Figure 3

Water maze latency learning curves. (A) During visible platform training, HFD mice swam slower than STD mice ****p < 0.0001 (ANOVA). There was also a diet × genotype interaction (F(1,786) = 3.995, p = 0.0460 (ANOVA)). On a STD, the swim speeds were not different in E4 than E3 mice. However, on a HFD swim speeds were lower in E4 than E3 mice. *p = 0.0257 (ANOVA). (B) During hidden platform training, there were effects of diet and genotype on swim speeds. Mice on a HFD swam faster than mice on a STD and E4 mice swam faster than E3 mice. ***p < 0.0001, **p = 0.008 (ANOVA). (C, D) When the time to reach the visible platform (latency) was analyzed, there were effects of genotype (F(1,380) = 7.138, p = 0.0079), diet (F(1,380) = 16.04, p < 0.0001), and a session x genotype interaction (F(3,380) = 3.992, p = 0.0081) (ANOVA). E4 mice performed poorer than E3 mice and HFD mice performed worse than mice on a regular diet. In E3 mice, the effect of HFD was most pronounced in the first two sessions. ***p = 0.0006 (ANOVA), on the first day of hidden platform training. In contrast, in E4 mice the effects of HFD were most pronounced in the third session, the first session on the second day of hidden platform training. **p = 0.0024 (ANOVA). (E–H) When the time to reach the hidden platform (latency) was analyzed, there was a session x treatment interaction (F(5,455) = 2.262, p = 0.047 (ANOVA)). The beneficial effects of insulin were most pronounced in the first and third hidden platform sessions. In the first hidden session, insulin-treated mice reached the hidden platform faster than saline-treated mice. *p = 0.047 (ANOVA). In the third hidden session, there was a genotype × diet × treatment interaction (F(1,91) = 4.075, p = 0.046 (ANOVA). There was a trend towards insulin improving the performance of E3 mice on a STD and E4 mice on a HFD. (I) When spatial memory retention was assessed in the probe trial by analyzing the time required to first swim to the platform was analyzed as performance measure, there was a genotype x diet x treatment interaction (F(1,91) = 11.146, p = 0.001) (ANOVA). In mice on a STD, there was an effect of genotype (F(1,37) = 112, p < 0.0001), an effect of insulin (F(1,37) = 4.227, p < 0.0469), and a treatment × genotype interaction (F(1,36) = 4.212, p = 0.0473) (ANOVA). In mice on a HFD, there was an effect of genotype (F(1,36) = 107.2, p < 0.0001), and trends towards an effect of insulin (F(1,36) = 4.039, p = 0.0520) and a trend towards a treatment × genotype interaction (F(1,36) = 4.045, p = 0.0518) (ANOVA). In E3 mice on a HFD and in E4 mice on a STD, insulin improved cognitive performance. **p = 0.0042, *p = 0.0496 (t test). In contrast, no beneficial effect of insulin on spatial memory retention was seen in E4 mice on a HFD or E3 mice on a STD. HFD: high fat diet; STD: standard diet. E3 males: n = 9 HFD and n = 12 STD; E3 females: n = 19 HFD and n = 13 STD; E4 males: n = 7 HFD and n = 11 STD; E4 females: n = 15 HFD and n = 14 STD.