Figure 1.
EEG re-referencing and filtering.
The same SSO is shown. The upper three traces are the raw EEG (the channels Cz, A1 and A2, all referenced to Fcz) signals. In the middle, Cz signal is shown after re-referencing to the average between the two earlobe potentials (A1 and A2). At the bottom, the re-referred Cz trace, band pass filtered in the range 0.1–4.4 Hz is plotted.
Figure 2.
Signals are scanned and SSOs are detected applying the Detection Criteria. A branch of concurrent SSOs, grouped in a event, are shown in the top left panel (blue and green lines, throughout all panels, the same color code has been used for each type of waves). The thickest blue trace identifies the ‘prototype’ of the event, namely the first SSO according to the temporal occurrence of its negative peak. In the middle panel on the left, instantaneous phases of the signals are plotted. In the bottom panel on the left, the cross-correlations functions between the prototype and the other traces are plotted. Applying the Likeness Constraint (dashed horizontal line), some sub-threshold SSOs are selected (red traces, in the bottom right panel). In fact, in the time-phase plane (middle panel on the right) well expressed and sub-threshold SSOs are superimposed. Also sub-threshold SSO signals are very similar to concurrent well-expressed SSOs (upper, right panel).
Figure 3.
Graphical definition of the extracted morphological features.
Each detected wave was characterized by means of seven features: three amplitude measures (the negative peak to positive peak amplitude - NP amplitude -, the negative peak amplitude - N amplitude - and the positive peak amplitude - P amplitude), two steepness measures (the absolute value of the slopes of the signal between the first zero crossing and the negative peak - slope 1 - and between the negative peak and the second zero-crossing - slope 2), and two different estimation of temporal width (the negative peak to positive peak time - NP time - and the first zero crossing to negative peak time interval - ZN time).
Figure 4.
Origin location map and Extent of propagation distribution.
A) The grand mean of the origin density map calculated from the individual maps is shown. It illustrates the frequency (reported as percentage) of SSO event origin for each electrode site. B) The thickest line represents the average (among subjects) cumulative distribution function (c.d.f.) of the extent of propagation and the two thinner lines correspond to the inter-subject confidence interval (p = 0.95).
Figure 5.
Extent of propagation versus Origin Location.
For the extent of propagation, contrasts between Origin Location areas have been done. On the left, the extent of propagation grand-average map is depicted. In order to obtain a map related to the Origin Location effect, events originating from each origin site (electrode) have been selected and the grand-mean of extent of propagation per electrode has been derived. On the right, the contrast results are graphically reported by a grayscale code. Each step in the grayscale indicates a significant difference (p<0.01) between contiguous tones. The grayscale is ordinal, namely lighter tones correspond to higher extent of propagation values.
Figure 6.
The extent of propagation in the area is shown as function of Detection Location and Sleep Stage. On the left of the first row, the map of detections (% of events affecting each electrode) is depicted; on the right the corresponding contrasts between areas for the Extent of propagation in the area are shown by a grayscale code. Each step in the grayscale indicates a significant difference (p<0.01) between contiguous tones. The grayscale is ordinal, namely lighter tones correspond to higher extent of propagation values. On the left of the second row, the difference between the detections in each stage (for each electrode, percentage of events in SWS minus same in stage 2) is depicted; on the right the corresponding contrasts between stages for the extent of propagation in the area are shown. The extent of propagation in the area is greater in NREM Sleep Stage 2 than SWS, except for SSOs detected in the posterior area. On the right column, a grayscale code indicates for each detection area if feature values are significant greater during SWS (white), or during sleep stage 2 (black) while grey indicates no difference between stages.
Figure 7.
Extent of propagation in the area as function of Origin Location.
On each row, map related to events originating in an area are shown. On the left, the maps of detections (% of events affecting each electrode) are depicted; on the right the corresponding contrasts between areas for the Extent of propagation in the area are shown.
Figure 8.
Continuity of event propagation.
For events originating in 5 exemplary origin sites (F3, F4, Cz, T5, T6), panels depict the grand-average maps of delays (left column) and the grand-average maps of detections. The number of detections at each electrode site was normalized to the number of events originating in the red-dot marked site (normalized units, n.u.). The hemispheric cortical separation does not seem to influence the radial propagation of SSO events.
Figure 9.
For each origin site, the black vector indicates the null hypothesis direction (isotropic radial propagation), while the red one indicates the estimated grand-mean direction and the yellow circular sector depicts the variability (95% confidence interval) between individuals. Data indicate a systematic tendency to drift toward frontal areas.
Table 1.
Individual mean values of each SSO feature.
Table 2.
Between-events effect analysis of SSO features.
Figure 10.
Morphological features as a function of Origin Location.
For features exhibiting an Origin Location effect, as reported in Table 2, contrasts between Origin Location areas is shown. Each row corresponds to a feature. On the left column, the feature grand-average map is depicted. Since maps correspond to the Origin Location effect, events originating from each origin site (electrode) have been selected and the grand-mean, for each electrode, has been derived. On the right column the contrast results are graphically reported by a grayscale code. Each step in the grayscale indicates a significant difference (p<0.01) between contiguous tones. The adopted grayscale is ordinal, namely lighter tones correspond to higher relative values for the feature.
Table 3.
Within-event effects analysis of SSO features.
Figure 11.
Morphological features as a function of Detection Location.
For features exhibiting a Detection Location effect, as reported in Table 3, contrasts between Detection Location areas have been done. In the figure, each row corresponds to a feature. On the left column, the feature grand-average map is depicted. Since maps correspond to the Detection Location effect, SSOs detected in each electrode site have been selected and the grand-mean, for each electrode, has been derived. On the right column the contrast results are graphically reported by a grayscale code. Each step in the grayscale indicates a significant difference (p<0.01) between contiguous tones. The grayscale is ordinal, namely lighter tones correspond to higher relative values for the feature.
Figure 12.
Morphological features as a function of Sleep Stage.
For features exhibiting a significant interaction between Detection Location and Sleep Stage, as reported in Table 3, the figure shows, for each area, comparisons between sleep stages. In the figure, each row corresponds to a feature. On the left column, grand-average maps derived subtracting Sleep Stage 2 values from SWS values are shown. On the right column, a grayscale code indicates for each detection area if feature values are significant greater during SWS (white), or during sleep stage 2 (black) while grey indicates no difference between stages.