Fig. 5: A neural network model reproduces emergence of engram-to-be cells and coactivity with common engram cells. | Nature Communications

Fig. 5: A neural network model reproduces emergence of engram-to-be cells and coactivity with common engram cells.

From: Parallel processing of past and future memories through reactivation and synaptic plasticity mechanisms during sleep

Fig. 5

A Schematic diagram of the neural network model sessions; (1) prelearning sleep, (2) context A, (3) postlearning sleep, and (4) context B. Green filled and blue hollow dots symbolize active and inactive cells, respectively. B Illustrative diagram of the neural model showing engram formation during learning (1), the plasticity mechanism occurring through sharp-wave ripples during sleep (2), and weight updating by synaptic scaling (3). Red, blue, and black arrows indicate long-term potentiation (LTP), long-term depression (LTD), and no change, respectively. C The matching ratio between session A and pre- or postlearning sleep in engram and nonengram cells. n = 5 simulations, statistical comparisons were made between engrams and nonengrams using two-way repeated-measures ANOVA, F (1, 4) = 2022, p = 0.0001, Sidak’s multiple comparisons test, p = 0.0001 (prelearning sleep), p = 0.0001 (postlearning sleep). D The matching ratio between session B and pre- or postlearning sleep in engram and engram-to-be, and other nonengram cells. n = 5 simulations, statistical comparisons were made between engram and engram-to-be vs other nonengram using two-way repeated-measures ANOVA, F (1, 4) = 275.4, p = 0.0001 (cell types), Sidak’s multiple comparisons test, p = 0.1530 (prelearning sleep, engram and engram-to-be vs other nonengram), p = 0.0001 (postlearning sleep, engram and engram-to-be vs other nonengram). E Pairwise correlations among engram, engram-to-be, and other nonengram cells in both pre- and postlearning sleep. n = 5 simulations, statistical comparisons were made between engram, engram-to-be, and other nonengram using two-way repeated-measures ANOVA, F (2, 8) = 97.32, p = 0.0001 (engram vs nonengram), Tukey’s multiple comparisons test, prelearning sleep; (engram vs engram-to-be, p = 0.0001), (engram vs other nonengram, p = 0.0001), (engram-to-be vs other nonengram, p = 0.0626). Postlearning sleep (engram vs engram-to-be, p = 0.0001), (engram vs other nonengram, p = 0.0001), (engram-to-be vs other nonengram, p = 0.0001). F The normalized coincidence ratio during postlearning sleep. n = 5 simulations, statistical comparisons were made between common engram, engram-to-be and other nonengram. Specific engram, engram-to-be and other nonengram using one-way ANOVA, F (3, 16) = 34.04, p = 0.0001, Tukey’s multiple comparisons test, common and engram-to-be; (vs specific and engram-to-be, p = 0.0001), (vs common and other nonengram, p = 0.0001), (vs specific and other nonengram, p = 0.0001). Specific and engram-to-be; (vs common and other nonengram, p = 0.3195), (vs specific and other nonengram, p = 0.8709). Common and other nonengram vs specific and other nonengram, p = 0.0916. ****p < 0.0001, n.s., nonsignificant. Data represent the mean ± s.e.m. Statistical comparisons were made using Repeated-measures two-way ANOVA with Sidak’s multiple comparisons test (C, D) and Sidak’s multiple comparisons test (E), one-way ANOVA with Tukey’s multiple comparison’s test (F). The Experiment was independently repeated five times. Source data are provided as a Source Data file. Detailed statistics are shown in Supplementary Data 1.

Back to article page