R172 0363-6119/91 $1.50 Copyright Oc 1991 the American Physiological Society
CARDIOVASCULAR RESPONSES AND BEHAVIOR

Integrating behavior and cardiovascular responses: the code

CLIFFORD A. ASTLEY, ORVILLE A. SMITH, ROGER D. RAY, EUGENE V. GOLANOV, MARGARET A. CHESNEY, VALERIE G. CHALYAN, DAVID J. TAYLOR, AND DOUGLAS M. BOWDEN Regional Primate Research Center and Department of Physiology and Biophysics, University of Washington, Seattle, Washington 98195; Department of Psychology, Rollins College, Winter Park, Florida 32789; Prevention Sciences Center, University of California, San Francisco, California 94105; Myasnikou Institute of Cardiology, Moscow; and Institute for Experimental Pathology and Therapy, Sukhumi, Union of Soviet Socialist Republics

cardiovascular dynamics; environmental changes; behavioral changes

ASTLEY, CLIFFORD A., ORVILLE A. SMITH, ROGER D. RAY, EUGENE V. GOLANOV, MARGARET A. CHESNEY, VALERIE G. CHALYAN, DAVID J. TAYLOR, AND DOUGLAS M. BOWDEN. Integrating behavior and cardiovascular responses: the code. AM. J. Physiol. 261 (Regulatory Integrative Comp. Physiol. 30): 11172-11181, 1991-The next revolution in biology is predicted to be in the integrative domain, and the need to involve physiologists in this kind of research has been recognized. This paper represents an approach to providing some of the tools required for dealing with integrative physiology at the behavioral level. Video tape recordings are made of the activities of a group of five baboons (Papio hamadryas) while simultaneous recordings of arterial blood pressure, heart rate, renal blood flow, and mesenteric or iliac blood flow are telemetered from two of the members of the group. The telemetered cardiovascular information is recorded on the two audio channels of the videotape. Subsequently the videotape is viewed, and a two-dimensional code is used to record the behavior of the two animals with the telemetry equipment. The first dimension of the code categorizes the behavior changes precisely regarding those aspects of behavior that are related to cardiovascular dynamics and does so with an accuracy of 16 ins. The second dimension codes relevant environmental changes. The paper describes the code and presents illustrations of how the code reflects the cardiovascular dynamics associated with the behavioral changes.

We videotaped the behavior of six social groups of baboons (2 adult males, 2 adult females, and an infant) while CV responses were telemetered from two of the animals (11). The CV data included blood pressure, renal blood flow, and blood flow from one other vascular bed, usually mesenteric or iliac. Heart rate was electronically derived from the pressure pulse. The behavior of the two animals equipped with telemetry was subsequently coded from the videotape using the stop and slow-motion capabilities of the video cassette recorder (VCR). A temporally continuous (i.e., not time-sampled) descriptive record of behavior was generated for subsequent merging with similarly continuous CV measures, thus preserving preceding behavioral and CV states for potential use as historical setting events for temporally localized analyses.

1nn Lie. Torso is resting in a horizontal position with minimal or no limb muscular support. This category includes leaning so far forward while sitting that the body is horizontal and nearly on the ground. This posture is often used while licking up food from the ground (code 111).

2nn Sit. Hindquarters are placed on a horizontal surface and support most of the body weight. Torso is usually vertical, although it can be leaning somewhat forward or to one side. Forelimbs may or may not be in contact with the surface.

3nn Stand on one or two legs. Animal assumes a stationary upright body position on a horizontal surface with weight supported by one or two hindlimbs. It often uses a vertical surface or another animal for support. One foot may move around during more active stands.

4nn Stand on three or four legs. Animal assumes a stationary horizontal torso position on a horizontal surface with weight supported by three or four limbs. Knees and/or elbows may be bent. A more active stand may include some deep knee bends or short steps with one or two feet.

5nn Walk on two legs. Locomotion across a horizontal surface with body in upright position, using only two hindlimbs. Contact with the surface is always maintained. At least one step with each hindlimb must be taken, and a vertical surface should not be used for support.

