Within-subjects ultra-short sleep-wake protocol for characterising circadian variations in retinal function

Prior studies suggest that visual functions undergo time-of-day variations. Under naturalistic entrainment, diurnal changes in physiology may be driven by circadian and/or homeostatic processes, and repeated measurements at different times of day are thus not suitable to draw unambiguous conclusions about circadian effects on visual function. In this study, we disentangle circadian and homeostatic effects on variations of retinal function. We examine the earliest stages of image-forming (temporal contrast sensitivity of the postreceptoral channels) and non-image forming visual functions (pupillary light response) by employing a short forced-desynchrony multiple-naps protocol lasting 40 hours. Participants (n=12, 50% female) will stay in a controlled time-isolating environment under dim light conditions and adhere to an ultra-short sleep-wake cycle, alternating between 2h30m of wake time in dim light and 1h15m hour of sleep in no light. During eleven intervals of wakefulness, participants will undergo psychophysical and pupillometric assessments with silent-substitution stimuli. We hypothesize that the sensitivity of retinal mechanisms undergoes circadian variations. This hypothesis will be investigated by separately determining psychophysical contrast thresholds to parafoveal silent-substitution stimuli targeting the post-receptoral pathways (isoluminant red-green, L-M; isoluminant blue-yellow, S; luminance, L+M+S). We will furthermore measure the pupillary light response to peripheral stimuli (annulus 10{degrees}-30{degrees}) in comparison to the response to stimuli isolating or including melanopsin stimulation. All stimuli will be delivered at constant retinal irradiance using a Maxwellian view system or artificially restricting pupil size. Additionally, we will quantify and report effects of our test stimuli on the circadian system by comparing the dim-light melatonin onset (DLMO) timing during two supplementary evening sessions, comparing dim-light conditions to such with experimental light exposure. Our work informs the fundamental biological mechanisms underlying the influence of light on the human circadian system. Based on our findings, current models about the sensitivity of the circadian system may need to be modified in order to account for the bidirectional influence of circadian function and photoreception.


Processing of light in the retina
Exposure to light enables vision and visual perception, allowing us to see the world in its colourful and spatial detail [1].Additionally, light exposure also influences our physiology and behaviour through circadian synchronization [2], melatonin suppression [3][4][5][6] and modulation of alertness of light [6].
The pathways underlying these distinct functions are also anatomically distinct and separated into image-forming and non-image-forming pathways.Signals in both pathways arise from the three classes of photoreceptors in the retina: rods, cones and the melanopsin-containing intrinsically photosensitive retinal ganglion cells (ipRGCs).
Image-forming signals primarily stem from rods and three types of cones with different spectral sensitivity.Phototransduction takes place at the photoreceptor level, but before the resulting signal is passed on to the retinal ganglion cells and eventually subcortical (lateral geniculate nucleus; LGN) and cortical targets (primary visual cortex; V1), the signal is recombined through retinal wiring.The coding of luminance is based on the summation of L-and M-cone signals, and two opponent channels encode colour: the red-green channel that is driven by the antagonism of L-cone versus M-cone signals (L-M), and the yellow-blue channel driven by the S-cone versus the sum of L-and M-cone signals (S-[L+M]) [7].
On the other hand, the non-image-forming pathway receives inputs from the ipRGCs expressing melanopsin.The effects of light on human circadian and neuroendocrine physiology are orchestrated by the suprachiasmatic nucleus (SCN) in the hypothalamus.Through the retinohypothalamic pathway, which relays information from the retina to the hypothalamus, the SCN encodes light to synchronize the internal circadian rhythm to the external environment, rendering light input an important circadian signal [8].
Across the day, light exposure has different and opposing circadian effects depending on the timing of exposure.Given by the phase response curve (PRC) to light, light exposure in the morning advances the circadian phase while evening light exposure will cause a delay [9][10][11].To what extent the PRC is driven by a modulation of sensitivity in the retina or mechanisms in the SCN is unknown.Studies investigating how the pupillary light reflex, which reflects the response of all photoreceptors to light [12,13], have revealed that there are a time-of-day differences in melanopsin-mediated pupil responses [14,15], indicating that at least some variability originates in the retina.
Only a few studies have documented the diurnal variation of basic visual functions, such as luminance perception, including temporal contrast [16] and spatial contrast sensitivity [17].Some studies have also investigated diurnal variation in colour vision [17,18].Additionally, there is strong evidence for diurnal variations of interocular pressure and corneal thickness, which could lead to differences in retinal function [19,20].
However, the emerging pattern of results is somewhat heterogeneous, and does not factor in that time-of-day variations in retinal and physiology cannot be uniquely attributed to the circadian clock.The goal of this study is to characterise the variability of retinal functions as a function of circadian phase.

