High-throughput visual assessment of sleep stages in mice using machine learning.

STUDY OBJECTIVES: Sleep is an important biological process that is perturbed in numerous diseases, and assessment of its substages currently requires implantation of electrodes to carry out electroencephalogram/electromyogram (EEG/EMG) analysis. Although accurate, this method comes at a high cost of invasive surgery and experts trained to score EEG/EMG data. Here, we leverage modern computer vision methods to directly classify sleep substages from video data. This bypasses the need for surgery and expert scoring, provides a path to high-throughput studies of sleep in mice. METHODS: We collected synchronized high-resolution video and EEG/EMG data in 16 male C57BL/6J mice. We extracted features from the video that are time and frequency-based and used the human expert-scored EEG/EMG data to train a visual classifier. We investigated several classifiers and data augmentation methods. RESULTS: Our visual sleep classifier proved to be highly accurate in classifying wake, non-rapid eye movement sleep (NREM), and rapid eye movement sleep (REM) states, and achieves an overall accuracy of 0.92 ± 0.05 (mean ± SD). We discover and genetically validate video features that correlate with breathing rates, and show low and high variability in NREM and REM sleep, respectively. Finally, we apply our methods to noninvasively detect that sleep stage disturbances induced by amphetamine administration. CONCLUSIONS: We conclude that machine learning-based visual classification of sleep is a viable alternative to EEG/EMG based scoring. Our results will enable noninvasive high-throughput sleep studies and will greatly reduce the barrier to screening mutant mice for abnormalities in sleep.

Normal sleep requires the astrocyte brain-type fatty acid binding protein FABP7.

Sleep is found widely in the animal kingdom. Despite this, few conserved molecular pathways that govern sleep across phyla have been described. The mammalian brain-type fatty acid binding protein (Fabp7) is expressed in astrocytes, and its mRNA oscillates in tandem with the sleep-wake cycle. However, the role of FABP7 in regulating sleep remains poorly understood. We found that the missense mutation FABP7.T61M is associated with fragmented sleep in humans. This phenotype was recapitulated in mice and fruitflies bearing similar mutations: Fabp7-deficient mice and transgenic flies that express the FABP7.T61M missense mutation in astrocytes also show fragmented sleep. These results provide novel evidence for a distinct molecular pathway linking lipid-signaling cascades within astrocytes in sleep regulation among phylogenetically disparate species.

Novel method for high-throughput phenotyping of sleep in mice.

Assessment of sleep in mice currently requires initial implantation of chronic electrodes for assessment of electroencephalogram (EEG) and electromyogram (EMG) followed by time to recover from surgery. Hence, it is not ideal for high-throughput screening. To address this deficiency, a method of assessment of sleep and wakefulness in mice has been developed based on assessment of activity/inactivity either by digital video analysis or by breaking infrared beams in the mouse cage. It is based on the algorithm that any episode of continuous inactivity of > or =40 s is predicted to be sleep. The method gives excellent agreement in C57BL/6J male mice with simultaneous assessment of sleep by EEG/EMG recording. The average agreement over 8,640 10-s epochs in 24 h is 92% (n = 7 mice) with agreement in individual mice being 88-94%. Average EEG/EMG determined sleep per 2-h interval across the day was 59.4 min. The estimated mean difference (bias) per 2-h interval between inactivity-defined sleep and EEG/EMG-defined sleep was only 1.0 min (95% confidence interval for mean bias -0.06 to +2.6 min). The standard deviation of differences (precision) was 7.5 min per 2-h interval with 95% limits of agreement ranging from -13.7 to +15.7 min. Although bias significantly varied by time of day (P = 0.0007), the magnitude of time-of-day differences was not large (average bias during lights on and lights off was +5.0 and -3.0 min per 2-h interval, respectively). This method has applications in chemical mutagenesis and for studies of molecular changes in brain with sleep/wakefulness.

Age-related changes in adenosine metabolic enzymes in sleep/wake regulatory areas of the brain.

The impact of age on the enzymatic activities of adenosine metabolic enzymes, i.e., adenosine deaminase, adenosine kinase, cytosolic- and ecto-5'-nucleotidase have been assessed in the brain sleep/wake regulatory areas of young, intermediate and old rats (2, 12 and 24 months, respectively). There were significant spatial differences in the distribution of enzymes of adenosine metabolism in the brain. Age did not impact on the enzymatic activity of adenosine deaminase. Adenosine kinase activity increased significantly in the cerebral cortex of old animals. However, there were no differences in the activity of adenosine kinase between young and intermediate aged rats. The largest age-related changes were in the activity of cytosolic- and ecto-5'-nucleotidase and there was a significant age-related increase in the activity of these enzymes in the sleep/wake regulatory areas. In addition, the activity of cytosolic- and ecto-5'-nucleotidase increased in the cerebral cortex of old and intermediate age rats when compared to young animals. An increase in the enzymatic activities in the cerebral cortex of adenosine kinase and 5'-nucleotideases was accompanied by an increase in the level of their mRNA. An increase in the activity of 5'-nucleotideases with age likely leads to an increase in adenosine levels in the brain.

