Figure 1
The experimental protocols. The experimental protocols are shown with an ordinate of 1 line per day and an abscissa of 24 h from midnight to midnight. Volunteers arrived in the laboratory at 09:30 on day 1. The ultra-short sleep-wake cycle, consisting of 30 min for sleep (black bars) followed by 60 min for wake (shaded bars) began at 10:30 and continued for 4.7 to 5.6 days. Three consecutive treatments (3 h bright light, yellow areas) were commenced after 38–54 h of baseline at one of 8 times. Circadian phase was assessed during the final 24 h of baseline preceding the first experimental treatment and for 24 h starting 6 h after the last treatment.
Figure 2
The light spectrum of the bright light treatment. The spectral content of the bright light treatment measured at eye level (standing) was averaged for the 3 subject rooms. The abscissa is wavelength in nanometers.
Figure 3
Profiles of aMT6s excretion. Examples of urinary aMT6s interpretation are plotted for one male participant, age 61 years (A and C), and one female participant, age 28 years (B and D). Panels A and B plot aMT6s (ng/h) in blue longitudinally during the two segments of continuous collection used for baseline and post-treatment phase assessment. The abscissa is h from the midnight commencing the first laboratory day (broken axis to omit 2 days of treatment). A cosine curve was fit to the 24 h immediately prior to the first light pulse (white bar) and again to the last 24 h. The horizontal red dotted lines represent the mesors (fitted means) associated with each cosine. Filled circles show the time of the cosine acrophases before (black), and after (grey) light treatment. Times of aMT6s onsets and offsets are represented respectively by upward and downward pointing arrows (black arrows for baseline and grey arrows for post-treatment.) The light-induced phase shifts in circadian aMT6s profiles are illustrated in panels C and D by replotting both baseline (1, black line) and post-stimulus curves (2, red line) on a noon-to-noon abscissa. In A and C, light given 8–11 PM, elicited phase delays of -5.0, -3.3, and -5.9 h, respectively, in the aMT6s acrophase, onset and offset. In B and D, the light stimulus given 5–8 AM produced phase advances of 1.2, 1.1, and 1.1 h, respectively. Note that the phase shifts were well-demonstrated despite the lower aMT6s excretion in the older participant.
Figure 4
Phase shifts of acrophase of aMT6s rhythm. A. Phase shifts in the aMT6s circadian rhythm resulting from light stimuli are shown for 106 participants. The ordinate shows the shift in h of the aMT6s acrophase, computed as the baseline aMT6s acrophase minus the acrophase after the 3 bright light treatments. Thus negative shifts indicate that the follow-up acrophase was later than the baseline acrophase, i.e., delayed in clock time. The abscissa represents the timing of the midpoints of the 3-h light stimuli, as referenced to the baseline aMT6s acrophase. Stimuli given with an abscissa near 0 were approximately centered at the baseline aMT6s acrophase. Black circles represent phase shifts of individual young adult participants, and red triangles represent phase shifts of older participants. The solid black horizontal line shows the mean of all points, approximating the phase shift resulting from the circadian free-running component. The black dashed and red lines represent the trends from 5-point moving averages for the young and older groups. Rectangles illustrate the average actigraphic home sleep times for the young and older groups, referenced to their aMT6s acrophases. B. The phase shifts in aMT6s acrophase were averaged to show the mean ± 1 SEM for 2-h bins of time-of-stimulation referenced to the aMT6s acrophase. "The abscissa (Circadian Clock Time) references the midpoint of 3 h light stimuli to the time of the baseline aMT6s acrophase, and then displays the environmental time scale corresponding to when the mean aMT6s acrophase occurred at baseline. Thus the Circadian Clock Time abscissa (Figs 4-6) also represents our best estimate of the mean environmental clock time at which bright light stimuli occurred, adjusted for each participant's baseline circadian phase (aMT6s acrophase or onset). The asterisks illustrate the mean aMT6s acrophase times for the young and older groups.
Figure 5
Phase shifts of onsets of aMT6s rhythm. A. Time shifts in aMT6s onsets are contrasted in young and older groups. The abscissa represents the time of bright light stimuli (midpoint of 3 h pulse) in reference to the time of the aMT6s onset at baseline. Trend lines for each age group represent 5-point moving averages. Relative to baseline aMT6s onsets, the inflection from delays to advances occurred earlier in the older adults. B. The shifts in aMT6s onsets were averaged to show the mean ± 1 SEM shift in onset times for non-overlapping bins of time-of-stimulation referenced to aMT6s onset. The abscissa (Circadian Clock Time) is our best estimate of the mean clock time at which bright light stimuli occurred, adjusted for each participant's aMT6s onset time. Asterisks represent the mean onsets for young and older groups.