Disentangling circadian and homeostatic process
The intercorrelated nature of processes controlled by the circadian clock and behavioural rhythms poses considerable challenges to the investigation of uniquely circadian effects, requiring well-controlled studies spanning many hours to multiple weeks.Briefly, daily variation in human physiology, behaviour, and experience at the 24-hour scale are governed by two independent processes, an endogenous circadian rhythm and a sleep homeostat [21,22].Under normal, entrained conditions, these February 23, 2024 2/23 processes are working in tandem to regulate the propensity for sleep and to create a regular cycles of sleep and wakefulness.
The two-process model postulates that the homeostatic "process S" gradually increases during periods of wakefulness, giving rise to an increasingly stronger drive for sleep the longer one stays awake, and affecting a variety of physiological and cognitive functions.During sleep, sleep pressure decreases exponentially, allowing for recovery and restoring balance before the next sleep-wake cycle.At the same time, the circadian "process C" regulates various physiological functions and signals which in turn influence sleep timing and propensity.Under entrained conditions, the sleep-wake cycle is aligned with the internal body clock and external environmental cues, making both processes correlated with one another [23].
In the laboratory, it is possible to disentangle their effects on neurobehavioural functions by systematically manipulating sleep timing, sleep duration and contextual factors such as ambient light, temperature and food intake.There are various within-person study designs that requiring repeated measurements of outcome parameters and careful monitoring and controlling of circadian signals and homeostatic influences [23][24][25].

Ultra-short sleep-wake forced-desynchrony protocol
Here, we will employ a 40-hour forced-desynchrony, multiple-naps protocol with a night-day/light-dark (LD) cycle totalling 3h45m.In this protocol, sleep pressure is kept at a constant low level.By rapidly alternating between sleep and wake periods (i.e.regular naps).Other homeostatic influences like food intake and activity will be adapted to the ultra-short rhythm.External influences including ambient temperature and dim light during wakefulness are kept constant to minimize the influence on the circadian rhythm.
In the literature, a similar protocol has been successfully implemented before, scheduling sleep and wake time in a 2:1 ratio of imposed day length of 3h45m [26].150 min (2h30m) wake intervals in dim light (< 10 lux) provide opportunity for This study will isolate circadian effects on early stages of visual processing from homeostatic effects with a short forced-desynchrony protocol.Participants will stay in the laboratory for 40 hours, which will cover, on average, 1.5 circadian periods.
They will adhere to an ultra-short sleep-wake rhythm schedule, comprised of 2h30m (150 minutes) of wake time and 1h15m (75 minutes) of sleep opportunity.During each scheduled instance of wakefulness, a series of (psycho-)physical parameters will be measured in order to determine sensitivity of photoreceptor mechanisms, circadian phase, sleep pressure, and other possible physiological outcomes assumed to undergo circadian change.Specifically, we will repeatedly collect data on: (1) psychophysical performance, and more specifically temporal contrast sensitivity (2) PLR to carefully crafted light stimuli, (3) intra-ocular pressure, (4) structural properties of the cornea and retina (5) salivary melatonin concentration, (6) mood, and (7) sleepiness; additionally, there will be continuous measurements of physiological functions (8) temperature in the core body and the proximal-distal gradient of skin temperature, (9) heart rate variability via electrocardiography (ECG), and (10) interstitial glucose concentration.with none of the conditions controlling for melanopsin modulation.

Prospective impact of the study
We will characterise circadian influences on photoreceptor sensitivity and sensitivity of post-receptoral mechanisms, thereby gaining further insights into mechanisms and extent of variations of visual performance.Studying circadian dependency in human visual function will deepen our mechanistic understanding of the link between image-forming and non-image-forming functions.Furthermore, our results have practical implications for professional contexts that involve driving or rely on colour discrimination and thus require high visual performance, and is of particular importance when circumstances impose circadian disruption and sleep disruption, such as in the medicine, aviation or military service.Finally, from a clinical perspective, the localisation of circadian variation in eye physiology may also have implications for the treatment and management of eye conditions such as glaucoma.
February Recruitment modalities and informed consent We will recruit participants using a multi-modal strategy using flyers, posters, mailing lists and other outlets, word-of-mouth and advertisements placed on the internet.The first point of contact will be an online screening.Upon completion of the online screening, their suitability for the study will be determined automatically, followed by an invitation or removal from the participant pool.Participants will undergo a thorough in-laboratory screening described in section 2.4.1.During screening, formal criteria for participation will be checked again with regard to their physical health and psychological stability.Participants will be informed about the study protocol and experimental procedures, and given opportunity to ask questions before giving written consent.

Participant remuneration
Participants will be remunerated for their time.There will be two in-laboratory evening sessions followed by an adaptation night in the lab (more details in section 2.4.2).For each of the evenings, participants will be compensated with €30 plus €20 for the adaptation nights.The second evening session and adaptation night will be immediately followed by the forced-desynchrony protocol starting in the morning.For the forced-desynchrony session of 40 hours, they will receive a compensation of €400.
Participants may leave at any point upon request, or will be excluded in case of reduced compliance to an extent that renders the data deficient.When returning the actigraph, February 23, 2024 5/23 participants will be rewarded €20.As participants are incentivised to provide complete data for actimetry and the sleep diary, they will be rewarded an additional €80 if less than 20% of data is missing.Maximum remuneration is therefore €600.

