HUB-DT: A tool for unsupervised behavioural discovery and analysis
Lindsay, A. J.
Show abstract
There has been an expansion in the diversity of tools used to measure various aspects of brain function in behaving animals. While these tools have great potential to transform our understanding of brain function, they are of little value if the behavior of interest is poorly defined or quantified. Traditional methods of behavioural labelling focus on easily quantified gross measure, such as velocity, gate crossing, nosepokes, etc. While these measures are specific and reproducible, they are crude descriptions of behaviour at best. Manually defined behaviours, while providing increased granularity and descriptive power over specific gross measures, suffer from being inexact and somewhat arbitrary. Consistent labelling between human observers is often difficult, and even if manually defined behaviours are subsequently labelled in an automated fashion (via a supervised learning algorithm) these behaviours need to be defined ahead of time, possibly biasing the range of behaviours of interest for a given task. Here we present HUB-DT, a behavioural discovery pipeline built on the frameworks of several existing tools and methods in the space of behavioural categorisation, the specifics of which will be highlighted in this report, and designed to address the requirements of behavioral discovery.
Matching journals
The top 5 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- RatInABox: A toolkit for modelling locomotion and neuronal activity in continuous environments 94%
- SpikeForest: reproducible web-facing ground-truth validation of automated neural spike sorters 93%
- MotorNet: a Python toolbox for controlling differentiable biomechanical effectors with artificial neural networks 93%
Similar papers in this journal
- Probing machine-learning classifiers using noise, bubbles, and reverse correlation 94%
- USE: An integrative suite for temporally-precise psychophysical experiments in virtual environments for human, nonhuman, and artificially intelligent agents 92%
- REVS: A New Open-Source Platform for High-Resolution Analysis of Rodent Wheel Running Behavior 92%
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.