Development and Reporting of Prediction Models: Guidance for Authors From Editors of Respiratory, Sleep, and Critical Care Journals.

Prediction models aim to use available data to predict a health state or outcome that has not yet been observed. Prediction is primarily relevant to clinical practice, but is also used in research, and administration. While prediction modeling involves estimating the relationship between patient factors and outcomes, it is distinct from casual inference. Prediction modeling thus requires unique considerations for development, validation, and updating. This document represents an effort from editors at 31 respiratory, sleep, and critical care medicine journals to consolidate contemporary best practices and recommendations related to prediction study design, conduct, and reporting. Herein, we address issues commonly encountered in submissions to our various journals. Key topics include considerations for selecting predictor variables, operationalizing variables, dealing with missing data, the importance of appropriate validation, model performance measures and their interpretation, and good reporting practices. Supplemental discussion covers emerging topics such as model fairness, competing risks, pitfalls of "modifiable risk factors", measurement error, and risk for bias. This guidance is not meant to be overly prescriptive; we acknowledge that every study is different, and no set of rules will fit all cases. Additional best practices can be found in the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) guidelines, to which we refer readers for further details.

Chronic Suppression of Hypothalamic Cell Proliferation and Neurogenesis Induces Aging-Like Changes in Sleep-Wake Organization in Young Mice.

Aging is associated with sleep-wake disruption, dampening of circadian amplitudes, and a reduced homeostatic sleep response. Aging is also associated with a decline in hypothalamic cell proliferation. We hypothesized that the aging-related decline in cell-proliferation contributes to the dysfunction of preoptic-hypothalamic sleep-wake and circadian systems and consequent sleep-wake disruption. We determined if cytosine-ß-D-arabinofuranoside (AraC), an antimitotic agent known to suppress hypothalamic cell proliferation and neurogenesis, causes sleep-wake instability in young mice. The sleep-wake profiles were compared during baseline, during 4 weeks of artificial cerebrospinal fluid (aCSF) + 5-bromo-2'-deoxyuridine (BrdU) or AraC+BrdU infusion into the lateral ventricle, and 8 weeks after treatments. The sleep-wake architecture after AraC treatment was further compared with sleep-wake profiles in aged mice. Compared to aCSF+BrdU, 4 weeks of AraC+BrdU infusion significantly decreased (-96%) the number of BrdU+ cells around the third ventricular wall and adjacent preoptic-hypothalamic area and produced a) sleep disruption during the light phase with decreases in non-rapid eye movement (nonREM) (-9%) and REM sleep (-21%) amounts, and increased numbers of shorter (<2 min; 142 versus 98 episodes/12 h) and decreased numbers of longer (>5 min; 19 versus 26 episodes/12 h) nonREM sleep episodes; and b) wake disruption during the dark phase, with increased numbers of shorter (138 versus 91 episodes/12 h) and decreased numbers of longer active waking (17 versus 24 episodes/12 h) episodes. AraC-treated mice also exhibited lower delta activity within nonREM recovery sleep. The sleep-wake architecture of AraC-treated mice was similar to that observed in aged mice. These findings are consistent with a hypothesis that a decrease in hypothalamic cell proliferation/neurogenesis is detrimental to sleep-wake and circadian systems and may underlie sleep-wake disturbance in aging.

Body temperature and sleep.

Sleep in mammals is accompanied by a decrease in core body temperature (CBT). The circadian clock in the hypothalamic suprachiasmatic nucleus regulates daily rhythms in both CBT and arousal states, and these rhythms are normally coupled. Reductions in metabolic heat production resulting from behavioral quiescence and reduced muscle tone along with changes in autonomic nervous system activity and thermoeffector activity contribute to the sleep-related fall in CBT. Reductions in sympathetic tone to the peripheral vasculature resulting in heat loss through the skin are reflected in a sleep-related increase in distal skin temperature that is a prominent feature of sleep onset in humans. Within a sleep episode, patterns of autonomic nervous system and thermoeffector activity and the ability to defend against heat and cold exposure differ during nonrapid eye movement (NREM) and rapid eye movement sleep. Anatomic and functional integration of the control of arousal states and thermoregulation occur in the preoptic/anterior hypothalamus. Subsets or warm-sensing neurons in the preoptic/anterior hypothalamus implicated in CBT regulation are spontaneously activated during sleep onset and NREM sleep compared to waking and may underlie sleep-related changes in autonomic nervous system and thermoeffector activity.

