OpenFieldAI - new open-source AI based software for tracking rodents and training open field test models

Open field test (OFT) is one of the widely used pre-clinical models for assessing the exploratory, locomotion and anxiety behavior of rodents. OFT parameters are often analyzed manually or by using automated systems. Although many softwares exist for OFT analysis, the steep learning curve and the cost of commercial software leads the researcher to opt for traditional approaches. Manual analysis is riddled with observer bias which can lead to ambiguity in behavior classification. This leads to the preference of automated system over manual observation in a scientific context. Herein we present OpenFieldAI, which is an open-source python-based software which has an intuitive graphical user interface (GUI) interface which makes it beginner friendly for researcher. The software utilizes YOLO algorithm for the detection and tracking of rodents in OFT apparatus. When pre-trained models are unable to infer sufficiently, the ability to train models according to user criteria can be helpful. Single/multiple pre-recorded videos (mp4) or live video with external web cam can be given as input to calculate the parameters like speed, distance, time spent in/out of region of interest (ROI) and entries/exits with the generation of box centroid graph, heat map and line path which are crucial information that give scientific insights into the rodents neurobehavior. The automatic detection of central/peripheral region by min/max calculation in 2D plane and manual drawing of ROI contribute to the ease in use and collection of intricate information. To validate the software, we compared the observation of 5 parameters - total distance, speed, entries into the central region and time spend in central and peripheral regions from the readings of ANY-maze (commercial software) using Pearson correlation coefficient. Correlation within all the three groups was found to be above 0.9 which indicates the reliability of the new software. Moreover, User's guide has been provided for proper utilization of the tool. OpenFieldAI which can be downloaded and installed in Windows OS for free at: https://sourceforge.net/projects/openfieldai/.