6nn Walk on three or four legs. Locomotion across a horizontal surface with torso in a horizontal position, using three or four limbs. At least three feet must take a step. Contact with surface is always maintained.

7nn Run. Rapid locomotion across a horizontal surface using three or four limbs to jog, gallop, trot, or move with faster gaits where all feet leave the surface simultaneously. This includes the rapid kinds of movements used during aggressive interactions (chasing, fighting, face-to-face noncontact jousting) and the short quick jumps used during a chase up and down vertical surfaces.

8nn Climb or jump. Locomotion up or down a vertical surface or a leap through the air between two surfaces. If this locomotion is used during a chase or fight it is treated as part of the running (7nn) and not coded separately.

9nn Hang. Stationary position on a vertical surface maintained by clinging with three or four limbs.

-nn Lunge. Rapid transference of weight to forelimbs while hindlimbs may or may not remain stationary. Often the whole body moves a few feet and then comes to an abrupt halt. It is also used to score the quick short lungelike scrambling moves used to gather food that lies within a few feet of the focal animal.

*nn Walk-stand sequence. Repetitive alternating walk-stand-walk-stand sequence in which the duration of each walk or stand is usually a few seconds but can last up to 10 or 15 s. This sequence is frequently seen when animals forage for food.

Primary Dimension: Behavior Domain
(Digits Two and Three)

This two-digit pair of the behavior domain identifies a goal-oriented behavior sequence, which frequently can be determined only in retrospect from video recordings. The first discernible body movement used in the sequence is considered to be the onset of the behavior. These two digits always follow a locomotion/posture digit (n), thus completing a three-digit code entry. Digit two identifies a general behavior, such as feeding, drinking, and grooming. Digit three is used to subdivide the general behavior. It may be an intensity or energy output modifier, or it may indicate a different quality of the behavior.

n1n Feeding. This category includes behaviors used for both gathering and eating food. This includes feeding behavior that is altered by concomitant orientation behavior but remains ongoing. In the latter case, first determine the intensity of feeding [i.e., calm (code 11), active (code 12), or excited (code 13)] and then determine whether the feeding behavior is altered by orientation. If yes, shift to the corresponding feeding intensity: 16, 17, or 18, respectively.

n11 Calm feeding. Animal slowly gathers food and places it in mouth or eats food stored in pouch. Animal shows few body posture changes, displays little head scanning, and focuses visual attention mostly on food. Animal shows no noticeable emotional display and an obvious lack of interest in other animal's or person's activities. (Note: if running or lunging is involved, then the feeding is not calm and must be coded at least a 12.)

n12 Active feeding. Animal gathers food quickly while in a sitting position or after a fast walk or slow run to approach the food. Eating is hurried, and animal displays frequent posture adjustments if sitting.

n13 Excited feeding. Animal gathers food in a frenzied manner (running, lunging, shuffling). Eating is frantic; often animal stuffs food into its cheek pouches to claim as much food as possible. Behavior frequently occurs when preferred foods are introduced in small quantities; as the food is thrown into the compound, the dominant animal lunges after and chases it to prevent others from getting it.

n15 Anxious feeding. Animal eats nervously, shifting attention repeatedly between food and other animals (usually of a more dominant status) to check for their approach or disapproval.

n16, n17, n18 Feeding and orienting. Same behavior as n11, n12, and n13 combined with intentional head orientation and visual fixation; body orientation is often directed toward a person or environmental stimuli. The orientation may have interruptions but still dominates the behavior. Casual glances at people during feeding do not warrant this code.

n2n Oral. Focal animal takes fluid into the mouth by sucking from a water delivery tube or by lapping water from a pan or pool or licks or bites an object or surface in the environment.

n2l Upright drinking. Animal drinks from a source located above the ground while standing on three or four legs with the torso horizontal.

n22 Head down drinking. Animal drinks from a source on the ground while standing on three or four legs with head down and the buttocks raised in the air.

n25 Licking. Animal licks object or surface but does not ingest food.

n26 Biting. Animal bites or chews on a nonedible object.

