Skip to main content
NCBI home page
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 2003 Feb 10;100(4):2094–2099. doi: 10.1073/pnas.0438002100

Operant behavior can be triggered by the position of the rat relative to objects rotating on an inaccessible platform

E Pastalkova *, E Kelemen †, J Bureš *,‡
PMCID: PMC149964  PMID: 12578961

Abstract

The present study describes a task testing the ability of rats to trigger operant behavior by their relative spatial position to inaccessible rotating objects. Rats were placed in a Skinner box with a transparent front wall through which they could observe one or two adjacent objects fixed on a slowly rotating arena (d = 1 m) surrounded by an immobile black cylinder. The direction of arena rotation was alternated at a sequence of different time intervals. Rats were reinforced for the first bar-press that was emitted when a radius separating the two adjacent objects or dividing a single object into two halves (pointing radius) entered a 60° sector of its circular trajectory defined with respect to the stationary Skinner box (reward sector). Well trained rats emitted 62.1 ± 3.6% of responses in a 60° sector preceding the reward sector and in the first 30° of the reward sector. Response rate increased only when the pointing radius was approaching the reward sector, regardless of the time elapsed from the last reward. In the extinction session, when no reward was delivered, rats responded during the whole passage of the pointing radius through the former reward sector and spontaneously decreased responding after the pointing radius left this area. This finding suggests that rats perceived the reward sector as a continuous single region. The same results were obtained when the Skinner box with the rat was orbiting around the immobile scene. It is concluded that rats can recognize and anticipate their position relative to movable objects.

Spatial navigation has thus far been studied mainly in experimental situations in which an animal has been moving through space. This movement was usually active locomotion as in the Morris water maze (1), on a radial arm maze (2), or on a circular arena (3–5). In these environments, animals demonstrated their spatial ability by direct navigation toward a particular place. Rarely, animals were moved passively through the environment. In the place-recognition task (6), rats sat on the edge of a circular rotating arena and were taught to bar-press whenever transported through a certain 60° sector of their circular trajectory.

In nature, however, there are situations when a stationary animal has to judge the position of other animals or objects moving in space. These may be predators, prey, members of the pack, sexual partners, or siblings. Although the position of the observing animal does not change, this task is spatial in its nature. It has been demonstrated that rats and birds can recognize and remember the position of another animal (7–9). For example, Ray and Heyes (8) let a rat (observer) observe another rat (tutor) while the tutor was pressing a bar for a reward. The bar could be pressed either up or down but only presses in one direction were rewarded. Observers tested for the same task afterward showed a tendency to press a bar in the same direction as the tutor but only when tested in the part of the apparatus where the tutor had worked.

We present an experimental paradigm to study the ability of rats to recognize the position of objects with respect to themselves. In our task the rat was confined within a Skinner box and observed objects rotating on a circular rotatable arena. The rat was rewarded by food for lever-presses emitted when objects on the arena were in a certain position defined with respect to the Skinner box.

Materials and Methods

Animals.

Three-month-old male Long–Evans rats (n = 8) obtained from the breeding colony of the Institute of Physiology, Czech Academy of Sciences, Prague, were used. Animals were housed in groups of three in a temperature-controlled room (20°C) with a 12 h light/12 h dark cycle (light on at 6:00 a.m.). Water was freely available; access to food was restricted to maintain rats at 85% of their free feeding weight. One animal died after training session 22. The treatment of the animals complied with Czech guidelines.

Apparatus and Procedures.

