Effects of Repeat Test Exposure on Gait Parameters in Naïve Lewis Rats

Rodent gait analysis has emerged as a powerful, quantitative behavioral assay to characterize the pain and disability associated with movement-related disorders. In other behavioral assays, the importance of acclimation and the effect of repeated testing have been evaluated. However, for rodent gait analysis, the effects of repeated gait testing and other environmental factors have not been thoroughly characterized. In this study, fifty-two naïve male Lewis rats ages 8 to 42 weeks completed gait testing at semi-random intervals for 31 weeks. Gait videos and force plate data were collected and processed using a custom MATLAB suite to calculate velocity, stride length, step width, percentage stance time (duty factor), and peak vertical force data. Exposure was quantified as the number of gait testing sessions. Linear mixed effects models were used to evaluate the effects of velocity, exposure, age, and weight on animal gait patterns. Relative to age and weight, repeated exposure was the dominant parameter affecting gait variables with significant effects on walking velocity, stride length, fore and hind limb step width, fore limb duty factor, and peak vertical force. From exposure 1 to 7, average velocity increased by approximately 15 cm/s. Together, these data indicate arena exposure had large effects on gait parameters and should be considered in acclimation protocols, experimental design, and subsequent data analysis of rodent gait data.

was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 19, 2023. ; https://doi.org/10.1101/2023.04.19.537488 doi: bioRxiv preprint 1. Introduction 56 Rodent gait analysis is a powerful preclinical assay to characterize behavioral 57 changes associated with movement-related pain and disability. Gait analysis for rodent 58 models has been particularly useful for identifying behavioral changes that manifest due  was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made week to once every four weeks. In total, each animal was tested at least 6 times. The . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made access to food and water in an atmosphere-controlled room with a 12h light dark cycle.   Data visualization and statistical analysis were conducted in R (version 4.0.2). For 144 stride length, duty factor, step width, and peak vertical force, linear mixed effects models 145 were used to simultaneously assess the fixed effects of velocity, weight, age, and 146 exposure. For velocity, a linear mixed effects model was used to simultaneously assess . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made the fixed effect of weight, age, and exposure. In all linear mixed effects models, animal 148 identifiers were used to capture the random effect of repeated testing on the same animal.

149
Models did not include interactions among the fixed effects, as goodness of fit (R 2 ) was 150 collected for each model. The significance of fixed effects was assessed using a type III 151 analysis of variance using Satterthwaite's method. When indicated by the analysis of 152 variance, comparisons of least square means estimates within a given fixed effect were 153 conducted using a pairwise, Tukey's HSD test for multiple comparisons. In all statistical 154 tests, a p-value of 0.05 or lower was considered statistically significant. Throughout the 155 paper, data are presented as mean ± 95% confidence intervals. . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The relationship between velocity and age was less uniform but still demonstrated 169 a slight decrease with each age group (p=0.0011, Fig. 1B). Here, mean velocity 170 decreased by 8.68 ± 6.65 cm/s from the youngest to oldest age group (p<0.05).

171
Velocity slightly decreased as weight increased; however, these effects were not 172 statistically significant (Fig. 1C).  There was no significant effect of weight on stride length, but generally, stride 185 length increased as weight increased (Fig. 2C). For hind limb step width, there was an effect of age (p<0.001) and weight (p<0.05) 189 and tended to be an effect of exposure (p=0.0533); for fore limb step width, there was a 190 signficant effect of exposure (p<0.05) and age (p<0.001, Fig. 3). Within fore step width, . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 19, 2023. ; https://doi.org/10.1101/2023.04.19.537488 doi: bioRxiv preprint 9 191 exposure 1 through 3 showed narrower step widths than exposure 6 (p<0.05), and 192 exposure 2 and 3 showed narrower step width than exposure 7 (p<0.05, Fig. 3A). There 193 were no distinct trends with exposure for hide step width.

194
For age, hind limb step width was larger in the older 36-57 week group compared 195 to all younger groups (p<0.001) with a 0.66 ± 0.23 cm difference between the 36-57 week 196 and 22-28 week age groups (Fig. 3B). Similarly, fore step width was larger in all older healthy animal. Hind limb duty factor was affected by weight (p<0.001); fore limb duty 208 factor was affected by exposure (p<0.05), age (p<0.001), and weight (p<0.01, Fig. 4).

209
However, despite a significant main effect, clear differences between specific numbers of 210 exposure were not found in post-hoc tests (Fig. 4A, 4D).

