Learning and Memory Mechanisms Involved in Compulsive Drug Use and Relapse - Drugs of Abuse: Neurological Reviews and Protocols - page 77

dependent habit, while inactivation of the dorsal striatum late in training
reveals that the hippocampus-dependent spatial strategy is still available, even
if normally unused (30) . The normal temporal transition from a cognitive to
a habitual mode of behavior is also clear in our own human behavior. For
example, driving a car along a complex route to a new workplace requires
considerable attention and a deliberate choice at each junction. With extended
practice, the task becomes automatic, performed at a low level of awareness—
the driver can simultaneously engage in an unrelated conversation, and no
delibrate choices need to be made.
A “habit” can be either a very “simple” response to a stimulus (e.g., a
lever-press) or a more complex, temporally extended action (as in the driving
example). In either case the component muscle group fi ring sequence, or
sequence of actions, becomes “chunked” (31) and thereafter performed in
a “closed-loop” fashion—smooth, effortless, without the need for online
monitoring. Having the ability to learn habits is useful because it allows
behaviors that have worked well in the past, over and over again, to be
performed without the need to expend time or effort in deciding the best course
of action. That is, habits allow evaluative decision-making to be skipped.
Avoiding the costs of decision-making does, however, produce an obvious
problem if circumstances are liable to change. While acting in a habitual mode
other, potentially more benefi cial actions will not even be properly considered.
To reduce this danger, habits are normally acquired slowly, thereby hopefully
refl ecting stable regularities of the world (invariances; 22 ). An automatic
program may even be engaged inappropriately unless care is taken—such
“actions-not-as-planned” are common in human behavior (32) . For example,
many people have had the experience of fi nding themselves driving to work
when another location was the originally intended goal. This refl ects the
“stimulus-response” nature of habits—no outcome is directly encoded as part
of the learned association (33) . In one recent experiment, rats were trained for
varying amounts of time in the win-stay task, using food rewards (34) . The
rats were then placed in another environment and the same food was paired
with a nausea-inducing injection (i.e., conditioned taste aversion) to “devalue”
that food. When subsequently returned to the win-stay task, rats with little
pretraining were slow to run on the task, as if they were less motivated than
before. In contrast, those rats that had received extensive pretraining ran swiftly
from the start to the food, despite the fact that after attaining the food, they
declined to eat it.
Such examples help to illustrate the critical difference between “reward”
and (positive) “reinforcement” (discussed in ref. 35 ). Usually, these are linked
processes—if an action produces consequences that are acutely evaluated as
pleasant (rewarding), the animal is persistently more likely to subsequently