The apparatus consisted of a circular arena (d = 1 m) surrounded by a 25-cm-wide belt. Both these parts could be independently rotated around the same axis. The belt carried a 0.75-m-high black cardboard cylinder, the floor of which was formed by the arena. A Skinner box (30 × 10 × 25 cm) was fixed to the belt. All of the walls of the Skinner box were nontransparent (black), except one that was formed by a wire meshwork. A long axis of the box was oriented toward the center of the arena and its wire mesh front wall fitted into a 10 × 25 cm window in the black cardboard cylinder. This arrangement made it possible either to have the belt stable and the arena rotating at one revolution per 33.3 s (RSPR, rotating scene place-recognition task; Fig. 1A) or to have the belt orbiting at the same angular velocity around the stable arena (ROPR, rotating observer place-recognition task; Fig. 1B). The direction of the rotation was alternated at a sequence of different time intervals. Under both conditions the rat sitting or standing in the front part of the box surveyed the entire surface of the arena. There were either two adjacent objects fixed on the arena at a distance of about two-thirds of the radius from the arena center (a blue-white striped cylinder, 10 cm high, 7 cm in diameter; and an adjacent green box, 3 × 15 × 15 cm; n = 4; Fig. 1) or a piece of red paper cut in the shape of a 60° arena sector with the peak at the arena center (n = 4). Results obtained from rats that were observing two objects and those that were observing the planar cue are reported together because we did not find any difference between their performance. Objects spent an equal amount of time in all parts of their trajectory when rotating clockwise (CW) and counterclockwise (CCW).

Figure 1.

Figure 1

Open in a new tab

(A) In the RSPR task, an animal could observe a rotating arena with two adjacent objects from a stationary Skinner box. (B) The Skinner box was orbiting around a stationary arena in the ROPR task.

Rats were reinforced for the first bar-press that was emitted when a radius separating two objects or dividing the planar object in two halves (pointing radius) entered a 60° sector of its circular trajectory defined with respect to the stationary Skinner box (reward sector). The animal was rewarded with ≈4 pasta pellets each ≈0.02 g in weight delivered from a feeder. During each 60-min session, the pointing radius passed 108 times through the reward sector (5.6 s each pass). Throughout the sessions, orientation of the pointing radius and the direction of its motion were recorded and written into a file once per second and each time a bar-press was made. All of the recorded files were examined off-line.

Rats received 36 training sessions. An extinction session, when no reward was delivered, was run after these training sessions.

Results

Recognition of the Spatial Relation Between the Skinner Box and the Arena Scene in the RSPR Task.

To characterize performance of rats the circular trajectory of the pointing radius was divided into 36 10° segments and the number of responses emitted in each segment was counted. Well trained rats (after 10 training sessions) increased their lever-pressing during approach of the pointing radius to the reward sector, and pressed rarely when the pointing radius was passing through the sectors distant from the reward sector (Fig. 2 B–D). One-way ANOVA with repeated measures (segments = 36, Nrats = 8) showed a significant main effect of orientation of the pointing radius on the arena on the number of emitted responses (F35,245 = 22.04, P < 0.0001; training session 10, Fig. 2B). A Newman–Keuls test indicated that the number of responses emitted within each of the six segments just preceding the reward sector (60° anticipatory sector) and within the first segment of the reward sector irrespective of the direction of movement differed from the number of responses emitted within any of the segments that were further than 10° from the anticipatory sector and from the first segment of the reward sector. Rats consumed their food reward immediately after it was delivered, which caused a steep decrease of lever-pressing within the reward sector.

Figure 2.

Figure 2

Open in a new tab

Average response rate (ordinate) corresponding to the position of a pointing radius on the circular trajectory (abscissa). Black and white squares denote CW and CCW sense of movement, respectively. Vertical bars depict borders of the 60° reward sector. (A) During the first training session rats responded at a constant rate and decreased their responding only after delivery of the reward within the reward sector (n = 8). (B–D) As training progressed, rats increased responding during approach of the pointing radius to the reward sector and responded rarely when the pointing radius was far from the reward sector [n = 8 (B and C) and 7 (D)].

Learning of the RSPR Task.

The circular trajectory of the pointing radius was divided into four 90° quadrants and the number of responses emitted in each of them was counted. The anticipation quadrant spanned 60° of the pointing radius trajectory just before entrance to the reward sector plus the first 30° of the reward sector. The postreward, opposite-to-reward, and prereward quadrants covered the rest of the trajectory of the pointing radius (Fig. 3 Upper). One-way ANOVA (day = 22, Nrats = 8) with repeated measures showed a significant main effect of days on the percentage of responses emitted within the anticipatory quadrant (F21,147 = 28.78, P < 0.0001). A Newman–Keuls test indicated that the percentage of responses emitted within the anticipatory quadrant was significantly lower on each of the first 5 training days than on any other training day between 7 and 22 (P < 0.01; day 1, 27.5 ± 0.49%; day 22, 62.1 ± 3.6%).

