A generalised framework for detailed classification of swimming paths inside the Morris Water Maze

Avgoustinos Vouros,Artur Janusz,Witold Konopka,Katarzyna Lukasiuk,Eleni Vasilaki,Carmen Sandi,Tiago V Gehring,Mike Croucher,Kinga Szydlowska,Zehai Tu

doi:10.1038/s41598-018-33456-1

Abstract

The Morris Water Maze is commonly used in behavioural neuroscience for the study of spatial learning with rodents. Over the years, various methods of analysing rodent data collected during this task have been proposed. These methods span from classical performance measurements to more sophisticated categorisation techniques which classify the animal swimming path into behavioural classes known as exploration strategies. Classification techniques provide additional insight into the different types of animal behaviours but still only a limited number of studies utilise them. This is primarily because they depend highly on machine learning knowledge. We have previously demonstrated that the animals implement various strategies and that classifying entire trajectories can lead to the loss of important information. In this work, we have developed a generalised and robust classification methodology to boost classification performance and nullify the need for manual tuning. We have also made available an open-source software based on this methodology.

Highlights

The Morris Water Maze (MWM), designed by Richard Morris, was first described in 1981 in a study regarding the spatial localisation of rats[1]
These simplistic measurements and statistics have been criticised as being insufficient to capture all of the different animal behaviours that are observed during MWM experiments[16,17]
We present an automatic boosted classification procedure based on majority voting, which improves on the classification error, and a validation framework which leads to conclusions with a high degree of confidence

Summary

Introduction

The Morris Water Maze (MWM), designed by Richard Morris, was first described in 1981 in a study regarding the spatial localisation of rats[1]. The categorisation method of Graziano et al was based on a number of path measures while in the work of Garthe et al a hierarchical classification algorithm was used and the categorisation of each swimming path was primarily based on the amount of time that the animal spent in each region of the arena. The latter method was used in more recent studies[23,24] (Illouz et al.[25] proposed a classification technique based on support vector machines (SVM)[26]. Similar to previously proposed classification methods, it did not have the ability to detect mixed animal behaviours within the same trial but assigns the whole swimming path of the animal during the trial into one behavioural class[17]

Methods

Results

Discussion

Conclusion