Fig. 4: Previous task context re-activates learned preparatory activity.

a, Possible outcomes of preparatory states for re-learning. b, Task contexts 1, 2 and 1′. Task contexts 1 and 1′ are identical. Time interval, 30.65 ± 8.73 days between task contexts 1 and 2, 57.19 ± 12.89 days between task contexts 1 and 1′; mean ± s.d. across fields of view. c, Mean deconvolved activities from an example field of view (n = 608 neurons). Neurons are sorted on the basis of their selectivity during the delay epoch in task context 1. d, Selectivity index in task context 1 (left), 2 (middle) and 1′ (right) for neurons showing significant trial-type selectivity (P < 0.001, two-tailed t-test) during the delay epoch. Red, neurons preferring lick left; blue, neurons preferring lick right. Significant selectivity and trial-type preferences are determined in task context 1. e, Schematic of movement-specific activity trajectories in activity space and CD_Delay vectors across task contexts. Red and blue shades, preparatory states for lick left and lick right, respectively. f, ALM activities from an example field of view projected on the CD_Delay from task context 1. Thick lines are the mean and thin lines show single trials. g, Same as f, but for the CD_Delay from task context 2. h, Decoding accuracy of the CD_Delay from task context 1 tested on task contexts 1, 2, 1′ and 2′. Grey circles and lines indicate fields of views imaged across task contexts 1, 2 and 1′ (n = 26 fields of view, 5 mice). Black circles and lines indicate fields of views imaged across task contexts 1, 2, 1′, and 2′ (n = 7 fields of view, 3 mice). Task context 1 versus 2, ***P = 1.12 × 10⁻¹⁰; task context 2 versus 1′, ***P = 3.46 × 10⁻⁷; task context 1′ versus 2′, **P = 0.0091. Two-tailed paired t-test. Data are mean ± s.e.m. i, Same as h, but for the CD_Delay from task context 2. Task context 1 versus 2, ***P = 4.87 × 10⁻¹⁰; task context 2 versus 1′, ***P = 7.40 × 10⁻⁹; task context 1′ versus 2′, *P = 0.0496. Two-tailed paired t-test.

Source Data