Differences in activity of cytochrome C oxidase in brain between sleep and wakefulness.

STUDY OBJECTIVES: Increased mRNA level of subunit 1 cytochrome c oxidase (COXI) during wakefulness and after short-term sleep deprivation has been described in brain. We hypothesized that this might contribute to increased activity of cytochrome oxidase (COX) enzyme during wakefulness, as part of the mechanisms to provide sufficient amounts of adenosine triphosphate to meet increased neuronal energy demands. DESIGN: COX activity was measured in isolated mitochondria from different brain regions in groups of rats with 3 hours of spontaneous sleep, 3 hours of spontaneous wake, and 3 hours of sleep deprivation. The group with 3 hours of spontaneous wake was added to delineate the circadian component of changes in the enzyme activity. Northern blot analysis was performed to examine the mRNA levels of 2 subunits of the enzyme COXI and COXIV, encoded by mitochondrial and nuclear DNA, respectively. SETTING: Laboratory of Biochemistry, Department of Animal Biology, and Center for Sleep and Respiratory Neurobiology, University of Pennsylvania. PARTICIPANTS: 2-month-old male Fischer rats (N = 21) implanted for polygraphic recording. MEASUREMENTS AND RESULTS: For COX activity, there was a main effect by analysis of variance of experimental group (P < .0001) with significant increases in COX activity in wake and sleep-deprived groups as compared to the sleep group. A main effect of brain region was also significant (P < .001). There was no difference between brain regions in the degree of increase in enzyme activity in wakefulness. Both COXI and COXIV mRNA were increased with wakefulness as compared to sleep. CONCLUSIONS: There is an increase in COX activity after both 3 hours of spontaneous wake and 3 hours of sleep deprivation as compared with 3 hours of spontaneous sleep in diverse brain regions, which could be, in part, explained by the increased levels of bigenomic transcripts of the enzyme. This likely contributes to increased adenosine triphosphate production during wakefulness. ABBREVIATIONS: ADP, adenosine diphosphate; ATP, adenosine triphosphate; COXI, cytochrome c oxidase subunit 1 mRNA; COX, cytochrome c oxidase (protein); CREB, cyclic AMP response element binding protein; DNA, deoxyribonucleic acid; EDTA, ethylenediaminetetraacetic acid; EEG, electroencephalography; EMG, electromyography; GABP, GA binding protein; HEPES, 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid; mRNA, messenger ribonucleic acid; NADH, nicotinamid adenine dinucleotide, reduced; NDII, NADH dehydrogenase subunit 2 mRNA; NRF, nuclear respiratory factor.

Enzymes of adenosine metabolism in the brain: diurnal rhythm and the effect of sleep deprivation.

Adenosine plays a role in promoting sleep, an effect that is thought to be mediated in the basal forebrain. Adenosine levels vary in this region with prolonged wakefulness in a unique way. The basis for this is unknown. We examined, in rats, the activity of the major metabolic enzymes for adenosine - adenosine deaminase, adenosine kinase, ecto- and cytosolic 5'-nucleotidase - in sleep/wake regulatory regions as well as cerebral cortex, and how the activity varies across the day and with sleep deprivation. There were robust spatial differences for the activity of adenosine deaminase, adenosine kinase, and cytosolic and ecto-5'-nucleotidase. However, the basal forebrain was not different from other sleep/wake regulatory regions apart from the tuberomammillary nucleus. All adenosine metabolic enzymes exhibited diurnal variations in their activity, albeit not in all brain regions. Activity of adenosine deaminase increased during the active period in the ventrolateral pre-optic area but decreased significantly in the basal forebrain. Enzymatic activity of adenosine kinase and cytosolic-5'-nucleotidase was higher during the active period in all brain regions tested. However, the activity of ecto-5'-nucleotidase was augmented during the active period only in the cerebral cortex. This diurnal variation may play a role in the regulation of adenosine in relationship to sleep and wakefulness across the day. In contrast, we found no changes specifically with sleep deprivation in the activity of any enzyme in any brain region. Thus, changes in adenosine with sleep deprivation are not a consequence of alterations in adenosine enzyme activity.