Figure 6
Phase shifts of onsets, offsets, and duration of aMT6s rhythm. A. The time shifts for aMT6s onsets (black triangles) and offsets (red triangles) are contrasted, after averaging the data in 2 h bins showing the mean ± 1 SEM shifts in time, referencing time-of-stimulation to aMT6s onset. The abscissa (Circadian Clock Time) represents our best estimate of the mean clock time at which bright light stimuli occurred, adjusted for each participant's aMT6s onset time, as in Figure 5. The asterisks show mean aMT6s onset and offset times for all 106 participants. B. The change in mean aMT6s duration resulting from unequal shifts in aMT6s onset and offset is plotted, using the same bins and abscissa as above.
Table 1
Baseline Characteristics of Participants (Mean ± SD)
| YOUNG ADULT | OLDER | AGE CONTRAST | |
| Wake-up (questionnaire) | 7:31 ± 1:24 | 5:55 ± 1:18 | P < 0.001 |
| Actigraphic wake time | 8:32 ± 2:16 | 6:39 ± 1:07 | P < 0.001 |
| Bedtime (questionnaire) | 23:48 ± 1:03 | 22:51 ± 1:05 | P < 0.001 |
| Bedtime (actigraph) | 00:33 ± 1:31 | 22:41 ± 1:24 | P < 0.001 |
| Sleep log total sleep time | 444 ± 65 min | 406 ± 63 min | P = 0.003 |
| Actigraphic total sleep time | 404 ± 70 min | 385 ± 68 min | NS |
| aMT6s onset | 23:04 ± 1:45 | 22:16 ± 1:46 | P = 0.02 |
| aMT6s acrophase | 3:47 ± 1:37 | 3:18 ± 1:58 | NS |
| aMT6s offset | 08:22 ± 1:28 | 08:20 ± 2:11 | NS |
| Cortisol acrophase | 9:59 ± 1:54 | 8:29 ± 2:51 | P = 0.008 |
| Oral Temperature bathyphase | 04:04 ± 2:10 | 04:44 ± 3:52 | NS |
| PSQI total score | 3.8 ± 2.3 | 3.6 ± 2.7 | NS |
| CESD at intake | 7.0 ± 7.0 | 3.5 ± 3.6 | P = 0.002 |
NS = Not Significant. Temperature bathyphase: the fitted minimum, 180° from the acrophase.
Table 2
PRC Bisection Test Inflection Phases and Clock Times of Inflection
| GROUP | INFLECTION referenced to acrophase | TIME (mean) | D | N | P |
| aMT6s phase shifts referenced to aMT6s acrophase | |||||
| All ages | 21.3° | 4:56 | 36.9 | 106 | <0.0001 |
| Young males | 21.3° | 5:29 | 37.4 | 19 | 0.004 |
| Young females | 45.3° | 6:37 | 38.7 | 31 | 0.007 |
| Older males | 29.1° | 5:23 | 28.8 | 28 | 0.087 |
| Older females | 1.6° | 3:49 | 45.35 | 28 | <0.0001 |
| cortisol phase shifts referenced to baseline aMT6s acrophase | |||||
| All ages | 4.5° | 3:49 | 35.6 | 73 | 0.0005 |
| oral temperature phase shifts referenced to baseline oral temperature acrophase | |||||
| All ages | 159.6° | 3:31 | 55.0 | 96 | <0.0001 |
| lab actigraphic sleep phase shifts referenced to baseline aMT6s acrophase | |||||
| All ages | 10.7° | 4:14 | 38.0 | 102 | 0.03 |
| Shifts in aMT6s onset, offset, and duration, with treatment time referenced to aMT6s acrophase | |||||
| aMT6s onset | 27.5° | 05:21 | 33.9 | 104 | <0.0001 |
| aMT6s offset | 14.1° | 04:28 | 22.5 | 104 | <0.02 |
| aMT6s duration* | 29.1° | 05:28 | 1.06 h | 104 | P < 0.01 |
* The bisection test was modified to determine if changes in aMT6s duration were random in phase