Establishing a reliable gait evaluation method for rodent studies

doi:10.1016/j.jneumeth.2017.03.017

Journal of Neuroscience Methods

Volume 283, 1 May 2017, Pages 92-100

https://doi.org/10.1016/j.jneumeth.2017.03.017 Get rights and content

Highlights

•
We developed a systematic method to reliably generate results with CatWalk, one of the most popular gait analysis tools.
•
We overcame many problems such as heel walking and poor compliance, which significantly compromised validity of results.
•
We manually corrected automation errors and isolated consistent stretches of walk cycles to generate more reliable results.
•
Stand Time, Duty Cycle, and Swing Speed were identified as reliable metrics for recovery evaluation after manual processing.
•
Print Area and Intensity parameters were found to be not trustable and should be used with extreme caution.

Abstract

Background

CatWalk is one of the most popular tools for evaluating gait recovery in preclinical research, however, there is currently no consensus on which of the many gait parameters captured by CatWalk can reliably model recovery. There are conflicting interpretations of results, along with many common but seldom reported problems such as heel walking and poor compliance.

New method

We developed a systematic manual classification method that overcomes common problems such as heel walking and poor compliance. By correcting automation errors and removing inconsistent gait cycles, we isolated stretches of recordings that are more reliable for analysis. Recovery outcome was also assessed by quantitative histomorphometric analysis of myelinated axons.

Results

While 40–60% of runs were erroneously classified without manual intervention, we corrected all errors with our new method, and showed that Stand Time, Duty Cycle, and Swing Speed are able to track significant differences over time and between experimental groups (all p < 0.05). The usability of print area and intensity parameters requires further validation beyond the capabilities of CatWalk.

Comparison with existing method(s)

There is currently no strategy that addresses problems such as heel walking and poor compliance, and therefore no standard set of parameters that researchers can rely on to report their findings.

Conclusion

Manual classification is a crucial step to generate reliable CatWalk data, and Stand Time, Duty Cycle, and Swing Speed are suitable parameters for evaluating gait recovery. Static parameters such as print area and intensity should be used with extreme caution.

Introduction

Peripheral nerve injury is a major burden to healthcare systems worldwide, affects 1.4 million patients, and costs over $150 billion dollars every year (Jia et al., 2014, Jiang et al., 2017, Jones et al., 2016, Taylor et al., 2008). Preclinical rodent studies are particularly valuable due to low cost and high translational potential (Kizilay et al., 2016, Wang et al., 2016), however, researchers must rely on postmortem pathological evaluation with only limited ways to evaluate functional recovery (Bervar, 2000, de Medinaceli et al., 1982). CatWalk has become a popular tool in rodent preclinical studies for evaluating functional recovery due to its ability to acquire a wide variety of gait parameters with its automatic “Auto Classify” function, and it is now widely used in mainstream research protocols in a wide variety of fields including peripheral nerve injury (Bozkurt et al., 2008, Deumens et al., 2007, Deumens et al., 2014, Johnson and Jia, 2016), spinal cord injury (Kjell et al., 2015, Salewski et al., 2015), traumatic brain injury (Kizilay et al., 2016), neurodegenerative diseases (Neckel, 2015) etc. Google Scholar generated 1520 search results with the search term “CatWalk Gait” and 490 results with “CatWalk Sciatic” since 2006.

While CatWalk is now a widely used tool, it has significant drawbacks. Researchers often choose from wide variety of gait parameters to report their findings without justification of selection preference (Freria et al., 2016, Fujimaki et al., 2016, Hausner et al., 2014, Huang et al., 2012), and there are conflicting interpretations of results across different injury models (Bozkurt et al., 2008, Hamers et al., 2006). The lack of a standard set of parameters and interpretations may be due to common but seldom mentioned problems following rat sciatic nerve injury such as preferential heel walking (Deumens et al., 2007, Deumens et al., 2014) and poor compliance (Neckel, 2015). While these problems significantly impair the ability of CatWalk to reliably automatically collect and analyze data, current literature has yet to propose any strategies on how to address them. Therefore, a standard method that effectively addresses these problems must first be established in order to properly identify a standard set of reliable CatWalk gait parameters.

