Revisiting the concept of behavior patterns in animal behavior with an example from food-caching sequences in wolves (Canis lupus), coyotes (Canis latrans), and red foxes (Vulpes vulpes).

We discuss the history, conceptualization, and relevance of behavior patterns in modern ethology by explaining the evolution of the concepts of fixed action patterns and modal action patterns. We present the movement toward a more flexible concept of natural action sequences with significant degrees of (production and expressive) freedom. An example is presented with the food caching behavior of three Canidae species: red fox (Vulpes vulpes), coyote (Canis latrans) and gray wolf (Canis lupus). Evolutionary, ecological, and neuroecological/neuroethological arguments are presented to explain the difference in levels of complexity and stereotypy between Canis and Vulpes. This article is part of a Special Issue entitled: Canine Behavior.

Developmental changes in the sequential behavior of interacting timber wolf pups.

We used information statistics to quantify first-order sequential dependencies in the social behavior of two sibling wolf pups (Canis lupus). Sequential dependencies in the behavior of the individual pups increased between the first sample (18-32 days of age) and two later samples taken from 34-53 days and 64-106 days of age. Sequential dependencies between pups were greatest during the second age sampled. We relate these findings to changes in the importance and style of interactions. In the first 32 days, social co-actions primarily involved mutual mouthing and pushing. Between 34 and 53 days, when interactions frequently became aggressive, pups were most responsive to the actions of their partner and tail-raising predicted biting. By 64 days, play-chasing was the mode. By this age pups had formed a relationship, were less attentive to the specific actions of their partner, and biting was no longer predicted by tail-raising. These results illustrate the importance of considering demand characteristics in distinguishing capabilities from performance, and suggest that changing social organization may underlie developmental changes in behavior. We also speculate that during these play-like interactions, pups may gain control over expressive displays.

The development of swimming behavior in the neurological mutant weaver mouse.

Weaver (wv/wv) mice have well-specified ontogenetic defects in both the cerebellum and striatum, but have not previously been evaluated systematically for patterns of motor development. In this study, the effects of the weaver mutation were evaluated through an examination of swimming behavior over the first 3 postnatal weeks. Detailed movement analyses of individual limb movements as well as interlimb coordination were used to evaluate the effects of the weaver mutation. Weaver mutant mice displayed a developmental lag in terms of swimming style relative to controls. They also displayed a generalized slowness in limb movements during the swim, which correlated with the developmental onset of use of a particular limb during the swim. However, basic motor patterns in weaver swimming continue to exhibit good overall coordination through the 3rd postnatal week, even though locomotor ataxia has become pronounced by this time. Our results indicate that specific and limited alterations in movement can be traced to very early in development (postnatal Day 3) in weaver mutant mice, a time at which the earliest biochemical and neuroanatomical deficits in these animals have been established. Our results also emphasize the need for systematic contextual analyses of movement to understand interlocking processes both in movement ontogeny and its disorders.

Separation of activation and pattern in grooming development of weaver mice.

The effects of environmental conditions and age on grooming behavior were examined in weaver mutant mice and control littermates. Due to deficits in both the cerebellum and the dopaminergic system, weaver mice provide an opportunity to investigate how both of these systems are involved in grooming. Although homozygous weaver (wv/wv mice display deficiencies in grooming behavior, our results indicate that these effects are both context and age dependent. Overall wv/wv mice spent less time grooming than did controls. However, during the post-swim period wv/wv, after day 13, reached the grooming levels of pre-swim control mice. After day 15 wv/wv mice showed a higher number of post-swim grooming bouts relative to pre-swim, and in fact exceeded the number of bouts performed by controls in either pre- or post-swim conditions. Although controls displayed longer bouts than mutants overall, during the post-swim period wv/wv mice, after day 13, produced bouts as long as the control animals did pre-swim. This could in part reflect activation by previous swimming. Our data indicate these activational effects can be separated from balance or posture problems. From examination of the individual grooming stroke types used by the two groups, it is evident that the strokes used by mutant animals clustered around the early grooming sequence phase. In contrast, some of the later strokes were never used by the wv/wv mice during the entire developmental period studied. Our results emphasize the importance of using multiple measures of an action sequence and testing under different conditions.

An acoustic basis for maternal recognition in timber wolves (Canis lupus)?

An in-den recording system was used to monitor the vocalizations and behavior of adult wolves tending to a litter of pups during the first five postnatal weeks. Two female adults, one of them the mother, tended to the pups on nonoverlapping schedules. The distributions of the fundamental frequencies of the adults' squeak vocalizations were largely nonoverlapping, suggesting that this feature may be available as an acoustic cue to individual recognition. Squeaks emitted outside the den, and which were associated with pup exit responses, had fundamental frequencies wholly within the range of the mother's, raising the possibility that the pups used this as a cue for maternal recognition.

Neurological dysfunction expressed in the grooming behavior of developing weaver mutant mice.

