Hyperactivity, headdorsiflexion, head-weaving, and circling [3?,44]. In no case were these behaviours seen in the sham-operated animals. In addition, the BVD animals in the 6 month condition were demonstrated to have significant memory impairment in the spatial forced alternation T maze task, as reported in Zheng et al. [5].Statistical AnalysisThe data were tested for normality and homogeneity of variance. If necessary, they were natural log transformed and then re-tested. A series of 1 or 2-way multivariate analyses of variance (MANOVAs), using surgery and time point (24 h, 72 h and 1 week data), surgery (1 month data) or surgery and training (6 month data) as factors, were performed in SPSS 20 for each individual hippocampal region, with the 8 proteins as dependent variables [41,42]. Pillai’s Trace statistic was used because it has been reported to be more robust against violation of assumptions than other MANOVA statistics [41]. Since the tissue processing for the early time points (24 h, 72 h and 1 week), the middle time point (1 month) and the late time point (6 months) was done at different times, the data from these 3 conditions were analysed using separate MANOVAs, followed by univariate ANOVAs in the case of a significant MANOVA. For the 1 and 6 month conditions, GluR4 and NR2A were not analysed. In order to further investigate the data, and determine whether combinations of variables were changing in addition to individual variables, weProtein Expression in the Hippocampus at 24 Hours, 72 Hours and 1 Week post-BVDThere was no significant effect of surgery on the protein expression for any hippocampal subregion, and no significant interaction between surgery and time point, for the 24 h, 72 h and 1 week conditions. However, there was a significant effect of time point on the protein expression in all cases (CA1: F(18,58) = 29.87, P#0.000; CA2/3: F(18,58) = 35.38, P#0.000; DG: F(18,58) = 57.30, P#0.000; Fig. 1). Using a LDA on the CA1, CA2/3 or DG data, no linear discriminant function could be identified that significantly predicted whether the brain tissue came from a BVD or a sham animal.Glutamate Receptors after Vestibular DamageFigure 4. Cluster analysis, using the correlation coefficient distance and Ward’s minimal variance GSK2256098 site algorithm, on the neurochemical data at 6 months post-op. for the CA1, CA2/3 and the DG data together, showing the relationship between the different neurochemical variables as z scores (zCaMKII etc) for the sham (A) and BVD (B) animals. Note that there is no difference between the clusters for the sham and BVD animals. doi:10.1371/journal.pone.0054527.gProtein Expression in the Hippocampus at 1 Month PostBVDThe results were similar for the 1 month data set: there 23977191 were no significant differences between BVD or sham animals either in the MANOVA or univariate ANOVAs for any hippocampal subregion (data not shown). Using a LDA, no linear discriminant function could be identified that significantly predicted whether the brain tissue came from a BVD or a sham animal.Protein Expression in the Hippocampus at 6 Months Post-BVDThe results for the 6 month data set were more complicated. For the MANOVA for CA1, surgery was still non-significant, as was the interaction between surgery and training. However, T maze training had a significant effect on the protein expression(F(7,13) = 16.60, P#0.000). Univariate ANOVAs indicated that spatial training significantly GSK343 site increased protein expression for CaMKIIa (F(1,19) = 43.Hyperactivity, headdorsiflexion, head-weaving, and circling [3?,44]. In no case were these behaviours seen in the sham-operated animals. In addition, the BVD animals in the 6 month condition were demonstrated to have significant memory impairment in the spatial forced alternation T maze task, as reported in Zheng et al. [5].Statistical AnalysisThe data were tested for normality and homogeneity of variance. If necessary, they were natural log transformed and then re-tested. A series of 1 or 2-way multivariate analyses of variance (MANOVAs), using surgery and time point (24 h, 72 h and 1 week data), surgery (1 month data) or surgery and training (6 month data) as factors, were performed in SPSS 20 for each individual hippocampal region, with the 8 proteins as dependent variables [41,42]. Pillai’s Trace statistic was used because it has been reported to be more robust against violation of assumptions than other MANOVA statistics [41]. Since the tissue processing for the early time points (24 h, 72 h and 1 week), the middle time point (1 month) and the late time point (6 months) was done at different times, the data from these 3 conditions were analysed using separate MANOVAs, followed by univariate ANOVAs in the case of a significant MANOVA. For the 1 and 6 month conditions, GluR4 and NR2A were not analysed. In order to further investigate the data, and determine whether combinations of variables were changing in addition to individual variables, weProtein Expression in the Hippocampus at 24 Hours, 72 Hours and 1 Week post-BVDThere was no significant effect of surgery on the protein expression for any hippocampal subregion, and no significant interaction between surgery and time point, for the 24 h, 72 h and 1 week conditions. However, there was a significant effect of time point on the protein expression in all cases (CA1: F(18,58) = 29.87, P#0.000; CA2/3: F(18,58) = 35.38, P#0.000; DG: F(18,58) = 57.30, P#0.000; Fig. 1). Using a LDA on the CA1, CA2/3 or DG data, no linear discriminant function could be identified that significantly predicted whether the brain tissue came from a BVD or a sham animal.Glutamate Receptors after Vestibular DamageFigure 4. Cluster analysis, using the correlation coefficient distance and Ward’s minimal variance algorithm, on the neurochemical data at 6 months post-op. for the CA1, CA2/3 and the DG data together, showing the relationship between the different neurochemical variables as z scores (zCaMKII etc) for the sham (A) and BVD (B) animals. Note that there is no difference between the clusters for the sham and BVD animals. doi:10.1371/journal.pone.0054527.gProtein Expression in the Hippocampus at 1 Month PostBVDThe results were similar for the 1 month data set: there 23977191 were no significant differences between BVD or sham animals either in the MANOVA or univariate ANOVAs for any hippocampal subregion (data not shown). Using a LDA, no linear discriminant function could be identified that significantly predicted whether the brain tissue came from a BVD or a sham animal.Protein Expression in the Hippocampus at 6 Months Post-BVDThe results for the 6 month data set were more complicated. For the MANOVA for CA1, surgery was still non-significant, as was the interaction between surgery and training. However, T maze training had a significant effect on the protein expression(F(7,13) = 16.60, P#0.000). Univariate ANOVAs indicated that spatial training significantly increased protein expression for CaMKIIa (F(1,19) = 43.
