- Email: [email protected]
Time allocations of various activities under multiple schedules in pigeons
113-l 24 0 1992 Elsevier Science Publishers B.V. All rights reserved 0376-6357/92/$05.00
Behavioural Processes, 26 (1992)
113
BEPROC 00403
s of various activities
Kazuchika Manabe, Shigeru Kuwata, Naomi Kurashige, Kaiuho Chino and Takashi Ogawa Department of Psychology and Pedagogy, Faculty of Humanities and Social Sciences, Meisei University, Hino, Tokyo, Japan
(Accepted 28 October 1991)
Abstract Time allocations of three pigeons’ various activities (pecking, turning, preening, wing flapping, etc.) were measured under usual m&pie schedules in which only key-peck response was reinforced. Positive behavioral contrast was found for both the number and the time allocation of key-peck response. Some time ailocations of the other activities were inversely varied to that of key-pecks, as stated in the competition thedry of behavioral contrast. However, the amount of reallocation of interim activities was not enough to complement the amount of increases or decreases of key-pecks in the unchanged component. Most of the amount of increases or decreases of key-pecks in the unchanged component was complemented with the inactive time. This result suggests that behavioral competition is not the primary mechanism for key-peck behavioral contrast in multiple schedules in pigeons. It is important to consider inactive time, which is emerged in the usual experimental setting, to clearly attribute the behavioral contrast to reallocation of time for other activities. Behavioral Key words: Behavioral competition; Multiple schedule; Pigeon
contrast;
Interim
activity;
Key peck;
Animals in experimental chambers show various activities other than activities reinforced by experimenters. Staddon (1977, 1983) divided those activities into two categories:
Correspondence to: K. Manabe, Department of Psychology and Pedagogy, Faculty of Humanities and Social Sciences, Meisei University,
rhino, Tokyo 191, Japan.
114
‘terminal’ activities directly controlled by scheduied reinforcement (such as lever-pressing for rats and key-pecking for pigeons); and ‘interim’ activities other than terminal activities. Furthermore, the interim activities are divided into two categories: ‘schedule-induced’ activities facilitated by a schedule of reinforcement without reinforcement to itself (such as drinking for rats and attack for pigeons: Azrir., Hutchison and Hake, 1966; Falk, 1961); and 'facultative' activities not induced by the schedule (such as running for rats and pacing for pigeons). Two important factors relate to the activities cited above, namely, time allocation and frequency of an activity. in usual behavioral experiments, a duration of session is fixed, thus a change in amount of time allocated to an activity may affect the amount of time allocated to other activities. That is, if an increase in rate of reinforcement for the activity increases the time allocation of a terminal activity, the time allocations of the other activities (interim activities) may decrease. This is because some amount of time is consumed by the terminal activities. Similarly, if the time allocations of interim activities increase (such as schedule-induced behavior: Azrin, Hutchinson and l-lake, 1966; Falk, 19611, the time allocation of the terminal activities may decrease. Hinson and Staddon (1978) demonstrated that when the rate of reinforcement in one component (changed component) of a multiple schedule decreases, the terminal activity (lever press) increases in the other component (unchanged component), that is, positive behavioral contrast occurs. On the contrary, the interim activity (wheel-running) increases in the changed component and decreases in unchanged component. Furthermore, the magnitude of behavioral contrast is larger when a wheel-running response is available than when it is not (Hinson and Staddon, 1978). Studies have shown that the availability of interim activi:y influences the magnitude of contrast, and this has prompted several researchers to consider the behavioral contrast as the result of a competition of terminal activity and interim activities (competition theory: Ettinger and Staddon, 1982; Hinson and Staddon, 1978; Staddon, 1982). According to this competition theory, the terminal activity in a changed component decreases because of a decrement in reinforcement rate. Consequently, the times that had been allocated to the terminal activity in the changed component become available to the interim activities in this component. When some interim activities in the unchanged component shift to the changed component, the time allocated to the interim activities becomes available to the terminal activity. As a result, the terminal activity increases in the unchanged component. Here it is shown that the positive behavioral contrast is caused by the reallocation of interim activities from the unchanged component to the changed component. Previous studies objectively measured and analyzed the frequency of interim activities concerned with the behavioral contrast. However, time allocation seems to be a more concrete measure of reallocation of interim activities. Furthermore, only rats served as subjects (Dougan, McSweeney and Farmer, 1985; Dougan, McSweeney and FarmerDougan, 1986; Hinson and Staddon, 1978; Jacquet, 1972). One of the present questions is whether the reallocations of interim activities from unchanged component to changed component could also be found in pigeons. The next question is what kind of interim activities are reallocated. The final question is whether reallocations of interim activities could be found in the usual experimental setting in which only key-peck is reinforced. In the present experiment, the time allocations of sixteen activities were measured with an interval recording.