The present study provides the first quantitative developmental analysis of movement in the weaver (wv/wv) mutant mouse. This autosomal recessive mutation affects both striatal and cerebellar circuitries that are related to motor performance. We report data on post-swim grooming behavior in 14 mutant and 14 control animals on alternate postnatal Days 13-20. Mutant animals showed a greater number, but shorter duration, of grooming bouts across this developmental period. The mutant animals also used external support during grooming, expressed various forms of ataxia, performed a higher proportion of smaller forelimb strokes than did the control animals, and failed to complete as many full stereotypic grooming sequences. These differences between mutant and control animals followed distinctive developmental courses. Our data demonstrate that previous anatomical and physiological characterizations of the weaver mutation have overt motor correlates.

History of developmental neuroethology: early contributions from ethology.

Ethology has its roots in the natural history of animal behavior. Questions of causation and function have historically been complementary, and each has rested upon a prior appreciation of the behavior of animals in nature. It is thus difficult to place a single time or place where ethology was born. Early evolutionary interests hinted at developmental constraints that continue to guide much research. It has only been relatively recently, however, that the explicit analysis of neural mechanisms in behavior has received the attention it deserves in developmental analyses. A mature developmental neuroethology will require a synthesis of the broad insights of ethology with refined neurobiological technique. Fundamental, however, is the primary focus upon behavior as it normally occurs.

Emergence of pattern in the development of mammalian movement sequences.

The manner in which behavior is patterned in space and over time represents a fundamental problem in both ethology and neuroscience. Prior to the analysis of mechanism it is important to be sensitive to issues involved in the provision of descriptive taxonomies. Often alternative modes of description lead to different perspectives and research strategies. In both the development of behavioral patterns and their expression a major question is how underlying organizational systems become self-organizing through the process of mutual interactions. It is clear that simple static dichotomies in both behavioral and developmental science must be replaced by more sophisticated models that emphasize the dynamics of pattern formation and control. Some of these perspectives are illustrated from our ongoing research on rodent movement patterns.

Food caching in captive coyotes: stereotypy of action sequence and spatial distribution of cache sites.

This paper describes some aspects of the food caching behaviour of four captive coyotes. Detailed observations of the actions used by coyotes to cache food revealed them to be strikingly similar to those previously described for timber wolves. The similarities included the identity of the movements used, their temporal sequencing, and their susceptibility to interruption. This suggests that there exists a stereotypy across canids in the action sequences used in caching. Second, an examination of the distribution of cache sites revealed that each coyote scattered cache sites widely within a wooded region of their enclosure and preferentially in terrains close to exposed tree roots.

Evaluation of immunoassay methods for the screening of cocaine metabolites in urine.

Immunoassay kits for urine cocaine (and metabolite) screening, obtained from two commercial sources, were examined for correlation of their results, expressed in terms of equivalent benzoylecgonine concentration, with the gas chromatography/mass spectrometry (GC/MS) concentration of benzoylecgonine. The correlation coefficients obtained, based on 62 (out of a total sample population of 3295) highly relevant samples, were 0.467 and 0.766 for Abuscreen (ARIA) and TDx (TDX), respectively. The preliminary screen cutoff values, which correspond to 150 ng/mL benzoylecgonine (as determined by GC/MS), were calculated based on the resulting regression equations and found to be 380 and 190 ng/mL for ARIA and TDX, respectively. With these cutoff values, ARIA generates 5 false negatives and 16 unconfirmed presumptive positives, while TDX results in 3 false negatives and 6 unconfirmed presumptive positives.

Food-caching in timber wolves, and the question of rules of action syntax.

This report presents data on the sequence of motor operations used by captive timber wolves to cache food. Videotapes were obtained of 151 caching episodes by 8 wolves. The vast majority of these episodes contained 3 distinct phases, each composed of movements unique to that phase. The excavation of the cache site was always done with the forefeet, and burying of the food was always done with the snout. Both the identity of the movements, and the serial order of phases were independent of the sex of the animal, the season in which the observations were made, and the nature of the substrate. A comparison of the temporal sequencing of these actions with the temporal stereotypy seen in rodent motor patterns (e.g. grooming) revealed a striking phenomenological similarity. The factors shaping the temporal sequencing in the two behaviors are, however, probably very different. This is because much, though not all, of the temporal stereotypy in the sequence of movements used by the wolf in caching is constrained by the logistics of the cache operation, while this is not the case for the phases of facial grooming in rodents. The implications of our data for the kinds of behavioral evidence required for ascription of such stereotypy to a central pattern generator are discussed.

A triggered hyperkinesia induced in rats by lesions of the corpus striatum.

The role of the corpus striatum (caudate, putamen, and globus pallidus) in movement control has been suggested to involve the modulation of sensory traffic to downstream motor mechanisms. We report that kainic acid lesions of the posterior corpus striatum, which preferentially spare fibers of passage while destroying striatopallidal neurons, produce a stimulus-sensitive movement pattern in rats that has a highly specific sensory trigger. The triggered choreic movement pattern is not a motor pathology per se, nor a response to diffuse states of arousal or stress, but rather is activated specifically in response to oral sensory stimulation. This sensory-specific hyperkinesia may be relevant to certain human sensorimotor pathologies.

Natural syntax rules control action sequence of rats.

