Abstract

Obesity is associated with increased discounting of future monetary rewards, suggesting that altered decision-making may contribute to the risk for obesity. Here we examined the relationship between measures of decision-making, obesity and dieting success using an online sample (Amazon’s Mechanical Turk service) and an undergraduate male sample. Participants completed monetary or food-based delay-discounting surveys, and a set of experiential foraging tasks (the Movie Row and Candy Row tasks). Rodent versions of the foraging tasks have also been used to demonstrate specific deficits in decision-making in drug-abstinent mice.

In the foraging tasks (Movie Row and Candy Row), participants navigated through a 3D virtual environment on a square track. Rewards (4-sec video clips displayed on a virtual movie screens in the Movie Row task or candy/snacks delivered to magazine for the Candy Row task) were available from four zones on the track. Only a subset of undergraduates completed the Candy Row task. As participants arrived at a zone, the reward category and a random delay were presented. Participants could accept the offer or move on to the next reward site. Behavior on the Candy Row task (using physical rewards) was similar to that of the Movie Row tasks (using videos as rewards), suggesting that these two versions of the task were sensitive to individual differences in rewards. Performance on the monetary and food-based versions of the delay-discounting measure were also positively correlated, but delay discounting measures were not well related to measures from the foraging tasks.

Obesity (determined from self-reported weight and height) was not related to delay discounting measures (k) for money or food, or to the number of rewards earned on the foraging tasks and delay thresholds (how long participants were willing to wait for a reward). Exploratory analyses found that obesity was associated with reduced regret on the Movie Row task. Self-reported change in body mass index (BMI) over the previous year was related to reported use of several dieting strategies (such as attempting to consume less food, avoiding carbohydrates), but not to measures of delay-discounting. However, several dieting strategies interacted with the discounting rate (k). Among participants who reported that they had attempted to eat less food to manage their weight, only participants with low discounting rates showed a reduction in BMI, while for participants who did not report using this strategy, there was no relationship between k and changes in BMI. These results suggest that obesity and dieting success may be related to several different disruptions in decision-making systems.

Abstract

Foraging tasks provide valuable insights into decision-making, as animals make decisions about how to allocate limited resources (such as time). In the “Restaurant Row” task, rodents obtain food rewards at several sites. In the “Web-Surf” task, humans are offered short videos from different categories. In both tasks, rewards are available after a variable delay. While rodents physically move between reward sites, humans transition to the next video type by pressing a button. Rodents and humans show similar patterns of behavior on these tasks, supporting their use to study decision-making across species. We tested a new human task (“Movie Row”), which combines elements of the Restaurant Row and Web-Surf tasks.

In the Movie Row task, participants navigate through a 3-d virtual environment on a rounded track. A movie screen is positioned at each corner of the track, where 4-sec video clips are available from one of four categories (kittens, bike accidents, landscapes, and dancing). As participants arrive at a screen, the category and loading time (delay) is presented. Participants accept offers by stepping onto a platform and looking at the screen, or skip an offer by moving to the next reward site. Delays range from 3 to 29 seconds (though longer delays were possible in the second version).

Two versions of the task were tested, the first in a sample of 30 male undergraduates who completed the Movie Row task in a laboratory. The second version was tested in an online sample of 51 participants (30 females, mean age = 42.2 years) recruited through Amazon’s Mechanical Turk.

Behavior on the Movie Row was similar to that of the Web-Surf task. Revealed video preferences, assessed by delay thresholds (the delay below which they would accept, above which they would skip an offer) were positively correlated with stated preferences, with more than 64% of participants having a correlation greater than +0.75 with the average video rankings and with post-experiment rankings of each video category. Decision consistency (the proportion of offers where participants deviated from their delay threshold) was lower on the first (in-person) version (M = 0.12, SD = 0.05) than the second (online) version (M = 0.20, SD = 0.11), higher than the reported mean for the Web-Surf task (M = 0.07), but similar to the Restaurant Row task (M = 0.12).

These data support the Movie Row task as translational tool to study foraging behavior in humans, providing convergent results with a rodent navigation task and a human stay/go foraging task. And, by including a virtual navigation component, the task may be useful in testing if behavioral measures identified in the rodent Restaurant Row task translate to human virtual navigation.