Sample size
This study will follow a within-person design.Each participant provides eleven repeated measurements.We will determine evidence for our hypotheses using a sequential Bayesian sampling strategy, such that data collection will continue until evidence strength suffices, or until a maximum of twelve participants.This maximum sample size is set by resource limitations.For the sequential sampling strategy, we will use Sequential Bayes Factor (BF) of the statistical model testing circadian modulation of melanopsin-dependent pupil constriction.The stopping criterion is BF > 10 or BF < 1/10 to quantify evidence strength for or against a circadian effect.

Study Design
The study is set up to be a within-participant design.We will invite one participant at a time.Each participant will take part in one evening session a week after starting a circadian stabilisation phase, then come in for another evening session that is followed by an adaptation night and the subsequent completions of the forced-desynchrony protocol, which encompasses 11 measurement blocks and spans 40 hours.Each measurement block encompasses various measurements.The order of measurements is constant between participants.In the pupillometric measurements, experimental February 23, 2024 6/23 conditions will be counterbalanced between participants, while for the psychometric measurements, the order of conditions will be fixed.
Repeating measurement blocks are interspersed with sleep opportunities to prevent sleep deprivation, thus systematically shortening a typical sleep-wake cycle from approximately 24 hours to 3h45m.During one period 24 hours, participants will thus complete 6.4 sleep-wake cycles.Additionally, throughout the entire study period, participants will provide a combination of regular ambulatory measurements and experience sampling starting up to two weeks before the measurements.This will yield longitudinal data spanning up to three weeks.

Study duration & timeline
Each person participating in the study will be enrolling for a total of three weeks, consisting of continuous field measurements and one in-laboratory screening and two separate visits for control and experimental sessions.
After getting in touch with the study team, participants will first complete an online screening via the online platform REDCap.The data from the online questionnaires or experience sampling is submitted to a REDCap server that is set up and maintained by the Chronobiology & Health team at TUM.It is set up as a virtual machine hosted by the Leibniz-Rechenzentrum der Bayerischen Akademie der Wissenschaften.If their answers comply with the inclusion and exclusion criteria, they will be invited for an in-person screening and, upon inclusion in the study, for two in-laboratory visits to take part in experiments.For an overview of the study participation, see Fig. 1.
The online screening will be complemented by an in-person screening by a psychologist to confirm eligibility for the study.Since the forced-desynchrony protocol places participants in an exceptional environment, our inclusion criteria, the screening process and procedures to ensure continued welfare are guided by suggestions from literature [27,28].
Upon inclusion in the study, participants will start a two-week long circadian stabilisation period.They will be invited for a first evening experimental session, the light exposure session, one week into the circadian stabilisation and one week before the forced-desynchrony session.Visual tests will be administered throughout the evening that achieve an experimental light exposure similar to the tasks administered during a wake period of within the forced-desynchrony session.The light exposure session also provides opportunity for the participant to get accustomed to the measurement instruments and prevent novelty or perceptual learning during the main study.After the evening session, participants will stay in the lab for full night of sleep as their first adaptation night.Circadian outcomes from this session (including DLMO and core body temperature) as well as vigilance tests will be compared to those of an evening session under dim-light conditions.The purpose of this comparison is to quantify the potency of the light exposure during the protocol to elicit a circadian phase shift (for more information, see section 2.4.2).
The second laboratory visit starts with one evening session under dim light conditions and another adaptation night when participants will have a full night of sleep (8 hours), followed by the forced-desynchrony session that involves 40 hours on a ultra-short sleep-wake cycles.Participants will enter the lab five hours before their habitual bedtime and complete a four-hour dim light session.The forced-desynchrony session starts 60 minutes after the end of the adaptation night.During the first 60 minutes, people have opportunity for personal hygiene and will be equipped with the sensors.40 hours will be spent in constant environmental conditions, with a temperature of 20°C, and no time cues are provided.The main room is lit with dim light at <10 lux, so that light exposure is constant except during light exposure due to visual tests and during sleep.For an overview of the study period, see Fig. 1.
The first ultra-short sleep-wake will start with a wake period.During wake time, all other tests will be carried out while sitting.To reach the devices, participants have to walk up within the laboratory for up to 30 meters.During wake time, light isocaloric snacks and water that are heated up to body temperature will be provided at three opportunities, mirroring breakfast, lunch and dinner.After experiments, a remaining wake time of 30 minutes can be spent freely.The 2h30m wake period is followed by a 1h15m hour sleep period, that will be spent in a recumbent position.Participants will put on a light-tight sleep mask.
An overview timeline during the forced-desynchrony session is shown in Fig. 2, and a more detailed timeline of single measurement blocks in Fig. 3

Experimental Procedure
All devices were purchased from internal funds of the investigator and were not sponsored by the manufacturer(s).

