Puppy Young Old

Figure 35.3 The combined sum of errors required to learn a delayed nonmatching-to-position task and an object discrimination and reversal task are plotted by age group. Aged dogs with scores greater than 2 standard deviations from the mean of the young dogs are considered impaired.

Puppy Young Old

Figure 35.3 The combined sum of errors required to learn a delayed nonmatching-to-position task and an object discrimination and reversal task are plotted by age group. Aged dogs with scores greater than 2 standard deviations from the mean of the young dogs are considered impaired.

young dogs to acquire the procedural learning tasks (Milgram et al., 1994). Thus, dogs with enriched experiences show a pattern similar to human learning, but having access to animals with restricted environments also revealed a dependency of learning ability on rearing history.

Discrimination learning A common measure of learning used in animals is the discrimination task—an associative learning problem in which subjects must attend to the relevant stimulus feature(s) and acquire an association between the stimulus parameters and a food reward (Sutherland and Mackintosh, 1971). In the object discrimination task, an animal must learn that one object is associated with food reward. The two objects used differ on multiple stimulus dimensions (i.e., size, color, brightness, shape). Studies in nonhuman primates (Voytko, 1999) and dogs (Head et al., 1998; Milgram et al., 1994) indicate that simple discrimination learning tasks, like the object discrimination, are insensitive to age. The discrimination learning task can be made more difficult when the differences between the objects are reduced. In the size discrimination task, the two objects are identical in appearance but differ only in size (Figure 35.2). The task now becomes more complex and age differences appear (Tapp et al., 2003a).

The importance of task complexity in evaluating discrimination learning in dogs is further demonstrated using an oddity discrimination task. In this task, the number of object choices is increased to make the task more difficult (Milgram et al., 2002b). The animal is presented with three objects, two identical and one different. To obtain reward, the animal is required to respond to the odd object. After reaching criterion, the animal is presented with three new oddity tasks where the odd object progressively becomes more similar in features to the other two identical objects, until four problems are solved. Old dogs are not only impaired at this task, but the number of total errors increases with each successive oddity problem, suggesting a lack of ability to learn the rule of selecting the odd object. Young animals, less than 6 years old, show rapid learning after the initial problem relative to old dogs, 9 to 13 years old, despite the increasing similarities in the appearance of the objects.

Object recognition learning Object recognition learning, the ability to classify and register the identity of objects, is examined using a delayed nonmatching to sample (DNMS) paradigm (Figure 35.2; Callahan et al., 2000). Originally developed by Mishkin and Delacour (1975) for nonhuman primates, the DNMS paradigm begins by presenting the dog with a single stimulus object located over the center food well. A response to the object is followed by a delay interval, after which, two objects are presented over the two lateral wells—the original object and a novel discriminandum. The correct response for the dog is to choose the novel stimulus. The location of the novel stimulus is randomized across trials, and different stimuli selected from a large pool of objects are used for each trial.

Initial attempts to test dogs on the DNMS task suggested that dogs were impaired at object recognition learning (Milgram et al., 1994). Using a weak criterion, only 4 of 10 young animals were successfully able to learn the task. Moreover, when tested on increasing delays, young dogs were able to respond with only 63% accuracy. None of the aged dogs could learn how to solve the problem. These results were initially interpreted as evidence that the canine visual system was not well suited for visual object recognition performance. A subsequent study, however, reported substantial improvements in object recognition learning in the dog when two important task-specific factors were modified (Callahan et al., 2000). The first was to control for the dogs' visual near point, the shortest distance from the dog's face for an object to be in focus, which is 25-30 cm. This ensured that objects were distinctly visible to the dog before making a choice on the nonmatch phase of the task. The second was to increase the visual processing time for the dog by introducing a 5-second pause between presenting the objects and allowing dogs to respond. Aged human subjects tested on delayed matching-to-sample tasks perform less accurately than young adults, but the level of performance in both groups improves when the sample stimulus duration is increased (Oscar-Berman and Bonner, 1985). With these two modifications, most young and some old dogs were able to solve the DNMS task at over 80% accuracy, but aged dogs generally performed more poorly than young animals (Callahan et al., 2000). A small proportion of severely impaired dogs, however, continued to fail the task, consistent with the strong DNMS impairments observed in Alzheimer's patients (Swainson et al., 2001).