n3n Grooming. One or more animals groom the focal animal or the focal animal grooms another. Grooming is the meticulous parting of hairs with the fingers and/or picking at the skin and hair with the fingers or teeth. This includes brief periods of suspended grooming while the groomee changes position if the grooming resumes.

n31 Groomed. Grooming by another animal begins. Groomee shows little muscular activity except for occasional body posture adjustments.

n33 Groomed while orienting. While being groomed, animal shifts visual attention and becomes fixed on a person or an environmental stimulus.

n35 Present for grooming. Animal presents body for grooming but grooming does not follow within 3-4 s.

n37 Grooms another. Focal animal grooms another animal.

n4n Maternal. This category is reserved for future development of categories that describe mother-infant interactions.

n5n Sexual. Intense visual attention and/or physical contact (nongrooming) by an adult male directed toward the hindquarters of an adult female or adolescent. Typically the male approaches the other animal, reaches out with a hand, touches or grabs the hindquarters, stands upright, mounts, thrusts, and copulates. Various parts of this sequence may be absent. The beginning of a sexual sequence is scored at the onset of the initial approach that leads directly to an identifiable sexual behavior.

n51 Approach female. Used for males only. This includes any anticipatory, approach, and mounting activities by a male that lead to sexual behavior toward an adult female. It usually begins with an uninterrupted visual fixation on a female (usually the hindquarters) or with the onset of the first movement of the approach toward the female. This period includes the mount and ends when thrusting begins or the male turns away and discontinues interest in the female.

n52 Mount with thrusting. Used for both a male that is mounting an adult female and an adult female that is being mounted by a male where thrusting without intromission occurs. Scored at the onset of the first thrust. If the absence of intromission cannot be confirmed, then thrusting episodes where both hind feet of the male are not on the female's lower legs are included.

n53 Copulation. Used for both male and adult female participants where thrusting with intromission occurs. It is scored at the onset of the first thrust that leads to copulation. If the presence of intromission cannot be confirmed, then only the thrusting episodes where both hind feet of the male securely grasp the lower legs of female are included.

n54 Toward adolescent. Used for male directing any one of the sexual behaviors toward an adolescent of either sex.

n56 Approached by male. Used for adult females only. This is the period of anticipation during the approach of a male that eventually displays sexual behavior toward the female. It begins with initiation of the first identifiable acknowledgment of the male's approach. This period ends when the female's hindquarters are presented to the male or the male's attention is redirected elsewhere.

n57 Present. Used for adult females only. Female presents hindquarters to the male, and the male responds with sexual behavior. It includes the period up to thrusting if mounting occurs or until the presentation ends.

n6n Submission. Focal animal acknowledges the dominance of another animal, backs up and gives ground to animal, person, or object, or shows a startle reaction to some stimulus.

n6l Low level. Animal leans away, grimaces, gives way to another animal that does not show aggression (walks away slowly), shows mild startle response, and/or retreats slowly from an object (food) or presence of a person who is not displaying aggression.

n62 Middle level. Animal crouches, freezes, gives way to or rapidly retreats from another animal that shows aggression, shows severe startle response, and/or retreats rapidly from an object (food) or a person who is not displaying aggression.

n63 High level, retreat from person. Rapid retreat from a person displaying aggression.

n65 High level, retreat from animal. Rapid retreat from pursuit by an aggressive animal without returning aggression; this does not include defensive face-to-face posturing even if ground is being lost.

n66 Captured. Animal is caught, grabbed, scratched, or bitten by an aggressive animal without returning aggressive action.

n68 Present to opposite sex. Animal presents to another animal of opposite sex without sexual response.

n69 Present to same sex. Animal presents to and/or is mounted by another animal of the same sex.

n7n Aggression/dominance. Focal animal displays dominance behaviors toward another animal or person.

n72 Middle level. Brief aggressive leap or lunge toward n82 another animal or a nonthreatening person or intentional rapid approach toward another animal resulting in hurried displacement and possibly subsequent following. Animal grabs aggressively at food held by a person or another animal.

n73 High level, attack person or other stimulus outside compound.