Screening
An online and an in-laboratory screening will proceed the forced-desynchrony session.
Participants will be included or excluded based on Tab 1 and section 2.1.1,respectively.
Online screening In the online screening, participants will be asked for their age, biological sex and gender identity, and employment status.They will fill out questions on whether they are taking any medication, are smoking, have a diagnosis of epilepsy, and female-born participants will be asked about their reproductive status and menstrual cycle based on the adapted Reproductive Status Questionnaire [29].
Additionally, participants will complete a series of psychological questionnaires, delivered via the online platform REDCap.Participants will complete the Alcohol Use Disorders Identification Test (AUDIT), an instrument to examine substance and alcohol abuse, in the self-report version [30].As an assessment of sleep habits, we will use the Pittsburgh Sleep Quality Index (PSQI), an instrument to assess quality of sleep based on usual sleep habits and sleep disturbances [31], and the Epworth Sleepiness Scale (ESS) to assess habitual sleepiness [32].The Munich Chronotype Questionnaire (MCTQ) will be used to determine chronotype [33].It determines habitual midpoint of sleep, corrected for sleep duration on free days (MSF SC ).Only slightly early to slightly late chronotypes are admissible for reasons of scheduling group-wise participation [34].Furthermore, we will administer the Sleep Regularity Questionnaire (SRQ) and will preferably select people with regular sleep patterns over those with more irregular patterns [35].
February 23, 2024 9/23 In-person screening During the in-person screenings, participants' physical health will be assessed using self-report questions, along with an examination by study personnel.The specific procedures encompass screening of their visual acuity using a Landolt C acuity test delivered on the Cambridge Research Systems Metropsis system, and their colour vision using HRR Pseudoisochromatic Plates.An ophthalmological exam including fundus photography and Optical Tomography images (macular image and optic disc image) will take place to confirm ocular and retinal health.Participants with a cup-to-disc ratio of > 0.3 will lead to exclusion of the participant, as well as an intra-ocular pressure (IOP) > 20 mmHg.
Alcohol and drug compliance At the beginning of the first experimental session, the light exposure session, abstinence from alcohol and drugs will be verified using a breathalyser (Alkoholtester, ACE GmbH, Freilassing, Germany) to determine blood alcohol content (BAC), and a urine sample to determine AMP/BZD/COC/MOR/OPI/THC content (nal von minden GmbH, Moers, Germany).
Participants will be asked to produce a urine sample in a collection device in a bathroom on-site.Experimenters will carry out the drug test by immersing a test strips in the urine sample.Any positive test will lead to immediate exclusion from the study.The urine sample, as well as the test strip, will be disposed of immediately on-site after testing.Documentation of the result is pseudonymised and will not contain any information for personal identification.

Quantification of an effect of the protocol on the circadian system
The experimental light exposure may affect the circadian system.However, indicators of the circadian systems are supposed to function as independent variable to predict visual outcomes.This may pose a problem of circularity, through which the effects of interest may be masked.The primary goal of the control sessions to estimate the effects of light exposure on circadian outcomes.During both control sessions, circadian outcomes described in section 2.4.6 will be measured.
To measure the effects of experimental light exposure on circadian outcomes, participants will visit the lab one week into the circadian stabilisation period, i.e. one week prior to the forced-desynchrony session for the light exposure session.Saliva collection starts four hours prior to the participant's bedtime, will be carried out every 30 minutes until 1 hour after habitual bedtime.For more information on the post-processing, go to section 2.4.6.During this time, participants will complete similar tasks as in the forced-desynchrony.The visual experiments serve to imitate the light exposure that will be experienced during the protocol and additionally provide comprehensive data on the participant's spatial and temporal contrast sensitivity functions (CSF).Simultaneously, they are introduced and familiarised with the experiments that will be performed during the forced-desynchrony session, and will be equipped with the continuous glucose monitor (section 2.4.3).
Furthermore, participants will enter the lab on the day before the forced-desynchrony session starts to spend an adaptation night in the lab before the protocol starts.They will enter lab approximately five hours before their habitual bedtime, 13 hours before the forced-desynchrony session.They will stay under constant dim light throughout the evening preceding the adaptation night (dim light session).
Data on salivary melatonin concentration and core body temperature will be collected throughout the evening.Comparing indicators of circadian phase measurements from the light exposure session and the evening dim light session, we will obtain an estimate of the potency of our stimulus protocol itself in affecting the circadian system.

