Science and Technology in Trondheim) have designed in-patient facilities where changes in light exposure are regulated automatically and where programmable lighting conditions form an integral part of the therapeutic milieu. These new units are exploring the impact of environmental lighting on the sleep and clinical outcomes of psychiatric in-patients, and some are also examining how technological innovations might be employed to monitor any treatment benefits.

The focus on lighting and novel non-invasive methods for monitoring patient sleep represent a significant shift from previous eras (van den Berg Reference van den Berg2005). Historically, acute psychiatric admission units have offered asylum and a place of safety, where ward routines and structured activities help to reduce arousal, regularise sleep–wake cycle patterns and improve self-esteem, and so on. However, the focus now is primarily on in-patient delivery of physical and pharmacological treatments that reduce symptoms and suicidality, enhance social functioning and sufficiently improve the individual's mental state to allow a timely return to out-patient or community care (Scott Reference Scott, Langsrud and Vethe2019). Although recent guidance on the design and structure of new psychiatric units has emphasised the importance of the quality and fabric of the buildings and furnishings (e.g. Department of Health 2013), there is a lack of recognition that ordinary indoor hospital lighting is unsatisfactory for daily living and does not support stabilisation of sleep–wake patterns or circadian rhythms (indeed, it may worsen them). As Feifel (Reference Feifel2008) points out, little or no attention has been given to the creation of a state-of-the-art in-patient milieu in psychiatry.

In laboratory studies, exposure to evening blue-depleted LED modules has been found to be less disruptive to the circadian system, to improve sleep and decrease arousal, although exposure to blue light during the day is important for a range of biological processes. Hence, the goal is not to create an environment that is devoid of blue light over 24 h. Rather, the aim is to create a lighting system that is programmed to produce blue-depleted evening light, which has an amber hue (Fig. 1). Ideally, this might be amalgamated with morning bright light (and/or change in the light spectrum and intensity throughout the unit during the course of the day) as this combined approach can increase zeitgeber strength, increase daytime wake time and strengthen the rest–activity rhythm (supplementary Appendix, available at https://doi.org/10.1192/bja.2020.47). However, some caution is needed regarding increasing the general daytime light intensity in units that regularly admit individuals presenting with acute mania. Furthermore, we await research to determine whether approaches will need further personalisation according to the circadian preference of each individual patient (chronotype).

FIG 1 Blue-depleted evening light in the new in-patient unit in Trondheim, Norway. The LED modules inside the light fittings contain a mix of red, green/white and blue diodes that can be programmed individually to emit different light intensities at different times of the day. To create evening blue-depleted lighting, only the green/white and red diodes emit light, and the blue diode is switched off. Blue-blocking window filters are also deployed in the evening, and all televisions have permanent blue-blocking filters.

As described in the publications from the Danish and Norwegian units, there are several options for introducing ‘circadian lighting’ (into existing or new-build units). In order of cost/complexity, these could include the following.

Implementation of environmental interventions: questions to be resolved

The above developments have predominantly been introduced or are planned for introduction in specialist wards such as ICUs, neurorehabilitation and in-patient affective disorders units. Findings from this research has been encouraging, although a recent small-scale RCT (with a more limited approach to and control of light exposure) showed only modest effects on sleep quality and other parameters (Okkels Reference Okkels, Jensen and Skovshoved2020). Before expecting mainstream psychiatry to consider investing in these lighting (and monitoring) systems, it is necessary as a minimum to address the following clinical questions:

1. (a) Does stabilisation of sleep–wake cycle disturbances hasten symptomatic and/or functional recovery, reduce the use of any classes of medication (such as night sedation) and, critically, is there any impact on hospital LOS (as in-patient treatment is the costliest element of psychiatric care)?

2. (b) Which subgroups of patients benefit from exposure to these new lighting environments and which (if any) are adversely affected?

3. (c) What side-effects do patients experienced (over and above those associated with any disorder-specific treatments)?