The goal of this study is to establish a standard data processing method to account for and address inherent problems such as heel walking and poor compliance, and to identify a set of CatWalk parameters that is capable of consistently evaluating injury and recovery. Current literature evaluating CatWalk only employ a single injury group and a sham group, so to emulate current research practices, we employed three experimental groups, each receiving nerve resection injury followed by different interventions leading to different levels of recovery: one with an autologous nerve graft (clinical gold standard) to serve as positive control, one with an empty nerve conduit to serve as negative control, and one with human neural crest stem cell (NCSC) implantation to serve as the study target (Bozkurt et al., 2008), which leads to improvements in recovery that is superior to empty nerve conduits but not as effective as autologous nerve grafts (Georgiou et al., 2015, Ni et al., 2013, Wang et al., 2015).

Section snippets

Animals

All animals were maintained according to NIH guidelines, and experimental protocols were approved by the IACUC of the University of Maryland School of Medicine. Every attempt was made to minimize the total number of animals used and their discomfort and pain. In this study, 36 athymic nude rats were used. Rats were individually housed in a controlled temperature (21 ± 1 °C) and humidity (50 ± 15%) throughout this study with a 12:12 h light/dark cycle to reduce stress. All rats had free access to food

Histomorphometric analysis and gastrocnemius muscle weight showed differentiated recovery across experimental groups over time

Overall, there was a significant difference in cell count per 100 um² when considering all groups and time points (p < 0.001 for all). See Fig. 1.

At 6 weeks, the cell count per 100 um² was statistically significantly greater for the autograft group (57.17 ± 9.84) than the conduit group (4.00 ± 1.67, p < 0.001) and the NCSC group (8.33 ± 1.76, p < 0.001). The NCSC group showed no significant difference with the conduit only group. Similarly, the cell count was statistically significantly greater for the

Discussion

CatWalk is a very popular tool for evaluating functional recovery following sciatic nerve injuries in rats, however, due to inherent problems such as heel walking and poor compliance, CatWalk is prone to error and fails to generate consistent and trustable results without extensive manual intervention. We tested and established a manual classification process that overcomes the limitations of CatWalk to address inherent problems, and were able to reliably model injury and recovery following

Conclusion

CatWalk is a popular system able to automatically detect many parameters of gait, however, it has many inherent problems that needed to be addressed. Though the automation feature of the software is convenient for researchers, it should not be trusted and relied upon. Further manual processing is a necessary step to improve the usability of results and generate meaningful data. We successfully established a manual processing method that yields reliable dynamic parameters of gait, namely Stand

References (35)

M. Bervar
Video analysis of standing–an alternative footprint analysis to assess functional loss following injury to the rat sciatic nerve
J. Neurosci. Methods
(2000)
A. Bozkurt et al.
CatWalk gait analysis in assessment of functional recovery after sciatic nerve injury
J. Neurosci. Methods
(2008)
A. Bozkurt et al.
Aspects of static and dynamic motor function in peripheral nerve regeneration: SSI and CatWalk gait analysis
Behav. Brain Res.
(2011)
R. Deumens et al.
The CatWalk gait analysis in assessment of both dynamic and static gait changes after adult rat sciatic nerve resection
J. Neurosci. Methods
(2007)
M. Georgiou et al.
Engineered neural tissue with aligned, differentiated adipose-derived stem cells promotes peripheral nerve regeneration across a critical sized defect in rat sciatic nerve
Biomaterials
(2015)
T. Hausner et al.
Inhibition of calpains fails to improve regeneration through a peripheral nerve conduit
Neurosci. Lett.
(2014)
W. Huang et al.
Regenerative potential of silk conduits in repair of peripheral nerve injury in adult rats
Biomaterials
(2012)
N.D. Neckel
Methods to quantify the velocity dependence of common gait measurements from automated rodent gait analysis devices
J. Neurosci. Methods
(2015)
D.H. Vrinten et al.
‘CatWalk' automated quantitative gait analysis as a novel method to assess mechanical allodynia in the rat; a comparison with von Frey testing
Pain
(2003)
L. de Medinaceli et al.
An index of the functional condition of rat sciatic nerve based on measurements made from walking tracks
Exp. Neurol.
(1982)

B.S. Boyd et al.

Deviations in gait pattern in experimental models of hindlimb paresis shown by a novel pressure mapping system

J. Neurosci. Res.

(2007)

R. Deumens et al.

Assessing motor outcome and functional recovery following nerve injury

Methods Mol. Biol.

(2014)

C.M. Freria et al.