Knowledge of the principles by which behavioral sequences are generated is essential to progress in our understanding of neural mechanisms. We describe here a set of natural principles or syntax rules that organize the components of grooming and feeding. The behavioral stream of facial grooming or of taste-elicited ingestive/aversive consummatory actions of rats can be viewed as a long series of individual movements linked together to form functional sequences. In order to ascertain the syntax rules that determine how these actions are linked together, many thousands of spontaneous grooming and elicited ingestive/aversive actions were videotaped and scored with a microcomputer. Techniques of information analysis of sequential stereotypy, tabulation of the sequential transitions between single actions and between action groups, and visual inspection for linear action chains, were employed to expose underlying rules of behavioral sequencing. These analyses revealed two global patterns: action perserveration and transitional reciprocation between sequential pairs and triplets, which together account for approximately 75% of all sequential transitions during grooming and ingestion/aversion. The pattern of transitional reciprocation could be divided further into patterns of alternation between individual actions on the one hand, and between perseverating bouts of actions on the other. Global syntax rules applied equally to actions emitted during grooming or during taste-elicited ingestion/aversion. In addition, a specific rule of linear chaining was found to apply only to facial grooming. These natural rules of action syntax provide insight into the sequential structure of behavior, and lend themselves well to analyses of neural mechanisms.

Deafferentation does not disrupt natural rules of action syntax.

Natural rules of action syntax control the sequential order of grooming and ingestive/aversive actions emitted by rats. Grooming and ingestive actions share a common feature in that all are performed with or directed towards the mouth, tongue, and face. This study examined the role of orofacial somatosensory cues and feedback in the generation of natural action syntax. Bilateral deafferentation of the mandibular and maxillary branches of the trigeminal nerve was used to eliminate tactile sensation from the rostral face and mouth while preserving motor function. Neither the overall degree of sequential stereotypy (H) of grooming or ingestive sequences, nor the generation of particular natural sequencing rules were affected by trigeminal deafferentation. These natural rules appear to be specified by the brain without need of somatosensory feedback.

Contextual control of trigeminal sensorimotor function.

Simple actions, such as rhythmic tongue protrusions, forelimb facial strokes, and forelimb flails, are emitted by rats both during taste-elicited ingestion/aversion and during postprandial grooming. This study combined peripheral trigeminal deafferentation with a computer-assisted video analysis of action form to examine the use of cutaneous feedback from the face in action production. Changes in action form after deafferentation were found to be context-dependent: Deformations characterized rhythmic tongue protrusions when emitted in ingestive but not in grooming contexts. The opposite was true for alterations in forelimb action. Further, postprandial grooming as a whole was found to comprise distinct sequentially defined phases. Actions occurring in one highly stereotyped sequence phase were protected from deafferentation effects, although the same actions occurring outside of this phase were not. The results suggest that behavioral context (e.g., grooming versus ingestive set, sequence phase) can shift the integration of sensory guided and endogenous mechanisms that pattern simple actions.

Early ontogeny of face grooming in mice.

A U-shaped function of behavioral growth characterizes the early ontogeny of face grooming in mice. In the first few postnatal days (0-100 hr), mice groom their face by using temporally isolated strokes, or bouts of strokes, which vary in amplitude and symmetry. Later on (100-200 hr), bouts disappear, asymmetry is eliminated, and the amplitude of strokes is gradually restricted; the infants engage in stereotyped, double-handed, temporally isolated strokes. Finally (200-300 hr), bouts reappear, including both short and long, symmetrical and asymmetrical, strokes. These changes, are accompanied by unidirectional changes such as an increasing participation of the trunk, the neck, and the head in grooming, which lead to the flexible organization of face grooming seen in adults. Instead of describing development in terms of emerging unitary "acts," we have recorded from high-speed films three simultaneous aspects of movement: the movements of the separate limb and body segments, the resultant paths which are traced by the forepaws, and the paths of contact which are traced on the face. This method of description discloses (a) reversible changes, and (b) a change in the relative stability of each of these aspects of face grooming, in the course of early ontogeny.

Trigeminal-taste interaction in palatability processing.

Peripheral transection of the sensory branches of the trigeminal nerve in rats unbalanced palatability, selectively reducing the ingestive actions elicited by preferred tastes but leaving unchanged the aversive actions elicited by unpreferred tastes. The reduction in the number of positive ingestive actions occurred even though the capacity to emit these actions remained unimpaired. These findings show that there is an interaction between somatosensation and gustation in the processing of palatability.

The development of coordination.

My concern in this paper is to provide a framework for analyzing the development of coordinated action systems. By emphasizing the general theme of pattern formation in coordinated action, attention is drawn to dual problems of establishing separable dimensions of action that are in turn combined into higher-order configurations. During development processes of differentiation and integration are combined to make coordinated action possible. The rules by which this is accomplished, however, are still poorly understood. The perspective offered here is that to understand the development of coordinated action it is valuable to seek relative degrees of continuity-discontinuity and change-stability from several complementary perspectives. This avoids unnecessarily simple "unit" concepts of coordinated action, and thereby provides the flexibility necessary to clarify underlying developmental pathways.