*2.3. AG Had No E*ff*ect on Object Recognition Memory*

The effect of AG on recognition memory was measured in the novel object test (Figure 3). AG had no effect on normal mice. However, the recognition memory in TBI model mice declined with AG treatment. The absence of a before–after change in the AG treated TBI group meant that AG prevented the adverse effect of TBI (Figure 3A). There was no significant effect of CMS and AG treatment on the recognition memory (Figure 3B)

**Figure 3.** Object recognition memory by AG in the TBI and CMS mice as measured by the recognition time index in the novel object test. (**A**) TBI model. There were before–after (within) effects [F(1,28) = 8.3, *p* = 0.008], treatment group (between) effects [F(3,28) = 0.67, *p* = 0.58], and within–between interaction effects [F(3,28) = 3.6, *p* = 0.026]. There was no before–after change in the control (*p* = 0.93), AG (*p* = 0.59) and TBI + AG (*p* = 0.97) groups, but there were decreases in the TBI (*p* = 0.001) group. (**B**) CMS model. There were no before-after effects [F(1,36) = 0.32, *p* = 0.57], treatment group effects [F(3,36) = 1.2, *p* = 0.31], or within–between interactions [F(3,36) = 0.12, *p* = 0.95]. All data were normally distributed and are represented as means ± S.E.M. Control: vehicle (DW) treated; AG: *Angelica gigas* 1 mg/kg; TBI: vehicle treated + traumatic brain injury; TBI + AG: *Angelica gigas* 1 mg/kg + traumatic brain injury; CMS: vehicle treated + chronic mild stress; CMS+AG: *Angelica gigas* 1 mg/kg + chronic mild stress. Repeated measures ANOVA, Tukey's HSD post-hoc test.