Impairment of toll-like receptors 2 and 4 leads to compensatory mechanisms after sciatic nerve axotomy

J. Neuroinflammation

(2016)

H. Fujimaki et al.

Oriented collagen tubes combined with basic fibroblast growth factor promote peripheral nerve regeneration in a 15 mm sciatic nerve defect rat model

J. Biomed. Mater. Res.

(2016)

L. Gan et al.

Cellulose/soy protein composite-based nerve guidance conduits with designed microstructure for peripheral nerve regeneration

J. Neural Eng.

(2016)

F.P. Hamers et al.

Automated quantitative gait analysis during overground locomotion in the rat: its application to spinal cord contusion and transection injuries

J. Neurotrauma

(2001)

F.P. Hamers et al.

CatWalk-assisted gait analysis in the assessment of spinal cord injury

J. Neurotrauma

(2006)

Cited by (34)

Coordination function index: A novel indicator for assessing hindlimb locomotor recovery in spinal cord injury rats based on catwalk gait parameters
2024, Behavioural Brain Research
In preclinical studies of spinal cord injury (SCI), behavioral assessments are crucial for evaluating treatment effectiveness. Commonly used methods include Basso, Beattie, Bresnahan (BBB) score and the Louisville swim scale (LSS), relying on subjective observations. The CatWalk automated gait analysis system is also widely used in SCI studies, providing extensive gait parameters from footprints. However, these parameters are often used independently or combined simply without utilizing the vast amount of data provided by CatWalk. Therefore, it is necessary to develop a novel approach encompassing multiple CatWalk parameters for a comprehensive and objective assessment of locomotor function. In this work, we screened 208 CatWalk XT gait parameters and identified 38 suitable for assessing hindlimb motor function recovery in a rat thoracic contusion SCI model. Exploratory factor analysis was used to reveal structural relationships among these parameters. Weighted scores for Coordination effectively differentiated hindlimb motor function levels, termed as the Coordinated Function Index (CFI). CFI showed high reliability, exhibiting high correlations with BBB scores, LSS, and T2WI lesion area. Finally, we simplified CFI based on factor loadings and correlation analysis, obtaining a streamlined version with reliable assessment efficacy. In conclusion, we developed a systematic assessment indicator utilizing multiple CatWalk parameters to objectively evaluate hindlimb motor function recovery in rats after thoracic contusion SCI.
Neurobehavioral deficits of mice expressing a low level of G127V mutant frataxin
2023, Neurobiology of Disease
Friedreich's ataxia (FRDA) is a neurodegenerative disease caused by reduced expression of the mitochondrial protein frataxin (FXN). Most FRDA patients are homozygous for large expansions of GAA repeats in intron 1 of FXN, while some are compound heterozygotes with an expanded GAA tract in one allele and a missense or nonsense mutation in the other. A missense mutation, changing a glycine to valine at position 130 (G130V), is prevalent among the clinical variants. We and others have demonstrated that levels of mature FXN protein in FRDA G130V samples are reduced below those detected in samples harboring homozygous repeat expansions. Little is known regarding expression and function of endogenous FXN-G130V protein due to lack of reagents and models that can distinguish the mutant FXN protein from the wild-type FXN produced from the GAA-expanded allele. We aimed to determine the effect of the G130V (murine G127V) mutation on Fxn expression and to define its multi-system impact in vivo. We used CRISPR/Cas9 to introduce the G127V missense mutation in the Fxn coding sequence and generated homozygous mice (Fxn^G127V/G127V). We also introduced the G127V mutation into a GAA repeat expansion FRDA mouse model (Fxn^GAA230/KO; KIKO) to generate a compound heterozygous strain (Fxn^G127V/GAA230). We performed neurobehavioral tests on cohorts of WT and Fxn mutant animals at three-month intervals for one year, and collected tissue samples to analyze molecular changes during that time. The endogenous Fxn G127V protein is detected at much lower levels in all tissues analyzed from Fxn^G127V/G127V mice compared to age and sex-matched WT mice without differences in Fxn transcript levels. Fxn^G127V/G127V mice are significantly smaller than WT counterparts, but perform similarly in most neurobehavioral tasks. RNA sequencing analysis revealed reduced expression of genes in oxidative phosphorylation and protein synthesis, underscoring the metabolic consequences in our mouse model expressing extremely low levels of Fxn. Results of these studies provide insight into the unique pathogenic mechanism of the FXN G130V mechanism and the tolerable limit of Fxn/FXN expression in vivo.
Melatonin promotes peripheral nerve repair through Parkin-mediated mitophagy
2022, Free Radical Biology and Medicine
Schwann cells (SCs) are the major glial cells in peripheral nervous system. They unsheathe and myelinate axons and play an essential role in peripheral nerve regeneration. Several studies report that Parkin-mediated mitophagy is associated with various diseases. Melatonin promotes proliferation of central glial cells. Little is known about the effect of melatonin and Parkin-mediated mitophagy on peripheral nerve repair.
In this study, using a rat model of a peripheral nerve injury (PNI) and in vitro model established by RSC96 cells treated with tert-butyl hydroperoxide (TBHP), we found that Parkin-mediated mitophagy can effectively reduce the production of mitochondrial reactive oxygen species (ROS), maintain the balance of mitochondrial membrane potential, maintain autophagic flux, and inhibit mitochondrial apoptosis. At the same time, we found that the increase of Parkin under stress is a manifestation of the RSC96 cells' resistance to oxidative stress to maintain RSC96 cells' balance. In our experiment, melatonin is similar to a Parkin agonist, up-regulating the expression of Parkin, enhancing all the positive results of Parkin in a stress state, such as inhibiting active oxygen production, maintaining autophagic flux, and inhibiting mitochondrial apoptosis. In addition, we design in vivo experiments to verify in In vitro experiments. In in vivo, melatonin promotes the expression of Parkin, maintains autophagic flux, inhibits apoptosis, promotes myelin regeneration, reduces the regeneration of collagen fibers around damaged tissues, and promotes peripheral nerve repair. When adenovirus was used to down-regulate the expression of Parkin, we found that all the positive effects of melatonin were attenuated.
