Algorithms for sleep–wake identification using actigraphy: a comparative study and new results

The aim of this study was to investigate two new scoring algorithms employing artificial neural networks and decision trees for distinguishing sleep and wake states in infants using actigraphy and to validate and compare the performance of the proposed algorithms with known actigraphy scoring algorithms. The study employed previously recorded longitudinal physiological infant data set from the Collaborative Home Infant Monitoring Evaluation (CHIME) study conducted between 1994 and 1998 [ http://dccwww.bumc.bu.edu/ChimeNisp/Main_Chime.asp ; Sleep 26 (1997) 553 ] at five clinical sites around the USA. The original CHIME data set contains recordings of 1079 infants <1 year old. In our study, we used the overnight polysomnography scored data and ankle actimeter (Alice 3) raw data for 354 infants from this data set. The participants were heterogeneous and grouped into four categories: healthy term, preterm, siblings of SIDS and infants with apparent life‐threatening events (apnea of infancy). The selection of the most discriminant actigraphy features was carried out using Fisher’s discriminant analysis. Approximately 80% of all the epochs were used to train the artificial neural network and decision tree models. The models were then validated on the remaining 20% of the epochs. The use of artificial neural networks and decision trees was able to capture potentially nonlinear classification characteristics, when compared to the previously reported linear combination methods and hence showed improved performance. The quality of sleep–wake scoring was further improved by including more wake epochs in the training phase and by employing rescoring rules to remove artifacts. The large size of the database (approximately 337 000 epochs for 354 patients) provided a solid basis for determining the efficacy of actigraphy in sleep scoring. The study also suggested that artificial neural networks and decision trees could be much more routinely utilized in the context of clinical sleep search.     

Electromyographic activity at the base and tip of the tongue across sleep–wake states in rats

Obstructive sleep apnea (OSA) patients have elevated tonic and phasic inspiratory activity in the genioglossus and other upper airway muscles during wakefulness; this protects their upper airway from collapse. In this group, sleep-related decrements of upper airway motor tone result in sleep-related upper airway obstructions. We previously reported that in the rat, a species widely used to study the neural mechanisms of both sleep and breathing, lingual electromyographic activity (EMG) is minimal or absent during slow-wave sleep (SWS) and then gradually increases after the onset of rapid eye movement sleep (REMS) due to the appearance of large phasic bursts. Here, we investigated whether sleep–wake patterns and respiratory modulation of lingual EMG depend on the site of EMG recording within the tongue. In nine chronically instrumented rats, we recorded from 17 sites within the tongue and from the diaphragm across sleep–wake states. We quantified lingual EMG in successive 10 s intervals of continuous 2 h recordings (1–3 p.m.). We found that sleep–wake patterns of lingual EMG did not differ between the base and tip of the tongue, and that respiratory modulation was extremely rare regardless of the recording site. We also determined that the often rhythmic lingual bursts during REMS do not occur with respiratory rhythmicity. This pattern differs from that in OSA subjects who, unlike rats, have collapsible upper airway, exhibit prominent respiratory modulation of upper airway motor tone during quiet wakefulness, retain considerable tonic and inspiratory phasic activity during SWS, and show nadirs of activity during REMS.     

Insomnia and accidents: cross‐sectional study (EQUINOX) on sleep‐related home,work and car accidents in 5293 subjects with insomnia from 10 countries

The link between sleepiness and the risk of motor vehicle accidents is well known, but little is understood regarding the risk of home, work and car accidents of subjects with insomnia. An international cross‐sectional survey was conducted across 10 countries in a population of subjects with sleep disturbances. Primary care physicians administered a questionnaire that included assessment of sociodemographic characteristics, sleep disturbance and accidents (motor vehicle, work and home) related to sleep problems to each subject. Insomnia was defined using the International Classification of Sleep Disorders (ICSD‐10) criteria. A total of 5293 subjects were included in the study, of whom 20.9% reported having had at least one home accident within the past 12 months, 10.1% at least one work accident, 9% reported having fallen asleep while driving at least once and 4.1% reported having had at least one car accident related to their sleepiness. All types of accident were reported more commonly by subjects living in urban compared to other residential areas. Car accidents were reported more commonly by employed subjects, whereas home injuries were reported more frequently by the unemployed. Car accidents were reported more frequently by males than by females, whereas home accidents were reported more commonly by females. Patients with insomnia have high rates of home accidents, car accidents and work accidents related to sleep disturbances independently of any adverse effects of hypnotic treatments. Reduced total sleep time may be one factor explaining the high risk of accidents in individuals who complain of insomnia.     