Abstract

Aim: Disorders such as Alzheimer’s Disease involve memory deficits and grey matter loss in the hippocampus and entorhinal cortex are predictors of future clinical diagnosis. Early identification of memory deficits is critical for treatment. Our lab has developed a navigation task, the Concurrent Spatial Discrimination Learning Task (CSDLT), to assess the use of hippocampal-dependent memory in healthy adults.

Methods: In the virtual navigation task, participants navigate on a 12-arm radial maze, to find objects hidden at the end of the arms of the maze. The twelve arms are divided into six pairs of adjacent arms, which are presented in a pseudorandom order. In each pair, one arm is rewarded, and participants are trained until they can correctly choose the rewarded arm in each pair. After reaching criterion, participants complete a set of probe trials in which they are presented with a recombined set of adjacent arms, in which the previously rewarded arm is paired with a new unrewarded arm. In the recombined pairs, participants relying on a hippocampal strategy will learn the location of the correct arm relative to distal cues, hence perform well on this probe trial. Participants using non-hippocampal strategies make more errors, tending to choose an arm based on its stimulus-response associations (e.g. “choose the left path”). In this study, we developed a short static version of the CSDLT task, in which participants are presented with images of the pairs of arms, but were not required to navigate in virtual reality to the end of the arm on each trial. Instead, participants click on a button to select the arm displayed on an image. Participants are given feedback whether their choice was correct or incorrect. To validate the short version of the CSDLT task, we tested participants recruited through Amazon’s Mechanical Turk (mTurk) service on an automated version of the original CSDLT task, and the new shortened version.

Results: Participants (n = 35 who completed both tasks) completed the short CSDLT task much faster (M = 7.4 minutes, SD = 4.2) compared to the original CSDLT (M = 18.2, SD = 7.3, t(34) = 8.3, p < 0.0001, d = 1.4). Performance on the probe trials was correlated across the two tasks (r(33) = 0.412, p = 0.014) and a similar relationship was obtained when the short version was administered as the first (r(18) = 0.47, p = 0.036) or second (r(13) = 0.49, p = 0.061) task.

Conclusion: These preliminary data support the use of the shortened CSDLT task to assess an individual’s preference for spatial strategies in navigation. While this task was not validated against MRIs, it may be helpful as a screening tool for an initial rapid assessment in clinics when time is limited, which can be followed by more extensive testing in at risk individuals.

Abstract

A previous study by Packard and McGaugh (1996) examined the use of spatial and egocentric navigation strategies by training rats to obtain a food reward in a plus-shaped maze. With one arm closed off to form a T-shaped maze, rats were trained to find food placed in one arm of the T. During the training phase, the rat began and retrieved food from the same locations. On a probe trial, rats were placed in the arm opposite to the starting location. If the rat could accurately locate the reward, it was demonstrating a place strategy. Conversely, if the rat turned in the same direction as it did in the training phase, it was navigating via a response strategy. However, the experimental paradigm used by Packard and McGaugh may have its limitations. When conducting a probe trial in a plus-shaped maze, it is difficult to confirm that animals are relying specifically on either a place or response strategy, as any choice made by the animal, even a random search, counts as a strategy: when approaching the food reward, rats were only able to turn in two directions. As a result, the rats have no other choice than to exhibit either a response or place strategy. In the present study, we introduced a modified maze design that included other areas where the rat could travel during a probe trial, using a symmetrical maze with four starting locations and four goal locations. Using this new maze configuration, we aimed to replicate the findings of Packard and McGaugh. We hypothesized that rats will adopt a place strategy, and transition to a response strategy after extended training. We expected also that two rats with excitotoxic lesions of the hippocampus would show a deficit in the use of a place strategy. We also will examine other behavioral measures (reaction time, latency to reach the food reward, and consistency of the path taken to the reward) to identify other measures (beyond accuracy) that identify the use of place and response strategies. We expect that rats using a place strategy to take significantly longer reaction time, and will take a more variable path to the food reward.