Circadian stabilisation and compliance
Procedure To stabilise their circadian system before entering the forced-desynchrony session, participants will start a two-week circadian stabilisation period.It is characterized by the maintenance of a strict 16:8-hour wake-sleep and light-dark schedule, and adherence to approximate meal times.Participants choose the sleep time based on their habitual sleep and wake time, and must then adhere to bed and wake times with a maximum deviation of 30 minutes.
Caffeine, drug, alcohol and napping abstinence Caffeine modulates the circadian response to light [36].To avoid possible influences, we ask participants to refrain from napping during daytime, as well as alcohol, recreational drug, and caffeine consumption seven days prior to lab visits.
Daily questionnaire The 15-item consensus sleep diary serves as confirmation of sleep/wake categorisation by the actigraph [37] and will be administered via REDCap.
Participants will be asked to plan notifications on their personal smartphones in order to receive prompts to complete the diary in the morning within approximately 1 hour after awakening.Participants will be asked to record their mealtime.For this purpose, they will be asked to answer questions about their meals (e.g.timing, size) and take a picture to submit it to the app.
Compliance Compliance to the circadian stabilisation will be monitored with daily logs, the sleep diary and actimetry data.Participants will be instructed to wear a light-weight actiwatch (35 g) by (ActTrust2, Condor Instruments, São Paulo, Brazil), on their non-dominant hand at all times, except for when it could get wet or damaged.
The device is equipped with an in-build accelerometer and light sensors.Activity and light exposure are measured in 1-min epochs.Light exposure is quantified in photopic illuminance and UV irradiance.Whether a participant was labelled awake or asleep depends on whether the activity exceeds a specified threshold for the activity counts.
The data will be used to confirm compliance with circadian stabilisation, and may further be used in exploratory analysis evaluating relationship of main outcomes with regularity of light exposure (Light Regularity Index; LRI) and sleep (Sleep Regularity Index, SRI) [38].
Influence on metabolic processes We will monitor metabolic activity starting the morning after the light exposure session until after the forced-desynchrony session for a total of 14 days.A continuous glucose monitoring (CGM) system (FreeStyle Libre 3, Abbott Laboratories, Alameda, California) is used to monitor glucose profile.The small sensor is placed on the skin of the back of the upper arm, and measures the glucose concentration in the interstitial fluid (ISF).As blood glucose levels seem to rise in food anticipation and reflect meal-time history of multiple days, they will likely not be in line with the meal-timing imposed during the ultra-short sleep/wake rhythm [39].
Participants will not have access to their glucose readings during wear time, as feedback may interfere with naturalistic eating behaviour.Glucose levels are quantified in mg/dL.During the forced-desynchrony session, participants will repeatedly undergo a battery of visual, physiological and psychometric tests during wakefulness.The experiments that will be administered during one repeatable block are described in

Pupillometric assessment of photoreceptor mechanism
Maxwellian-view pupillometry The apparatus used for pupillometry is a custom-made six-primary binocular Maxwellian view light stimulation system [40].The system is optimized to target different photoreceptors in isolation, including melanopsin.Light is produced by LEDs of different wavelengths (420, 450, 470, 520, 590, and 630 nm) with a bandwidth (FWHM) of approximately 20 nm each.The produced light stimulus is an annular field of 10-30°that is presented in monocular viewing conditions to the dominant eye.Participant movement will be limited using an elastic headband.
The device measures pupil diameter over time based on video recordings of the respective eyes.Our main outcome is pupillary light response, i.e. pupil constriction upon monocular stimulation with visual stimuli of different spectral composition.
Calibration of the stimulation device is done using a spectroradiometer (spectraval 1511, JETI, Jena, Germany) focused on a diffuser surface at the location of the entry pupil, and an optical power and energy meter (PM100D, Thorlabs, Bergkirchen, Germany).In detail, spectra at 22 different intensities were scaled based on the power measured with the power, and used to calibrate the device.
Silent substitution Spectral composition of the light stimuli is varied to achieve silent substitution of the photoreceptor mechanisms of interest.Briefly, this means the activity of one or more photoreceptors is selectively modulated while the activation of other photoreceptor is kept constant [13,[41][42][43].While rods are saturated photopic light levels [44,45], stimuli will modulate melanopsin and cone activation for five conditions: Comparing the listed conditions, we will determine the unique circadian variation of the photoreceptor mechanisms.
Experimental procedure In each block, participants will start by adapting to a photopic background (60 cd/m 2 ) for four minutes.Participants will see light flashes from one of the five conditions, with a stimulus presentation of five seconds, followed by a period of baseline stimulation for 13 second, white stimulus, yielding a trial duration of 17 seconds.One tests consists of 25 trials, five repetitions of each condition; in the trial sequence, conditions are counterbalanced using an m-sequence.The whole experiment encompassing two tests with 25 trials each will take approximately 30 minutes.The outcome is the relative pupil constriction in response to a receptor-or mechanism-isolating stimulus.To obtain a measure of relative pupil constriction, we will obtain a maximum pupil constriction per trial from the pupil trace.The amplitude is defined by the maximum difference between pupil diameter one second prior to stimulus onset and the pupil diameter during the 5-second stimulus presentation.We will calculate the mean pupil response amplitude across trials and test repetitions for each condition.The final outcome measure is the average pupil response amplitude per condition relative to the average pupil response amplitude under light flux, as specified in Eq. (1).
indicating that they cannot enter an MRI scanner.Temperature data will be used to estimate the person-specific circadian period τ with available data dim light session, the subsequent adaptation night and the forced-desynchrony session as described in section 2.5.1.