115
Subjects
Three adult male homing pigeons (Columba livia domestica) were maintained at 80% of their free-feeding weight. The birds had free access to grit and water in their home cages. Apparatus
The dimension of experimental chamber was 30 X 30 x 30 cm. A response key was mounted on one wall behind a 2-cm diameter hole at a height of 20 cm from the floor. A force of approximately 0.15 N activated a microswitch behind this key. The key could be transilluminated by red and green lights. Beneath the response key was a food hopper that contained grain. The reLlforcer was a 3-s period of access to this grain. Three walls other than the front panel and the ceiling were translucent acrylic panels. The activities of subjects were monitored and filmed with two cameras and a video recording system. One camera was mounted above the chamber and the other was situated in the back half, left from the chamber. The two images from the two cameras were filmed simultaneously in one plane of image. The image from above was filmed in the left half of a frame and the other was in the right. A timer, indicating the time elapsed from the beginning of the session, was filmed in the left top of the picture. The unit of time for the timer was 1 /IO0 s. A microcomputer, located in an adjacent roorn, controlled all experimental conditions. Procedure
After the completion of magazine training, the key-peck response was shaped by the successive approximation method. Afterwards, the subjects were exposed to a series of multiple schedules. This series consisted of two baseline conditions (multiple variable interval variable interval) and a contrast condition (multiple variable interval extinction). The contrast condition was inserted between the first baseline condition and the second baseline condition. The values of variable intervai (VI) were 90 s for subject 8404, 30 s for subject 8002, and 15 s for subject 8602. These values were used because the behavioral contrast was found at least within the range of the above values of VI. Therefore, if the competition theory is valid, the reallocation of interim activities from the unchanged component to the changed component should be observed whatever the value of VI is. Furthermore, the amount of interim activities may be different across the above three schedule values (Jacquet, 1972). This difference in the amount of interim activities may affect the amount of behavioral contrast, Thus, a different schedule value was assigned for each subject. The color of key light in one component (unchanged component), in which the schedule remained constant, was red for subjects 8002 and 8602, and green for subject 8404. The key color in the other component (changed component), in which the schedule was changed to extinction during contrast condition, was green for subject 8404, red for subjects 8002 and 8602. The duration of the components was 30 s and the key color was altered according to the Gellerman method (1933). When a reinforcer was presented in the component, the duration of the component was prolonged by the duration of reinforcement. A 2-s intertrial interval was inserted every 30 s. Reinforcers that became available but were not collected before a component changed were held over for collection when the component resumed. The intervals of VI schedules were according to the Fleshier and Hoffman method (1962). Sessions were conducted 6 days per week and were terminated when a fixed number of components had been presented. This number of presentation of two components was varied across schedules to keep the number of
116 TABLE 1 The categories of activities measured and the
definition.
Activity
Definition
Air-peck
Pecking movement directed at the key, but the beak did not come
into
contact with the key. Wall-peck
‘c and the beak comes into contact Pecking movement directed at the wall-,
Floor-peck
with the walls. Pecking movement directed at the floor, and the beak comes into contact
Flapping
A vigorous up and down movement of the bird’s wings.
Wing-stretching
A movement in which bird’s wing is stretched with a stretching foot.
Preening
Any movement in which the beak comes into contact with the feathers on
Scratching
Scratching head with the bird’s claws.
Half-turn
A turning response in which the bird’s direction is changed from 0” to 135”
Full-turn
A turning
with the floor.
the bird’s body.
through 225” away from facing the front panel. response in which the bird’s direction
is changed from 0” to
more lhan 225” away from facing the front panel. Locomotion
The bird walks about in no particular direction.
Head-shaking
A response in which the bird’s head is shaken right
Yawning
A
Erect feather
A response in which the feathers on the bird’s body are erected. The
and left very quickly. response in which the bird’s beak is opened without pecking movement. response should be continued at least for 2 s. ’
Head-insert
into
filled feeder Head-insert
into
blank feeder a Observers
temporarily
A response in which at least the beak or more of the bird’s head is inserted into the magazine feeder opening when the feeder is up. A response in which at ieast the beak or more of the bird’s head is inserted into the magazine feeder opening when the feeder is dowr;. marked the sign when the feathers were erected, and erased the sign
when the response did not continue for 2 s. If the response was continued for 2 s, the sign was not erased.
reinforcers presented in the session roughly constant. The number of presentations of each component was 30 times for VI 90 s (subject 84041, 20 times for VI 30 s (subject 80021, and 10 times for VI 15 s (subject 8602). Each condition was performed for 15 sessions. Measuring of time allocation After all conditions were accomplished, three observers measured the time allocation of various activities from the film. The categories of activities measured and the definitions are shown in Table 1. Before regular observations, the categories were chosen according to the following procedure. Three observers watched the three video tapes filmed in the last sessions of each of the three conditions for three pigeons’ activities in the present experiment. When a category was found at least once, this category was chosen for the regular observation. In the regular observation, video tapes were replayed at 10 times slower than the real speed. Fifteen categories of activities were scored in every lo-s-interval play. Observers marked a sign for each occurrence of the category of activities on a check sheet every 10-s scene in which the original l-s scene was recorded. The duration Qf the category of activities was defined as 1 S. If two categories were observed successively
117
during the origin,al 1 -s interval, observers marked two signs on the check sheet in the order of occurrence. The duration was defined as 0.5 s respectively. Furthermore, if three categories were observrld successively duri;?g the original l-s interval, observers marked three signs on the chec k sheet in the order of occurrence. Each duration was defined as 0.33 S. The rate of cxrespondence among observers (the number of l-s-intervals that three observers agreed with each other/the total number of 1-s-intervals) was set at above 0.85 in every observation. Further, the check sheets of three observers were checked by themselves every 2 mill. If the sheet was different in sign from the other two sheets, every scene that did not coryspond to each other was observed again. In the re-observation, the scenes were replayed until three observers agreed with each other.
Fig. 1 shows the frequency per 1 min and percentaeq of time allocation ((time/(the time subtracted total 111time and total reinforcement time from total session time)) x 100) for key-peck response in tile last session of each condition. Frequency and time allocation of the changed component were decreased under contrast condition, while those of the unchanged component were increased under contrast condition (namely, positive behavioral contrast occurred). Fig. 2 shows the proportion of time allocations of key-pecking and that of the other 15 activities measured under the three conditions. The open blocks show the sum of the amount of other activities other than the 15 categories and the amount of time in which subjects were inactive. Because subjects were less frequently engaged in activities other than the 15 activities measured, the open blocks can be considered as inactive time. The experimental conditions are shown from top to bottom for each block. As the competition theory predicts, the total proportion of the 15 activities decreased in the unchanged
8404(Vl9Os)
I
m
8802(Vl”Itis)
8002W3Os)
m
I
u-
vIw/
n
I
q
VlEXT/
II
q
VIVD
Fig. 1. Frequency (upper ha10 and time allocation (lower halO for key-peck. Filled circles indicate those in changed component,
open circles those in unchanged component.
