Article Figures & Data
Figures
- Figure 1.
RI30/RI60 and RR10/RR20 produce different operant behavior and different response-per-reinforcer ratios. A, Schematic showing the relationship between response rate and reinforcer delivery rate under four example schedules. The red dotted line indicates the mean response rate a subject must perform to achieve equal reinforcer delivery rates under RI30 or RR10 schedules. The value is the same for achieving equal reinforcer delivery rates between RI60 and RR20. B, Mean reinforcer delivery rates across training, split by schedule. C, Mean response rates across training, split by schedule. D, Mean responses-per-reinforcer ratio during each phase of training. (Note: animals on RR20 can have mean response/reinforcer ratio larger than 20 because responses only counted once the previous reinforcer had been collected.) Individual data points are shown separated by sex within each group (Females: left, Males: right). RI: N = 9 mice, RR: N = 10 mice, * indicates independent Student’s t test p < 0.05. Data are shown as mean ± SEM (error bars are occluded by mean symbol in panel B). See also Extended Data Figure 1-1.
- Figure 2.
Designing RR schedules to match the responses-per-reinforcer ratios observed in mice following RI schedules under distinct levels of food restriction. A, Schematic showing the training strategy for three groups distinguished by their level of food restriction. Each mouse underwent 3 d of RI30, 4 d of RI60, and 2 d of outcome devaluation testing. B, Plot showing that each pair of response rate/reinforcer delivery rate can be achieved by a specific RR (shown as a color map). The response rate/reinforcer delivery rate values observed for each phase/restriction group are depicted as circles color-coded to match A. C, Summary of B showing the RR that matches each RI based on the performance of each RI group. D, Schematic showing the training strategy for three groups distinguished by their level of food restriction. Each group followed the RR schedule identified in B, C. E, Mean responses-per-reinforcer ratios for each restriction/schedule group, separated by phase of training. Individual data points are shown separated by sex within each group (Females: left, Males: right). RI-noRestriction N = 10 mice, RR-noRestriction N = 9 mice, RI-MildRestriction N = 14 mice, RR-MildRestriction N = 15 mice, RI-strongRestriction N = 14 mice, RR-strongRestriction N = 10 mice; * indicates two-way ANOVA p < 0.05 (E). ns indicates two-way ANOVA p > 0.05 (E) and post hoc Tukey’s HSD test p > 0.5 (E). Data are shown as mean ± SEM. See also Extended Data Figure 2-1.
- Figure 3.
Food restriction increases response rates more effectively in mice following RR schedules than in mice following RI schedules. A, Mean response rates across training, split by restriction and schedule. B, Mean response rates on the last day of training summarizing the data used to perform a two-way ANOVA. C, Mean ratio of response rate on last session over first session (after FR1 acquisition) summarizing the data used to perform a two-way ANOVA. Individual data points are shown separated by sex within each group (Females: left, Males: right). RI-noRestriction N = 10 mice, RR-NoRestriction N = 9 mice, RI-MildRestriction N = 14 mice, RR-MildRestriction N = 15 mice, RI-strongRestriction N = 14 mice, RR-strongRestriction N = 10 mice; black * indicates two-way ANOVA p < 0.05 (B, C); gray * indicates p < 0.05 post hoc Tukey’s HSD test for matched RR-RI pairs (B, C). ANOVA results and post hoc Tukey’s HSD tests are reported in the statistical table (Extended Data Fig. 1-1). Data are shown as mean ± SEM. See also Extended Data Figure 3-1.
- Figure 4.