Figure 3.

Figure 3

Open in a new tab

Rats largely reduced the number of responses within quadrants distant from the reward sector (prereward and opposite-to-reward quadrants) with training and slightly increased the number of responses emitted within the anticipatory quadrant. Rats consumed their reward immediately after it was delivered, which caused a constantly low number of responses emitted in the postreward quadrant.

This increase in proportion of responses within the anticipation quadrant may be caused by an increase in the absolute number of responses emitted in the anticipation quadrant, a decrease in the number of responses emitted outside the anticipation quadrant, or a combination of both. Two-way ANOVA (quadrants × days, 4 × 22, Nrats = 8) with repeated measures on both factors showed a significant main effect of quadrants on the number of emitted responses (F3,21 = 101.644, P < 0.0001), a significant main effect of training (F21,147 = 2.93, P < 0.0001), and a significant interaction (F63,441 = 10.41, P < 0.0001) (Fig. 3.). A Newman–Keuls test indicated that the number of responses emitted within quadrants distant from the reward sector (the prereward and opposite-to-reward quadrant) was significantly higher at the beginning of training (in each of the first 5 days) than during the later stages of training (each day between 6 and 22) (P < 0.01). The number of responses emitted within the anticipatory quadrant was significantly lower during each of the first 3 days than on days 19, 21, and 22 (P < 0.05). The number of responses emitted in the postreward quadrant did not change with training, possibly because rats usually eat the reward in this quadrant at all stages of training. In addition, one-way ANOVA (day = 22, Nrats = 8) with repeated measures showed a significant main effect of days on the total number of emitted responses (F21,147 = 3.44, P < 0.0001). A Newman–Keuls test indicated that the number of responses emitted on the third day was higher than on each day from days 8 to 22 (P < 0.05). These data suggest that increased performance during training was caused primarily by decreased lever-pressing when objects were far from the reward sector (Fig. 3).

Two-way ANOVA (day × sense, 22 × 2, Nrats = 8) with repeated measures on both factors showed a significant main effect of the direction of movement on the percentage of responses emitted within the anticipatory quadrant (F1,7 = 7.22, P < 0.05), a significant main effect of days (F21,147 = 29.16, P < 0.01), and a significant interaction (F21,147 = 2.06, P < 0.01). A Newman–Keuls test indicated that rats responded differently when objects moved CW or CCW on day 13 only.

What Strategy Did Rats Use to Solve the RSPR Task?

Rats can learn and use a fixed time interval to solve operant tasks (10). To avoid use of this strategy in the RSPR task, CW and CCW direction of movement of the pointing radius was alternated at a sequence of different time intervals. As a result, intervals between two subsequent passes of the pointing radius through the reward sector varied in a constant sequence (41, 24, 33, 41, 24, 33, 58, 8, 33, 58, 8, and 33 s) throughout an experiment. To test whether rats learned to increase responding after a fixed time interval from the last reward, we plotted the number of responses as a function of time from a preceding pass of the pointing radius through the reward sector separately for each of the inter-reward intervals (8, 24, 33, 41, and 58 s; Fig. 4).

Figure 4.

Figure 4

Open in a new tab

(Insets) Arrows show two longest trajectories of the pointing radius [41 s (A); 58 s (B)] between its subsequent passes through the reward sector (gray triangle). Trajectories are depicted for CCW/CW sequence only. Histograms: Each line depicts distribution of responses in the time period between two following passes of the pointing radius through the center of the reward sector. Note that the sense of movement was reversed in the middle of that time period. Rats increased responding only after the pointing radius approached the reward sector independently of time elapsed from the preceding reward. A decrease in response within the reward sector (at the end of each line) was caused by the food consumption.

Rats reduced their responding about 4 s after the pointing radius left the reward sector and pressed at a low rate when it was far from the reward sector independently of time elapsed from the preceding reward. Finally, rats increased lever pressing about 4 s (40°) before the pointing radius entered the reward sector [anticipatory activity (11)]. In addition to this pattern, rats increased their responding when the pointing radius was in the middle of the longest trajectory (58 s; Fig. 4B). At that time the pointing radius was approaching the reward sector but stopped 15° from the border of the sector and started to move in the opposite direction. Rats spontaneously decreased responding about 3 s after the direction of movement had been reversed.

