Sleep spindles and sleep consciousness
Joshua J. LaRocque , Francesca Siclari , Bradley R. Postle , and Giulio Tononi
2,3
2
Neuroscience Training Program , Departments of Psychiatry and Psychology , University of Wisconsin - Madison
Sleep spindles are negatively
associated with sleep consciousness
No conscious
experience (NCE)
Conscious experience
with recall (CE)
Conscious experience
without recall (CWR)
1 uV2/Hz
0.5
n=82
n=45
- Two healthy participants
- 16 overnight hd-EEG recordings (256 electrodes)
- 290 awakenings (222 in N2 and N3)
- Many awakenings, few subjects to account for interindividual variation in sleep
consciousness and EEG
- Assessment of sleep consciousness - categories:
• No sleep consciousness
• Sleep consciousness without recall of content
• Sleep consciousness with recall of content
- Assessment of richness and length of dream
- Time frequency transform of hd-EEG signal during two minutes preceding awakening
- Spindle detection algorithm for identification of individual sleep spindles2
- Multivariate pattern analysis (L2regularized logistic regression) for EEG
signal within sigma band (12-15.5 Hz)
- For spindle correlation analyses,
an anterior medial ROI was used (purple
in ROI schematic)
11
10.5
9.5
9
8.5
0.6
0.5
0.4
0.3
0.2
0.1
−2.5
−2
−1.5
−1
−0.5
0
0.5
1
r = -.24
p = .03
12
11
10
9
8
7
6
9.5
9
8.5
8
No conscious
experience
Conscious
experience
No conscious
experience
Conscious
experience
12
11
10
9
0.8
5
0.6
No conscious
experience
Time before waking (s)
-20
0.4
Left Ant Lat
0.7
0.6
0.5
0.4
0.3
0.2
−2
−1.5
−1
−0.5
0
0.5
Number of spindles vs. estimate of length of
last dream experience
15
r = -.3
p = .018
−1.5
−1
−0.5
0
0.5
1
5
0
1.5
Sleep spindle activity correlates
with quantitative estimates of
the richness and length of the
last dream experience. However,
when both subjective measures
are included in a multiple regression, only the time estimate
remains significantly correlated
with spindle power.
−2
−1.5
−1
−0.5
0
Time estimate (log scale)
0.5
Time since last spindle vs. estimate of
length of last dream experience
60
r = .37
p = .003
50
40
30
20
10
0
−2
−1.5
−1
−0.5
0
0.5
Conclusions
subject1
subject2
- The coordinated thalamocortical activity of sleep spindles disrupts
the ability of the brain to generate consciousness during sleep
- The amount of sleep spindle activity is associated with subjective,
qualitative characteristics of conscious experience during sleep
- Initial results of pattern classification indicate that classification is
more successful with spindle power averaged over long durations
References
1. Tononi, G. Consciousness as Integrated Information: A Provisional Manifesto. Biol Bull 2008; 15(3); 216-242.
2. Ferrarelli F, Huber R, Peterson MJ, Massimini M, Murphy M, Riedner BA, Watson A, Bria P, Tononi G. Reduced Sleep Spindle Activity in
Schizophrenia Patients. Am J Psychiatry 2007;164:483-492.
0.5
Waking
Tone
0.8
Conscious
experience
0.6
0.5
r = -.42
p = .0005
0.9
Time estimate (log scale)
0.7
-40
−2
7
0.8
0.7
1
10
10
8
6
Spindle power vs. estimate of length of last
dream experience
0.1
1.5
Number of spindles vs. estimate of richness
of last dream experience
0
−2.5
0.9
-60
0.7
10
Conscious
experience
No conscious
experience
0.9
-80
0.8
Richness estimate (log scale)
1
-100
r = -.19
p = .09
0.9
10.5
10
8
Subject 2
Initial results from MVPA of EEG spectral power in
spindle range
0.4
-120
ROI schematic
11
Number of spindles
There is a non-significant
trend in an anterior medial
ROI for more spindles
during sleep preceding a
report of no conscious experience (p = 0.08). There
is a similar, non-significant
trend in total spindle duration (p = 0.14). Though
these trends are nonsignificant, they agree
qualitatively with the
spindle power analyses
above.
Subject 1
Total duration of spindles (s)
Methods
1
-0.1
Classification Accuracy
Finally, as a proof of principle, we applied multivariate pattern
analysis to the electroencephalogram in order to predict presence of
conscious experience from the power density in the 11-15 Hz frequency band.
Spindle power vs. estimate of richness of last
dream experience
0.1 uV2/Hz
0
Number of spindles
We also collected subjective reports of features of the conscious experience, including richness and length of the experience. We predicted that sleep spindle activity would correlate with reduced levels
of these subjective reports.
Sleep spindles correlate with
quantitative features of conscious
experience during sleep
15
Total duration of spindles (s)
Sleep spindles are fusiform, 11-15 Hz oscillations characteristic of
non-REM sleep. They are generated by reciprocal interactions between inhibitory thalamic reticular neurons and bursting thalamocortical relay neurons. The highly stereotyped, synchronized neural
activity underlying sleep spindles implies a reduced array of potential thalamocortical firing patterns.1 Therefore, we hypothesized that
sleep spindles would be associated with reduced reports of consciousness during sleep.
0
Difference
(CWR-CE)
Difference
(NCE-CE)
Awakenings with a
report of conscious
experience were preceded by lower spindle
power in an anterior
medial cluster
(p=.003). All topoplots
show 12-15.5 Hz
power from 2 min
prior to waking.
Classification Accuracy
1s
n=95
[email protected]
Spindles power (12-15.5 Hz)
Conscious experience during sleep, also known as dreaming, is a
ubiquitous phenomenon that is incompletely understood. Many
theories of dreaming emphasize a distinction between rapid-eyemovement (REM) sleep and non-REM sleep. However, reports of
conscious experience can be elicited upon awakening from any
stage of sleep. We seek to understand the neural substrates of conscious experience during non-REM sleep.
3
Number of spindles
Background
2
Spindles power (12-15.5 Hz)
1
Number of spindles
2
Time since last spindle (s)
1
Ant Med
Right Ant Lat
Left Post Lat
Post Med
Right Post Lat
Scalp ROI
Wisconsin Rath Distinguished Graduate Fellowship (J.J.L.), Swiss National Foundation Grant No. 139778 (F.S.),
NIH MH064498 (B.R.P.), NIH Mental Health Grant 5P20MH077967 (G.T.)