Abstract

Mindfulness training (mindfulness-based stress reduction, MBSR) has been found to be effective in reducing stress and improving cognitive function, and to produce changes in regional grey matter density in brain areas including the hippocampus. Most studies of MBSR’s effects on cognitive function have focused on improvements in attention and working memory, but some studies have found improvements in declarative memory following MBSR training. Other work has shown that even in participants who have not completed mindfulness training, regional grey matter density in the hippocampus is positively associated with measures of dispositional (trait) mindfulness. Here, we investigated if dispositional mindfulness was associated with the use of hippocampally-dependent navigation strategies in a set of virtual navigation tasks. Participants (n = 111) were recruited through Amazon’s Mechanical Turk service after completing the Five Facet Mindfulness Scale (FFM). Participants completed up to four new, automated versions of standard navigation tasks: the 4 on 8 virtual maze (Iaria, Petrides, Dagher, Pike, & Bohbot, 2003), the Concurrent Spatial Discrimination Task (CSDLT, Etchamendy, Konishi, Pike, Marighetto, & Bohbot, 2012), a wayfinding task set in a virtual town and a virtual water maze. Results from the behavioral tasks fit our expectations based on previous research (performance on the wayfinding and water maze task was strongly related, and flexible performance on the CSDLT was associated with better performance on the wayfinding and water maze task). Only the 4 on 8 task was associated with trait mindfulness, with participants who reported using hippocampus-dependent spatial strategies also reporting higher scores on the Describe FFM subscale but not other FFM subscales. In a second study, participants completed the Calgary Symptoms of Stress Inventory (C-SOSI), and participants (n = 96) with high (C-SOSI > 50) and low (C-SOSI < 50) stress scores were recruited to complete the FFM and the 4 on 8 virtual maze. C-SOSI scores were negatively correlated with several FFM subscales, and participants reporting high levels of stress were less likely to use hippocampus-dependent spatial strategies (though, a small number of participants reporting very high scores, C-SOSI > 100, were more likely to use spatial strategies). These data provide some support for the claim that hippocampus-dependent spatial strategies may be associated with both dispositional mindfulness and self-reported stress. Programs which aim to reduce stress by increasing mindfulness (as in MBSR programs) may thus be associated with an increase in the use of hippocampally-dependent navigation strategies.

Abstract

Abstract

Abstract

Abstract

Many addictive drugs facilitate release of dopamine in the basal ganglia, and drug addiction may involve the recruitment of brain systems that support habitual behaviors. Work with operant tasks has shown that relative to natural rewards, rats develop habits for actions rewarded with drugs such as cocaine and alcohol. Using post-training injections of cocaine, our lab has shown that cocaine can act directly to facilitate habit learning, and that the infralimbic cortex is required for habit facilitation by cocaine. The present study addressed if cocaine infusion directly into the infralimbic cortex was sufficient for the facilitation of habit learning. Twenty seven Sprague-Dawley rats were implanted with bilateral cannula targeting the infralimbic cortex. Rats were trained after surgery to perform two actions (lever pressing) which were rewarded with either grain or sucrose flavored pellets. After acquiring both lever presses, rats received three training sessions on each lever, rewarded on a random interval 30 second schedule (RI-30s). After each RI-30s training session, rats received infusions of cocaine (200 micograms/microliter) or vehicle. Using sensory-specific satiety, extinction tests revealed that for the lever paired with cocaine infusions, lever pressing was not sensitive to devaluation of the reward while the lever paired with saline injections were sensitive to devaluation. These data suggest that drugs such as cocaine can act directly during memory consolidation in brain areas required for habitual behavior in order to facilitate habit learning.

Abstract

Many addictive drugs facilitate release of dopamine in the basal ganglia, and drug addiction may involve the recruitment of brain systems that support habitual behaviors. Work with operant tasks has shown that relative to natural rewards, rats develop habits for actions rewarded with drugs such as cocaine and alcohol. Using post-training injections of cocaine, our lab has shown that cocaine can act directly to facilitate habit learning, and that the infralimbic cortex is required for habit facilitation by cocaine. The present study addressed if cocaine infusion directly into the infralimbic cortex was sufficient for the facilitation of habit learning. Sixteen Sprague-Dawley rats were implanted with bilateral cannula targeting the infralimbic cortex. Rats were trained after surgery to perform two actions (lever pressing) which were rewarded with either grain or sucrose flavored pellets. After acquiring both lever presses, rats received three training sessions on each lever, rewarded on a random interval 30 second schedule (RI-30s). After each RI-30s training session, rats received infusions of cocaine (200 g/L) or vehicle. Using sensory-specific satiety, extinction tests revealed that for the lever paired with cocaine infusions, lever pressing was not sensitive to devaluation of the reward while the lever paired with saline injections were sensitive to devaluation. These data suggest that drugs such as cocaine can act directly during memory consolidation in brain areas required for habitual behavior in order to facilitate habit learning.

Abstract

Abstract

Abstract

Abstract

Abstract