To summarize, rats did not respond at any fixed time interval from the previous reward.

Extinction Session: Did the Rats Recognize the Whole Reward Sector or only Its Boundaries?

Rats increased lever-pressing shortly before the pointing radius entered the reward sector and stopped pressing and started to eat after they received the food reward. As a result a sharp peak of activity was observed at the boundaries of the reward sector. This distribution of responses could indicate that the rats recognized the boundaries of the reward sector rather than the reward sector as a whole. To decide whether this was the case an extinction session was performed. During extinction, reward, and consequently the decrease in lever pressing induced by eating, was eliminated.

During the first 15 min of the extinction session, the rats responded mainly when the pointing radius was close to or within the former reward sector. One-way ANOVA with repeated measures (segments = 36, Nrats = 7) showed a significant main effect of orientation of the pointing radius on the number of responses (F35,210 = 22.91, P < 0.0001). A Newman–Keuls post hoc test indicated that rats pressed significantly more when the pointing radius was closer than 60° to the center of the former reward sector than when it was further than 70° away from the center of the former reward sector (P > 0.05; Fig. 5).

Figure 5.

Figure 5

Open in a new tab

Comparison of distribution of responses during one session with reinforcement (white squares) and one session without reinforcement (extinction session, black triangles). In sessions with reinforcement the rats pressed the bar when the pointing radius was approaching the reward sector, and then stopped responding after the reward was delivered. During the first 15 min of the extinction session rats maintained the high responding when the radius was passing through the entire sector. Rats spontaneously reduced response after the pointing radius left the reward sector [CW (A) and CCW (B) sense of movement, respectively].

During the first 15 min of the extinction session, lever-presses were distributed approximately symmetrically around the center of the former reward sector. In this respect, the extinction differed from sessions with reinforcement when lever-presses were most abundant just before the reward sector and suppressed in its later part. Two-way ANOVA (segments × reinforcement, 36 × 2, Nrats = 7) with repeated measures on both factors showed a significant main effect of reinforcement (F1,6 = 14.91, P < 0.05), a significant main effect of segments (F35,210 = 27.3, P < 0.01), and a significant interaction (F35,210 = 12.44, P < 0.01). A Newman–Keuls test indicated that during extinction the rats responded less than during reward when the pointing radius was 10° before the border of the reward sector and more when it was passing through the distant half of the former reward sector. Two-way ANOVA (sense × sector, 2 × 36, Nrats = 7) with repeated measures did not show any significant main effect of the sense of movement (F1,6 = 0.09, P > 0.05) but revealed a significant main effect of segments (F35,210 = 18.13, P < 0.01) and a significant interaction (F35,210 = 2.395, P < 0.01). The Newman–Keuls test indicated that rats responded more when objects were moving CW and passing through the 20° segment at the exit from the reward sector than when passing through the same segment and entering the reward sector (Fig. 5).

Thus, rats responded during extinction with the same increase of bar-pressing when the pointing radius was approaching the reward sector as during the regular session. Importantly, rats continued to respond at high rates during the whole time the pointing radius was inside the former reward sector and spontaneously reduced their activity after it left the sector. Responses were distributed approximately symmetrically over the reward sector for both CW and CCW direction of movement. This finding suggests that rats perceived the reward sector as a continuous single region.

ROPR Task: The Rat Passively Transported Around the Stable Arena with Objects.

In the previous experiment, the rat was stationary relative to the room and observed the rotating scene. An inverse condition was used in the ROPR task, in which the belt carrying the Skinner box with a rat and the black cylinder was rotated around the stable arena with objects. In addition, the whole arena-belt assembly was continuously slowly rotated in one direction throughout the session so that the reward sector defined by the Skinner box-scene relation was not at a stable position in the room. Visual aspect remained the same in both tasks (identity and position of objects, velocity of movement, and randomization cycle). Rats (n = 4) in the ROPR task were exposed to additional stimuli caused by movement of the box and changes in the direction of the movement of the box (vestibular, proprioceptive stimuli) and auditory stimuli from uncontrolled room-based landmarks.

