Differentiating acute from chronic insomnia with machine learning from actigraphy time series data.

Acute and chronic insomnia have different causes and may require different treatments. They are investigated with multi-night nocturnal actigraphy data from two sleep studies. Two different wrist-worn actigraphy devices were used to measure physical activities. This required data pre-processing and transformations to smooth the differences between devices. Statistical, power spectrum, fractal and entropy analyses were used to derive features from the actigraphy data. Sleep parameters were also extracted from the signals. The features were then submitted to four machine learning algorithms. The best performing model was able to distinguish acute from chronic insomnia with an accuracy of 81%. The algorithms were then used to evaluate the acute and chronic groups compared to healthy sleepers. The differences between acute insomnia and healthy sleep were more prominent than between chronic insomnia and healthy sleep. This may be associated with the adaptation of the physiology to prolonged periods of disturbed sleep for individuals with chronic insomnia. The new model is a powerful addition to our suite of machine learning models aiming to pre-screen insomnia at home with wearable devices.

A machine learning model for multi-night actigraphic detection of chronic insomnia: development and validation of a pre-screening tool.

We propose a novel machine learning-based method for analysing multi-night actigraphy signals to objectively classify and differentiate nocturnal awakenings in individuals with chronic insomnia (CI) and their cohabiting healthy partners. We analysed nocturnal actigraphy signals from 40 cohabiting couples with one partner seeking treatment for insomnia. We extracted 12 time-domain dynamic and nonlinear features from the actigraphy signals to classify nocturnal awakenings in healthy individuals and those with CI. These features were then used to train two machine learning classifiers, random forest (RF) and support vector machine (SVM). An optimization algorithm that incorporated the predicted quality of each night for each individual was used to classify individuals into CI or healthy sleepers. Using the proposed actigraphic signal analysis technique, coupled with a rigorous leave-one-out validation approach, we achieved a classification accuracy of 80% (sensitivity: 76%, specificity: 82%) in classifying CI individuals and their healthy bed partners. The RF classifier (accuracy: 80%) showed a better performance than SVM (accuracy: 75%). Our approach to analysing the multi-night nocturnal actigraphy recordings provides a new method for screening individuals with CI, using wrist-actigraphy devices, facilitating home monitoring.

Network-based Responses to the Psychomotor Vigilance Task during Lapses in Adolescents after Short and Extended Sleep.

Neuroimaging studies of the Psychomotor Vigilance Task (PVT) have revealed brain regions involved in attention lapses in sleep-deprived and well-rested adults. Those studies have focused on individual brain regions, rather than integrated brain networks, and have overlooked adolescence, a period of ongoing brain development and endemic short sleep. This study used functional MRI (fMRI) and a contemporary analytic approach to assess time-resolved peri-stimulus response of key brain networks when adolescents complete the PVT, and test for differences across attentive versus inattentive periods and after short sleep versus well-rested states. Healthy 14-17-year-olds underwent a within-subjects randomized protocol including 5-night spans of extended versus short sleep. PVT was performed during fMRI the morning after each sleep condition. Event-related independent component analysis (eICA) identified coactivating functional networks and corresponding time courses. Analysis of salient time course characteristics tested the effects of sleep condition, lapses, and their interaction. Seven eICA networks were identified supporting attention, executive control, motor, visual, and default-mode functions. Attention lapses, after either sleep manipulation, were accompanied by broadly increased response magnitudes post-stimulus and delayed peak responses in some networks. Well-circumscribed networks respond during the PVT in adolescents, with timing and intensity impacted by attentional lapses regardless of experimentally shortened or extended sleep.

Evening types demonstrate reduced SSRI treatment efficacy.

Selective serotonin reuptake inhibitors (SSRIs) have a profound effect on the circadian system's response to environmental light, which may impact treatment outcomes for patients depending on their habitual light exposure patterns. Here, we investigated the relationship between time-of-day preference, depressive symptoms and self-reported antidepressant treatment response. Evening types reported having taken a higher number of antidepressant medications in the previous 5 years and lower SSRI efficacy than morning types. While undergoing SSRI treatment, evening types also reported more depressive symptoms and suicidality. It is concluded that time-of-day preference may prove informative in predicting SSRI treatment responses.

A protocol for isolating Xenopus oocyte nuclear envelope for visualization and characterization by scanning electron microscopy (SEM) or transmission electron microscopy (TEM).

