Table 2.
Studies examining sleep spindle features in typical development in relation to motor, memory, or cognition
| First author, year | Sample characteristicsa | Study design, total sleep time mean (SD)b | Sleep recordingc | Sleep measure(s)d | Domain(s), taske | Main findings |
|---|---|---|---|---|---|---|
| Horváth, 2018 | 3 months n=45;15M, 30F sleep analysis n=15 |
Experimental nap: 41 (21.5) min | PSG, 6 EEG channels (F3, F4, C3, C4, O1, O2) | Spectral power (10–15 Hz), automated spindle detection | Memory, habituation (visual paired comparison) | Infants with higher frontal spindle density processed visual information faster (decreased habituation time). |
| Simon, 2017 | 6.5 (6-7) months n=37 sleep analysis n=21 |
Experimental nap: 64 (33.3) min |
PSG, 6 EEG channels (F3, F4, C3, C4, O1, O2) | Spectral power (10-15 Hz) | Memory, encoding & retention (language) | No significant correlations between sigma power and retention. |
| Friedrich, 2017 | 7 (6-8) months, n=107; 60M, 47F sleep analysis n=69 |
Experimental nap: 28 - 49 min | PSG, 8 EEG channels (F3, Fz, F4, C3, C4, P3, Pz, P4) | Spectral power (12-15 Hz), automated spindle detection-number, density, amplitude, mean length | Memory, encoding & retention (formation of word meaning) | Infants with longer naps displayed a primary effect for more generalized words. Infants with a longer duration of N2 displayed a greater inverse in the N400 priming effect. |
| Satomaa, 2020 | 7.9 (7.4 - 9) months n=56; 24M, 32F |
Observational overnight: 543 (50.3) min | PSG, 6 EEG channels (F4, C4, O2, F3, O1, C3) | Spectral power slow: 10-12.5 Hz fast: 12.5-15 Hz | Fine motor, cognitive
ability (Bayley-III) |
Right occipital slow sigma was associated with fine motor ability. Left central and occipital slow sigma correlated with cognitive ability. |
| Friedrich, 2015 | 12 (9-16) months n=90; 52M, 38F sleep analysis n=36 |
Experimental nap: 45 (21) min | PSG, 6 EEG channels (F3, F4, C3, C4, P3, P4) | Spectral power (10-15 Hz), automated spindle detection-density, mean peak-to-peak amplitude, mean length | Memory, retention & generalization (object-word pairs) | Spindle power was associated with higher generalization for specific words (N400). |
| Friedrich, 2019 | 14-16 months n=30; 15M, 15F |
Experimental nap learn: 63 (23.7)
min no-learn: 55 (19.8) min |
PSG, 8 EEG channels (F3, Fz, F4, C3, C4, P3, Pz, P4) | Spectral power, automated spindle detection-number, density, amplitude | Memory, encoding (object-word pairs) | The number and density of centroparietal fast spindles were associated with encoding for object word pairs. When similar unknown words were not generalized, there was higher spindle activity. |
| Friedrich, 2020 | 14-17 months n=60; 34M, 26F sleep analysis n=30 |
Experimental nap: 57 (21.8) min | PSG, 8 EEG channels (F3, FZ, F4, C3, C4, P3, PZ, P4) | Automated spindle detection-spindle number, density, amplitude, length | Memory, encoding & retention (object-word pairs) | Infants who napped after encoding retained more object-word pairs. Infants with higher frontal fast spindle (13-15 Hz) amplitude had better memory for the context of a word. |
| Page, 2018 | 20 (12-30) months n=30:14M, 16F |
Observational nap: 77 min |
128 EEG channels | Spectral power (10-17 Hz), automated spindle detection-frequency, amplitude, duration, and density | Fine motor, cognitive ability (MSEL) | Spindle density, duration, and frequency changed with age. When controlling for age, spindle features were not associated with fine motor or cognitive ability. |
| Kurdziel, 2013 | 46.8 (36-67) months n=40: 5M, 31F sleep analysis n=14 |
Experimental nap: 73 (19.9) min | PSG, 4 EEG channels (F3, F4, C3, C4) | Spectral power (9-15 Hz), manual spindle detection- peak frequency, density, amplitude | Memory, delayed & immediate recall (Visuo-spatial learning) | Spindle density showed a negative correlation with immediate recall and a positive correlation with delayed recall. |
| Doucette, 2015 | 4.3 years n=10: 5M, 5F |
Observational overnight: 545 (77.2) min | 128 EEG channels | Spectral power, manual detection slow: 10-13 Hz fast: 13.25-17 Hz | Cognitive processing, reaction time (computer task) | Higher parietal slow sigma power was associated with faster processing speed. |
| Lokhandwala, 2021 | 51.2 (36-71) months n=22: 15M, 7F |
Experimental nap: 94 (13.2) min | PSG, 32 EEG channels | Automated spindle detection (C3) | Memory, encoding & recall (storybook task) | Sleep spindle density was not associated with post-nap, nor 24-hour recall of a storybook sequence. |
| Joechner, 2021 | 70 months n=24:11M, 13F |
Experimental overnight base: 586
min learn:592 min |
PSG, 7 EEG channels (F3, F4, C3, Cz, C4, Pz, Oz) | Spectral analysis, (9-16 Hz), time frequency analysis, automated spindle detection- frequency, density, amplitude | Memory, encoding & recall (scene-word) | Increased fast centroparietal spindle density was associated with memory gains for low-quality memories. Slow frontal spindle density, amplitude and child’s age were related to memory maintenance for medium-quality memories. |
| Ujma, 2016 | 6.2 (3.8-8.4) years n=28:13M, 15F |
Observational overnight: 540 (45.2) min | PSG, 19 or 32 EEG channels | Automated spindle detection-individual average slow and fast spindle frequency density, duration, amplitude | Cognitive ability (Raven CPM) |