Two-way ANOVA (task × segment, 2 × 36, NRSPR = 4, NROPR = 4) with repeated measures on segments revealed no significant main effect of the task (F1,6 = 0.03, P > 0.05), a significant main effect of segments (F35,210 = 19.66, P < 0.01), and no significant interaction (F35,210 = 1.03, P > 0.05) (compare Fig. 6A with Fig. 2B). Two-way ANOVA (task × day, 2 × 10, NRSPR = 4, NROPR = 4) with repeated measures on days showed no significant main effect of the task (F1,6 = 1.15, P > 0.05), a significant main effect of days (F9,54 = 33.74, P < 0.01), and no significant interaction (F9,54 = 1.15, P > 0.05) (Fig. 6B).

Figure 6.

Figure 6

Open in a new tab

(A) Average response rate (ordinate) corresponding to the position of an orbiting Skinner box on its circular trajectory (abscissa) in training session 10 of the ROPR task. Black and white squares denote CW and CCW sense of movement, respectively. Vertical bars depict borders of the 60° reward sector. Distribution of responses in the ROPR task did not differ from that one in the RSPR task (compare with Fig. 2B). (B) Percentage of responses (ordinate) emitted within the anticipatory sector increased with the same rate in the RSPR and the ROPR tasks during training.

Thus, rats that observed the stationary scene from the orbiting Skinner box (ROPR) learned the task as well and at the same rate as rats that observed a rotating scene from the stationary Skinner box (RSPR).

Discussion

We have developed a task in which rats observed a rotating inaccessible arena with one or two objects from an adjacent Skinner box and pressed a lever when a radius separating the two adjacent objects or dividing a single object into two halves (pointing radius) entered a 60° sector of its circular trajectory defined with respect to the stationary Skinner box (reward sector).

Do Rats Recognize the Position of Objects?

We have shown that rats responded according to the orientation of the pointing radius: rats pressed rarely when it was far from the reward sector and significantly increased responding when it was approaching the reward sector. We have excluded the possibility that rats used a fixed time interval strategy (i.e., that they responded at a constant time interval after the preceding reward). Because no auditory or olfactory cues indicated the position of objects on the arena, we conclude that rats relied on the position of the objects relative to each other or relative to the box.

During an extinction session, when responses were not reinforced, rats bar-pressed throughout the time the pointing vector was passing through the reward sector. This finding suggests that rats recognized the whole 60° reward sector.

Rats Do Not Generate Locomotion-Derived Cues (Proprioceptive, Vestibular Cues, Efference Copies) Relevant to the Task.

To recognize position within an environment and to navigate, animals may use perceptible distal cues (visual and auditory), local cues (olfactory and tactile), and path integration information (using vestibular, proprioceptive cues, efference copies, and visual flow). Several experimental paradigms have been developed to study the role of different types of isolated sensory inputs. In these studies, some sensory inputs were eliminated or devalued and some others spared and used during spatial behavior. For example, the Morris water-maze paradigm (1) was developed with the intention to study navigation in rats with proximal cues (olfactory and tactile) eliminated. In a place-recognition task (6), restrained rats are passively transported through an environment and taught to bar-press for reward in a certain location. Rats recognize the place by using visual, auditory, and vestibular cues; proximal (olfactory and tactile) and self-movement-derived cues (proprioception and efference copies) are irrelevant to the task. In the active allothetic place-avoidance task (12), rats walk on a slowly rotating arena and are trained to avoid a 60° sector of the arena defined in the room frame. Similarly to the place-recognition task, arena movement forces the animals to rely on extra-arena cues (visual and auditory). The intra-arena (olfactory and tactile) and self-movement-derived cues have to be ignored. Stuchlik et al. (13) and Stuchlik and Bures (14) studied the ability of rats to navigate by using either pure idiothetic cues or by using idiothetic and proximal cues (olfactory and tactile).

