Fig. 4

Assessment of pattern separation ability (A) Model separation test chamber simulation diagram (B) The learning process and the time required before the TUNL test. C TUNL test design. TUNL was divided into sampling and selection phases. The mouse's selection of the newly presented square during the choice phase was counted as correct. D Two kinds of experiments were designed in the selection phase. The smaller the separation interval, the harder it is to distinguish and the higher the requirement for pattern separation ability. E TUNL accuracy. At the high separation level, AD mice showed impaired pattern separation. EA improved pattern separation ability, interfered with Wnt5a expression, aggravated pattern separation ability, and the effect of EA was weakened. At the low separation, the pattern separation of AD mice decreased, and EA and interference with wnt5a did not affect the pattern separation ability. That is, the pattern separation ability reached the upper limit. All data are presented as mean ± SD; N = 10 mice; ^*P < 0.05 and ^***P < 0.001 vs. WT group; ^##P < 0.01 vs. 5 × FAD group; ^&P < 0.05 vs EA group