This protocol details methods for the isolation of oocyte nuclear envelopes (NEs) from the African clawed toad Xenopus laevis, immunogold labeling of component proteins and subsequent visualization by field-emission scanning electron microscopy (FESEM) and transmission electron microscopy (TEM). This procedure involves the initial removal of the ovaries from mature female X. laevis, the dissection of individual oocytes, then the manual isolation of the giant nucleus and subsequent preparation for high-resolution visualization. Unlike light microscopy, and its derivative technologies, electron microscopy enables 3-5 nm resolution of nuclear structures, thereby giving unrivalled opportunities for investigation and immunological characterization in situ of nuclear structures and their structural associations. There are a number of stages where samples can be stored, although we recommend that this protocol take no longer than 2 d. Samples processed for FESEM can be stored for weeks under vacuum, allowing considerable time for image acquisition.

Generation of cell-free extracts of Xenopus eggs and demembranated sperm chromatin for the assembly and isolation of in vitro-formed nuclei for Western blotting and scanning electron microscopy (SEM).

This protocol details methods for the generation of cell-free extracts and DNA templates from the eggs and sperm chromatin, respectively, of the clawed toad Xenopus laevis. We have used this system with scanning electron microscopy (SEM), as detailed herein, to analyze the biochemical requirements and structural pathways for the biogenesis of eukaryotic nuclear envelopes (NEs) and nuclear pore complexes (NPCs). This protocol requires access to female frogs, which are induced to lay eggs, and a male frog, which is killed for preparation of the sperm chromatin. Egg extracts should be prepared in 1 d and can be stored for many months at -80 degrees C. Demembranated sperm chromatin should take only approximately 2-3 h to prepare and can be stored at -80 degrees C almost indefinitely. The time required for assembly of structurally and functionally competent nuclei in vitro depends largely on the quality of the cell-free extracts and, therefore, must be determined for each extract preparation.

Visualization of the nucleus and nuclear envelope in situ by SEM in tissue culture cells.

Our previous work characterizing the biogenesis and structural integrity of the nuclear envelope and nuclear pore complexes (NPCs) has been based on amphibian material but has recently progressed into the analysis of tissue-culture cells. This protocol describes methods for the high resolution visualization, by field-emission scanning electron microscopy (FESEM), of the nucleus and associated structures in tissue culture cells. Imaging by fluorescence light microscopy shows general nuclear and NPC information at a resolution of approximately 200 nm, in contrast to the 3-5 nm resolution provided by FESEM or transmission electron microscopy (TEM), which generates detail at the macromolecular level. The protocols described here are applicable to all tissue culture cell lines tested to date (HeLa, A6, DLD, XTC and NIH 3T3). The processed cells can be stored long term under vacuum. The protocol can be completed in 5 d, including 3 d for cell growth, 1 d for processing and 1 d for imaging.

The effects of total sleep deprivation on cerebral responses to cognitive performance.

We review the findings from a study utilizing functional magnetic resonance imaging (FMRI) to examine the effects of total sleep deprivation (TSD) on verbal learning, arithmetic, and divided attention. For verbal learning and divided attention, TSD was associated with increased activation in the bilateral prefrontal cortex and parietal lobes. Increased sleepiness after TSD and lower levels of memory impairment were correlated with increased activation in specific regions of the prefrontal cortex and parietal lobes, respectively. The arithmetic task led to significantly decreased activation in the bilateral prefrontal cortex and parietal lobes. Based on this and other data, we hypothesize an adaptive cerebral response during cognitive performance following TSD with the specific pattern of adaptation depending on the specific cognitive processes performed. We discuss the need to test the hypothesis in a variety of ways.

Increased cerebral response during a divided attention task following sleep deprivation.

We recently reported that the brain showed greater responsiveness to some cognitive demands following total sleep deprivation (TSD). Specifically, verbal learning led to increased cerebral activation following TSD while arithmetic resulted in decreased activation. Here we report data from a divided attention task that combined verbal learning and arithmetic. Thirteen normal control subjects performed the task while undergoing functional magnetic resonance imaging (FMRI) scans after a normal night of sleep and following 35 h TSD. Behaviourally, subjects showed only modest impairments following TSD. With respect to cerebral activation, the results showed (a) increased activation in the prefrontal cortex and parietal lobes, particularly in the right hemisphere, following TSD, (b) activation in left inferior frontal gyrus correlated with increased subjective sleepiness after TSD, and (c) activation in bilateral parietal lobes correlated with the extent of intact memory performance after TSD. Many of the brain regions showing a greater response after TSD compared with normal sleep are thought to be involved in control of attention. These data imply that the divided attention task required more attentional resources (specifically, performance monitoring and sustained attention) following TSD than after normal sleep. Other neuroimaging results may relate to the verbal learning and/or arithmetic demands of the task. This is the first study to examine divided attention performance after TSD with neuroimaging and supports our previous suggestion that the brain may be more plastic during cognitive performance following TSD than previously thought.

