Time-restricted light, food and exercise altered diurnal activity and enhanced central clock amplitude.

a-e) Schematics of treatment schedules and double-plotted ambulatory activity (Amb, count/hour) of WT mice. Control (n = 12), time-restricted light (TRL, n = 15), time-restricted feeding and light (TRFL, n = 13), time-restricted exercise and light (TREL, n = 10), time-restricted light, feeding and exercise (LiFE, n = 16). Blue bar: light exposure; grey bar: dark; orange bar: access to food; purple bar: access to running wheels. Numbers indicate Zeitgeber time (ZT, hours after light on). f) Comparison of ambulatory activity between control and TRL, yellow and grey bars on the x axis indicate light and dark. g) Comparison of ambulatory activity of TRL, TRFL and TREL. Orange and purple bars indicate access to food and wheels, yellow and grey bars on the x axis indicate light and dark. h) Comparison of ambulatory activity of TRL and LiFE. i) Daily total ambulatory activity counts (one-way ANOVA with Bonferroni correction, *P<0.05, **P<0.01) j) Food anticipatory activity (total activity during 2 hours before light-off) (Browns-Forsythe test with Dunnett’s T3 correction, *P<0.05). k) illustrations showing ZT times(colored dots) of peak values measured in Per2 luminescence, blank bars indicate light phase, grey bars indicate dark phase. l) Amplitude of Per2 luminescence in SCN cultures from control(n = 10), TRL (n = 5) and LiFE (n = 13) treated mice(mixed-effect model with Bonferroni correction, *P<0.05; **P<0.01). Data are represented as means ± SEM; see statistical details in the method section.

Effects of LiFE treatment on sleep time and structure.

a) Percentage of total sleep in electroencephalogram (EEG) recordings in WT mice (n =12) habituated to control and TRL schedule in 1h bin. Yellow and grey bars on the x axis indicate light and dark. b) Percentage of total sleep in EEG recordings in WT mice (n =12) habituated to TRL and LiFE schedule in 1h bin. Orange and purple bars indicate access to food and wheels. c) Calculations of average daily sleep percentage (paired one-way ANOVA with Bonferroni correction, ***P<0.001). d) Calculations of average sleep percentage in the dark phase (paired one-way ANOVA with Bonferroni correction, ***P<0.001). e) Calculations of average sleep percentage in the light phase (paired one-way ANOVA with Bonferroni correction, *P<0.05). f) Percentage of NREM sleep in EEG recordings in WT mice (n =12) habituated to control and TRL schedule in 1h bin. g) Percentage of NREM sleep in EEG recordings in WT mice (n =12) habituated to TRL and LiFE schedule in 1h bin. h) Calculations of average daily NREM sleep percentage (paired one-way ANOVA with Bonferroni correction, **P<0.01). i) Calculations of average NREM sleep percentage in the dark phase (paired one-way ANOVA with Bonferroni correction, **P<0.01). j) Calculations of average NREM sleep percentage in the light phase. k) Percentage of REM sleep in EEG recordings in WT mice (n =12) habituated to control and TRL schedule in 1h bin. l) Percentage of REM sleep in EEG recordings in WT mice (n =12) habituated to TRL and LiFE schedule in 1h bin. m) Calculations of average daily REM sleep percentage (paired one-way ANOVA with Bonferroni correction, *P<0.05). n) Calculations of average REM sleep percentage in the dark phase. o) Calculations of average REM sleep percentage in the light phase. p) Counts of microawakenings in control, TRL and LiFE treated WT mice. q) Counts of stage shift in control, TRL and LiFE treated WT mice. r) Average bout lengths of NREM sleep in control, TRL and LiFE treated WT mice. s) Average bout lengths of REM sleep in control, TRL and LiFE treated WT mice (paired one-way ANOVA with Bonferroni correction, **P<0.01). t) Average bout lengths of REM sleep in the dark phase. u) Average bout lengths of REM sleep in the light phase (paired one-way ANOVA with Bonferroni correction, *P<0.05, **P<0.01). Data are represented as means ± SEM; see statistical details in the method section.

LiFE treatment reduced baseline blood glucose level and fluctuation.

a) Bodyweight of WT mice on control (n = 18), TRL (n = 12) and LiFE (n = 18). b) Daily food consumption of WT mice (n = 12) on control, TRL and LiFE schedule. c) Food consumption percentage of WT mice on control (n = 6), TRL (n = 12) and LiFE (n = 18) schedules in a 4 hour bin. Orange and purple bars indicate access to food and wheels, yellow and grey bars on the x axis indicate light and dark. (mixed-effect model with Bonferroni correction, *P<0.05, **P<0.01, ***P<0.001, ****P<0.001, green stars: LiFE vs. control; blue stars: TRL vs. control; +: LiFE vs. TRL). d) Averaged daily baseline blood glucose calculated from e) (one-way ANOVA with Bonferroni correction, ***P<0.001). e) Baseline blood glucose of WT mice on control (n = 11 at time 12, 16, 20; n = 23 at time 0, 4, 8), TRL (n = 12) and LiFE (n = 23) schedule (mixed-effect model with Bonferroni correction, *P<0.05, **P<0.01, ***P<0.001). f) Difference between maximum and minimum baseline blood glucose calculated from e) (one-way ANOVA with Bonferroni correction, **P<0.01). g) Glucose tolerance test of WT mice (n = 12) on control, TRL and LiFE schedule at dark + 4 hours (mixed-effect model with Bonferroni correction, *P<0.05). h) Area under curve calculated from g) (one-way ANOVA with Bonferroni correction, *P<0.05). Data are represented as means ± SEM; see statistical details in the method section.

