Attention and Learning

A dog's ability to concentrate selectively on specific aspects of the environment and to exclude others is a faculty of tremendous importance for effective dog training and behavior modification. Historically, the study of attention was neglected due to a widely held belief that the scientific investigation of behavior ought to be restricted to the study of measurable units resulting from the interaction of external events (i.e., stimuli and responses). This general doctrine known as radical behaviorism rejected cognitive phenomena like attention as inaccessible and irrelevant to a scientific understanding of animal behavior. Radical behaviorists also rejected explanations that employed physiological hypotheses and constructs.

The artificial dissection of behavior from its cognitive and physiological underpinnings was an unfortunate stratagem—one that logically precluded from the outset the possibility of a complete and holistic theory of behavior and learning. Interestingly, though, in spite of the outward rejection of attention as a worthy subject of scientific psychology, the radical behavioristic tradition has indirectly provided a rich and useful foundation and methodology for its investigation by contemporary cognitive psychologists (Cohen and O'Donnell, 1993). It should be stressed, however, that not all learning theorists historically rejected attention as a subject for study. Thorndike, for example, posited a subordinate law that he called the prepotency of elements in which certain features of the environment are selectively attended to by an animal on the basis of their prepotency or usefulness for solving a problem. Also, many modern learning theorists have contributed significantly to the study of attention; espe cially relevant here is the work of Rescorla, Kamin, and Macintosh (Hall, 1991). in addition, the dogmatic viewpoint regarding atten-tional behavior has yielded to some thoughtful revisions in recent textbooks covering the topic of attention and attending behavior (Schwartz, 1989; Catania, 1992; Lieberman, 1993).

Attention is perhaps the most basic class of behavior in which both classical and instrumental elements closely cooperate in the mediation of effective perception and action. What a dog pays attention to from moment to moment involves the participation of a complex cognitive gateway or interfacing mechanism processing information from within the animal (e.g., motivational states) and coordinating it with events and opportunities occurring outside of the animal within the changing circumstances of the environment. This cognitive gateway is regulated by a variety of motivational, perceptual, and motor components in constant interaction. in a broad sense, attentional activities specify a dog's intentions, reveal a dog's motivational state, and to some extent define what a dog is prepared to learn—that is, attentional activities reflect a dog's overall disposition to learn.

At the most basic level, all learning requires that an animal exhibit an active attention toward the training situation. As noted in the previous chapter, surprising or startling reinforcers produce the strongest effect on behavior. Such events also evoke the keenest interest and attentional focus—that is, startle and surprise serve to emotionally mediate and potentiate attentional behavior. Lieberman (1993) has called such events markers, suggesting that surprising/startling events produce dramatic effects on learning. Stimuli that lack surprising or startling qualities tend to drift into the background, become progressively irrelevant, and are eventually ignored. in the case of classical conditioning, conditioned stimulus (CS) salience or interest depends on its predicting some element of surprise. Once a CS exactly predicts the extent of the unconditioned stimulus (US), interest in the CS as a source of information gradually diminishes. This does not mean that the CS is considered to be irrelevant (ir relevance occurs when the CS occurs independently of the US), but that it is no longer actively followed or paid much attention to because there is not much more to be learned from its occurrence. Instead of actively attending to the CS, such well-conditioned stimuli are responded to in a progressively mechanical and automatic way. However, if suddenly a larger-than-expected US (e.g., consider the case of a bonus or jackpot) occurs in the presence of the diminished CS (e.g., "Good"), then new interest and attention will be generated by the future presentation of the conditioned reinforcer "Good."

Besides facilitating classical conditioning, markers also appear to play a very significant role in instrumental learning. Surprising events potentiate learning abilities, even promoting learning occurring under adverse conditions. As has been previously discussed, any delay of reinforcement usually has a deleterious effect on learning. However, Lieberman and his associates (1979) studied various situations in which these adverse effects could be overcome by utilizing a marking event. For example, they placed rats in a T maze where a food reward was delivered after a minute delay, provided that the rat chose the right arm of the maze. if the rat chose the left arm instead, no reward was delivered. The rats, as one might predict, failed to learn the correct response required to obtain the belated reward. However, the experimenters found that if the subjects were picked up immediately after they made their choice (whether correct or wrong) and were then placed back into the maze to complete their chosen route, the handled subjects learned the correct route much more effectively than nonhandled controls. During testing, the handled rats chose the correct arm in 90% of the trials, whereas the controls were only correct 50% (chance) of the time. Similar effects were observed in the case of other surprising/startling stimuli (e.g., light or noise) that were presented immediately after the rat's choice (correct or wrong) was made. Again, the food reward associated with the correct choice was always delayed. Lieberman and colleagues speculate that markers enhance the functioning of at-tentional processes and memory coding of relevant cues, with the marker evoking extra attention to events occurring immediately prior to its presentation. The overall effect is to make marked events more likely to be remembered and associated with remote outcomes, such as the delay of reinforcement as in the foregoing experiment.