Altered brain response to verbal learning following sleep deprivation.

The effects of sleep deprivation on the neural substrates of cognition are poorly understood. Here we used functional magnetic resonance imaging to measure the effects of 35 hours of sleep deprivation on cerebral activation during verbal learning in normal young volunteers. On the basis of a previous hypothesis, we predicted that the prefrontal cortex (PFC) would be less responsive to cognitive demands following sleep deprivation. Contrary to our expectations, however, the PFC was more responsive after one night of sleep deprivation than after normal sleep. Increased subjective sleepiness in sleep-deprived subjects correlated significantly with activation of the PFC. The temporal lobe was activated after normal sleep but not after sleep deprivation; in contrast, the parietal lobes were not activated after normal sleep but were activated after sleep deprivation. Although sleep deprivation significantly impaired free recall compared with the rested state, better free recall in sleep-deprived subjects was associated with greater parietal lobe activation. These findings show that there are dynamic, compensatory changes in cerebral activation during verbal learning after sleep deprivation and implicate the PFC and parietal lobes in this compensation.

Sleep deprivation-induced reduction in cortical functional response to serial subtraction.

Thirteen normal volunteers were studied with fMRI during arithmetic performance after a normal night of sleep and following sleep deprivation (SD). Aims included determining whether the prefrontal cortex (PFC) and the parietal lobe arithmetic areas are vulnerable to the effects of SD. After a normal night of sleep, activation localized to the bilateral PFC, parietal lobes and premotor areas. Following SD, activity in these regions decreased markedly, especially in the PFC. Performance also dropped. Data from the serial subtraction task are consistent with Horne's PFC vulnerability hypothesis but, based on this and other studies, we suggest the localized, functional effects of SD in the brain may vary, in part, with the specific cognitive task.

The sleep of abstinent pure primary alcoholic patients: natural course and relationship to relapse.

Sleep in male pure primary alcoholic inpatients was examined at a mean of 16 days (n = 29), 19 weeks (n = 29), 14 months (n = 9), and 27 months (n = 4) of abstinence. Results were as follows: (1) the sleep of abstinent alcoholic patients is short, fragmented, and shallow early in abstinence; (2) a patient's sleep improves slowly over at least the first year of abstinence; however, (3) some facets of a patient's sleep remain abnormal even after 27 months of abstinence; (4) insomnia and sleep fragmentation after approximately 5 months of abstinence may be related to relapse by 14 months. The mechanism underlying the relationship between sleep and withdrawal in alcoholic patients is not well understood, and the issue of treating sleep problems as an adjunct to prevention of relapse warrants further investigation.

Alpha sleep and information processing, perception of sleep, pain, and arousability in fibromyalgia.

This study examined the relationship between alpha sleep and information processing during sleep, perception of sleep, musculoskeletal pain, and arousability in patients with fibromyalgia. Patients (n = 20) were allowed to sleep undisturbed for the first 60 minutes of the study to assess amount of alpha sleep and were classified as high or low alpha generators based on quantitative analyses of alpha activity during this period. The groups were compared for performance on two memory tasks, perceptions of polysomnographically-defined sleep and EEG arousals in response to auditory stimuli. Correlations between symptoms of fibromyalgia and alpha activity were also examined. Alpha activity during sleep in fibromyalgic patients was associated with the perception of shallow sleep and an increased tendency to arouse in relation to auditory stimuli. Alpha activity was not associated with increased memory for auditory stimuli presented during sleep, sleep state misperception, or with myalgia symptoms. Alpha sleep appears to be, electrophysiologically, a shallow form of sleep. Our results suggest that it is perceived as such phenomenologically and that it is also associated with increased arousability.

Sustained facial muscle activity during REM sleep and its correlation with depression.

Recently, we proposed that the coupling of cognitive activation and diminished arousal during REM sleep may have a mood regulating effect. Conversely, increased arousal during REM sleep may be associated with mood dysregulation. In this paper, the desensitization model is described, and data are presented on the association between motor activity during REM sleep, wakefulness and severity of depression. Motor activity sleep EEG data and two measures of depressive severity (BDI and HRSD) were obtained from 23 depressed patients. BDI scores were significantly correlated with motor activity only during REM sleep. HRSD scores were correlated with motor activity only during quiet wakefulness. These findings are consistent with the theoretical perspective that dysregulation in arousal mechanisms during REM sleep may promote mood disturbance during the depressive episode.