Comparison of apnoea–hypopnoea index and oxygen desaturation index when identifying obstructive sleep apnoea using type‐4 sleep studies

The concordance of different indices from type‐4 sleep studies in diagnosing and categorising the severity of obstructive sleep apnoea is not known. This is a critical gap as type‐4 sleep studies are used to diagnose obstructive sleep apnoea in some settings. Therefore, we aimed to determine the concordance between flow‐based apnoea–hypopnoea index (AHI_flow50%) and oxygen desaturation index (ODI_3%) by measuring them concurrently. Using a random sub‐sample of 296 from a population‐based cohort who underwent two‐channel type‐4 sleep studies, we assessed the concordance between AHI_flow50% and ODI_3%. We compared the prevalence of obstructive sleep apnoea of various severities as identified by the two methods, and determined their concordance using coefficient Kappa(κ). Participants were aged (mean ± SD) 53 ± 0.9 years (48% male). The body mass index was 28.8 ± 5.2 kg m^?2 and neck circumference was 37.4 ± 3.9 cm. The median AHI_flow50% was 5 (inter‐quartile range 2, 10) and median ODI_3% was 9 (inter‐quartile range 4, 15). The obstructive sleep apnoea prevalence reported using AHI_flow50% was significantly lower than that reported using ODI_3% at all severity thresholds. Although 90% of those with moderate–severe obstructive sleep apnoea classified using AHI_flow50% were identified by using ODI_3%, only 46% of those with moderate–severe obstructive sleep apnoea classified using ODI_3% were identified by AHI_flow50%. The overall concordance between AHI_flow50% and ODI_3% in diagnosing and classifying the severity of obstructive sleep apnoea was only fair (κ = 0.32), better for males (κ = 0.42 [95% confidence interval 0.32–0.57] versus 0.22 [95% confidence interval 0.09–0.31]), and lowest for those with a body mass index ≥ 35 (κ = 0.11). In conclusion, ODI_3% and AHI_flow50% from type‐4 sleep studies are at least moderately discordant. Until further evidence is available, the use of ODI_3% as the measure of choice for type‐4 sleep studies is recommended cautiously.     

CBT for late‐life insomnia and the accuracy of sleep and wake perceptions: Results from a randomized‐controlled trial

Subjective and objective estimates of sleep are often discordant among individuals with insomnia who typically under‐report sleep time and over‐report wake time at night. This study examined the impact and durability of cognitive‐behavioural therapy for insomnia on improving the accuracy of sleep and wake perceptions in older adults, and tested whether changes in sleep quality were related to changes in the accuracy of sleep/wake perceptions. One‐hundred and fifty‐nine older veterans (97% male, mean age 72.2 years) who met diagnostic criteria for insomnia disorder were randomized to: (1) cognitive‐behavioural therapy for insomnia (n = 106); or (2) attention control (n = 53). Assessments were conducted at baseline, post‐treatment, 6‐months and 12‐months follow‐up. Sleep measures included objective (via wrist actigraphy) and subjective (via self‐report diary) total sleep time and total wake time, along with Pittsburgh Sleep Quality Index score. Discrepancy was computed as the difference between objective and subjective estimates of wake and sleep. Minutes of discrepancy were compared between groups across time, as were the relationships between Pittsburgh Sleep Quality Index scores and subsequent changes in discrepancy. Compared with controls, participants randomized to cognitive‐behavioural therapy for insomnia became more accurate (i.e. minutes discrepancy was reduced) in their perceptions of sleep/wake at post‐treatment, 6‐months and 12‐months follow‐up (p < .05). Improved Pittsburgh Sleep Quality Index scores at each study assessment preceded and predicted reduced discrepancy at the next study assessment (p < .05). Cognitive‐behavioural therapy for insomnia reduces sleep/wake discrepancy among older adults with insomnia. The reductions may be driven by improvements in sleep quality. Improving sleep quality appears to be a viable path to improving sleep perception and may contribute to the underlying effectiveness of cognitive‐behavioural therapy for insomnia.     