Salivary hormone concentrations
Melatonin and cortisol concentration will be obtained using from saliva samples that are sampled approximately every 45 minutes throughout the forced-desynchrony, except during sleep periods.Saliva will be sampled approximately every 30 minutes starting 4 hours prior to bedtime until habitual bedtime.Salivette ® collection devices were chosen for saliva samples and bioassaying for the extraction of hormone concentration saliva collection.In brief, participants chew on a cotton swab for 5 minutes and place it into a plastic collection tube.The Salivette ® will then be centrifuged for 3 minutes at 3,000 rpm and frozen at -20°for temporary storage.Biological samples will be collected initially in the laboratory and then sent to external laboratory companies for further processing.The saliva samples will be sent to Munich, and assaying will be performed using enzyme-linked immunoabsorbent assay (ELISA) in the bioassaying facilities of the Prevention Centre at the Technical University of Munich as commissioned work.All samples will be destroyed after completion of the hormone analyses.
We will determine circadian phase by estimating the DLMO using the hockeystick algorithm [46].

Homeostatic processes
We will collect data from different modalities to capture homeostatic state.
Subjective sleepiness As measure of subjective fatigue, participants asked to report their sleepiness via the Karolinska Sleepiness Scale (KSS) at every measurement block.
Distal-proximal temperature gradient Wireless iButton devices will be used to monitor skin body temperature at two pre-defined locations on the skin, one proximal, e.g.near the neck, and one distal, e.g. on one ankle.The thermometry devices will also be attached to the skin.From the sensor readings, the temperature gradient between the distal and proximal skin (DPG) will be calculated, as it as been shown to vary with sleep propensity [47].Sampling frequency for skin temperature is 2 Hz.Information derived from skin temperature may feed into exploratory analyses separating effects of the circadian rhythm statically and homeostatic processes or the sleep-wake rhythm, statistically.
Cardiovascular and respiratory monitoring For blood pressure monitoring during the experimental sessions, we will employ the ABPMpro device (SOMNOmedics AG, Randersacker, Germany).Blood pressure measurements will be taken twice per measurement block.The ABPMpro can be extended by a 3-Channel electrocardiography (ECG) sensor (Holter).We will attach the surface electrodes to the participant's chest and rips for impedance cardiography throughout the 40 hours forced-desynchrony.We may derive outcomes such as heart rate, heart rate variability, and breathing frequency from measurements, to evaluate activity of the autonomous nervous system in exploratory analyses.

Alertness and Vigilance
In every block, participants will complete a 10-minute visual psychomotor vigilance test (PVT, CWE, Inc., USA).The test will be delivered on the PVT-192 Psychomotor Vigilance Task Monitor (Ambulatory Monitoring, Inc., Ardsley, USA; [48]) and serves as a measure of alertness, vigilance or conversely, sleepiness.As part of the test, participant are tracking a visual stimulus, a digit counter that appears at random on the device display and starts counting up in milliseconds.
The display acts as both, the stimulus and the performance feedback display.
Participants have to respond as fast as possible to the stimulus with a button press with the thumb of their dominant hand.Time between stimulus presentation and response constitutes the participant's response time if they are below or equal 500 ms.Median reaction time, lapse probability (reaction time > 500 ms) per session, and mean of fastest 10% reaction times are used as main outcome measures of alertness.

Ocular structure and function
All measurements of ocular physiology will be performed on the left and the right eye.Data on various aspects starting from light incidence at the retina to structure of the retinal layers will be collected.All applied procedures are standard procedures that are routinely carried out during a visit to the ophthalmologist.The ophthalmologist of the team will review concerning results and write a physician's letter in case of incidental findings.
Photography of fundus and anterior segment During the first session, coloured images of the anterior segment (including iris) and the fundus will be made.
OCT measurements We will repeatedly capture structural images of the retina and the cornea (without concomitant fundus photography due to light exposure), to obtain estimates of corneal and macular thickness using time-domain optical coherence tomography system (OCT-2000, Topcon Corporation, Tokyo, Japan).
Intra-ocular pressure For measurements of intra-ocular pressure (IOP), we will use a Non-Contact Tonometry (NT-2000 NC Tonometry, NIDEK CO., LTD., Japan).The outcome variable for one measurement block is the arithmetic mean of three pressure estimates, quantified in mmHg.Non-contact tonometry involves administration of an air-puff is to your eyes which may lead to a temporary sensation of dryness or discomfort in the eyes.However, evidence about time-of-day and posture dependent changes in IOP warrant careful control of circadian effects [49].