775 High level, attack animal. Chases another animal. The locomotion used during chases is usually a complex combination of rapidly changing high energy movements (run, lunge, climb, jump, and others). Because run always dominates this combination, the first digit of a chase code by definition is 7 (run) to avoid routinely breaking the chase into multiple small units of changing locomotion.

776 Fight. Catches another animal and/or fights by biting, grabbing, and scratching or displays noncontact aggressive face-to-face jousting. This includes the period when a submissive animal being chased turns, holds its ground, and returns aggression. The locomotion is always coded as 7 (run) for reasons explained above (code 775).

n79 Hindquarters, same sex. Focal animal directs undivided attention toward the hindquarters of a same-sex animal, grabs, and/or mounts the animal.

n8n Idle/nonspecific. Behavioral state with no ongoing active interaction with another animal, food, or object, attention i's focused on self or unfocused (not directed toward anything that can be identified). This can include aimless visual scanning or very casual uninvolved staring at other animals or the environment. If a definite visual and/or body orientation toward an environmental stimulus or person is the only dominating behavior, then shift to categories 86, 87, and 88. Note that these three categories represent progressive levels of intensity of the behavior state, not orientation intensity

n8l Low level. Baseline activity state for each given posture/locomotion with no identifiable emotional behavioral component. During lie, sit, stand and hang, animal assumes stationary posture, is calm, displays only small casual head or limb movements. This may include occasional small movements across a surface or turning of the body. During walk and climb, animal displays casual unhurried locomotion without obvious purpose or direction and uses only the low-energy movements of the head and limbs required for locomotion. During jump and run, animal uses only the minimal effort required to accomplish these locomotions. Motions are usually preceded and/or followed by a low-intensity locomotion/posture. Jumps in this category are not airborne but are more steplike, i.e., contact is maintained with one surface during the entire change in location. Lunge has no baseline equivalent here.

n82 Middle level. This category is used when activity is above category 81 baseline but not extreme enough to warrant the 83 code. In effect, it contains behaviors with a wide activity continuum between the extreme states. Borderline cases that are difficult to assign are always coded as 82 to preserve the purity of the extreme categories.

n83 High level. Animal is extremely agitated during a posture or locomotion, with obvious addition of an unfocused and probably unidentifiable emotional component (irritation, fear, uncontrollable curiosity, and others). Posture/locomotion changes are frequent and include rapid abrupt movements.

n86, n87, n88 Idle/nonspecific and orienting. Same as n8l n82 and n83, but the idle behavior is combined with an intentional head orientation and visual fixation, and often body orientation is directed toward a person or an environmental stimulus. The orientation may have interruptions but must still dominate the behavior. Casual glances at people do not warrant this category.

n9n Other. These categories include behaviors for which there is no appropriate category or of which the frequency of occurrence is too low to warrant a single category. The n99 category is a catchall for low-frequency or unusual behaviors and provides a marker for possible later analysis. These data are not lost.

n91 Huddling.

n92 Play interaction with an adolescent in either an active or passive role.

n98 Significant or interesting "other" type behavior that has been referenced and described in a separate handwritten log for easy future reference.

n99 All other behaviors for which there is no appropriate category.

Secondary Dimension: Environmental Setting
Conditions Domain (Digits One and Two)

Categories of this domain are used to record the onset and in some cases the end of changing environmental conditions that are not uniquely reflected in the three digit locomotion/posture behavior codes. This dimension is recorded as a separate two-digit code entry beginning with a +. The + in the first digit position tells the computerized data acquisition system to record and complete the code entry after the second digit instead of the third.

+1 Entrance opened. The living compound entrance is opened to release an animal from the back room or trapping run into the living compound.

+2 Food brought into environment. The onset of the first behavioral change that indicates that the focal animal has become aware that food is coming. This category must always be terminated by +3 or +4.

+3 Food delivered. The person feeding makes the first piece of food available to the focal animal by placing it in the living compound, or the animal reaches outside the compound and pulls the food in. This marks the beginning of the actual feeding period (gathering and eating food). A single +3 condition is considered to be continuous through successive food deliveries as long as the animal continues to exhibit expectations of being given more food. This category must always be terminated by a +4.