The effects of food restriction and schedule on performance during devaluation procedures are dominated by extinction. A, Diagram describing the devaluation procedure, extinction test, and preference test. B, Summary of response rates during devaluation sessions, grouped by devaluation (valued/food, devalued/sucrose). Valued conditions consisted of 1-h access to regular chow and devalued 1-h access to 10% sucrose. The two test sessions were performed on consecutive days and the order counterbalanced within each group (valued-RI-MildRestriction: 4.3 ± 0.56 responses/min, devalued-RI-MildRestriction: 4.95 ± 0.91 responses/min, N = 8; valued-RR-MildRestriction: 4.88 ± 1.08 responses/min; devalued-RR-MildRestriction: 6.40 ± 1.21 responses/min, N = 10; valued-RI-StrongRestriction: 6.54 ± 0.92 responses/min; devalued-RI-StrongRestriction: 6.74 ± 1.16 responses/min, N = 14; valued-RR-StrongRestriction: 9.18 ± 1.46 responses/min; devalued-RR-StrongRestriction: 10.82 ± 2.05 responses/min, N = 9). C, Same data as shown in B summarized as a devaluation index for each mouse [valued response rate/(valued response rate + devalued response rate)]. Means for the mice grouped by schedule, restriction and testing order are also shown on the right. D, Same data as shown in B, grouped by session number (first extinction session, second extinction session; first-RI-MildRestriction: 4.95 ± 0.74 responses/min, second-RI-MildRestriction: 4.30 ± 0.77 responses/min, N = 8; first-RR-MildRestriction: 6.48 ± 1.08 responses/min; second-RR-MildRestriction: 4.80 ± 0.62 responses/min, N = 10; first-RI-StrongRestriction: 9.01 ± 1.05 responses/min; second-RI-StrongRestriction: 4.27 ± 0.47 responses/min, N = 14; first-RR-StrongRestriction: 12.91 ± 1.99 responses/min; second-RR-StrongRestriction: 7.09 ± 0.66 responses/min, N = 9). E, Same data as shown in D summarized as an extinction index for each mouse [second session response rate/(first session response rate + second session response rate)]. Means for the mice grouped by schedule and by restriction are also shown on the right. * indicates two-way ANOVA p < 0.05 (C, E). ns indicates two-way ANOVA p > 0.05 (C, E). Data are shown as mean ± SEM. Individual data points are shown separated by sex within each group (Females: left, Males: right). See also Extended Data Figure 4-1.
- Figure 5.
The effects of food restriction on extinction are not apparent within a session. A, Response rates during the first extinction session. Data are shown across the duration of the session in 1 min time bins. B, Response rates during the first minute of the extinction session show similar results as by the end of acquisition. C, Graph showing the ratio of the number of responses in the second minute over the number of responses during the first and second minutes capturing the magnitude of the decrease in response rate. D, Graph showing the number of responses executed before the inter-response interval became larger than mean ± 2 SD based on a rolling window of five responses. * indicates two-way ANOVA p < 0.05 (B). ns indicates two-way ANOVA p > 0.05 (B–D). Data are shown as mean ± SEM. Individual data points are shown separated by sex within each group (Females: left, Males: right). RI-MildRestriction: N = 8, RR-MildRestriction: N = 10, RI-StrongRestriction: N = 14, RR-StrongRestriction: N = 9).
Extended Data
Extended Data Figure 1-1
Statistical table. Download Figure 1-1, XLSX file.
Extended Data Figure 2-1
No difference in time to acquisition across restriction/schedule and in the efficacy of food restriction across sex. A, Graph showing the time to acquisition across training schedules and restriction groups [RI-NoRestriction: 4.1 ± 0.5 sessions (N = 10), RI-MildRestriction: 4.29 ± 0.62 sessions (N = 14), RI-StrongRestriction: 3.86 ± 0.49 sessions (N = 14), RR-NoRestriction: 3.78 ± 0.52 sessions (N = 9), RR-MildRestriction: 3.93 ± 0.45 sessions (N = 15), RR-StrongRestriction: 3.4 ± 0.45 sessions (N = 10)]. A two-way ANOVA revealed no effects of restriction group (two-way ANOVA F = 0.45, df = 2, p = 0.64) and no effect of task schedule (two-way ANOVA F = 0.77, df = 1, p = 0.38). These results are expected as we assigned mice to groups post FR1 acquisition to avoid any differences. B, Graph showing the effects of food restriction across time as the fraction of initial body weight [data on session 7: Females/NoRestriction: 1.02 ± 0.02 (N = 9), Females/MildRestriction: 0.99 ± 0.03 (N = 10), Females/StrongRestriction: 0.79 ± 0.01 (N = 12), Males/NoRestriction: 1.02 ± 0.01 (N = 10), Males/MildRestriction: 0.92 ± 0.02 (N = 10), Males/StrongRestriction: 0.76 ± 0.01 (N = 12)]. A two-way ANOVA on the last day (session 7) revealed a significant effect of restriction group (two-way ANOVA F = 90.1, df = 2, p = 2.3e-18) but no effect of sex (two-way ANOVA F = 3.48, df = 1, p = 6.7e-2). Download Figure 2-1, TIF file.
Extended Data Figure 3-1
Food restriction and schedule influenced reinforcer delivery rates and response rate variability. A, Mean reinforcer delivery rates across training, split by restriction and schedule. B, Mean reinforcer delivery rates on the last day of training summarizing the data used to perform a two-way ANOVA. C, Mean coefficient of variation of the inter-response interval across training, split by restriction and schedule. D, Mean coefficient of variation on the last day of training summarizing data used to perform a two-way ANOVA. RI-noRestriction N = 10 mice, RR-NoRestriction N = 9 mice, RI-MildRestriction N = 14 mice, RR-MildRestriction N = 15 mice, RI-strongRestriction N = 14 mice, RR-strongRestriction N = 10 mice; black * indicates two-way ANOVA p < 0.05 (B, D); grey * indicates p < 0.05 post hoc Tukey’s HSD test for matched RR-RI pairs (B, D). ANOVA results and post hoc Tukey’s HSD tests are reported in the statistical table. Data are shown as mean ± SEM. Individual data points are shown separated by sex within each group (Females: left, Males: right). Download Figure 3-1, TIF file.
