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.

The present study performed a detailed analysis of behavior in a rat model of epilepsy using both established and novel methodologies to identify behavioral impairments that may differentiate between animals with a short versus long latency to spontaneous seizures and animals with a low versus high number of seizures. Temporal lobe epilepsy was induced by electrical stimulation of the amygdala. Rats were stimulated for 25 min with 100-ms trains of 1-ms biphasic square-wave pluses that were delivered every 0.5 s. Electroencephalographic recordings were performed to classify rats into groups with a short latency (< 20 days, n = 7) and long latency (> 20 days, n = 8) to the first spontaneous seizure and into groups with a low number of seizures (62 ± 64.5, n = 8) and high number of seizures (456 ± 185, n = 7). To examine behavioral impairments, we applied the following behavioral tests during early and late stages of epilepsy: behavioral hyperexcitability, open field, novel object exploration, elevated plus maze, and Morris water maze. No differences in stress levels (e.g., touch response in the behavioral hyperexcitability test), activity (e.g., number of entries into the open arms of the elevated plus maze), or learning (e.g., latency to find the platform in the Morris water maze test during training days) were observed between animals with a short versus long latency to develop spontaneous seizures or between animals with a low versus high number of seizures. However, we found a higher motor activity measured by higher number of entries into the closed arms of the elevated plus maze at week 26 post-stimulation in animals with a high number of seizures compared with animals with a low number of seizures. The analysis of the Morris water maze data categorized the strategies that the animals used to locate the platform showing that the intensity of epilepsy and duration of epileptogenesis influenced swimming strategies. These findings indicate that behavioral impairments were relatively mild in the present model, but some learning strategies may be useful biomarkers in preclinical studies.

K-Means is one of the most used algorithms for data clustering and the usual clustering method for benchmarking. Despite its wide application it is well-known that it suffers from a series of disadvantages; it is only able to find local minima and the positions of the initial clustering centres (centroids) can greatly affect the clustering solution. Over the years many K-Means variations and initialisation techniques have been proposed with different degrees of complexity. In this study we focus on common K-Means variations along with a range of deterministic and stochastic initialisation techniques. We show that, on average, more sophisticated initialisation techniques alleviate the need for complex clustering methods. Furthermore, deterministic methods perform better than stochastic methods. However, there is a trade-off: less sophisticated stochastic methods, executed multiple times, can result in better clustering. Factoring in execution time, deterministic methods can be competitive and result in a good clustering solution. These conclusions are obtained through extensive benchmarking using a range of synthetic model generators and real-world data sets.

Lay summary: Rats selected for different glucocorticoid responses to stress show differences in spatial learning and cognition. While the high responding animals have a better long-term spatial memory retention, the Low responding rats have an inferior spatial memory and impaired reversal learning abilities.

No abstract