118 8602
VllSs
VllSs
EXT
Vl15s
Vl15s
fi; keypeck I n9 El actlvltles ineasured
8002 CHANOED
0
UNCHANGED
I
the
other or
8404 CHANDED
I
UNCHANGED Vl9Os
Vl9Os
VI905
Fig. 2. Proportion of time allocations of key-peck and that of the other 15 activities. See text for details.
and increased in the changed component in contrast condition. The total amount of the 15 activities of the two components (changed and unchanged components) in the last baseline condition for each subject. On the other hand, the total amount of inactive time of the two components was inc:eased in contrast condition and was decreased in baseline condition, except for that in the first baseline condition for subject 8002. The total inactive time in the first baseline condition for subject 8002 was almost the same as that in the contrast condition. Fig. 3 shows the percentage of time allocated to ten categories of activities. Because the other five categories of activities were less frequently observed, the data are not shown. Y axes were different in value among the ten activities. As the competition theory predicts; in the contrast condition, the time allocation of locomotion in the changed component for subject 8404 was increased in the changed component and was decreased in the unchanged component. Similar results were found for the following activities: (a) headshaking for subject 8404 (VI 90 s); (b) half turn for subjects 8002 (VI 30 s) and 8602 (VI 15 s); (cl full turn for subject 8002; and Cd) preening for subject 8602. The preening for subject 8404 occurred almost throughout in the changed component in the contrast condition while it occurred infrequently in the unchanged component and the two baseline conditions. Similar results were found for the following activities: (a) flapping for subjects 8002 and 8602; and (b) wall-pecking for subject 8602. The time allocation of air-pecking in unchanged component for subject 8404 was increased in the contrast
component
vlvl/I-
VlEXT/
vlw
-CHANGED COMPONENT o--_-o UNC,,ANGED COMPONENT
Fig. 3. The percentage of the time allocated to ten categories of interim activities. See text for details.
TABLE 2 The increases or decreases of ratio to the session time in the unchanged component across the time allocations in the contrast condition and those in the baseline conditions for key-peck, for activities that showed reairocation tendency, and for inactive time. The column ‘Contrast condition’ shows the increases or decreases of the ratios when the condition was altered from first baseline condition to contrast condition. The column ‘2nd-baseline condition’ shows those when the condition was altered from contrast condition to second base!ine condition. The ‘Sums’ indicate the total ratio of activities shown and inactive time. A ’ +’ indicates an increase of the ratio in unchanged component in the condition and a ’ .- ’ a decrease, respectively. -Subject
Activity
Contrast condition
Znd-baseline condition
Key-peck Preening Half-turn Inactive Sum
+ 0.463 - 0.011 - 0.031 - 0.239 +0.182
-0.198 + 0.018 + 0.028 +0.146 - 0.006
Key-peck Half-turn Full-turn Inactive Sum
+ 0.336 - 0.042 - 0.035 -0.217 + 0.042
-0.109 +0.016 + 0.037 + 0.070 +0.014
Key-peck Locomotion Head-shaking Inactive Sum
+ 0.611 - 0.070 - o.cin1 -0.121 -I-o.419
- 0.251 + 0.054 + 0.001 +0.160 - 0.036
8602
8002
8404
120
condition. The results for the other activities were not consistent with the prediction of competition theory. The total time allocated to half-turn for subject 8404 in both components decreased in all three conditions. Similar results were found for the following activities: (a) full-turn for subjects 8404 and 8602; (b) air-pecking for subjects 8002 and 8602; (c) head-shaking for subjects 8002 and 8602; and (d) preening for subject 8002. Table 2 shows the increases or decreases in the ratio to the session time in the unchanged component between the time allocation in the contrast condition and those in the baseline conditions for key-peck, for activities that showed reallocation tendency, and for inactive time. The column ‘Contrast condition’ shows the increases or decreases in the ratios when the condition was altered from first baseline condition to contrast condition. The column ‘2nd-baseline condition’ shows those ratios when the condition was altered from contrast condition to second baseline condition. The ‘Sums’ indicate the total of increases or decreases in the ratios for activities shown and that for inactive time. The ‘Sums’ in the ‘2nd-baseline condition’ are less than 0.05 for all subjects. This is also found in the contrast condition for subject 8002. That is, increases or decreases of key-peck in the unchanged component can be attributed almost to these activities and inactive time except for those of subjects 8602 and 8404 in the contrast condition. For subject 8602, the ratio! to session time for full-turn decreased by 0.157 in the contrast condition. For subject 8404, total ratio to session time for full- and half-turns decreased by 0.426 in the contrast condition. It is indicated that the above two exceptions were based on the large decrement of turnings. The amounts of increases or decreases of activities that showed reallocation tendency were less than a quarter of those of key-peck, except for those in the 2nd-baseline condition for subject 8002. On the other hand, the amounts of increases or decreases of inactive times were larger than half of those for key-peck, except for those in the contrast condition for subject 8404. The magnitude of contrast for key-peck was larger in the following order; (1) subject 8404 IV1 90 s); (2) subject 8602 (VI 15 s); and (3) subject 8002 WI 30 9. The amount of contrast was not related to the rate of reinforcement in the present experiment.