Characteristics of sleep-active neurons in the medullary parafacial zone in rats.

Growing evidence supports a role for the medullary parafacial zone in non-rapid eye movement (non-REM) sleep regulation. Cell-body specific lesions of the parafacial zone or disruption of its GABAergic/glycinergic transmission causes suppression of non-REM sleep, whereas, targeted activation of parafacial GABAergic/glycinergic neurons reduce sleep latency and increase non-REM sleep amount, bout duration, and cortical electroencephalogram (EEG) slow-wave activity. Parafacial GABAergic/glycinergic neurons also express sleep-associated c-fos immunoreactivity. Currently, it is not clear if parafacial neurons are non-REM sleep-active and/or REM sleep-active or play a role in the initiation or maintenance of non-REM sleep. We recorded extracellular discharge activity of parafacial neurons across the spontaneous sleep-waking cycle using microwire technique in freely behaving rats. Waking-, non-REM sleep-, and REM sleep-active neuronal groups were segregated by the ratios of their discharge rate changes during non-REM and REM sleep versus waking and non-REM sleep versus REM sleep. Parafacial neurons exhibited heterogeneity in sleep-waking discharge patterns, but 34 of 86 (40%) recorded neurons exhibited increased discharge rate during non-REM sleep compared to waking. These neurons also exhibited increased discharge prior to non-REM sleep onset, similar to median preoptic nucleus (MnPO) and ventrolateral preoptic area (VLPO) sleep-active neurons. However, unlike MnPO and VLPO sleep-active neurons, parafacial neurons were weakly-moderately sleep-active and exhibited a stable rather than decreasing discharge across sustained non-REM sleep episode. We show for the first time that the medullary parafacial zone contains non-REM sleep-active neurons. These neurons are likely functionally important brainstem compliments to the preoptic-hypothalamic sleep-promoting neuronal networks that underlie sleep onset and maintenance.

Neuronal substrates of sleep homeostasis; lessons from flies, rats and mice.

Sleep homeostasis is a fundamental property of vigilance state regulation that is highly conserved across species. Neuronal systems and circuits that underlie sleep homeostasis are not well understood. In Drosophila, a neuronal circuit involving neurons in the ellipsoid body and in the dorsal Fan-shaped body is a candidate for both tracing sleep need during waking and translating it to increased sleep drive and expression. Sleep homeostasis in rats and mice involves multiple neuromodulators acting on multiple wake- and sleep-promoting neuronal systems. A functional central homeostat emerges from A1 receptor mediated actions of adenosine on wake-promoting neurons in the basal forebrain and hypothalamus, and A2A adenosine receptor-mediated actions on sleep-promoting neurons in the preoptic hypothalamus and nucleus accumbens.

The role of adenosine in the maturation of sleep homeostasis in rats.

Sleep homeostasis in rats undergoes significant maturational changes during postweaning development, but the underlying mechanisms of this process are unknown. In the present study we tested the hypothesis that the maturation of sleep is related to the functional emergence of adenosine (AD) signaling in the brain. We assessed postweaning changes in 1) wake-related elevation of extracellular AD in the basal forebrain (BF) and adjacent lateral preoptic area (LPO), and 2) the responsiveness of median preoptic nucleus (MnPO) sleep-active cells to increasing homeostatic sleep drive. We tested the ability of exogenous AD to augment homeostatic responses to sleep deprivation (SD) in newly weaned rats. In groups of postnatal day (P)22 and P30 rats, we collected dialysate from the BF/LPO during baseline (BSL) wake-sleep, SD, and recovery sleep (RS). HPLC analysis of microdialysis samples revealed that SD in P30 rats results in significant increases in AD levels compared with BSL. P22 rats do not exhibit changes in AD levels in response to SD. We recorded neuronal activity in the MnPO during BSL, SD, and RS at P22/P30. MnPO neurons exhibited adult-like increases in waking neuronal discharge across SD on both P22 and P30, but discharge rates during enforced wake were higher on P30 vs. P22. Central administration of AD (1 nmol) during SD on P22 resulted in increased sleep time and EEG slow-wave activity during RS compared with saline control. Collectively, these findings support the hypothesis that functional reorganization of an adenosinergic mechanism of sleep regulation contributes to the maturation of sleep homeostasis. NEW & NOTEWORTHY: Brain mechanisms that regulate the maturation of sleep are understudied. The present study generated first evidence about a potential mechanistic role for adenosine in the maturation of sleep homeostasis. Specifically, we demonstrate that early postweaning development in rats, when homeostatic response to sleep loss become adult like, is characterized by maturational changes in wake-related production/release of adenosine in the brain. Pharmacologically increased adenosine signaling in developing brain facilitates homeostatic responses to sleep deprivation.