Long-term LiFE treatment improves learning in Morris water maze and novel object recognition.

a) Schematics of experiment timeline of WT, 5xFAD and 5xFAD/PS19 mice with long-term LiFE treatment. b) Latency to platform on day 1 measured in MWM in WT mice (control: n = 14; LiFE: n = 13). c) Latency to platform on acquisition days in MWM in WT mice. d) Percentage of time spent in target quadrant on probe day of MWM in WT mice. e) Swim speed of WT mice in MWM. f) Recognition index (see method) of WT mice in NOR (control: n = 12; LiFE: n = 11, Student’s t-test, *P<0.05). g) Latency to platform on day 1 in MWM in 5xFAD mice (control: n = 13; LiFE: n = 11). The dashed line indicates the level of WT controls. (mixed-effect model with Bonferroni correction, *P<0.05) h) Latency to platform on acquisition days in MWM in 5xFAD mice. i) Percentage of time spent in target quadrant on probe day of MWM in 5xFAD mice. j) Swim speed of 5xFAD mice in MWM (Student’s t-test). k) Recognition index of 5xFAD mice in NOR(control: n = 13; LiFE: n = 9, Student’s t-test). l) Latency to platform on day 1 in MWM in 5xFAD/PS19 mice (n = 10). m) Latency to platform on acquisition days in MWM in 5xFAD/PS19 mice. n) Percentage of time spent in target quadrant on probe day of MWM in 5xFAD/PS19 mice. o) Swim speed of 5xFAD/PS19 mice in MWM. p) Recognition index of 5xFAD/PS19 mice in NOR(n = 12, Student’s t-test). Data are represented as means ± SEM; see statistical details in the method section.

Long-term LiFE treatment reduced AD pathology in the hippocampus.

a) Representative images of D54D2 immunohistochemistry (IHC) staining in 5xFAD mice, 10X objective lens. b) Quantification of area fractions of D54D2 staining in 5xFAD mice (n = 9). c) Quantification of area fractions of D54D2 staining in 5xFAD/PS19 mice (control: n = 9, LiFE: n = 10 DG, CA1, LC, n = 8 MC; Student’s t-test). d) Representative images of D54D2 and AT8 IHC staining in 5xFAD/PS19 mice, 10X objective lens. e) Quantification of area fractions of AT8 staining in 5xFAD/PS19 mice (control: n = 9 DG, CA1, MC, n = 5 LC; LiFE: n = 10 DG, n = 9 CA1, n = 7 LC, n = 10 MC; black p-value: Student’s t-test; green p-value: Kolmogorov–Smirnov test). Data are represented as means ± SEM; see statistical details in the method section.

Characterizing behavior of WT mice in LiFE treatment.

a) Bodyweight of WT mice on control (n = 10) and LiFE (n = 16) schedules normalized to the control weight of day 1. b) Food consumption in every 24 hours measured with WT mice (n =6) on control and LiFE schedule. c) Wheel running activity in a 1 hour bin of WT mice under control (n = 9) and LiFE (n = 10) schedule. d) Total wheel running counts of data in a) (Student’s t-test, ****P<0.0001).

Comparison of cognitive behavior in WT and Alzheimer’s mouse models.

a) Latency to the platform of WT (n = 14), 5xFAD (n = 13) and 5P (n = 9) control mice on day 1 learning in MWM (mixed-effect model with Bonferroni correction, ****P<0.0001). b) Swim speed of WT (control n = 14, LiFE n = 12), 5xFAD (control n = 13, LiFE n = 11) and 5P (control n = 9, LiFE n = 10) mice in MWM test in control and LiFE treatment group (two-way ANOVA with Bonferroni correction, **P<0.01, ***P<0.001).c) Latency to the platform of WT, 5xFAD and 5P control mice on acquisition days in MWM (mixed-effect model with Bonferroni correction, ****P<0.0001). d) Time-spent in target quadrant of WT, 5xFAD and 5P control mice on probe day in MWM. e) Recognition index of WT (n = 12), 5xFAD (n = 13) and 5P (n = 12) control mice in NOR (one-way ANOVA with Bonferroni correction, *P<0.05, **P<0.01).