211
For age, fore limb duty factors were higher in the 8-14 weeks group compared to 212 all older groups (p<0.05) (Fig. 4E). From the 8-14 week to 36-57 week group, there was 213 a 2.46 ± 1.86% decrease in fore limb duty factor.
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made For weight, hind limb duty factors were greater in the >500g group compared to all 215 other groups accounting for a 3.17 ± 1.71% increase from the <350g to >500g group 216 (p<0.05, Fig. 4F). Similarly, fore limb duty factors were highest in the three heaviest 217 weight groups compared to the <350g group with a 3.06 ± 1.70% increase from the <350g 218 to the >500g group (p<0.01, Fig. 4C).

219
Insert Figure 4 here less consistent than other gait parameters, but generally showed increases from earlier 225 to later exposures (Fig. 5A). Here, exposures 2 and 7 had the lowest velocity-peak vertical 226 force relationship while exposures 4 through 6 had the highest (p<0.05).

227
Although age did not have a substantial effect on peak vertical force, younger 228 animals tended to exert less force than older age groups (Fig. 5B).

229
On the other hand, despite peak vertical force being normalized to weight, the 230 heavier weight groups tended to produce lower weight-normalized peak vertical forces 231 than the lighter weight groups (Fig. 5C); however. this fixed effect of weight was not 232 significant in our model.  The first objective of this study was to assess the effects of repeated arena 236 exposures on several gait parameters. Most markedly, velocity substantially increased . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made generally increased across all exposures with a sizable 3.47 ± 2.53% increase from 241 exposure 2 to 4. For step width and duty factor, there was no clear across exposure, but 242 exposure remained a significant predictor for fore limb step width and duty factor.

243
Considering these effects together, exposure plays a considerable role in several gait 244 parameters with marked increases after repeated testing for velocity, stride length, and 245 peak vertical force.

246
The second objective was to assess the effects of age and weight in naïve animals was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made   The effects of exposure, however, were comparable or larger than age and weight 272 in velocity, stride length, and peak vertical force. The nearly 15cm/s increase in velocity 273 after 7 exposures is about twice the size of the velocity decrease across age groups and 274 three times the size of the decrease across weight groups, indicating exposure is the 275 strongest factor for velocity. For stride length and peak vertical force, exposure, age, and 276 weight had approximately equal shifts indicating exposure was comparable to age and 277 weight for these parameters. As mentioned previously, the original purpose of this data 278 collection was to establish a large age-and weight-matched database for future studies. 279 We expected the effects of weight and age to be dominant; however, our data suggest 280 the magnitude of exposure effects is just as critical as age and weight.

281
The effects of repeated test exposure have been demonstrated in other gait and 282 behavioral testing, but the reason rodents alter their gait after repeated tests remains . CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 19, 2023.  To control for size effects related to age and weight, historical data on age-and was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made are needed to account for exposure effects. One method of control is rodent acclimation.

307
Most parameters were highly affected between the first few testing exposures and 308 showed slowed or minimal change after the 3 rd testing session. By completing three 309 testing sessions before experimental data is collected, the experimenter can further was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 19, 2023. ; https://doi.org/10.1101/2023.04.19.537488 doi: bioRxiv preprint changes with exposure did not stabilize, even after seven exposures. Although the 330 magnitude and significance of change for most gait parameters declined substantially by 331 exposure 4, 5, 6 and 7, a definitive asymptote may not be reached for all gait parameters.

332
Additionally, age, weight, and exposure are inevitably autocorrelated in the model. 333 However, as noted in Supplemental Table 1, the age of first exposure ranged from 12 to 334 52 weeks old and weight at first exposure ranged from 346-662g. Despite this large range 335 of starting ages and weights, exposure remained a dominant effect. We also selected 336 type III analysis of variance models due to this autocorrelation, rather than a sequential 337 type I analysis; the robustness of the model was also confirmed by switching the order 338 the factors that appear in the model and confirming that the statistical outcomes did not 339 change.

340
In conclusion, these findings establish repeated testing exposure as a substantial 341 factor affecting gait testing in rodents. With effects comparable or greater than the effects 342 of age and weight, it is recommended that researchers design experiments, acclimation 343 procedures, and data analysis with the effects of repeated exposure in mind.

350
The authors have no conflicts of interest to disclose.

351
. CC-BY 4.0 International license available under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (which this version posted April 19, 2023. ; https://doi.org/10.1101/2023.04.19.537488 doi: bioRxiv preprint  was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made