Suppression of preoptic sleep-regulatory neuronal activity during corticotropin-releasing factor-induced sleep disturbance.

Corticotropin releasing factor (CRF) is implicated in sleep and arousal regulation. Exogenous CRF causes sleep suppression that is associated with activation of at least two important arousal systems: pontine noradrenergic and hypothalamic orexin/hypocretin neurons. It is not known whether CRF also impacts sleep-promoting neuronal systems. We hypothesized that CRF-mediated changes in wake and sleep involve decreased activity of hypothalamic sleep-regulatory neurons localized in the preoptic area. To test this hypothesis, we examined the effects of intracerebroventricular administration of CRF on sleep-wake measures and c-Fos expression in GABAergic neurons in the median preoptic nucleus (MnPN) and ventrolateral preoptic area (VLPO) in different experimental conditions. Administration of CRF (0.1 nmol) during baseline rest phase led to delayed sleep onset and decreases in total amount and mean duration of non-rapid eye movement (NREM) sleep. Administration of CRF during acute sleep deprivation (SD) resulted in suppression of recovery sleep and decreased c-Fos expression in MnPN/VLPO GABAergic neurons. Compared with vehicle controls, intracerebroventricular CRF potentiated disturbances of both NREM and REM sleep in rats exposed to a species-specific psychological stressor, the dirty cage of a male conspecific. The number of MnPN/VLPO GABAergic neurons expressing c-Fos was reduced in the CRF-treated group of dirty cage-exposed rats. These findings confirm the involvement of CRF in wake-sleep cycle regulation and suggest that increased CRF signaling in the brain 1) negatively affects homeostatic responses to sleep loss, 2) exacerbates stress-induced disturbances of sleep, and 3) suppresses the activity of sleep-regulatory neurons of the MnPN and VLPO.

Neuronal activity in the preoptic hypothalamus during sleep deprivation and recovery sleep.

The preoptic hypothalamus is implicated in sleep regulation. Neurons in the median preoptic nucleus (MnPO) and the ventrolateral preoptic area (VLPO) have been identified as potential sleep regulatory elements. However, the extent to which MnPO and VLPO neurons are activated in response to changing homeostatic sleep regulatory demands is unresolved. To address this question, we continuously recorded the extracellular activity of neurons in the rat MnPO, VLPO and dorsal lateral preoptic area (LPO) during baseline sleep and waking, during 2 h of sleep deprivation (SD) and during 2 h of recovery sleep (RS). Sleep-active neurons in the MnPO (n = 11) and VLPO (n = 13) were activated in response to SD, such that waking discharge rates increased by 95.8 ± 29.5% and 59.4 ± 17.3%, respectively, above waking baseline values. During RS, non-rapid eye movement (REM) sleep discharge rates of MnPO neurons initially increased to 65.6 ± 15.2% above baseline values, then declined to baseline levels in association with decreases in EEG delta power. Increase in non-REM sleep discharge rates in VLPO neurons during RS averaged 40.5 ± 7.6% above baseline. REM-active neurons (n = 16) in the LPO also exhibited increased waking discharge during SD and an increase in non-REM discharge during RS. Infusion of A2A adenosine receptor antagonist into the VLPO attenuated SD-induced increases in neuronal discharge. Populations of LPO wake/REM-active and state-indifferent neurons and dorsal LPO sleep-active neurons were unresponsive to SD. These findings support the hypothesis that sleep-active neurons in the MnPO and VLPO, and REM-active neurons in the LPO, are components of neuronal circuits that mediate homeostatic responses to sustained wakefulness.

Animal care practices in experiments on biological rhythms and sleep: report of the Joint Task Force of the Society for Research on Biological Rhythms and the Sleep Research Society.