Spatial learning One of the most consistent cognitive deficits in old dogs is the ability to acquire and use spatial information. In humans, spatial learning impairments are a common feature of aging and become more severe in neurodegen-erative disorders (Freedman and Oscar-Berman, 1989). Spatial learning, the ability to locate objects in space, can be accomplished in two ways: (1) by reference to the observer's body position (egocentric learning) or (2) by reference to the position of an external referent or landmark (allocentric learning).

In the dog, we have assessed egocentric spatial learning with a delayed nonmatching to position (DNMP) task. Two versions of this task were developed, a 2-choice DNMP (Head et al., 1995) and 3-choice DNMP (Chan et al., 2002). In the 2-choice DNMP (2cDNMP) task (Figure 35.2), dogs are presented with a single object located over one of two lateral food wells, the sample phase. After a brief delay, a second identical object is presented over the remaining lateral food well, the match phase. Only responses to the nonmatch location are rewarded with food. Some old dogs perform as well as young dogs on this task while others show marked impairments.

The 3-choice DNMP (3cDNMP) task uses three food wells (two lateral and one medial) instead of two to prevent the use of positional strategies in solving the 2cDNMP. Similar age-related deficits were obtained using the 3cDNMP task, but this version of the DNMP procedure is much more difficult for all dogs (Chan et al., 2002). Compared to only 18% of old dogs that failed the 2cDNMP (Adams et al., 2000 a,b; Head et al., 1995), 83% were unable to complete the 3cDNMP task. Young dogs never failed the 2cDNMP, but 12% were unable to solve the 3cDNMP.

Allocentric spatial learning in the dog is assessed with a landmark discrimination test originally developed in nonhuman primates (Pohl, 1973). In this task, subjects are rewarded for selecting one of the two identical objects that is closest to an external landmark. Solving the task relies solely on the use of allocentric spatial cues, because information about the correct response is provided only by the location of the landmark. Tests of allocentric learning in the dog indicate that landmark discrimination is far more difficult for both young and old dogs compared to the DNMP tasks. Initial attempts to train dogs on the landmark task were unsuccessful. The only way dogs could learn this task was to teach them to attend to the landmark. This was achieved by placing the landmark on top of one of the two identical objects and progressively move the landmark away from the object. Once the animals learned to attend to the landmark, allocentric spatial learning was observed in young and old dogs with age-related decrements in performance. Performance on this task declines in the old dogs as the distance between the landmark and the rewarded discriminanda increase (Milgram et al., 1999; 2002a).

More recently, our lab developed an egocentric discrimination learning task to assess spatial learning where dogs were rewarded for responding to an object farthest to the left or right side of the tray (Christie et al., 2005). Acquisition of the initial discrimination did not depend on age but the reversal task produced age-dependent declines in performance and correlated with performance on the landmark task. Thus, although the egocentric discrimination is similar to a visual discrimination, there is a common component with the allocentric spatial task suggesting that egocentric and allocentric based learning share some common neural substrates but are also subserved by unique brain regions.

Together, these results suggest that: (1) dogs in general may be predisposed to spatial learning based on an egocentric frame of reference, and (2) both egocentric and allocentric spatial learning are impaired with age.

Blood Pressure Health

Blood Pressure Health

Your heart pumps blood throughout your body using a network of tubing called arteries and capillaries which return the blood back to your heart via your veins. Blood pressure is the force of the blood pushing against the walls of your arteries as your heart beats.Learn more...

Get My Free Ebook

Post a comment