+4 Food removed from environment. This code is used to mark the end of condition +3, indicating that the focal animal no longer expects more food to be delivered to the living compound. Ongoing orientation and interest toward a person or the environment stops.

+7 Threat. A person jumps into the animal's view unexpectedly, yells, displays a threatening face with direct eye contact, and then after 2-4 s retreats out of view.

+9 Exit opened. The living compound exit is opened to move the animal into the backroom or trapping run.

Examples of the Code at Work

The ability of the locomotion/posture domain of the primary dimension to distinguish a CV pattern associated with a simple posture shift is illustrated in Fig. 2, which presents the CV responses associated with a change from sitting (code 281) to walking (code 681, shaded areas). A more dynamic change in behavior involving the behavior domain is shown in Fig. 3, which illustrates a dramatic shift in CV response accompanying an act of aggression (code 772). The ability to link behavior to CV responses precisely in time reveals that the CV responses associated with the aggression actually occur several seconds before any somatic movement is discernible. The influence of an environmental variable (secondary dimension) on the long-term general level of CV responses is illustrated in Fig. 4; the anticipation of food is designated as +2, and the actual providing of food is indicated as +3. Figure 5 illustrates the ability to determine the average CV response accompanying a particular behavior regardless of the immediately preceding or following behavior. In this case, the randomization of all preceding behaviors results in the presit CV response approaching a straight line, whereas the CV response of the summed 106 instances of sitting (code 281) emerges as a measurable signal with a characteristic wave form.

FIG. 2. Cardiovascular (CV) responses during 3 repetitions of walking (behavioral code 681; shaded areas) followed by sitting (behavioral code 281). Note similar CV changes at each instance of walking.

FIG. 3. CV responses associated with an attack by dominant male against subordinate male. Recording is of dominant male. First indication of somatic movement on the part of the attacking animal is indicated by line at time 0. Note increases in heart rate and blood pressure and rapid decreases in renal and mesenteric blood flows that occur several seconds before any somatic movement can be perceived.

FIG. 4. Effect of environmental setting conditions (secondary dimension) on general baselines of CV variables. For first 3 min and 20 s animal is sitting quietly. At +2, animal shows first signs of increased attention and alertness to food preparation outside the compound. At +3, food is introduced into compound. Coding of dynamic behavior (primary dimension) is not presented.

FIG. 5. Computer averages (means ± SD) of 106 instances of sitting behavior (code 281). First body movement indicative of sitting (usually lowering of hips) is represented at time 0. Magnitude of SD reflects fact that these trials are unselected, i.e., any kind of behavior might have occurred during 60 s before initiation of sit at time 0.

The use of the time-code generator in conjunction with the stop-frame capability of the VCR allowed the timing of a behavioral action correlated with a CV change to be determined within a single heartbeat. This permitted us to determine processes such as the anticipation of an aggressive act from the CV data long before the behavioral reaction occurred (Fig. 3).

Because the physical size of the telemetry equipment requires a large animal to be able to carry it, the code has been developed and used primarily with male baboons. Recent technical advances in miniaturizing the equipment now allow it to be used on females. We believe that by reserving the n4n category in the primary dimension for maternal behavior, we have made the code generic enough to code female baboon behavior accurately. Many behaviors ordinarily considered to be gender specific are found in both sexes among the nonhuman primates.

In some instances the code calls for interpretation by the observer. This is especially true for the periods of food anticipation (+2, +4, and +5 categories of the secondary dimension). The inference that an animal is anticipating food is clearly hazardous and subject to error. However, anticipation of an event is a major factor that influences CV dynamics. For example, in Fig. 3 there are dramatic changes in the CV system of the dominant male before any overt behavioral signs of aggression appear. Also, in Fig. 4 the major change in the absolute baseline level of the CV variables for all behaviors at the onset of the +2 condition and maintenance of that level during the +3 condition is readily visible. One can use these categories of the secondary dimension only by viewing the videotape, determining the outcome of a situation, and then returning to the 150 most probable point of initiation or change in the situation. Even so, this remains a major source of potential error.

