Motivation

Motivation is a modulating and coordinating influence on the direction, vigor, and composition of behavior. This influence arises from a wide variety of internal, environmental, and social sources and is manifested at many levels of behavioral and neural organization. As an illustration of motivational influence, consider the cyclical changes in sexual interest and receptivity shown by a female rat (McClintock 1984). Two days following the onset of her last period of sexual receptivity, she encounters one of the male rats with which she shares a communal burrow. The rats sniff each other indifferently and continue on their separate ways. Two days later, their paths cross again. Levels of gonadal hormones in the female’s blood  have increased markedly since her last encounter with the male, and she now responds in a strikingly different fashion, approaching, nuzzling, and crawling over him. Turning away, she runs a short distance and stops. The male follows hesitantly, but then spies a puddle and stops to drink, appearing to lose interest in the female. Undeterred, the female returns and repeats the pattern of approach and contact followed by turning, running away, and stopping. She soon succeeds in attracting and holding the attention of the male, and he follows her on a tortuous high-speed chase. The female then halts abruptly, and coitus ensues. During the first encounter, the female treats the male as a neutral stimulus, whereas during the second, she treats him as a valuable goal object. Had he grasped her flanks with his forepaws during the first encounter, the female would have kicked him away, but when he delivers similar sensory stimulation during the second encounter, she responds by adopting a posture that allows him to mount her and mate. During the first encounter, the female’s gait is similar to that of the male, whereas during the second encounter, her gait includes a distinctive combination of darts and hops followed by pauses that may be accompanied by vigorous ear wiggling. Finally, her prodding of the male, distinctive motor patterns, and postural response vary in intensity as a function of her hormonal status and recent experience. Such coordinated changes in the evaluation of goal objects (Shizgal forthcoming), the impact of external stimuli, the prepotency of sensorimotor units of action (Gallistel 1980), and the vigor of performance can be said to reflect a common motivational influence; the set of internal conditions responsible for this common influence can be said to constitute a motivational state.

The response of the male to the solicitation of the female illustrates the contribution of external as well as internal inputs to the genesis and maintenance of a motivational state (Bindra 1969). Such external inputs, called “incentive stimuli,” are important both to behavioral continuity and change. Positive feedback between incentive stimuli and motivational states tends to lock in commitment to a particular course of action. The more the male interacts with the female, the more he exposes himself to olfactory, tactile, and visual stimuli that increase the likelihood of further interaction. Thus his initial hesitancy gives way to vigorous pursuit. Moreover, sufficiently powerful incentive stimuli incompatible with a current objective can trigger an abrupt, self-reinforcing switch in the direction of the solicited behavior, as when the male rat is sidetracked from slaking his thirst by the intervention of the female.

Motivational states not only modulate the stimulus control of behavior, they also act as intermediaries in its temporal control by transducing internal and external signals indicating the season and time of day into changes in the likelihood of initiating different goal-directed activities. Timing signals make it possible for behavior to anticipate physiological need states, thus lessening the risk that supplies will be depleted and that the physiological imbalance will compromise the capacity of the animal to procure additional resources. For example, migrating birds eat voraciously and greatly increase their body weight prior to their departure. Nonetheless, anticipatory intake may prove insufficient to meet later needs or may exceed subsequent expenditures. Thus the contribution of motivation to the regulation of the internal environment depends both on signals that predict future physiological states and on signals that reflect current ones (Fitzsimons 1979).

To illustrate the regulatory challenges addressed by the motivational modulation of behavior, let us revisit the female rat as she begins to wean her litter, about six weeks after she has mated successfully. The total weight of her pups now exceeds her own. First via her bloodstream and then via her milk, she has succeeded in providing the necessary resources without seriously compromising her own viability. Accomplishing this feat has required dramatic alteration in her intake patterns. For example, her caloric intake and calcium consumption during lactation will have reached 2–3 times postweaning levels, reflecting both hormonally driven anticipatory changes and feedback from the physiological consequences of increased expenditures (Millelire and Woodside 1989).

The motivational modulation of preferences can extend beyond the point of neutrality, rendering previously repulsive stimuli attractive and vice versa (Cabanac 1971). Prior to her first pregnancy, a female rat will treat a rat pup as an aversive stimulus, positioning herself as far away from it as possible when placed together with the pup in an enclosure. In contrast, when she is in the maternal state, the female will actively retrieve pups, even if they are not her own (Fleming 1986). Changes in motivational state are expressed at many levels of behavioral and neural organization. For example, the posture adopted by a receptive female rat during copulation reflects the highly stereotyped operation of a spinal reflex (Pfaff 1982). Provided the female is sexually receptive, the reflex can be triggered in response to pressure on the flanks regardless of whether the stimulus is applied by the forepaws of the male rat or the hand of a human. Although facilitation from brain stem neurons is necessary for execution of the reflex (Pfaff 1982), the integrity of the cerebral cortex is not (Beach 1944). In contrast, the organization of the individual components of solicitation into the pattern of approach, contact, withdrawal, and pausing is contextsensitive, flexible (McClintock 1984), and dependent on cortical integrity (Beach 1944). The solicitation behavior of the intact female is directed preferentially at an appropriate sexual partner, and each of the constituent acts may be varied in intensity and duration or omitted entirely, depending on the response of the male. Following removal of the CEREBRAL CORTEX, components of the behavior survive, but their patterning is disrupted and is no longer well coordinated with the behavior of the male.

At higher levels of behavioral and neural organization, motivational states interact with cognitive processes in

influencing behavior. For example, the information directing ongoing behavior may be drawn from COGNITIVE MAPS of the environment rather than from current sensory input (Gallistel 1990; Marlow and Tollestrup 1982;). Another point of contact between motivation and cognition is the control of ATTENTION (Simon 1993). Changes in motivational state alter the likelihood that a stimulus will attract attentional resources, and directing these resources at a stimulus can boost its incentive effects (Shizgal 1996). By gating input to WORKING MEMORY, attention can restrict the set from which goals are selected and control the access of goal-related information to the processes involved in PLANNING. In the examples provided above, the objects of evaluation and goal selection are physical resources and activities. In humans, and perhaps in other animals as well, abstractions can serve as goals and as the objects of evaluation . For example, our evaluations of ourselves have profound motivational consequences (Higgins, Strauman, and Klein 1986), and our objectives may be defined with respect to current and projected self-concepts (Cantor and Fleeson 1993). Nonetheless, the psychological and neural foundations for such abstract expressions of motivational influence may have much in common with mechanisms, perhaps highly conserved across animal species, that modulate pursuit of concrete biological goals. Motivational influences are incorporated in some artificial intelligence models. For example, such signals provide contextual information in an important model of REINFORCEMENT LEARNING (Barto 1995), although the manner in which motivational signals are processed to modulate the impact of rewards and to guide action tends to be left unspecified in such models. A hierarchical account of motor control (Gallistel 1980) and recent modeling (Shizgal 1997, forthcoming) of the neural and computational processes underlying goal evaluation and selection  represent early steps toward formal description of the motivational influence on behavior.