William A. Roberts’s research while affiliated with Western University and other places

Three experiments are reported that used a new test of spatial memory in rats. The apparatus used was dual eight-arm radial mazes that were connected at one arm of each maze, with a start arm and doors to each maze. Rats could be forced to go to one maze or the other or could make a free choice between mazes. In Experiment 1, rats formed reference memory for the arm containing food on one maze but had food randomly placed on different arms over trials on the other maze. In Experiment 2, rats formed working memory for the arm containing food on one maze but not the other. In Experiment 3, food location changed randomly among trials on both mazes, but one maze contained a cue for the location of food. Rats used reference and working memory to go directly to the food arm on one maze but found food only after searching several arms on the other maze. Most importantly, when given free-choice trials rats developed a significant preference for the maze where they knew the location of food reward or found the cue indicating the location of reward. We suggest these findings may be best interpreted by rats applying two successive rules: (1) choose the maze that leads to the most immediate reward, and (2) use extramaze or intramaze cues to find reward location on the maze.

In the past 20 years, research in animal cognition has challenged the belief that complex cognitive processes are uniquely human. At the forefront of these challenges has been research on mental time travel and future planning in jays. We tested whether Canada jays (Perisoreus canadensis) demonstrated future planning, using a procedure that has produced evidence of future planning in California scrub-jays. Future planning in this procedure is cach-ing in locations where the bird will predictably experience a lack of food in the future. Canada jays showed no evidence of future planning in this sense and instead cached in the location where food was usually available, opposite to the behaviour described for California scrub-jays. We provide potential explanations for these differing results adding to the recent debates about the role of complex cognition in corvid caching strategies.

We tested dogs for a violation of independence from irrelevant alternatives, which would indicate irrational behavior. In Experiment 1, we offered 10 dogs' choices among alternative passages. The target passage led to more food than the competitor passage but required dogs to enter a narrower passage. The decoy passage asymmetrically dominated the competitor passage because although it contained a larger amount of food, it was narrower than the target passage. We found that dogs increased their preference for the target passage in the presence of the decoy passage, which violated the assumption of independence from irrelevant alternatives. Our second experiment controlled for energetic hunger state because previous findings had suggested that the violation effect might arise from changes in energetic state (Schuck-Paim, Pompilio, & Kacelnik, 2004). We provided supplementary feedings to each dog between each trial such that each dog consumed the same amount of food on each trial. The violation of independence from irrelevant alternatives effect persisted, though to a lesser degree than in Experiment 1. Cognitive implications are discussed.

Although pigeons do not naturally cache and recover food items as found in members of the corvid and parid families, an operant analog of food caching and recovery in pigeons was studied in four experiments. Pigeons were trained to peck a caching key that added a fixed increment of time to the final duration of reinforcement obtained by pecking a payoff key. The same key served as the caching and payoff keys in Experiment 1, but separate caching and payoff keys were used in Experiments 2-4. In Experiments 2-3, each peck on a left red caching key added 0.5 s of reinforcement earned by pecking a right white payoff key. In Experiment 4, red or green caching keys appeared on different trials, with 0.5 s of reinforcement earned for pecking the red key and 1.0 s of reinforcement earned for pecking the green key. Pigeons showed an increased number of pecks on the caching key over ten sessions in Experiments 1-3 and more pecks on the green caching key than on the red caching key in Experiment 4.

A previous study failed to find evidence that dogs could use olfactory cues to discriminate between 1 and 5 hot dog slices presented on a single trial (Horowitz et al., Learning and Motivation, 44, 207–217, 2013). In the experiments reported here, multiple trials were used to test dogs’ ability to use olfaction to choose one of two opaque containers under which a larger number of food items was placed. In Experiment 1, dogs chose between 1 and 5 hot dog slices. In Experiments 2 and 3, we examined dogs’ ability to discriminate between numbers of hot dog slices that varied in the numerical distance and the ratio between the smaller and larger quantities. Experiment 4 explored olfactory discrimination between quantities of a different food, dog kibble. Experiments 1–3 all showed that dogs used olfactory stimuli to choose the larger number of hot dog slices, but Experiments 2 and 3 revealed no effects of distance or ratio between numerical quantities. In Experiment 4, dogs failed to discriminate between 1 and 5 pieces of dog kibble. Factors that allow dogs to use olfactory cues to discriminate between quantities are discussed.

Rats' working memory for locations previously visited and not visited was tested on the radial maze. Trials consisted of a study phase followed by a test phase. In the study phase, rats were forced to visit half the arms on the maze, with the other half of the arms blocked. In the test phase, rats chose among all arms, with food found only on the arms not visited in the study phase. When different patterns of arms visited in the study phase were used, it was found that rats remembered an alternating pattern better than an adjacent pattern or a random pattern and that this effect became more pronounced at longer retention intervals. In addition, rats remembered isolated nonvisited arms in the random pattern that were sandwiched between visited arms better than nonvisited arms that were not isolated between visited arms. Several hypotheses were examined and tested to explain this isolation effect, but no clear theoretical account was found. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

We tested the information preferences of three different species; pigeons, rats and dogs. Eight animals of each species received forced trials that produced one of two stimulus sequences. In the first sequence, response to an initial stimulus led to one of two other stimuli, one of which guaranteed a food reward was coming and the other of which guaranteed no food reward was coming. In the second sequence, response to an initial stimulus led to one of two other stimuli, both of which predicted food reward on 50% of the trials. The net reinforcement rate for both of the sequences was 50%. On probe test trials, both initial stimuli were presented, and the subject chose between the informative and the non-informative cue, and the percent choice of the information sequence, in which stimuli predicted food or no food reliably, was recorded for each species across 10 sessions. Statistical tests showed that although pigeons showed a preference for the information sequence, neither rats nor dogs showed this preference. Experimental and ecological explanations are discussed.

Disagreement has arisen in the scientific literature regarding the relative olfactory ability of humans relative to other mammals, specifically canines and rodents. A series of experiments are reported in which memory for multiple olfactory discriminations was measured in dogs, rats, and humans. Participants from all three species learned a sequence of 20 different discriminations between an S + odor and an S- odor. Choice of the S+ odor was rewarded with food for dogs and rats and with positive verbal feedback for humans. After learning the discriminations, an initial memory test was given that involved presentation of all 20 S + and S- pairs. A subsequent mix-and-match test was given in which each S + odor was presented with three different S- odors. The memory tests revealed that dogs were superior to rats and that dogs and rats were superior to humans. The relatively poor performance of humans contrasts with prior findings of high recognition memory for odors followed by slow forgetting. We attribute the low accuracy of humans in our experiments to the requirement that participants had to remember the outcome associated with S + (correct) and S- (incorrect) cues and not just their familiarity. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

Howard, Avargues-Weber, Garcia, Greentree, and Dyer (Science Advances, 5,1–6, 2019) report experiments in which honeybees initially shown a number of shapes could subsequently choose a pattern that added or subtracted one from that number. Further, the operations of addition and subtraction were cued by the colors of the shapes.

Past research has shown that testing memory in the same context in which the memory was encoded leads to improved retention relative to testing memory in a new context. Context-dependent memory is directly related to the extent to which the encoding context can be reproduced. An experiment with pigeons is reported in which the context was a colored house-light that completely enveloped the learning and testing contexts. Under this condition, perfect retention of a visual discrimination that reversed at midsession was shown. Beyond reactivation of memory, new research with pigeons suggests that context provides access to different working and reference memory systems. Finally, experiments are reported that suggest context may selectively access information about features from the different dimensions of place, color, and time.