Disclrssio In the present experiment, the real!ocations of some of interim activities from the unchanged component to the changed component were found in pigeons in the usual experimental setting in which only key-peck is reinforced. The amount of increases or decreases for key-peck in the unchanged component did not reflect the amount of reallocation of the interim activities. The key-peck showed a large amount of increases or decreases while those of interim activities were small. Most of the increases or decreases of key-peck in the unchanged component were complemented with inactive time. The competition theory (Ettinger and Staddon, 1982; Hinson and Staddon, 1978; Staddon, 19C2) attributes the increases or decreases of key-peck in the unchanged component primarily to the reallocation of interim activities across the unchanged component anIl 1:~ changed component in the contrast condition. Therefore, the competition theory cannot fully account for the present results, basically the influence of inactive time on reallocation of activities. Most of the session time is probably fulfilled with the activities in both experimental settings In which terminal activity and other activity are explicitly reinforced (such as in Hinson and Staddon (1978)) and in which one activity is reinforced with rather a high rate
121
of reinforcement. For example, Dougan, McSweeney and Farmer (1985) reported that rats spent almost all session time either bar-pressing or licking in the condition in which bar-press was reinforced with food and a water bottle was mounted in the wall of the chamber. In these experimental settings, behavioral competition could be the primary mechanism for contrast, because any change in time allocation for one activity reliably varies from those for the other activities. However, a considerable amount of inactive time emerges in the usual setting in which only one activity is reinforced with the moderate rate of reinforcement (such as in the present experiment). In the usual setting, a change in time allocation for one activity does not reliably vary from those for the other activities because of the change in inactive time. If behavioral contrast is caused by any reallocation of time other than that for the terminal activity, the inactive time should be considered more carefully in the usual experimental setting. The present results offer the importance of reallocation of inactive time. Evidence of the local contrast across contrast and baseline conditions might have occurred in this study. The total amount of the interim activities and the inactive time were iower in the second baseline condition than in the first baseline condition for all subjects. On the contrary, the total time allocation for key-peck was higher in the second baseline condition than in the first baseline condition. The larger amount of time allocation for key-peck might be caused by the alternation from lower total rate of reinforcement condition (contrast condition) to higher rate condition (second baseline condition). However, typical increase in the rate of key-peck response (Nevin and Shettleworth, 1966) was not found for all subjects just after the condition was altered from a relatively low rate of reinforcement to a high rate of reinforcement. It is unlikei;! that the larger total time allocation for key-peck in the second baseline condition was evidence of the local contrast of a longer time range. Another possibility is that the decrease in the total amount of interim activities and the inactive time might be caused by a kind of differentiation in performance across experimental chamber and home cage develope as the experiment proceeded. That is, the experimental chamber is a ‘working area’ in which food is obtained and the home cage is a ‘rest area’. In a personal observation, pigeons are almost inactive in the home cage except for pre- and post-experimental sessions and mealtimes. Some interim activities and the inactive time that did not relate to food might be reallocated from experimental chamber to home cage as the experiment proceeded. To test this hypothesis, the activities in the home cage should be measured in future. The present data on individual interim activities were varied among individual subjects. This variability might be caused by some factors, such as an accidental reinforcement in which the activity was accidentally followed by a key-peck reinforced or a temporal discrimination in which response chaining was shaped (cf. Ferster and Skinner, 1957; Dews, ‘1965a, 1965b). For subject 8404 in the changed component in contrast condition, half- and full-turns occurred in the initial and locomotion occurred in the latter. This response pattern may be a response chaining complemented extinction phase. Another factor is a difference in the rate of reinforcement (VI value). Staddon (1977) showed that rates or probabilities of various interim activities were varied across various rates of reinforcement. On the other hand, the present data on the individual difference in interim activities across sr,Ibjects may be based on the difference in the rate of reinforcement. However, this issue goes beyond the scope of the present experimental settinr7s. In summary, some interim activities were reallocated from the unchanged component to the changed component in contrast condition in the pigeon. However, the amount of
122 reallocation
was not enough
to complement
the amount
key-peck
in the unchanged component. Half of the amount
key-peck
in the unchanged
behavioral
component
was complemented
contrast is caused by any reallocation
activity, the primary mechanism activities but the reallocation
for behavioral
of Education,
was supported
of increases or decreases of with
the inactive
If
contrasts is not the reallocation
of interim
setting in which
is reinforced.
by a Grant-in-Aid
Science and Culture,
time.
of time other than that of the terminal
of inactive time in the usual experimental
only one activity (such as key-peck)
This research
of increases or decreases of
for Scientific
Research from
the Ministry
lapan.
eferemces Azrin, N.H., Hutchinson, R.R., and Hake, D.F. 1966. Extinction-induced aggression. Journal of the Experimental Analysis of Behavior, 9: 191-204. Dougan, J.D., McSweeney, F.K., and Farmer, V.A. 1985. Some parameters of behavioral contrast and allocation of interim behavior in rats. Journal of the Experimental Analysis of Behavior, 44: 325-335. Dougan, I.D., McSweeney, F.K., and Farmer-Dougan, V..4. 1986. Behavioral contrast in competitive and noncompetitive environments. Journal of the Experimental Analysis of Behavior, 46: 185-l 97. Dews, P.B. 1965a. The effect of multiple S-’ periods on responding on a fixed-interval schedule: II. In a primate. Journal of the Experimental Analysis of Behavior, 8: 53-54. Dews, P.B. 1965b. The effect of multiple S’ periods on responding on a fixed-interval schedule: III. Effect of changes in pattern of interruptions, parameters and stimuli. Journal of the Experimental Analysis of Behavior, 8: 427-435. Ettinger, R.H., and Staddon, J.E.R. 1982. Behavioural competition, component duration and multiplc-schedule contrast. Behavioural Analysis Letters, 2: 31-38. Falk, J.L. 1961. Production of polydipsia in normal rats by an intermittent food schedule. Science, 133: 195-196. Ferster, C.B and Skinner, B.F. 1957. Schedules of reinforcement. NJ: Prentice-Hall. Fleshler, M., and Hoffman, H.S. 1962. A progression for generating variable-interval schedules. Journal of the Experimental Analysis of Behavior, 5: 529-530. Gellarman, L.W. 1933. Chance orclers of alternating stimuli in visual discrimination experiments. Journal of Genetic Psychology, 42: 207-208. t-‘zarst, E. and Jenkins, H.M. 1974. Sign-tracking: the stimulus-reinforcer relation and direct action. Austin, Texas: The Psychonomic Society. Hinson, J.M., and Staddon, J.E.P. 1978. Behavioral competition: A mechanism for schedule interactions. Science, 202: 432-434. Jacquei, Y.F. 1972. Sc hed u Ie-induced licking during multiple schedules. Journal of the Experimental Analysis of Behavior, 17: 413-423. Nevin, J.A., and Shettleworth, S.J. 1966. An analysis of contrast effects in multiple schedules. Journal of the Experimental Analysis of Behavior, 9: 305-315. Staddon, J.E.R. 1977. Schedule-induced behavior. In W.K. Honig and J.E.R. Staddon (Eds.), Handbook of operant behavior (pp. 125-l 52). NJ: Prentice-Hall.
123 Staddon, J.E.R. 1982. Behavioral competition, contrast and matching. In M.L. Commons, R.]. Herrnstein, and H. Rachlin t.Eds.1, Quantitative analysis of behavior: Vol. 2. Matching and maximizing accounts (pp. 243-261). Cambridge; MA: Ballinger. Staddon, J.E.R. 1983. Adaptive behavior and learning. Cambridge: Cambridge University Press.