Extended Data Figure 4-1
Prefeeding selectively and effectively reduced mice’s preference for the prefed food. A, Graphs showing how much food [left; RI-MildRestriction: 1.08 ± 0.24 g (N = 8), RR-MildRestriction: 0.94 ± 0.17 g (N = 10), RI-StrongRestriction: 1.25 ± 0.09 g (N = 14), RR-StrongRestriction: 1.52 ± 0.16 g (N = 10); two-way ANOVA main effect of restriction group df = 1, F = 4.35, p = 0.044; main effect of schedule df = 1, F = 0.15, p = 0.70; interaction df = 1, F = 2.39, p = 0.13) or sucrose (right; RI-MildRestriction: 1.62 ± 0.44 g (N = 8), RR-MildRestriction: 1.65 ± 0.40 g (N = 10), RI-StrongRestriction: 2.9 ± 0.22 g (N = 14), RR-StrongRestriction: 3.16 ± 0.18 g (N = 10); two-way ANOVA main effect of restriction group df = 1, F = 18.6, p = 1.15e-4; main effect of schedule df = 1, F = 0.24, p = 063; interaction df = 1, F = 0.15, p = 0.70] mice consumed during the 1 h prefeeding session preceding the extinction test. B, Graphs showing consumption during the postdevaluation preference test, during which mice have free access to both food and sucrose for 10 min, for food [left; RI-MildRestriction_postSucDeval: 0.52 ± 0.04 g (N = 5), RI-MildRestriction_postFoodDeval: 0.24 ± 0.07 g (N = 5), RR-MildRestriction_postSucDeval: 0.48 ± 0.09 g (N = 4), RR-MildRestriction_postFoodDeval: 0.27 ± 0.11 g (N = 4), RI-StrongRestriction_postSucDeval: 0.45 ± 0.05 g (N = 4), RI-StrongRestriction_postFoodDeval: 0.15 ± 0.05 g (N = 4), RR-StrongRestriction_postSucDeval: 0.55 ± 0.05 g (N = 2), RR-StrongRestriction_postFoodDeval: 0.10 ± 0.10 g (N = 2); three-way ANOVA main effect of restriction group df = 1, F = 1.68, p = 0.21; main effect of schedule df = 1, F = 0.014, p = 0.91; main effect of prefeed df = 1, F = 30.8, p = 1.4e-5; interactions all showed p > 0.05) and for sucrose (right; RI-MildRestriction_postSucDeval: 0.24 ± 0.11 g (N = 5), RI-MildRestriction_postFoodDeval: 1.26 ± 0.28 g (N = 5), RR-MildRestriction_postSucDeval: 0.22 ± 0.05 g (N = 4), RR-MildRestriction_postFoodDeval: 1.22 ± 0.36 g (N = 4), RI-StrongRestriction_postSucDeval: 0.10 ± 0.04 g (N = 4), RI-StrongRestriction_postFoodDeval: 1.35 ± 0.28 g (N = 4), RR-StrongRestriction_postSucDeval: 0.15 ± 0.15 g (N = 2), RR-StrongRestriction_postFoodDeval: 0.65 ± 0.55 g (N = 2); three-way ANOVA main effect of restriction group df = 1, F = 0.59, p = 0.45; main effect of schedule df = 1, F = 0.59, p = 0.45; main effect of prefeed df = 1, F = 33.5, p = 8.0e-6; interactions all showed p > 0.05]. C, Preference index based on data shown in B showing all groups preferred the nondevalued option equally. Negative values indicate preference for food and positive values indicate preference for sucrose [following sucrose devaluation, RI-MildRestriction: −0.59 ± 0.12 (N = 5), RR-MildRestriction: −0.45 ± 0.15 (N = 4), RI-StrongRestriction: −0.66 ± 0.05 (N = 4), RR-StrongRestriction: −0.76 ± 0.24 (N = 2); indices following food devaluation are the opposite, two-way ANOVA main effect of restriction group df = 1, F = 1.49, p = 0.25; main effect of schedule df = 1, F = 0.15, p = 0.70; interaction df = 1, F = 0.74, p = 0.41]. Data are shown as mean ± SEM. Individual data points are shown separated by sex within each group (Females: left, Males: right). Download Figure 4-1, TIF file.
In this issue