Collectively, these findings indicate that melatonin upregulates Parkin-mediated mitophagy and promotes peripheral nerve repair. The results provide a basis for development of effective drugs for PNI treatment.
Local vibration therapy promotes the recovery of nerve function in rats with sciatic nerve injury
2022, Journal of Integrative Medicine
Citation Excerpt :
The nerve conduction velocity (NCV; in m/s) was calculated as follows [32]: the distance between two stimulation electrodes (in mm) divided by the difference between the action potential latency of the two stimulation points (in ms). On the 28th day after surgery, the gastrocnemius muscles were excised bilaterally, and the wet weights were measured using a high-precision electronic balance [33]. The wet weight ratio (%) of the gastrocnemius was calculated as the wet weight of the muscle on the injured side divided by the wet weight of the muscle on the contralateral side × 100%.
It has been reported that local vibration therapy can benefit recovery after peripheral nerve injury, but the optimized parameters and effective mechanism were unclear. In the present study, we investigated the effect of local vibration therapy of different amplitudes on the recovery of nerve function in rats with sciatic nerve injury (SNI).
Adult male Sprague-Dawley rats were subjected to SNI and then randomly divided into 5 groups: sham group, SNI group, SNI + A-1 mm group, SNI + A-2 mm group, and SNI + A-4 mm group (A refers to the amplitude; n = 10 per group). Starting on the 7th day after model initiation, local vibration therapy was given for 21 consecutive days with a frequency of 10 Hz and an amplitude of 1, 2 or 4 mm for 5 min. The sciatic function index (SFI) was assessed before surgery and on the 7th, 14th, 21st and 28th days after surgery. Tissues were harvested on the 28th day after surgery for morphological, immunofluorescence and Western blot analysis.
Compared with the SNI group, on the 28th day after surgery, the SFIs of the treatment groups were increased; the difference in the SNI + A-2 mm group was the most obvious (95% confidence interval [CI]: [5.86, 27.09], P < 0.001), and the cross-sectional areas of myocytes in all of the treatment groups were improved. The G-ratios in the SNI + A-1 mm group and SNI + A-2 mm group were reduced significantly (95% CI: [−0.12, −0.02], P = 0.007; 95% CI: [−0.15, −0.06], P < 0.001). In addition, the expressions of S100 and nerve growth factor proteins in the treatment groups were increased; the phosphorylation expressions of ERK1/2 protein in the SNI + A-2 mm group and SNI + A-4 mm group were upregulated (95% CI: [0.03, 0.96], P = 0.038; 95% CI: [0.01, 0.94], P = 0.047, respectively), and the phosphorylation expression of Akt in the SNI + A-1 mm group was upregulated (95% CI: [0.11, 2.07], P = 0.031).
Local vibration therapy, especially with medium amplitude, was able to promote the recovery of nerve function in rats with SNI; this result was linked to the proliferation of Schwann cells and the activation of the ERK1/2 and Akt signaling pathways.
An analysis of the CatWalk XT and a composite score to assess neurofunctional deficits after photothrombosis in mice
2021, Neuroscience Letters
Citation Excerpt :
The automatic analyses provide independence of an observer’s subjective judgement, thereby increasing reliability. Nevertheless, several limitations apply including differences in the animal’s compliance to traverse, deficits in the automatic classification of the footprints, locomotion speed that impacts on the analysis of limb kinematics and interrelated speed-depending parameters, all of which may result in errors [13,18–20]. We here evaluate the ability and reproducibility of the CatWalk gait assessment to depict motor deficits after cortical photothrombotic stroke in mice and relate it to a composite neuroscore [21].
The purpose of this study was to evaluate CatWalk’s capability for assessing the functional outcome after photothrombotic stroke affecting the motor cortex of mice. Mice were tested up to 21 days after photothrombosis or sham surgery using CatWalk, and a composite score assessing functional deficits (neuroscore).
The neuroscore demonstrated deficits of the contralateral forelimb for more than two weeks after stroke. There were no asymmetric or coordinative dysfunctions of limbs detected by CatWalk. However, CatWalk data revealed impairment of locomotion speed and its depending parameters for one-week after stroke in strong correlation to the neuroscore.
Data suggest that the composite neuroscore allows to more sensitively and precisely specify and quantify photothrombosis-induced hemisyndromes than CatWalk.
Use of the CatWalk gait analysis system to assess functional recovery in rodent models of peripheral nerve injury – a systematic review
2020, Journal of Neuroscience Methods
Injuries of the peripheral nervous system are common among the population affecting around 3% of all trauma patients. This high clinical need in the field of peripheral nerve injury and regeneration has been steadily driving experimental and epidemiological research. Thereby, it is crucial to determine the exact degree of recovery of end-organ function. Regeneration after nerve injuries is assessed by a wide variety of techniques and pre-clinical model systems, where rodent models are among the most widely used. However, results from rodents are difficult to translate to human patients in general, and reproducible and comparable assessment of functional recovery is of highest importance. Computerized gait analysis allows comprehensive acquisition of locomotor function. As the animals cross the recording device voluntarily, functional recovery is assessable with a minimum degree of human interference on their behavior. This article aims to give a detailed overview on the existing literature on CatWalk gait analysis in rodent models of peripheral nerve injuries of upper and lower extremities, e.g. axonotmesis, neurotmesis or fibrosis, with special emphasis on differences between models. Researchers interested in assessment of locomotor function in such models will especially benefit from this work as it will provide them with an overview of the various experimental setups and expected outcomes. This work also addresses potential pitfalls and hurdles in order to promote well designed, comparable studies allowing for accelerated development of therapeutic strategies in peripheral repair and regeneration.