In summary, there are experimental models that permit the study of spatial behavior driven mainly by distal (visual and auditory; refs. 1, 6, 12, and 15) or proximal cues (tactile and olfactory; ref. 16). There is a task in which rats navigate by using only self-motion-derived cues (14). None of these tasks makes it possible to study spatial behavior when all of these idiothetic cues are irrelevant to the task. In the present task (RSPR), rats neither walked nor were transported; therefore, this task allows for the study of spatial behavior mediated only by visual cues.

Rats Did Not Have to Walk Over the Arena to Recognize the Position of Objects on the Arena.

This result is in agreement with an observation that rats trained for a Morris water-maze task in a water tank with a Plexiglas barrier (which prevented them from entering a part of the maze not containing the escape platform) were able to find a platform when released from the previously protected area after it was gradually made accessible.§ In contrast, the navigation of the rats in the same task was impaired when the barrier was removed at once (18). According to Matthews et al.,§ inability of rats to navigate within the earlier protected part of the maze after a sudden removal of a barrier was caused by the sudden absence of a silent cue (barrier). Gradual removal of the barrier§ caused the exploratory activity of rats first but rats learned quickly to ignore the instability of the barrier.

Further evidence that rats can facilitate their navigation within an environment by observing the area without active exploration was reported by Sutherland and Linggard (19) and Keith and McVety (20). They have shown that exposure of a naive rat to a submerged platform within a water basin improved performance of this rat in a subsequent swim toward the submerged platform. Improved performance was found even 4 h after the placement (21). In addition, Whishaw (21) has demonstrated that one 30-s exposure of a rat to the platform location in a novel environment is effective in improving the performance of the animal in the Morris water-maze task after the rat acquired the procedural aspects of the task.

Rats Do Not Perform Differently When Observing a Rotating Arena with Objects from a Stationary Skinner Box or When Transported in This Box Around a Stable Arena.

In the present study (RSPR), the Skinner box was stationary and so was the black cardboard cylinder with the rotating floor. The immobile walls of the cylinder supported the impression of a stable surround contrasting with the clearly visible movement of the objects. Interpretation of the observed movement of the arena as movement of the Skinner box seems improbable. The subject may try to connect the activation of the lever with the changed position of the objects seen on the arena.

To compare the RSPR task with experiments in which rats were transported through the environment, we used a control experiment (ROPR) inspired by the place-recognition task (6): rats were transported around the stationary arena in the Skinner box oriented to the arena center. A new group of rats learned the ROPR task at the same rate as the original group learned the RSPR task. This result is not surprising if we take into account the fact that there was no visual difference between the two situations and that the movement of the orbiting box was indicated only by vestibular and proprioceptive signals generated during transportation and changes of the direction of rotation.

Although the operant behavior in both tasks was essentially similar, these tasks may allow involvement of different neural mechanisms. More insight could be obtained by recording place-specific activity of hippocampal neurons (22) in the RSPR and ROPR tasks. Location of a firing field could depend on the position of the subject relative to the start (idiothetic cues) or relative to the landmarks on the inner arena (allothetic reference frame). It is conceivable that confined rats trained in the place-recognition task will pay full attention to the inaccessible inner arena and that the dissociation of the idiothetic and allothetic reference frames by isolated rotation of the arena or of the Skinner box will demonstrate a decisive role of allothesis in the control of the firing fields. Place-recognition training could replace the place-clamp technique (17, 23) that tried to keep the rat in the allothetically defined place field by rotating the ring in a direction opposite to the spontaneous locomotion of the rat. Application of the field clamp to naive rats led in most cases to dissipation of strong firing fields probably caused by the conflicting information received from the idiothetic and allothetic inputs. In contrast to the place-clamp training, rats trained for the RSPR task do not generate any task-relevant idothetic cues. If place-recognition training increases the allothetic control of place cells, then their activity will reflect the current position of the Skinner box relative to the arena.

Acknowledgments

We thank Andre A. Fenton and Daniel Klement for their suggestions and comments, Lynn Nadel for critical reading of the manuscript, and Colleen Dockery for help with training the rats. This work was supported by Grant Agency of the Czech Republic Grant 309/00/1656 and Academy of Sciences Research Project 5011922.

Abbreviations

CW

clockwise

CCW

counterclockwise

RSPR

rotating scene place recognition

ROPR

rotating observer place recognition

Footnotes

§

Matthews, D. B., White, A. M., Brusch, E. D. & Best, P. J. (1995) Soc. Neurosci. Abstr. 21, 2086.

