Introduction. The suboptimal choice procedure consists on giving an organism a choice between a discriminative alternative with a low probability of reinforcement, and a non-discriminative alternative with higher probability of reinforcement. In most experiments, pigeons choose the discriminative alternative and rats choose the non-discriminative alternative. On the other hand, the economic concept of essential value (EV), derived from fitting an exponential model of demand to data obtained from exposing an organism to increasing fixed-ratio schedules, has been shown to predict choice among different primary and conditioned reinforcers. Objective. The main goal of this study was to calculate the EV of each of the alternatives of the suboptimal choice procedure to assess if it can predict choice. Method. Eight pigeons and 8 rats were exposed in different conditions to each of the alternatives of the suboptimal choice procedure. Throughout sessions the FR requirement in the initial link for each alternative was increased from 1 to 120 in order to obtain the demand curve and calculate the essential value. Results. Pigeons showed higher EVs in the discriminative alternative, whereas rats showed higher EVs in the non-discriminative alternative. Discussion. Results of this study support the hypothesis that EV of each alternative may predict results in the suboptimal choice procedure, so further research manipulating variables that can affect choice should be tested.

Changes in unconditioned stimulus (US) probability affect the level of conditioned responding. However, these changes are confounded with changes in uncertainty and total amount of reward. In the present experiment, we aimed to differentiate between these elements influencing subjects’ behavior. Five groups of mice varied in their probability of reinforcement (.25, .50, .50x2, .75, and 1.0). Two groups (.50 and.50x2) shared the same probability of reinforcement but group .50x2 received twice the reward amount on every reinforced trial. Thus, group .50x2 received the same total number of rewards as group 1.0. Groups .25 and .75 had the same uncertainty, even if differing in their probability of reinforcement and the amount of reward. Groups .50 and .50x2 had the highest level of uncertainty, while group 1.0 had the least uncertainty. After this phase, all subjects went through 4 sessions of extinction with no reinforcement delivery. Response rates, latencies and probability of responding are influenced by probability of reinforcement. In addition, latencies are modulated by total amount of reward. During extinction, subjects with higher probability of reinforcement showed quicker extinction (i.,e. Partial Reinforcement Extinction Effect; PREE). These results suggest that probability plays a big role in the acquisition and extinction of appetitive conditioning and this role appears to be independent of the uncertainty with which the CS signals the US. Furthermore, latencies appear to be regulated by the total amount of reward predicted by the CS.

Video analysis is an important way to record and observe animal behavior. Although object tracking such as open-field test is easy to be implemented, tracking of body parts or other details still needs intrusive markers which could affect behavior. With deep learning, researchers at Harvard University developed an efficient marker-less tracking software – DeepLabCut. Since it is a universal software that could track any parts of the body without markers and get coordinate data of each track point, pose information still needed process. Here we developed a user-friendly tool to convert coordinate data into pose. All 9 kind of pose (Rearing, Sniffing, Grooming, Scratching, Freezing, Eating, Sign-Tracking, Goal-Tracking) could be converted from data to pose automatically by comparing to human pose estimation. With pose information, researchers could analyze the inner connections between posture and other indexes easier.

When traveling, the vertical dimension is negligible compared with the horizontal one for most animals. Therefore, a volumetric spatial representation is unnecessarily complicated and not really needed. Nevertheless, surface-travelers may face slopes, branches, rocks, and multilayer structures. It was suggested that surface-bounded animals encode the horizontal plane in detail whereas vertical information is more roughly encoded as contextual, non-metric information. Here we posed the question of how the vertical domain is perceived by rodents, depending on their motor capacities and dwelling habitat. For this we compared how rodents living in flatlands cope with a three-dimensional (3D) apparatuses (pyramids, slant surfaces) compared with rodents living in structured habitats. We found that as long as rodents could maintain the natural posture of their head (parallel to substrate and perpendicular to gravity) they kept the usual exploratory behavior seen on a horizontal surface. Rodents that live in flatlands favored to descent down to the floor whereas those living in structured habitats ascent to top parts of the apparatus. When on slant surface, rodent progressed in a winding path, and on steep surface they progressed either straight upward or parallel to the ground. Therefore, surface-bounded rodents try to maintain the natural horizontal posture of their head when coping with 3D environments. When forced to veer away from the natural posture, rodents attempted to keep similar altitude of both head sides (and thereby both posture of their internal ears). Altogether, the results demonstrate how behavior in 3D environments accords with motor capacities and natural habitats.