People with nothing but their own anecdotal experience about the subject trying to (and failing at) criticizing a scientific paper. No one has pulled off a "correlation != causation", but most other usual suspects are already there.
Disclaimer: 1) I work in neuroscience , but not in sleep research. This (hopefully) makes me qualified to assess the seriousness of the methods, even though I don't have much background in circadian rhythms. 2) I'm biased in this case since I know personally the first author of the paper and some of the other researchers involved (and they are among the smartest and more conscientious people that I know of).
If there were any methodology error lay people would be able to spot in the paper, it wouldn't have been published in any respected journal in the field (and most people who commented here are lay people regarding (sleep) science). This is Science Magazine... They have the best reviewers in the world ("best" as in "smarter than you can probably imagine").
There are shortcomings inherent to this kind of study, of course. You cannot control experimental parameters like you can in physics or chemistry. That's why you have to 1) build your research on solid ground (i.e. solid results already published and reproduced) and 2) be as careful as possible in designing your protocol in order to have your bases covered. Schmidt et al. used the gold standard, they're frankly out of reach of such "low hanging" criticism.
</rant>
I enjoin you to have a look at the paper and the supporting material.
_
Full text: http://www.innovatieforganiseren.nl/wp-content/uploads/2009/...
Supporting Online Material: http://www.sciencemag.org/cgi/data/324/5926/516/DC1/1
_
Here's the beginning of the latter:
_
1. Material and methods
Subjects.
[Ethical committee disclaimer]. [All subjects] were screened for morningness or eveningness according to their timing preferences as defined by two questionnaires (MEQ (1) and MCTQ (2)). The two groups were matched according to age, sex and educational level and did not differ in their anxiety and depression levels as well as in sleep quality and day time sleepiness (all ps > 0.1; Table S1). Morning and evening types significantly differed in their scores on the two chronotype questionnaires (MEQ and MCTQ). Exclusion criteria were reports of medical, psychiatric and sleep disorders, medication or drug consumption, alcohol abuse, excessive caffeine consumption or physical activity, shift work within the three past months, and transmeridian travel or disturbances in the sleep-wake cycle within one month before the experiment.
Design and Procedures.
An overview of the study design is illustrated in Fig. 1 of the main text. Individual times were scheduled according to each volunteer’s preferred sleep and wake timing. Criteria for such timing preferences included sleep schedules adopted on free days as assessed by the MCTQ (2) and after interviewing the subject to ensure that the scheduled timing was as close as possible to the schedule that he or she would spontaneously adopt. In a second step, the screened subjects came to the sleep facility for a habituation night. After this night, they were asked to follow the sleep schedule (± 30 minutes) they would spontaneously adopt while free from any social and professional constraints. Target bedtimes and wake times were determined for a sleep duration of about 8 h (± 30 minutes). To assess the subjects’ compliance to the selected rest-activity patterns, motor activity of the non-dominant arm was recorded using actimeters the week prior to the experimental sessions along with sleep-wake logs. After this week under actimetry recording, subjects came to the sleep laboratory for 2 consecutive nights. The precise schedule of each session was individually adapted according to the subject’s habitual bedtime on the basis of the mean timing of the subject’s sleep midpoint derived from actimetric data of the preceding week.
Subjects reported to the laboratory 7 hours before habitual lights off on day 1. After the hook-up of the electrodes, they continuously stayed under controlled conditions in dim light (< 10 lux) in order to avoid the influence of bright light on circadian rhythmicity parameters (3) and in the aim to equalize pre-scan conditions between subjects. Subjective sleepiness (Visual Analogue Scale (VAS) and Karolinska Sleepiness Scale (KSS (4))) and objective vigilance (a modified version of the PVT (5)) were assessed at hourly intervals while awake. Furthermore, hourly collected saliva samples were assayed for melatonin using a direct double-antibody radioimmunoassay validated by gas chromatography-mass spectroscopy with an analytical at least detectable dose of 0.65 pg/ml (6). Circadian phase was estimated by the calculation of the mid-range crossing time of salivary melatonin (7, 8). For each individual curve, the maximum value and the minimum value was averaged (mid-range value pg/ml) and taken to determine the mid-range crossing time (time of day, h) on the abscissa.
Polygraphic data (see below) were recorded during the nights preceding fMRI sessions. After lights off, subjects were allowed to sleep for 8 hours. Then, 1.5 (morning session) and 10.5 (evening session) hours after wake up of scheduled sleep timing, they underwent a fMRI session during the practice of various cognitive tests, including the psychomotor vigilance task on which we focus here. For half of the subjects, the morning session followed the first experimental night and the evening session the second night, whereas for the other half of the volunteers the morning session followed the second experimental night and the evening session the first night. Subjects were allowed to leave the facility between the two experimental nights. They stayed in the laboratory under dim light conditions (<10 lux) for at least 4 hours before the scanning sessions (see dashed line in Fig. 1). They wore protective glasses avoiding excessive light input when going to the scanner room.
The order of selected cognitive tasks was counterbalanced across subjects and sessions. Before the start of the experimental protocol, all subjects underwent a short habituation scan session in order to familiarize them with the noise and the body positions associated with the fMRI environment.