Behavioural Processes, 26 (1992)
113
BEPROC 00403
s of various activities
Kazuchika Manabe, Shigeru Kuwata, Naomi Kurashige, Kaiuho Chino and Takashi Ogawa Department of Psychology and Pedagogy, Faculty of Humanities and Social Sciences, Meisei University, Hino, Tokyo, Japan
(Accepted 28 October 1991)
Abstract Time allocations of three pigeons’ various activities (pecking, turning, preening, wing flapping, etc.) were measured under usual m&pie schedules in which only key-peck response was reinforced. Positive behavioral contrast was found for both the number and the time allocation of key-peck response. Some time ailocations of the other activities were inversely varied to that of key-pecks, as stated in the competition thedry of behavioral contrast. However, the amount of reallocation of interim activities was not enough to complement the amount of increases or decreases of key-pecks in the unchanged component. Most of the amount of increases or decreases of key-pecks in the unchanged component was complemented with the inactive time. This result suggests that behavioral competition is not the primary mechanism for key-peck behavioral contrast in multiple schedules in pigeons. It is important to consider inactive time, which is emerged in the usual experimental setting, to clearly attribute the behavioral contrast to reallocation of time for other activities. Behavioral Key words: Behavioral competition; Multiple schedule; Pigeon
contrast;
Interim
activity;
Key peck;
Animals in experimental chambers show various activities other than activities reinforced by experimenters. Staddon (1977, 1983) divided those activities into two categories:
Correspondence to: K. Manabe, Department of Psychology and Pedagogy, Faculty of Humanities and Social Sciences, Meisei University,
rhino, Tokyo 191, Japan.
114
‘terminal’ activities directly controlled by scheduied reinforcement (such as lever-pressing for rats and key-pecking for pigeons); and ‘interim’ activities other than terminal activities. Furthermore, the interim activities are divided into two categories: ‘schedule-induced’ activities facilitated by a schedule of reinforcement without reinforcement to itself (such as drinking for rats and attack for pigeons: Azrir., Hutchison and Hake, 1966; Falk, 1961); and 'facultative' activities not induced by the schedule (such as running for rats and pacing for pigeons). Two important factors relate to the activities cited above, namely, time allocation and frequency of an activity. in usual behavioral experiments, a duration of session is fixed, thus a change in amount of time allocated to an activity may affect the amount of time allocated to other activities. That is, if an increase in rate of reinforcement for the activity increases the time allocation of a terminal activity, the time allocations of the other activities (interim activities) may decrease. This is because some amount of time is consumed by the terminal activities. Similarly, if the time allocations of interim activities increase (such as schedule-induced behavior: Azrin, Hutchinson and l-lake, 1966; Falk, 19611, the time allocation of the terminal activities may decrease. Hinson and Staddon (1978) demonstrated that when the rate of reinforcement in one component (changed component) of a multiple schedule decreases, the terminal activity (lever press) increases in the other component (unchanged component), that is, positive behavioral contrast occurs. On the contrary, the interim activity (wheel-running) increases in the changed component and decreases in unchanged component. Furthermore, the magnitude of behavioral contrast is larger when a wheel-running response is available than when it is not (Hinson and Staddon, 1978). Studies have shown that the availability of interim activi:y influences the magnitude of contrast, and this has prompted several researchers to consider the behavioral contrast as the result of a competition of terminal activity and interim activities (competition theory: Ettinger and Staddon, 1982; Hinson and Staddon, 1978; Staddon, 1982). According to this competition theory, the terminal activity in a changed component decreases because of a decrement in reinforcement rate. Consequently, the times that had been allocated to the terminal activity in the changed component become available to the interim activities in this component. When some interim activities in the unchanged component shift to the changed component, the time allocated to the interim activities becomes available to the terminal activity. As a result, the terminal activity increases in the unchanged component. Here it is shown that the positive behavioral contrast is caused by the reallocation of interim activities from the unchanged component to the changed component. Previous studies objectively measured and analyzed the frequency of interim activities concerned with the behavioral contrast. However, time allocation seems to be a more concrete measure of reallocation of interim activities. Furthermore, only rats served as subjects (Dougan, McSweeney and Farmer, 1985; Dougan, McSweeney and FarmerDougan, 1986; Hinson and Staddon, 1978; Jacquet, 1972). One of the present questions is whether the reallocations of interim activities from unchanged component to changed component could also be found in pigeons. The next question is what kind of interim activities are reallocated. The final question is whether reallocations of interim activities could be found in the usual experimental setting in which only key-peck is reinforced. In the present experiment, the time allocations of sixteen activities were measured with an interval recording.
115
Subjects
Three adult male homing pigeons (Columba livia domestica) were maintained at 80% of their free-feeding weight. The birds had free access to grit and water in their home cages. Apparatus
The dimension of experimental chamber was 30 X 30 x 30 cm. A response key was mounted on one wall behind a 2-cm diameter hole at a height of 20 cm from the floor. A force of approximately 0.15 N activated a microswitch behind this key. The key could be transilluminated by red and green lights. Beneath the response key was a food hopper that contained grain. The reLlforcer was a 3-s period of access to this grain. Three walls other than the front panel and the ceiling were translucent acrylic panels. The activities of subjects were monitored and filmed with two cameras and a video recording system. One camera was mounted above the chamber and the other was situated in the back half, left from the chamber. The two images from the two cameras were filmed simultaneously in one plane of image. The image from above was filmed in the left half of a frame and the other was in the right. A timer, indicating the time elapsed from the beginning of the session, was filmed in the left top of the picture. The unit of time for the timer was 1 /IO0 s. A microcomputer, located in an adjacent roorn, controlled all experimental conditions. Procedure
After the completion of magazine training, the key-peck response was shaped by the successive approximation method. Afterwards, the subjects were exposed to a series of multiple schedules. This series consisted of two baseline conditions (multiple variable interval variable interval) and a contrast condition (multiple variable interval extinction). The contrast condition was inserted between the first baseline condition and the second baseline condition. The values of variable intervai (VI) were 90 s for subject 8404, 30 s for subject 8002, and 15 s for subject 8602. These values were used because the behavioral contrast was found at least within the range of the above values of VI. Therefore, if the competition theory is valid, the reallocation of interim activities from the unchanged component to the changed component should be observed whatever the value of VI is. Furthermore, the amount of interim activities may be different across the above three schedule values (Jacquet, 1972). This difference in the amount of interim activities may affect the amount of behavioral contrast, Thus, a different schedule value was assigned for each subject. The color of key light in one component (unchanged component), in which the schedule remained constant, was red for subjects 8002 and 8602, and green for subject 8404. The key color in the other component (changed component), in which the schedule was changed to extinction during contrast condition, was green for subject 8404, red for subjects 8002 and 8602. The duration of the components was 30 s and the key color was altered according to the Gellerman method (1933). When a reinforcer was presented in the component, the duration of the component was prolonged by the duration of reinforcement. A 2-s intertrial interval was inserted every 30 s. Reinforcers that became available but were not collected before a component changed were held over for collection when the component resumed. The intervals of VI schedules were according to the Fleshier and Hoffman method (1962). Sessions were conducted 6 days per week and were terminated when a fixed number of components had been presented. This number of presentation of two components was varied across schedules to keep the number of
116 TABLE 1 The categories of activities measured and the
definition.
Activity
Definition
Air-peck
Pecking movement directed at the key, but the beak did not come
into
contact with the key. Wall-peck
‘c and the beak comes into contact Pecking movement directed at the wall-,
Floor-peck
with the walls. Pecking movement directed at the floor, and the beak comes into contact
Flapping
A vigorous up and down movement of the bird’s wings.
Wing-stretching
A movement in which bird’s wing is stretched with a stretching foot.
Preening
Any movement in which the beak comes into contact with the feathers on
Scratching
Scratching head with the bird’s claws.
Half-turn
A turning response in which the bird’s direction is changed from 0” to 135”
Full-turn
A turning
with the floor.
the bird’s body.
through 225” away from facing the front panel. response in which the bird’s direction
is changed from 0” to
more lhan 225” away from facing the front panel. Locomotion
The bird walks about in no particular direction.
Head-shaking
A response in which the bird’s head is shaken right
Yawning
A
Erect feather
A response in which the feathers on the bird’s body are erected. The
and left very quickly. response in which the bird’s beak is opened without pecking movement. response should be continued at least for 2 s. ’
Head-insert
into
filled feeder Head-insert
into
blank feeder a Observers
temporarily
A response in which at least the beak or more of the bird’s head is inserted into the magazine feeder opening when the feeder is up. A response in which at ieast the beak or more of the bird’s head is inserted into the magazine feeder opening when the feeder is dowr;. marked the sign when the feathers were erected, and erased the sign
when the response did not continue for 2 s. If the response was continued for 2 s, the sign was not erased.
reinforcers presented in the session roughly constant. The number of presentations of each component was 30 times for VI 90 s (subject 84041, 20 times for VI 30 s (subject 80021, and 10 times for VI 15 s (subject 8602). Each condition was performed for 15 sessions. Measuring of time allocation After all conditions were accomplished, three observers measured the time allocation of various activities from the film. The categories of activities measured and the definitions are shown in Table 1. Before regular observations, the categories were chosen according to the following procedure. Three observers watched the three video tapes filmed in the last sessions of each of the three conditions for three pigeons’ activities in the present experiment. When a category was found at least once, this category was chosen for the regular observation. In the regular observation, video tapes were replayed at 10 times slower than the real speed. Fifteen categories of activities were scored in every lo-s-interval play. Observers marked a sign for each occurrence of the category of activities on a check sheet every 10-s scene in which the original l-s scene was recorded. The duration Qf the category of activities was defined as 1 S. If two categories were observed successively
117
during the origin,al 1 -s interval, observers marked two signs on the check sheet in the order of occurrence. The duration was defined as 0.5 s respectively. Furthermore, if three categories were observrld successively duri;?g the original l-s interval, observers marked three signs on the chec k sheet in the order of occurrence. Each duration was defined as 0.33 S. The rate of cxrespondence among observers (the number of l-s-intervals that three observers agreed with each other/the total number of 1-s-intervals) was set at above 0.85 in every observation. Further, the check sheets of three observers were checked by themselves every 2 mill. If the sheet was different in sign from the other two sheets, every scene that did not coryspond to each other was observed again. In the re-observation, the scenes were replayed until three observers agreed with each other.
Fig. 1 shows the frequency per 1 min and percentaeq of time allocation ((time/(the time subtracted total 111time and total reinforcement time from total session time)) x 100) for key-peck response in tile last session of each condition. Frequency and time allocation of the changed component were decreased under contrast condition, while those of the unchanged component were increased under contrast condition (namely, positive behavioral contrast occurred). Fig. 2 shows the proportion of time allocations of key-pecking and that of the other 15 activities measured under the three conditions. The open blocks show the sum of the amount of other activities other than the 15 categories and the amount of time in which subjects were inactive. Because subjects were less frequently engaged in activities other than the 15 activities measured, the open blocks can be considered as inactive time. The experimental conditions are shown from top to bottom for each block. As the competition theory predicts, the total proportion of the 15 activities decreased in the unchanged
8404(Vl9Os)
I
m
8802(Vl”Itis)
8002W3Os)
m
I
u-
vIw/
n
I
q
VlEXT/
II
q
VIVD
Fig. 1. Frequency (upper ha10 and time allocation (lower halO for key-peck. Filled circles indicate those in changed component,
open circles those in unchanged component.
118 8602
VllSs
VllSs
EXT
Vl15s
Vl15s
fi; keypeck I n9 El actlvltles ineasured
8002 CHANOED
0
UNCHANGED
I
the
other or
8404 CHANDED
I
UNCHANGED Vl9Os
Vl9Os
VI905
Fig. 2. Proportion of time allocations of key-peck and that of the other 15 activities. See text for details.
and increased in the changed component in contrast condition. The total amount of the 15 activities of the two components (changed and unchanged components) in the last baseline condition for each subject. On the other hand, the total amount of inactive time of the two components was inc:eased in contrast condition and was decreased in baseline condition, except for that in the first baseline condition for subject 8002. The total inactive time in the first baseline condition for subject 8002 was almost the same as that in the contrast condition. Fig. 3 shows the percentage of time allocated to ten categories of activities. Because the other five categories of activities were less frequently observed, the data are not shown. Y axes were different in value among the ten activities. As the competition theory predicts; in the contrast condition, the time allocation of locomotion in the changed component for subject 8404 was increased in the changed component and was decreased in the unchanged component. Similar results were found for the following activities: (a) headshaking for subject 8404 (VI 90 s); (b) half turn for subjects 8002 (VI 30 s) and 8602 (VI 15 s); (cl full turn for subject 8002; and Cd) preening for subject 8602. The preening for subject 8404 occurred almost throughout in the changed component in the contrast condition while it occurred infrequently in the unchanged component and the two baseline conditions. Similar results were found for the following activities: (a) flapping for subjects 8002 and 8602; and (b) wall-pecking for subject 8602. The time allocation of air-pecking in unchanged component for subject 8404 was increased in the contrast
component
vlvl/I-
VlEXT/
vlw
-CHANGED COMPONENT o--_-o UNC,,ANGED COMPONENT
Fig. 3. The percentage of the time allocated to ten categories of interim activities. See text for details.
TABLE 2 The increases or decreases of ratio to the session time in the unchanged component across the time allocations in the contrast condition and those in the baseline conditions for key-peck, for activities that showed reairocation tendency, and for inactive time. The column ‘Contrast condition’ shows the increases or decreases of the ratios when the condition was altered from first baseline condition to contrast condition. The column ‘2nd-baseline condition’ shows those when the condition was altered from contrast condition to second base!ine condition. The ‘Sums’ indicate the total ratio of activities shown and inactive time. A ’ +’ indicates an increase of the ratio in unchanged component in the condition and a ’ .- ’ a decrease, respectively. -Subject
Activity
Contrast condition
Znd-baseline condition
Key-peck Preening Half-turn Inactive Sum
+ 0.463 - 0.011 - 0.031 - 0.239 +0.182
-0.198 + 0.018 + 0.028 +0.146 - 0.006
Key-peck Half-turn Full-turn Inactive Sum
+ 0.336 - 0.042 - 0.035 -0.217 + 0.042
-0.109 +0.016 + 0.037 + 0.070 +0.014
Key-peck Locomotion Head-shaking Inactive Sum
+ 0.611 - 0.070 - o.cin1 -0.121 -I-o.419
- 0.251 + 0.054 + 0.001 +0.160 - 0.036
8602
8002
8404
120
condition. The results for the other activities were not consistent with the prediction of competition theory. The total time allocated to half-turn for subject 8404 in both components decreased in all three conditions. Similar results were found for the following activities: (a) full-turn for subjects 8404 and 8602; (b) air-pecking for subjects 8002 and 8602; (c) head-shaking for subjects 8002 and 8602; and (d) preening for subject 8002. Table 2 shows the increases or decreases in the ratio to the session time in the unchanged component between the time allocation in the contrast condition and those in the baseline conditions for key-peck, for activities that showed reallocation tendency, and for inactive time. The column ‘Contrast condition’ shows the increases or decreases in the ratios when the condition was altered from first baseline condition to contrast condition. The column ‘2nd-baseline condition’ shows those ratios when the condition was altered from contrast condition to second baseline condition. The ‘Sums’ indicate the total of increases or decreases in the ratios for activities shown and that for inactive time. The ‘Sums’ in the ‘2nd-baseline condition’ are less than 0.05 for all subjects. This is also found in the contrast condition for subject 8002. That is, increases or decreases of key-peck in the unchanged component can be attributed almost to these activities and inactive time except for those of subjects 8602 and 8404 in the contrast condition. For subject 8602, the ratio! to session time for full-turn decreased by 0.157 in the contrast condition. For subject 8404, total ratio to session time for full- and half-turns decreased by 0.426 in the contrast condition. It is indicated that the above two exceptions were based on the large decrement of turnings. The amounts of increases or decreases of activities that showed reallocation tendency were less than a quarter of those of key-peck, except for those in the 2nd-baseline condition for subject 8002. On the other hand, the amounts of increases or decreases of inactive times were larger than half of those for key-peck, except for those in the contrast condition for subject 8404. The magnitude of contrast for key-peck was larger in the following order; (1) subject 8404 IV1 90 s); (2) subject 8602 (VI 15 s); and (3) subject 8002 WI 30 9. The amount of contrast was not related to the rate of reinforcement in the present experiment.