View all citing articles on Scopus

^☆: Supported by Maryland Stem Cell Research Fund (2013-MSCRFE-146-00) (to XJ), and R01HL118084 from NIH (to XJ).

View full text

Original research paperEstablishing a reliable gait evaluation method for rodent studies☆

Highlights

Abstract

Background

New method

Results

Comparison with existing method(s)

Conclusion

Introduction

Section snippets

Animals

Histomorphometric analysis and gastrocnemius muscle weight showed differentiated recovery across experimental groups over time

Discussion

Conclusion

J. Neurosci. Methods

J. Neurosci. Methods

Behav. Brain Res.

J. Neurosci. Methods

Biomaterials

Neurosci. Lett.

Biomaterials

J. Neurosci. Methods

Pain

Exp. Neurol.

Deviations in gait pattern in experimental models of hindlimb paresis shown by a novel pressure mapping system

J. Neurosci. Res.

Assessing motor outcome and functional recovery following nerve injury

Methods Mol. Biol.

Impairment of toll-like receptors 2 and 4 leads to compensatory mechanisms after sciatic nerve axotomy

J. Neuroinflammation

Oriented collagen tubes combined with basic fibroblast growth factor promote peripheral nerve regeneration in a 15 mm sciatic nerve defect rat model

J. Biomed. Mater. Res.

Cellulose/soy protein composite-based nerve guidance conduits with designed microstructure for peripheral nerve regeneration

J. Neural Eng.

Automated quantitative gait analysis during overground locomotion in the rat: its application to spinal cord contusion and transection injuries

J. Neurotrauma

CatWalk-assisted gait analysis in the assessment of spinal cord injury

J. Neurotrauma

Original research paper
Establishing a reliable gait evaluation method for rodent studies☆