Figure 1.
Interval training leads to habitual responding.
A) Mice were trained to lever press for food pellets across 14 days. On days 1–2 the ratio group was trained on a RR5 schedule while the interval group was trained on a RI15 schedule. On days 3–4 the ratio group was trained on a RR10 schedule while the interval group was trained on a RI30 schedule. On days 5–14 the ratio group was trained on a RR20 schedule while the interval group was trained on a RI60 schedule. Both groups showed an increase in responding across training days. B) Mice trained on RR schedules showed a selective reduction in responding on the lever leading to the devalued outcome indicating that behavior was goal-directed. Mice trained on RI schedule did not exhibit a selective reduction in responding indicating that behavior was habitual. Error bars represent ± SEM. *p<0.05.
Figure 2.
Reward cues motivate habitual responses more than goal-directed actions.
A) Mice underwent Pavlovian conditioning for 8 days where conditional stimuli (CSs) were paired with the same food rewards used in instrumental training. Both ratio and interval groups showed an increase in conditional responding (magazine entries) across training days. The data are expressed as an elevation ratio = (CS entries/ITI entries). B) During the transfer test, mice trained on ratio schedules did not show an increase in lever pressing when the CSs were presented. In contrast, mice trained on interval schedules showed a significant increase in lever pressing during the CS presentations relative to the ITI period. This increase was not selective as it was observed on the lever leading to the same outcome as the CS and the lever leading to a different outcome (i.e. general transfer). C) Mice underwent additional Pavlovian training for 8 days. D) A second transfer test was conducted and this time mice trained on ratio schedules showed a significant increase in lever pressing when the CSs were presented. This increase was not selective and observed on both levers. Mice trained on interval schedules showed increased responding during the CS presentations that was also non-selective. The amount of transfer was significantly larger in the interval group compared to the ratio group. Error bars represent ± SEM. *p<0.05.