Subjective well-being
Participants stay under laboratory conditions for several days.This touches upon most aspects of everyday like.Eating habits during this period are also observed.However, participants are allowed to have a minimum of 30 minutes of their own time during each awake period.Additionally, their emotional state is closely tracked with mood questionnaire (I-PANAS-SF) [50].
Enrichment During the experimental sessions, participants will not be able to use their phones, or any other device indicating time of day, not will they be allowed to receive visitors.Relatives or close ones may be given a phone number to call the study team on-site in the event of an emergency.They may bring their own media or reading material, or may choose from pre-selected music, podcasts or audiobooks.
Physical needs During the lab visits, participants will be provided water ad libitum and shakes containing a balanced amount of the macro nutritions (carbohydrates, protein, fats, and fibre) as well as micronutrients.Shake intake is scheduled during February 23, 2024 16/23 . wakefulness.To determine caloric input during the 40-hour forced-desynchrony session, a participant's base metabolic rate (BMR) will be determined using the Mifflin-St.Jeor formula, taking into account their sex, body weight, and body height [51].

Estimation of circadian period
First, we will use a harmonic regression model with d = 3 harmonics, primarily to estimate the period of the circadian pacemaker τ from CBT: with t is the time or block, measured µ is the average CBT.The three parameters to be fitted for each individual participant are A r and B r which are the amplitude of the sine and cosine component, respectively, and τ , the circadian period.The point estimate of τ will be propagated into the confirmatory models testing for a circadian rhythm.

Confirmatory analysis
For our confirmatory hypotheses, we will examine two types of models.One model will test confirmatory hypotheses H1a-H1d investigating circadian rhythms in the PLR, and one model will be used to test hypotheses H2a-H2c concerned with circadian rhythms in the psychophysical thresholds.
To test our hypotheses regarding the PLR, will examine the ratio R of response amplitudes as a function of time since the start of the experimental session t by calculating the Bayes factor of the comparison between a full model: and the null model: M is the overall intercept or MESOR (Midline Statistic Of Rhythm); θ t,p is a function representing a circadian rhythm that varies over time t with person-specific random effects; u p is the person-specific random intercept; and ε t,p is the error term associated with the response at time t and the grouping variable person p.
The circadian rhythm θ t,p with a sine-cosine parameterisation using the point estimate τ from Eq. ( 2): where a and b are the amplitude of the sine and cosine component, respectively, with ), a < 0, b < 0 A full model for contrast sensitivity threshold C of a specific condition proposing effects of temporal frequency, circadian effects and interaction is formalised as: where M is the overall intercept or fixed effect; θ t,p is a function representing a circadian rhythm that depends on t with person-specific random effects, detailed below; c is the weight of the fixed effect of temporal frequency f with f ∈ [0.4,4.0]; u p is the person-specific random intercept with u ∼ N (0, σ 2 u ); and ε t,p is the error term associated with the response at time t and the grouping variable person p with ϵ ∼ N (0, σ 2 ϵ ).The null model missing the proposed effects, explaining the data only with an intercept and noise is as follows: Due to the complexity of the full model, we will build the model starting from the null model and adding predictors with the highest assumed explanatory power.Models will be evaluated be Bayesian information (BIC) criteria after adding one predictor at a time.This way, we will optimize for goodness of fit without compromising the parsimony of the model.
If the full model is selected, we will perform parameter estimation using the estimation procedure implemented for mixed effects cosinor model [52].This model is particularly adept at accommodating the complex interactions between circadian parameters and temporal frequency.Furthermore, it is able to immediately evaluate amplitude (A) and acrophase (ϕ).If model selection suggests a model without an interaction term, we will evaluate with regard to the presence of a circadian effect with the procedure described above.
We will not enter sex, age, or demographic variables in our confirmatory analysis, as we do not see sufficient empirical evidence for confounding effects.

Exploratory analysis
Effects found in the previously mentioned models will be further explored in additional analyses.Building upon the results from hypotheses H1a-H1d and H2a-H2c, we may further explore circadian effects specific to temporal frequencies or comparing circadian effects of specific temporal frequencies of different conditions.Additionally, we may confirm circadian effects in ocular structures and physiological parameters, to validate our procedure.Trajectories of ocular structures or summary statistics of physiological parameters may be included in post-hoc analysis as confounding factors in the models described in Eq. (3) for hypotheses H1a-H1d, and in Eq. ( 7) for hypotheses H2a-H2c.

Ethical Approval
The protocol has been reviewed and approved by the Medical Ethics Committee of the Technical University of Munich (2023-369-S-SB).All research will be conducted in accordance with the Declaration of Helsinki.3 Discussion

Strengths and weaknesses of the protocol
Our protocol is uniquely geared to measuring circadian variations of retinal mechanisms while minimizing homeostatic effects.In evaluating our protocol, we highlight several points.