Many physiological and molecular processes are strongly rhythmic and profoundly influenced by sleep. The continuing effort of biological, medical, and veterinary science to understand the temporal organization of cellular, physiological, behavioral and cognitive function holds great promise for the improvement of the welfare of animals and human beings. As a result, attending veterinarians and IACUC are often charged with the responsibility of evaluating experiments on such rhythms or the effects of sleep (or its deprivation) in vertebrate animals. To produce interpretable data, animals used in such research must often be maintained in carefully controlled (often constant) conditions with minimal disruption. The lighting environment must be strictly controlled, frequent changes of cages and bedding are undesirable, and daily visual checks are often not possible. Thus deviations from the standard housing procedures specified in the Guide for the Care and Use of Laboratory Animals are often necessary. This report reviews requirements for experiments on biological rhythms and sleep and discusses how scientific considerations can be reconciled with the recommendations of the Guide.

Adenosine A(2A) receptors regulate the activity of sleep regulatory GABAergic neurons in the preoptic hypothalamus.

The median preoptic nucleus (MnPN) and the ventrolateral preoptic area (VLPO) are two hypothalamic regions that have been implicated in sleep regulation, and both nuclei contain sleep-active GABAergic neurons. Adenosine is an endogenous sleep regulatory substance, which promotes sleep via A1 and A2A receptors (A2AR). Infusion of A2AR agonist into the lateral ventricle or into the subarachnoid space underlying the rostral basal forebrain (SS-rBF), has been previously shown to increase sleep. We examined the effects of an A2AR agonist, CGS-21680, administered into the lateral ventricle and the SS-rBF on sleep and c-Fos protein immunoreactivity (Fos-IR) in GABAergic neurons in the MnPN and VLPO. Intracerebroventricular administration of CGS-21680 during the second half of lights-on phase increased sleep and increased the number of MnPN and VLPO GABAergic neurons expressing Fos-IR. Similar effects were found with CGS-21680 microinjection into the SS-rBF. The induction of Fos-IR in preoptic GABAergic neurons was not secondary to drug-induced sleep, since CGS-21680 delivered to the SS-rBF significantly increased Fos-IR in MnPN and VLPO neurons in animals that were not permitted to sleep. Intracerebroventricular infusion of ZM-241385, an A2AR antagonist, during the last 2 h of a 3-h period of sleep deprivation caused suppression of subsequent recovery sleep and reduced Fos-IR in MnPN and VLPO GABAergic neurons. Our findings support a hypothesis that A2AR-mediated activation of MnPN and VLPO GABAergic neurons contributes to adenosinergic regulation of sleep.

Maturation of sleep homeostasis in developing rats: a role for preoptic area neurons.

The present study evaluated the hypothesis that developmental changes in hypothalamic sleep-regulatory neuronal circuits contribute to the maturation of sleep homeostasis in rats during the fourth postnatal week. In a longitudinal study, we quantified electrographic measures of sleep during baseline and in response to sleep deprivation (SD) on postnatal days 21/29 (P21/29) and P22/30 (experiment 1). During 24-h baseline recordings on P21, total sleep time (TST) during the light and dark phases did not differ significantly. On P29, TST during the light phase was significantly higher than during the dark phase. Mean duration of non-rapid-eye-movement (NREM) sleep bouts was significantly longer on P29 vs. P21, indicating improved sleep consolidation. On both P22 and P30, rats exhibited increased NREM sleep amounts and NREM electroencephalogram delta power during recovery sleep (RS) compared with baseline. Increased NREM sleep bout length during RS was observed only on P30. In experiment 2, we quantified activity of GABAergic neurons in median preoptic nucleus (MnPN) and ventrolateral preoptic area (VLPO) during SD and RS in separate groups of P22 and P30 rats using c-Fos and glutamic acid decarboxylase (GAD) immunohistochemistry. In P22 rats, numbers of Fos(+)GAD(+) neurons in VLPO did not differ among experimental conditions. In P30 rats, Fos(+)GAD(+) counts in VLPO were elevated during RS. MnPN neuronal activity was state-dependent in P22 rats, but Fos(+)GAD(+) cell counts were higher in P30 rats. These findings support the hypothesis that functional emergence of preoptic sleep-regulatory neurons contributes to the maturation of sleep homeostasis in the developing rat brain.

Hypothalamic versus neocortical control of sleep.