References

  • 1.Morris R B. Learn Motiv. 1981;12:239–261. [Google Scholar]
  • 2.Olton D S, Samuelson R J. J Exp Psychol Anim Behav Processes. 1976;2:97–116. [Google Scholar]
  • 3.Whishaw I Q, Maaswinkel H. J Neurosci. 1998;18:3050–3058. doi: 10.1523/JNEUROSCI.18-08-03050.1998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Maaswinkel H, Whishaw I Q. Behav Brain Res. 1999;99:143–152. doi: 10.1016/s0166-4328(98)00100-4. [DOI] [PubMed] [Google Scholar]
  • 5.Rossier J, Kaminsky Y, Schenk F, Bures J. Behav Neurosci. 2000;114:273–284. [PubMed] [Google Scholar]
  • 6.Klement D, Bures J. Proc Natl Acad Sci USA. 2000;97:2946–2951. doi: 10.1073/pnas.040576197. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Heyes C M, Ray E D, Mitchell C J, Nokes T. Learn Motiv. 2000;31:83–98. [Google Scholar]
  • 8.Ray E D, Heyes C M. Anim Cognit. 2002;5:245–252. doi: 10.1007/s10071-002-0154-7. [DOI] [PubMed] [Google Scholar]
  • 9.Akins C K, Klein E D, Zentrall T R. Anim Learn Behav. 2002;30:275–281. doi: 10.3758/bf03192836. [DOI] [PubMed] [Google Scholar]
  • 10.Dews P. In: The Theory of Reinforcement Schedules. Schoenfeld W N, editor. New York: Appleton-Century-Crofts; 1970. pp. 43–61. [Google Scholar]
  • 11.Ferster C B, Skinner B F. Schedules of Reinforcement. New York: Appleton-Century-Crofts; 1957. [Google Scholar]
  • 12.Cimadevilla J M, Kaminsky Y, Fenton A A, Bures J. J Neurosci Methods. 2000;102:155–164. doi: 10.1016/s0165-0270(00)00288-0. [DOI] [PubMed] [Google Scholar]
  • 13.Stuchlik A, Fenton A A, Bures J. Proc Natl Acad Sci USA. 2002;98:3537–3542. doi: 10.1073/pnas.051630498. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Stuchlik A, Bures J. Behav Brain Res. 2002;128:179–188. doi: 10.1016/s0166-4328(01)00314-x. [DOI] [PubMed] [Google Scholar]
  • 15.Simpson E L, Gaffan E A. Q J Exp Psychol B. 1999;52:1–29. doi: 10.1080/713932691. [DOI] [PubMed] [Google Scholar]
  • 16.Fenton A A, Wesierska M, Kaminsky Y, Bures J. Proc Natl Acad Sci USA. 1998;95:11493–11498. doi: 10.1073/pnas.95.19.11493. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Bures J, Fenton A A, Kaminsky Y, Zinyuk L. Proc Natl Acad Sci USA. 1997;94:343–350. doi: 10.1073/pnas.94.1.343. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Sutherland R J, Chew G L, Baker J C, Linggard R C. Psychobiology. 1987;15:48–57. [Google Scholar]
  • 19.Sutherland R J, Linggard R. Behav Neural Biol. 1982;36:103–107. doi: 10.1016/s0163-1047(82)90101-7. [DOI] [PubMed] [Google Scholar]
  • 20.Keith J R, McVety K M. Psychobiology. 1988;16:146–151. [Google Scholar]
  • 21.Whishaw I Q. Q J Exp Psychol B. 1991;43:83–103. [PubMed] [Google Scholar]
  • 22.O'Keefe J, Dostrofsky J. Brain Res. 1971;34:171–175. doi: 10.1016/0006-8993(71)90358-1. [DOI] [PubMed] [Google Scholar]
  • 23.Bures J, Fenton A A, Kaminsky Y, Rossier J, Sacchetti B, Zinyuk L. Philos Trans R Soc London B. 1997;352:1515–1524. doi: 10.1098/rstb.1997.0138. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES