Sleep Spindle Characteristics in Children with Neurodevelopmental Disorders and Their Relation to Cognition
Table 1
(a) Characteristics of sleep spindles and associations between sleep spindles and cognitive function in children with attention deficit hyperactivity disorder (ADHD) versus typically developing children. (b) Characteristics of sleep spindles and associations between sleep spindles and cognitive function in children with autism spectrum disorder (ASD) versus typically developing children. (c) Characteristics of sleep spindles and associations between sleep spindles and cognitive function in children with developmental dyslexia (DD) versus typically developing children. (d) Characteristics of sleep spindles and associations between sleep spindles and cognitive function in children with intellectual disability (ID) versus typically developing children. (e) Characteristics of sleep spindles and associations between sleep spindles and cognitive function in typically developing (TPD) children.
ADHD participants: more REM sleep and shorter SWS latency
() Overall reduced recognition accuracy in ADHD group. () Enhanced recognition accuracy in the sleep condition, but greater in control compared to ADHD
Positive association between () duration of non-REM sleep and () slow oscillation power and sleep-associated memory consolidation in healthy controls but not in ADHD
ASD participants: () More transitions from a waking epoch to REM sleep compared to healthy controls. () Made less transitions from a S2 epoch to REM sleep compared to healthy controls. () Lower S2 spindle density.
Sustained and selective attention (RT) working memory (span score) Declarative episodic memory (# of recalled figures) Sensory-motor procedural memory (contact time, # of trials) Procedural memory (time)
() Lower REM density. () Lower SWS% () Higher S1% in lower S2 spindle density (C3) in AS compared to controls
Poorer performance in ASD compared to controls on sustained attention working memory. Sensory-motor procedural memory
() Positive association in controls but not in ASD between the SWS% and declarative memory immediate recall () Negative associations in both groups between SWS% and learning the sensory-motor procedural memory task () Negative associations in ASD but not in healthy controls between S1% and declarative memory immediate recall; sensory-motor and cognitive procedural memory tasks and selective attention () S2 spindle density (C3) is negatively correlated with RT in the selective attention only in healthy controls () S2 spindle density (C3) negatively correlated with sensory-motor procedural memory during the learning phase only in ASD group
Experimental Case-Control Repeated Measures Design training session in the morning 30 minutes after habitual wake time after the night of sleep recording and a testing session in the evening
PSG. EEG (F1, F2, C3, Cz, C4, O1, and O2), EOG, and EMG.
Spectral power calculated for slow wave oscillation (0.5–1 Hz), delta (1–4 Hz), theta (4–7 Hz), alpha (8–11 Hz), sigma (12–15 Hz), and beta (16–20 Hz) frequency ranges. S2 sleep spindles (Cz): 12–15 Hz, lasting 0.5–2.0 s, no amplitude criteria
Sleep-dependent memory consolidation (difference in performance (in %) between training and testing phases)
Less REM sleep in ASD compared to control group
() In both conditions ASD group performed more poorly than control group () Both groups recalled better in the sleep condition compared to the wake condition
Slow oscillation power correlated with sleep dependent memory consolidation in the ASD group only
PSG. EEG (C3, C4, F3, F4, O1, and O2), EOG, and EMG.
S2 sleep spindles
Intelligence WISC-III. Performance IQ (PIQ), verbal IQ (VIQ), and full scale IQ (FSIQ) scores.
Lower S2 spindle density (Fp1) in ASD group compared to control group
() VIQ positively correlated with S2 spindle duration (C4) in control group but not in ASD group () VIQ is associated with spindle density only in ASD group () FSIQ is associated with spindle density only in ASD group
Power spectra calculated for delta, theta, alpha, sigma, and beta Sleep spindles Scored during N2.
() Memory and learning transfer () Reading test (speed and accuracy) () Writing test (accuracy) () Intelligence WISC-IIIR
Compared to healthy controls DD participants had the following: () Less stage-shifts per hour () Lower N2% () Less N3% () Increase in power of frequency bands 0.5–3 Hz and 11-12 Hz during N2 and 0.5–1 Hz during N3 () Increased spindle density
Mean IQ DD: FSIQ: 93.9 VIQ: 90 PIQ: 98.9
Only in DD group: (1) Sigma power band in N2 was positively correlated with the memory and learning transfer reading test reading test () Spindle density was positively correlated with performance on the word reading test
PSG. EEG (F3-C3, F4-C4, F3-A1, F4-A2, C3-A1, C4-A2, O1-A1, and O2-A2), EOG, EMG, and ECG
S2 sleep spindles
DQ IQ
() 42% had a ratio of more than 2.00 (# spindles longer than 0.4s/# of spindles shorter than 0.4s) () 15.6% had a ratio of 1.99–1.00 and 6% had a ratio of 0.99–0.50 () 22% had a ratio of less than 0.50 () 14% had no sleep spindles
() Subjects with a ratio of less than 0.50 had significantly lower DQs than subjects with a ratio of more than 2.00 () Subjects with no sleep spindles at all had significantly lower DQs than subjects with a ratio of more than 2.00 () Subjects with a ratio of less than 0.50 or no sleep spindles had significantly lower DQs than subjects with a ratio of 1.99–1.00
CP: 4 months–5 years Non-CP patients: 4 months–12 years
CP: 23 Non-CP patients: 39
Descriptive study
PSG (F3-A1, C3-A1, and O1-A1), EOG, ECG, respiration, and EMG
Not specified
DQ IQ
() 13% of non-CP patients had no spindles or extremely low incidence of spindles () 8% of CP cases showed no sleep spindles () 22% of CP cases showed extremely low incidence of spindles () 4.3% of CP cases showed extreme spindles
CP patients with indistinguishable sleep stages had lower DQs than CP patients with normal EEG patterns and non-CP patients with normal EEG patterns
PSG. EEG (F3-C3, F4-C4, F3-A1, C3-A1, F4-A2, and C4-A2), EOG, ECG, respiration, and EMG
Spindles measured during S2. Extreme spindles: 9-10 c/s, 50–120 μV
DQ
() 2.3% had no REM sleep () 30% had no sleep spindles () 4.6% had high voltage fast activity (20–30 c/s, 100–200 uV) without sleep spindles in wakefulness, S1, S2, and REM () 7% had low voltage activity without spindles throughout nocturnal sleep () 2.3% had indistinguishable stages with the presence of sleep spindles () 2.3% had extreme spindles
() Patients with abnormal EEG patterns throughout S1–S4 had lower DQs than patients with abnormal EEG patterns only in S1-S2 and patients with normal EEG patterns () Patients with abnormal EEG patterns in S1-S2 had lower DQs than patients with normal EEG patterns
Average duration of slow waves at FPz was negatively correlated with density of fast spindles and positively correlated with density of slow
Accuracy was better if the interval contained sleep
() Higher SWS% was associated with increases accuracy () Children with higher density of slow spindles had lower overall speed and accuracy whereas children with higher density of fast spindles had higher overall speed () Children with higher density of slow spindles showed a stronger overnight increase in accuracy but not speed. () Children with a longer average duration of slow waves had stronger overnight increase in accuracy but not speed.
All intelligence and neurocognitive functioning was in the normal range
() Mean central spindle frequency was negatively correlated with nonverbal working memory, planning, and fine motor functioning () Fast spindle density was positively correlated with narrative memory and negatively correlated with sensorimotor functioning and fine motor functioning
Correlational study Sleep was recorded over 2 nights separated by 1 or 2 weeks in a sleep laboratory
PSG. EEG 10–20 International system EOG and EMG
Mean all-night power spectra of stage N2 sleep.
Intelligence WISC-IV Attention task (RT)
Subjects had normal sleep structure
All results were in the normal range
() Sleep spindle peak frequency correlated negatively with full scale IQ () Individual relative sigma power correlated positively with full scale IQ and fluid IQ