PURPOSE OF REVIEW: Regions of the neocortex most strongly activated during waking exhibit increased sleep intensity during subsequent sleep. The novel concept that aspects of sleep homeostasis are determined locally in the cortex contrasts with the established views that global changes in neocortical activity during sleep are achieved through inhibition of ascending arousal systems that originate in the brainstem and hypothalamus. RECENT FINDINGS: Experiments in animals and humans document asymmetries in neocortical electroencephalogram (EEG) slow-wave activity (SWA), a marker of homeostatic sleep need, as a result of functional activity during waking. In addition to local, use-dependent augmentation of EEG SWA and evoked potentials, expression of plasticity-related genes and of sleep-regulatory cytokines and neuromodulators have been shown to be elevated in a use-dependent manner in neocortex. The functional consequences of local sleep are hypothesized to involve regulation of synaptic plasticity, synaptic homeostasis and energy balance. SUMMARY: The evidence for use-dependent modulation of neocortical activity during sleep is compelling and provides novel insights into sleep function. However, local changes in neocortex are generally expressed on a background of global sleep. It remains to be determined if events initiated in the cortex have global sleep-promoting effects and how neocortical and hypothalamic mechanisms of sleep control interact.

Growth hormone-releasing hormone activates sleep regulatory neurons of the rat preoptic hypothalamus.

We examined whether growth hormone-releasing hormone (GHRH) may promote non-rapid eye movement (NREM) sleep via activation of GABAergic neurons in the preoptic area. Male Sprague-Dawley rats were implanted with EEG, EMG electrodes and a unilateral intracerebroventricular cannula. Groups of rats received injections (3 microl icv) with gonadotropin-releasing hormone (GHRH) (0.1 nmol/100 g body wt) or equal volume of physiological saline at the onset of the dark period and were permitted spontaneous sleep for 90 min. Separate groups of rats were sleep deprived by gentle handling for 90 min, beginning at the time of GHRH or saline injection, at the onset of the dark period. Other groups of rats received intracerebroventricular octreotide (somatostatin analog OCT) injections, intracerebroventricular injection of one of two doses of competitive GHRH antagonist, or intracerebroventricular saline injection at light onset and were then permitted 90 min spontaneous sleep-waking. Rats were killed immediately after the 90-min sleep/wake monitoring period. Brain tissue was processed for immunohistochemistry for c-Fos protein and glutamic acid decarboxylase (GAD). Single c-Fos and dual Fos-GAD cell counts were determined in the median preoptic nucleus (MnPN), and in the core and the extended parts of the ventrolateral preoptic nucleus (cVLPO and exVLPO). Intracerebroventricular GHRH elicited a significant increase in NREM sleep amount. Double-labeled Fos+GAD cell counts were significantly elevated after GHRH injection in the MnPN and VLPO in both undisturbed and sleep-deprived groups. OCT and GHRH antagonist significantly decreased NREM sleep amount compared with control rats. OCT injection increased single c-Fos-labeled cell counts in the MnPN, but not in the VLPO. Double-labeled cell counts were significantly reduced after OCT and the high dose of GHRH antagonist injection in all areas examined. These findings identify GABAergic neurons in the MnPN and VLPO as potential targets of the sleep-regulatory actions of GHRH.

Role of adenosine A(1) receptor in the perifornical-lateral hypothalamic area in sleep-wake regulation in rats.

The perifornical-lateral hypothalamic area (PF-LHA) has been implicated in the regulation of arousal. The PF-LHA contains wake-active neurons that are quiescent during non-REM sleep and in the case of neurons expressing the peptide hypocretin (HCRT), quiescent during both non-REM and REM sleep. Adenosine is an endogenous sleep factor and recent evidence suggests that adenosine via A(1) receptors may act on PF-LHA neurons to promote sleep. We examined the effects of bilateral activation as well as blockade of A(1) receptors in the PF-LHA on sleep-wakefulness in freely behaving rats. The sleep-wake profiles of male Wistar rats were recorded during reverse microdialysis perfusion of artificial cerebrospinal fluid (aCSF) and two doses of adenosine A(1) receptor antagonist, 1,3-dipropyl-8-phenylxanthine (CPDX; 5 microM and 50 microM) or A(1) receptor agonist, N(6)-cyclopentyladenosine (CPA; 5 microM and 50 microM) into the PF-LHA for 2 h followed by 4 h of aCSF perfusion. CPDX perfused into the PF-LHA during lights-on phase produced arousal (F=7.035, p<0.001) and concomitantly decreased both non-REM (F=7.295, p<0.001) and REM sleep (F=3.456, p<0.004). In contrast, CPA perfused into the PF-LHA during lights-off phase significantly suppressed arousal (F=7.891, p<0.001) and increased non-REM (F=8.18, p <0.001) and REM sleep (F=30.036, p<0.001). These results suggest that PF-LHA is one of the sites where adenosine, acting via A(1) receptors, inhibits PF-LHA neurons to promote sleep.