Fig 1.
Chamber schedule and the impact of meal timing on subjects’ RERs.
See S3 Fig and S2A Table and S1 Data for underlying data and analyses. (A) Protocol for "Breakfast" Session versus "Snack" Session. Subjects experienced two separate 56-h continuous sessions with constant metabolic monitoring by indirect calorimetry, each session lasting 56 h. The Breakfast Session included a breakfast (B), lunch (L), and dinner (D), while the Snack Session contained a lunch, dinner, and late-evening snack meal (S). The late-evening snacks were of equivalent caloric and nutritional value to the breakfast meals (approximately 700 kcal; see S1 Table for details). Note from S2 Fig that the daily phasing of sleep for the subjects prior to entry into the metabolic chamber was the same as the "lights-off" interval during the 56-h time course, so the subjects did not experience a phase shift of their daily cycle when they entered the experimental conditions. (B) Breakfast Session: blue line indicates the average hourly RER over the entire 56-h time course among all 6 subjects when a breakfast, lunch, and dinner were presented. Error bars are the standard deviation. Letters indicate time and type of meals, and gray shaded areas indicate the lights-off periods. Green shaded areas indicate meals that were given at the same time in both Breakfast and Snack Sessions (lunch and dinner). Blue shaded areas indicate when breakfast was given, and gray shading indicates the lights-off period. See S3 Fig for data of all subjects individually. (C) Snack Session: the red line indicates the average hourly RER over the entire 56-h time course among all subjects when a lunch, dinner, and late-evening snack were presented. Red shaded areas indicate when late-evening snacks were given. Breakfasts and late-evening snacks contained the same number of calories and the same lipid, carbohydrate, and protein content (S1A and S1B Table). Error bars are the standard deviation (n = 6). (D) Average difference in RER over the entire 56-h time course for the Breakfast Session subtracted from the Snack Session. Deviation from zero (horizontal black line) indicates where differences in RER occurred between subjects. Error bars indicate standard deviation in the differences. In Panels B, C, and D, times of meals are indicated by letters (B = breakfast, L = lunch, D = dinner, S = late-evening snack), and gray areas are lights-off (sleep) intervals. Mealtimes are shaded as in panels A–C; breakfasts and snacks occurred only in their respective sessions. All RER data were collected minute by minute, and in this figure, the minute-by-minute data were binned and averaged for all 60 values within an hour. Abscissa are clock time. RER, respiratory exchange ratio.
Table 1.
Subjects involved in this study.
Fig 2.
See S3 and S4 Figs and S2A and S2D Table for underlying data and analyses. (A) MR by indirect calorimetry for a representative participant (Subject #3). The data for Subject #3 are plotted as a moving average using 180 data points (= 3 h) after aligning all time points to clock time and integrated on a 24-h scale. (B) Average MR for all subjects plotted modulo-24 h. Data were averaged into 1-h bins with error bars indicating standard deviation. See S4 Fig for data of all subjects' MRs individually plotted. (C) Average hourly pairwise comparison of (breakfast − snack) MR values for all subjects. Error bars indicate 95% confidence intervals and values are based on a mixed-model analysis. Asterisks indicate significant differences (p-value < 0.05) between breakfast and snack values for the indicated 1-h bins. See S2D Table for the hour-by-hour statistical comparison of the Breakfast versus the Snack Sessions. (D) RER (VCO2/VO2) by indirect calorimetry of a representative individual (subject #3). The data for Subject #3 are plotted as a moving average using 180 data points (= 3 h) after aligning all time points to clock time and integrated on a 24-h scale. (E) Average RER for all subjects plotted modulo-24 h. Data were averaged into 1-h bins with error bars indicating standard deviation. See S3 Fig for data of all subjects' RER individually plotted. (F) Average hourly pairwise comparison of (breakfast − snack) RER values for all subjects. Error bars indicate 95% confidence intervals, and values are based on a mixed-model analysis. Asterisks indicate significant differences (p-value < 0.05) between breakfast and snack values for the indicated 1-h bins. See S2A Table for the hour-by-hour statistical comparison of the Breakfast versus the Snack Sessions and S3 Table for a statistical comparison of peak/trough amplitude. All panels: the blue line indicates values during the subjects’ Breakfast Sessions and the red line for the subjects’ Snack Sessions. Shading indicates meals and lights off as in Fig 1B and 1C. Error bars indicate ± standard deviation. MR, metabolic rate; RER, respiratory exchange ratio.
Fig 3.
Meal timing alters substrate oxidation; see S5 and S6 Figs and S2E and S2F Table for underlying data and analyses.
(A) CO data of a representative subject (#3) calculated from indirect calorimetry measurements as described [26,27]. Data are plotted as a 3-h moving average (180-min data points). (B) Average of all subjects for daily CO calculated from indirect calorimetry measurements as described [26,27], and the 56-h time course data are plotted on a modulo-24 h scale. Averaged data for all subjects are organized in 1-h bins. The p-value of 0.130 refers to a pairwise comparison of the average (breakfast − snack) difference values over the full 56-h time course for CO. See S5 Fig for data of all subjects' CO rates plotted individually. (C) Average hourly pairwise comparison of (breakfast − snack) difference CO values for all subjects. Error bars indicate 95% confidence intervals, and values are based on a mixed-model analysis. Asterisks indicate significant differences (p-value < 0.05) between breakfast and snack values for the indicated 1-h bins. See S2E Table for the hour-by-hour statistical comparison of the breakfast versus the snack sessions. (D) LO data of a representative subject (#3) calculated from indirect calorimetry measurements as described [26,27]. Data are plotted as a 3-h moving average (180-min data points). (E) Average of all subjects for daily LO calculated from indirect calorimetry measurements as described [26,27] and the 56-h time course data are plotted on a modulo-24 h scale. Averaged data for all subjects are organized in 1-h bins. The p-value of 0.028 refers to a pairwise comparison of the average (breakfast − snack) difference values over the full 56-h time course for LO. See S6 Fig for data of all subjects' LO rates plotted individually. (F) Average hourly pairwise comparison of (breakfast − snack) difference LO values for all subjects. Error bars indicate 95% confidence intervals, and values are based on a mixed-model analysis. Asterisks indicate significant differences (p-value < 0.05) between breakfast and snack values for the indicated 1-h bins. See S2F Table for the hour-by-hour statistical comparison of the breakfast versus the snack sessions. All panels: the blue line indicates values during the subjects’ Breakfast Sessions and the red line for the subjects’ Snack Sessions. Shading indicates meals and lights off as in Fig 1B and 1C. Error bars indicate ± standard deviation. CO, carbohydrate oxidation; LO, lipid oxidation.
Fig 4.
Schematic: Late-evening snacking interacts with the circadian rhythm of metabolism to inhibit LO.
(A and B) Hour-by-hour oxidation rates for carbohydrates (panel A) and lipids (panel B) in the two sessions. These curves are smoothed versions of the experimental data in Fig 3. (C) Cumulative food intake on the Breakfast versus Snack Sessions. (D and E) Cumulative oxidation rates over the 24-h cycle derived from the curves in panels A and B and the experimental data of Fig 3. Panel D shows cumulative CO, while panel E shows cumulative LO. The horizontal dashed lines indicate the daily total intake of carbohydrates (D) and lipids (E) for comparison with the cumulative respective oxidations. (F) Approximate net relative daily storage of carbohydrates and lipids inferred from the data of Fig 3 and the analyses depicted in the other panels of this figure (arbitrary units). Positive values indicate the extent of substrate accumulation/storage, and negative values indicate the extent of substrate oxidation ("burning"). CO, carbohydrate oxidation; LO, lipid oxidation.