Disclrssio In the present experiment, the real!ocations of some of interim activities from the unchanged component to the changed component were found in pigeons in the usual experimental setting in which only key-peck is reinforced. The amount of increases or decreases for key-peck in the unchanged component did not reflect the amount of reallocation of the interim activities. The key-peck showed a large amount of increases or decreases while those of interim activities were small. Most of the increases or decreases of key-peck in the unchanged component were complemented with inactive time. The competition theory (Ettinger and Staddon, 1982; Hinson and Staddon, 1978; Staddon, 19C2) attributes the increases or decreases of key-peck in the unchanged component primarily to the reallocation of interim activities across the unchanged component anIl 1:~ changed component in the contrast condition. Therefore, the competition theory cannot fully account for the present results, basically the influence of inactive time on reallocation of activities. Most of the session time is probably fulfilled with the activities in both experimental settings In which terminal activity and other activity are explicitly reinforced (such as in Hinson and Staddon (1978)) and in which one activity is reinforced with rather a high rate
121
of reinforcement. For example, Dougan, McSweeney and Farmer (1985) reported that rats spent almost all session time either bar-pressing or licking in the condition in which bar-press was reinforced with food and a water bottle was mounted in the wall of the chamber. In these experimental settings, behavioral competition could be the primary mechanism for contrast, because any change in time allocation for one activity reliably varies from those for the other activities. However, a considerable amount of inactive time emerges in the usual setting in which only one activity is reinforced with the moderate rate of reinforcement (such as in the present experiment). In the usual setting, a change in time allocation for one activity does not reliably vary from those for the other activities because of the change in inactive time. If behavioral contrast is caused by any reallocation of time other than that for the terminal activity, the inactive time should be considered more carefully in the usual experimental setting. The present results offer the importance of reallocation of inactive time. Evidence of the local contrast across contrast and baseline conditions might have occurred in this study. The total amount of the interim activities and the inactive time were iower in the second baseline condition than in the first baseline condition for all subjects. On the contrary, the total time allocation for key-peck was higher in the second baseline condition than in the first baseline condition. The larger amount of time allocation for key-peck might be caused by the alternation from lower total rate of reinforcement condition (contrast condition) to higher rate condition (second baseline condition). However, typical increase in the rate of key-peck response (Nevin and Shettleworth, 1966) was not found for all subjects just after the condition was altered from a relatively low rate of reinforcement to a high rate of reinforcement. It is unlikei;! that the larger total time allocation for key-peck in the second baseline condition was evidence of the local contrast of a longer time range. Another possibility is that the decrease in the total amount of interim activities and the inactive time might be caused by a kind of differentiation in performance across experimental chamber and home cage develope as the experiment proceeded. That is, the experimental chamber is a ‘working area’ in which food is obtained and the home cage is a ‘rest area’. In a personal observation, pigeons are almost inactive in the home cage except for pre- and post-experimental sessions and mealtimes. Some interim activities and the inactive time that did not relate to food might be reallocated from experimental chamber to home cage as the experiment proceeded. To test this hypothesis, the activities in the home cage should be measured in future. The present data on individual interim activities were varied among individual subjects. This variability might be caused by some factors, such as an accidental reinforcement in which the activity was accidentally followed by a key-peck reinforced or a temporal discrimination in which response chaining was shaped (cf. Ferster and Skinner, 1957; Dews, ‘1965a, 1965b). For subject 8404 in the changed component in contrast condition, half- and full-turns occurred in the initial and locomotion occurred in the latter. This response pattern may be a response chaining complemented extinction phase. Another factor is a difference in the rate of reinforcement (VI value). Staddon (1977) showed that rates or probabilities of various interim activities were varied across various rates of reinforcement. On the other hand, the present data on the individual difference in interim activities across sr,Ibjects may be based on the difference in the rate of reinforcement. However, this issue goes beyond the scope of the present experimental settinr7s. In summary, some interim activities were reallocated from the unchanged component to the changed component in contrast condition in the pigeon. However, the amount of
122 reallocation
was not enough
to complement
the amount
key-peck
in the unchanged component. Half of the amount
key-peck
in the unchanged
behavioral
component
was complemented
contrast is caused by any reallocation
activity, the primary mechanism activities but the reallocation
for behavioral
of Education,
was supported
of increases or decreases of with
the inactive
If
contrasts is not the reallocation
of interim
setting in which
is reinforced.
by a Grant-in-Aid
Science and Culture,
time.
of time other than that of the terminal
of inactive time in the usual experimental
only one activity (such as key-peck)
This research
of increases or decreases of
for Scientific
Research from
the Ministry
lapan.
eferemces Azrin, N.H., Hutchinson, R.R., and Hake, D.F. 1966. Extinction-induced aggression. Journal of the Experimental Analysis of Behavior, 9: 191-204. Dougan, J.D., McSweeney, F.K., and Farmer, V.A. 1985. Some parameters of behavioral contrast and allocation of interim behavior in rats. Journal of the Experimental Analysis of Behavior, 44: 325-335. Dougan, I.D., McSweeney, F.K., and Farmer-Dougan, V..4. 1986. Behavioral contrast in competitive and noncompetitive environments. Journal of the Experimental Analysis of Behavior, 46: 185-l 97. Dews, P.B. 1965a. The effect of multiple S-’ periods on responding on a fixed-interval schedule: II. In a primate. Journal of the Experimental Analysis of Behavior, 8: 53-54. Dews, P.B. 1965b. The effect of multiple S’ periods on responding on a fixed-interval schedule: III. Effect of changes in pattern of interruptions, parameters and stimuli. Journal of the Experimental Analysis of Behavior, 8: 427-435. Ettinger, R.H., and Staddon, J.E.R. 1982. Behavioural competition, component duration and multiplc-schedule contrast. Behavioural Analysis Letters, 2: 31-38. Falk, J.L. 1961. Production of polydipsia in normal rats by an intermittent food schedule. Science, 133: 195-196. Ferster, C.B and Skinner, B.F. 1957. Schedules of reinforcement. NJ: Prentice-Hall. Fleshler, M., and Hoffman, H.S. 1962. A progression for generating variable-interval schedules. Journal of the Experimental Analysis of Behavior, 5: 529-530. Gellarman, L.W. 1933. Chance orclers of alternating stimuli in visual discrimination experiments. Journal of Genetic Psychology, 42: 207-208. t-‘zarst, E. and Jenkins, H.M. 1974. Sign-tracking: the stimulus-reinforcer relation and direct action. Austin, Texas: The Psychonomic Society. Hinson, J.M., and Staddon, J.E.P. 1978. Behavioral competition: A mechanism for schedule interactions. Science, 202: 432-434. Jacquei, Y.F. 1972. Sc hed u Ie-induced licking during multiple schedules. Journal of the Experimental Analysis of Behavior, 17: 413-423. Nevin, J.A., and Shettleworth, S.J. 1966. An analysis of contrast effects in multiple schedules. Journal of the Experimental Analysis of Behavior, 9: 305-315. Staddon, J.E.R. 1977. Schedule-induced behavior. In W.K. Honig and J.E.R. Staddon (Eds.), Handbook of operant behavior (pp. 125-l 52). NJ: Prentice-Hall.
123 Staddon, J.E.R. 1982. Behavioral competition, contrast and matching. In M.L. Commons, R.]. Herrnstein, and H. Rachlin t.Eds.1, Quantitative analysis of behavior: Vol. 2. Matching and maximizing accounts (pp. 243-261). Cambridge; MA: Ballinger. Staddon, J.E.R. 1983. Adaptive behavior and learning. Cambridge: Cambridge University Press.