Obviously, stimulating and controlling at-tentional behavior is of considerable interest to the trainer/behaviorist. Dogs pay attention to occurrences that are significant to them and learn to ignore occurrences that are irrelevant. Stimuli that have been associated with hedonically significant events or fear tend to attract more attention than neutral stimuli not having undergone such conditioning. In addition, previously conditioned stimuli tend to overshadow neutral stimuli occurring coin-cidentally in the training situation, thus blocking an associative connection from developing between them and the relevant US—a classical conditioning process that has been investigated in detail by Kamin (1968). Those elements of the environment that do not hold a dog's active attention are of little significance to the learning process. In short, selective attention allows dogs to focus on relevant stimuli while ignoring irrelevant occurrences competing for their attention. Without this ability to attend selectively to environmental events, dogs would not only be unable to learn, they would be thoroughly incapacitated by a disorganized influx of chaotic stimulation. Clearly, attention plays a very significant role here in terms of transforming raw experience into informative input about the environment.

Since attention is highly correlated with reinforcement (both positive and negative), it follows that animals should become more attentive with experience. This analysis implies that reinforcement of attention in one situation should improve attending behavior in other more remote training situations. Attending behavior may be reasonably interpreted, therefore, as a higher-order class of behavior that contains a large subclass of behaviors in which attention plays an instrumental role. In addition, since attending behavior is present in most successful learning situations, it may be considered the most dominant class of higher-order behavior, under which all other classes of instrumental behavior are subsumed according to their relative frequency and probability. According to this line of analysis, the most likely behavior in any given learning situation is an atten-tional orientation (physical and perceptual) toward significant training stimuli. It is astonishing to consider that the most dominant class of instrumental behavior—attending— has been the least carefully studied.

The sort of stimuli that attracts a dog's attention frequently reveals an underlying biological significance and purpose being served by them as well as past learning. Many of these stimuli and events naturally attract and hold a dog's attention even prior to learning. For example, the vaginal discharge produced by an estrous female attracts the intense attention of an intact male even before his first sexual experience. The discharge contains pheromones (i.e., chemical signals) that trigger or mediate interest via olfactory stimulation of appropriate brain centers controlling sexual activity. These hardwired connections are established prior to actual sexual contact with a receptive female. A dog's sensory faculties also predispose it to react to certain stimuli in a relatively fixed manner. visual cliff-avoidance reactions can be observed in young puppies (reliably after 28 days of age) prior to their experiencing any actual falls. Also, loud noises elicit intense startle reactions as soon as the ear canals open at about 3 weeks of age. The startle reaction to noise occurs without actually having been associated with previous aversive stimulation. Some sensory stimuli are attended to more keenly than others. Clearly, while color perception is of some usefulness to dogs, they are more apt to attend visually to movement, shapes, and shades of gray than to discriminate objects based on their color. Also, dogs appear to select certain classes of stimuli preferentially over others as cues when learning discrimination tasks. For example, Lawicka's (1964) discrimination studies have demonstrated that dogs prefer spatially lateralized discriminative stimuli or cues when learning a directional task and qualitatively differentiated cues (different tone frequencies) when learn ing go/no go discriminations. Also, Mc-Connell (1990) has reported that dogs selectively respond to auditory stimulation depending on its characteristics. Her studies suggest that active behaviors such as coming when called may be more easily associated with signals composed of rapidly repeated sounds (e.g., hand claps and smooches), whereas staying in place may be more easily obtained with continuous, drawn-out signals (see Chapter 5).

Another important influence on a dog's attending behavior is the dog's changing motivational state. Motivation has a pronounced influence on the class of stimuli that will attract and maintain a dog's attention. Hungry dogs are especially attracted to odors and conditioned stimuli that have been associated with food in the past. Highly social or socially deprived dogs will likely search the environment for opportunities to make contact with other dogs or people. Aggressive dogs scan the environment for evidence of people or other dogs that they might challenge or attack. Fearful dogs often engage in hypervigi-lant searching behavior in an effort to identify and avoid potential threats.

Dog Training 101

Dog Training 101

If you own a dog, you will have to do more than just feed him and give him toys to play with. Just about every dog, no matter what breed, will have some bad habits and behaviors that you will need to break.

Get My Free Ebook


Post a comment