Age and spindle amplitude were correlated. When age was controlled, slow and fast spindle amplitude were associated with cognitive ability in females. |
| Chatburn, 2013 | 8.2 (4.1-12.7) years n=27:13M, 14F |
Observational overnight: 438 (34.4) min | PSG, 2 EEG channels (C3, C4) | Automatic and manual detection-number, frequency, density, duration | Fine motor, sensorimotor (NEPSY); Cognitive ability (Stanford Binet Intelligence Scale) | Mean central spindle frequency negatively correlated with fine motor ability. Number of fast spindles was negatively associated with sensorimotor. Fast spindle density was negatively associated with fine motor and sensorimotor. Mean central spindle frequency was negatively correlated with nonverbal working memory and planning domains. Fast spindle density was associated with better narrative memory. |
| Gruber, 2013 | 8.7 (7-11) years n=29:15M, 14F |
Observational overnight: 513 (58.1) min | PSG, 8 EEG channels (F3, F4, C3, C4, P3, P4, O1, O2) | Bandpass filter (11-15 Hz) Automated spindle detection-frequency, density, amplitude, duration |
Cognitive ability (WISC-IV) | Spindle frequency negatively correlated with reasoning and memory domains. Other spindle features were not associated with cognitive ability. |
| Hoedlmoser, 2014 | 9.4 (8-11) years n=54: 29M, 25F |
Observational overnight base: 559.4
min learn: 554.3 min |
PSG, 12 EEG channels | Automated & semi-automated spindle detection-mean spindle activity (duration x amplitude) | Memory, cued & delayed recall (word pair); Cognitive ability (WISC-IV) | Sleep spindle activity was related to learning efficiency, but not associated with improvement in declarative memory. Higher slow spindle activity correlated with higher cognitive abilities. |
| Sulkamo, 2021 | 9.6 (8.9-10.8) years n= 17: 7M, 10F |
Observational overnight: 527 min | PSG, 6 EEG channels (Fp1, Fp2, C3, C4, O1, O2) | Automated spindle detection-local, bilateral, and diffuse spindle density | Fine motor & sensorimotor (NEPSY); psychomotor, speed & accuracy (CogniSpeed); cognitive ability (WISC-III) | Spindles were not associated fine motor or sensorimotor. Diffuse and bilateral central spindles were negatively associated with reaction time and accuracy. Local central spindles were positively associated with accuracy. Bilateral central spindle density was associated with object assembly. |
| Zinke, 2017 | 10 (8-12) years n=25: 11M, 14F |
Experimental overnight: 576 (8.5) min | 9 EEG channels (F3, Fz, F4, C3, Cz, C4, P3, Pz, P4) | Spectral power (11-15 Hz) number, density | Memory accuracy, speed & retrieval (serial reaction time test); Vigilance, reaction time (psychomotor vigilance task) | Spindle number and density were positively correlated with sequence recall. Mean sigma density was associated with retrieval time. |
| Geiger, 2011 | 10.5 (9.1-12.5) years n=14: 8M, 6F |
Observational, overnight base: 442.2
min post: 477.3 min |
128 EEG channels | Spectral analysis, individual relative sigma | Cognitive ability (WISC-IV) | Sigma power was positively correlated with full-scale and fluid intelligence. Individual spindle peak frequency was negatively correlated with full scale intelligence. |
| Geiger, 2012 | 10.5 (9.1-12.5) years n=14: 8M, 6F |
Observational overnight sleep: 457 (28.3) min | 128 EEG channels | Spectral power slow:10-13.25 Hz fast: 15.5-20 Hz |
Cognitive ability (WISC-IV) | Power (10-13.25 Hz; 15.5-20 Hz) in central and
parietal areas was negatively correlated with full-scale
intelligence. Power (13-20 Hz) in frontal and parietal regions showed a positive associated with fluid intelligence. |
| Astill, 2014 | 10.7 years n=30:11M, 19F | Experimental overnight: 432 (4.6) min | PSG, 2 EEG channels (FPz, Cz) | Spindle detection-amplitude, density, duration | Memory, speed & accuracy (finger tapping) | More slow spindles were associated with the overnight improvement in accuracy. Children with increased fast spindle density had the fastest performance. |
Studies are ordered by mean participant age.
a
Sample characteristics- Mean age (range, when provide); sample size, sex: M, males; F, females; Unless indicated, the reported study sample reflects the numbers of participants included in the sleep analysis (EEG/PSG).
b
Study design, total sleep time- Research studies are listed as experimental or observational, and followed by a description of when sleep was collected (nap or overnight sleep), and if the study collected multiple sleep recordings this is also shown. The total sleep time is shown in minutes, and listed as mean and standard deviation (when provided) or the range.
c
Recording setting- PSG, polysomnography; EEG, electroencephalogram, EEG channels used in the analysis
d
Sleep measures- Only the sleep spindle features relevant for this review are reported.
e
Behavioral measures- Only motor, memory and/or cognitive domains relevant for this review are reported. Bayley-III, Bayley Scales of Infant and Toddler Development; MSEL, Mullen Scales of Early Learning; Raven CPM, Raven’s Colored Progressive Matrices; NEPSY, Neuropsychological Developmental Assessment; WISC-IV, Wechsler Intelligence Scale for Children; WISC-III, Wechsler Intelligence Scale for Children