An additional caveat is that the code is based on a very simple environmental situation. The addition of elaborate devices to the environment may elicit different kinds of behavior that may not be included in the code. However, the code is flexible enough to accommodate additions. A related point is that the code has been developed for use with P. hamadryas and P. cynocephalus. The use of the code with other species will most probably require some modification.

The results demonstrate clear phenomenological properties of the behavioral process that are related to physiologically significant perturbations in the dynamics of CV processes. Establishing such external validation criteria, in the form of the CV events, is also an important step toward clarifying the methodological issues attending behavioral transduction. The significance of concerns about the "arbitrariness" of transducing analog behavioral events into digitized taxonomic description systems seems diminished by the present approach. Likewise, behavioral descriptions appear to offer a potent and promising tool for illuminating somatic-autonomic relations under relatively complex circumstances.

1. BOBBITT, R. A., G. D. JENSEN, AND B. N. GORDON. Behavioral elements (taxonomy) for observing mother- infant-peer interaction in Macaca nemestrina. Primates 5: 71-80, 1964.

2. COELHO, A. M., AND C. BRAMBLETT. Behaviour of the genus Papio: Ethogram, taxonomy, methods, and comparative measures. In: Perspectives in Primate Biology, edited by P. K. Seth and S. Seth. New Delhi, India: Today & Tomorrow's, 1989, vol. 3, p. 117-140.

3. HAUSFATER, G. Dominance and reproduction in baboons (Papio cynocephalus). A quantitative analysis. Contrib. Primatol. 7: 1-150, 1975.

4. JOHNSTON, J., AND H. S. PFNNYPACKER. Strategies and Tactics of Human Behavioral Research. Hillsdale, NJ: Erlbaum, 1980.

5. KAUFMAN, I., AND L. A. ROSENBLUM. A behavioral taxonomy for M. nemestrina and M. radiata based on longitudinal observations of family groups in the laboratory. Primates 7: 205-258, 1966.

6. LONG-RANGE PLANNING COMMITTEE OF THE AMERICAN PHYSIOLOGICAL SOCIETY. What's past is prologue. A "white paper" on 10 the future of physiology and the role of the American Physiological Society in it. Physiologist 33: 161-180,1990.

7. RAY. R. D.. AND D. J. DELPRATO. Behavioral systems analysis: methodological strategies and tactics. Behav. Sci. 34: 81-127, 1989.

8. RAN. R. D.. J. D. UPSON, AND B, J. HENDERSON. A systems approach to behavior. III. Organismic pace and complexity in time-space fields. Psychol. Rec. 27: 649-682, 1977.

9. SACKETT. G. P.. E. STEPHENSON, AND G. C. RUPPENTHAL. Digital data acquisition systems for observing behavior in laboratory and field settings. Behav. Res. Methods Instrum. 5: 344-348, 1973.

10. SMITH, 0. A., C. A. ASTLEY, M. A. CHESNEY, D. J. TAYLOR, AND F. A. SPELMAN. Personality, stress and cardiovascular disease: human and nonhuman primates. In: Neural Mechanisms and Cardiovascular Disease, edited by B. Lown, A. Malliani, and M. Prodocimi. Padua, Italy: Liviana, 1986, vol. 5. (Fidia Res. Set.)

11. SPELMAN, F. A., C. A. ASTLEY, 0. A. SMITH, Y. GOLANOV, J. J. CUPAL, E. FONZO, T. G. SUSOR, G. MCNIORROW, AND D. M. BOWDEN. A system to acquire and record data remotely from nonhuman primates. IEEE Trans. Biomed. Eng. In press.

This work was supported by National Heart, Lung, and Blood Institute (NHLBI) Grants HL-16910, HL-24416, and Division of Research Resources Grant RR-00166. The interactions between the Soviet and American scientists were supported by the US-USSR Exchange in the Cardiovascular Area, Program Area 7, Hypertension (NHLBI).