Integrative body–mind–spirit intervention for concurrent sleep and mood disturbances: sleep‐specific daytime functioning mediates sleep and mood improvements

Sleep disturbances and depressive symptoms are associated closely with daytime dysfunctions, yet few studies have investigated their temporal relationship in a randomized controlled trial. We investigated the inter‐relationships among sleep, depressive symptoms and daytime functioning following an integrative body–mind–spirit (I‐BMS) intervention. One hundred and eighty‐five participants (mean age 55.28 years, 75.1% female) with co‐existing sleep and depressive symptoms were randomized to I‐BMS or waiting‐list. Daytime functioning variables included the daytime dysfunction items of the Pittsburg Sleep Quality Index (PSQI‐day), Somatic Symptom Inventory, Hospital Anxiety Depression Scale and Short Form Health Survey collected at baseline, post‐treatment and 3‐month follow‐up. Sleep and depressive symptoms were measured by the sleep items of the PSQI (PSQI‐night) and Center for Epidemiological Studies Depression Scale (excluding the sleep item) (CESD‐M). Regression and path analyses were used to understand the role of daytime functioning in sleep and depressive symptoms. We found significant group and time effects on almost all daytime variables and significant group × time interactions on PSQI‐day and somatic symptoms. The adjusted regression model showed that CESD‐M was associated with all daytime variables. However, PSQI‐night was associated only with PSQI‐day. Path analyses indicated that PSQI‐day bridged PSQI‐night and CESD‐M in a two‐way direction after the I‐BMS intervention. The conclusion was that, following I‐BMS intervention, improvement in daytime functioning was related predominantly to improvement in depressive symptoms. Night‐time sleep related only to daytime dysfunction that was specific to sleep disturbances. Therefore, ‘sleep‐specific daytime impairment’ could be regarded as a major link from night‐time sleep to depressive symptoms. More studies are required to understand the concept of ‘sleep‐specific daytime impairment’.     

Sleep timing is associated with diet and physical activity levels in 9–11‐year‐old children from Dunedin,New Zealand: the PEDALS study

It is well documented that short sleep duration is associated with excess body weight and poor food intake in children. It has been suggested that sleep timing behaviour may also be an important predictor of weight and other related behaviours, independent of sleep duration; however, there is a lack of research investigating these relationships. The present study investigated sleep timing in association with diet and physical activity levels in 439 children aged 9–11 years old from New Zealand. Sleep and physical activity data were collected using accelerometry, and food choice using a short food‐frequency questionnaire. Participants were classified into one of four sleep timing behaviour categories using the median split for sleep‐onset and ‐offset times. Differences between sleep timing groups for weekly consumption frequency of selected food groups, dietary pattern scores and minutes of moderate‐to‐vigorous physical activity were examined. Children in the late sleep/late wake category had a lower ‘Fruit & Vegetables’ pattern score [mean difference (95% CI): ?0.3 (?0.5, ?0.1)], a lower consumption frequency of fruit and vegetables [mean weekly difference (95% CI): ?2.9 (?4.9, ?0.9)] and a higher consumption frequency of sweetened beverages [mean weekly difference (95% CI): 1.8 (0.2, 3.3)] compared with those in the early sleep/early wake category. Additionally, children in the late sleep/late wake category accumulated fewer minutes of moderate‐to‐vigorous physical activity per day compared with those in the early sleep/early wake category [mean difference (95% CI): ?9.4 (?15.3, ?3.5)]. These findings indicate that sleep timing, even after controlling for sleep duration, was associated with both food consumption and physical activity.