Known limitations in silent substitution
The silent-substitution stimuli we use are nominally designed to stimulate specific photoreceptor classes.The precise contrast seen by the photoreceptor classes in the actual stimulus conditions can vary due to individual differences, which difficult to characterize for an individual.As a consequence, our stimuli may contain some small residual contrast on the unstimulated photoreceptor classes.This is a well-known problem [42] but does not challenge the validity of silent-substitution measurements at large contrasts.
Artificial laboratory conditions Due to the controlled nature of the experiment, participants are placed in an artificial environment, unable to fully represent real-world conditions.In particular, the experimental light conditions may have an impact on psychological variables in such a way that behaviour diverges from behaviour in natural settings.This is a limitation of any laboratory study, but a precondition for experimentally manipulating the mechanisms under investigation.

Sample considerations
In our study, we aim to recruit a total of twelve participants.
As our study follows a strong within-subjects design, we believe that this sample size is appropriate to capture within-participant circadian influences on the outcome measures.Importantly, different to many chronobiological studies, the present study is unique in that it strives to have a representative sample regarding distribution of sex (50% female participants).Due to unknown variations introduced by hormonal fluctuations, data required for comparisons will be collected at similar stages in the menstrual cycle, specifically during the mid-follicular phase (see section 2.4.2).

Dissemination
We are dedicated to ensuring the robustness and transparency of the findings of this research, and its dissemination.First, we will present our results at relevant academic conferences, fostering discussion and feedback to ensure adequate data processing.All resources, including code, data, recruitment materials will be made openly available on GitHub in order to promote transparency and reproducibility.Finally, where applicable, we are committed to engage with the public to convey the importance and applicability of the presented research, as well as to foster a better understanding of the underlying biological phenomena, such as circadian rhythms, sleep-wake behaviour and visual performance.

Protocol amendments, deviations and termination
In the publications describing the results arising from this protocol, any termination of the protocol will be described transparently and published for the scientific community, openly discussing conditions and feasibility of the presented study protocol, as well as relevant learnings.Any deviations from or ad-hoc amendments of the protocol will be described, along with the rationale.At the data analysis stage, all available data will be utilized.In the expected rare occurrence of termination by the participant, we will February 23, 2024 19/23 describe circumstances and if available reasons for termination, and where possible and in line with the inclusion criteria, we will include partial data in our analysis.
measurements and are followed by 75 min (1h15m) sleep intervals in darkness.The protocol lasts 40 hours, yielding 11 repeating LD cycles or measurement blocks.
Building upon prior evidence for a time-of day dependency in image-forming (i.e., luminance and colour perception) and non image-forming functions (i.e., pupillary light February 23, 2024 3/23 response, PLR), we hypothesize that photoreceptor sensitivity and retinal mechanisms are modulated by the circadian rhythm.Our hypotheses regarding variations in non-image forming function investigate circadian effects in the PLR to different conditions.The variable is the relative PLR amplitude normalized with the PLR amplitude to a joint modulation of all cones and melanopsin at equal contrast (light flux stimulation; Light flux condition) to obtain a ratio R. Hereafter, the PLR refers to the normalized response amplitude.The detailed confirmatory hypotheses state that the PLR shows circadian effects in the response to 4. modulation of the colour opponent process L-M with constant activation of S and melanopsin (L-M) [H1a]; 5. modulation the S-cones with constant activation of L-cone, M-cone, and melanopsin activation (S) [H1b]; 6. joint modulation of L-, M-, and S-cones (L+M+S [H1c]; 7. selective modulation of melanopsin activation (Mel, constancy of cone activation) [H1d] To investigate circadian effects isolated from homeostatic effects in the image-forming pathway by evaluating temporal contrast sensitivity threshold to two different frequencies, focusing on the following photoreceptor mechanisms using silent-substitution stimuli: activation of the luminance channel (L+M+S), post-receptoral red-green channel (L-M), activation of the S-cones, activation of the melanopsin in the ipRGCs.We will evaluate the contrast sensitivity to: 1. modulation of the colour opponent process L-M with constant activation of S and melanopsin (L-M) [H2a]; 2. modulation the S-cones with constant activation of L-cone, M-cone, and melanopsin activation (S) [H2b]; 3. joint modulation of L-, M-and S-cones (L+M+S) [H2c].

Fig 1 .
Fig 1. Overview over study period including circadian stabilisation period and all experimental sessions.Timeline relative to the forced desynchrony session

A = a 2 + b 2 February 23
up ).We will sample from the posterior distributions to obtain parameter estimates and derive amplitude (A) and acrophase (ϕ) of the proposed underlying circadian rhythm with

Table 2 .
Exclusion criteria . Timeline during 40-hour forced-desynchrony session with ultra-short sleep-wake rhythm.Start time depends on participant's sleep time; exemplary start time.