- Email: [email protected]
Rats’ memory for event duration in delayed matching-to-sample with nonspatial comparison response alternatives
Available online at www.sciencedirect.com
Behavioural Processes 78 (2008) 1–9
Rats’ memory for event duration in delayed matching-to-sample with nonspatial comparison response alternatives Patrick Van Rooyen, Neil McMillan, Angelo Santi ∗ Department of Psychology, Wilfrid Laurier University, Waterloo, Ontario, Canada N2L 3C5 Received 8 June 2007; received in revised form 24 October 2007; accepted 29 November 2007
Abstract Previous research has suggested that using stationary and moving levers as nonspatial response alternatives can significantly enhance the speed of acquiring a temporal discrimination in rats. In Experiment 1, rats were trained to discriminate 2 and 8 s of magazine light illumination by responding to either a stationary lever or a moving lever with a cue light illuminated above it. Rats learned to discriminate event durations at a high level of accuracy after 25 sessions of training. During subsequent delay tests, rats exhibited a strong choose-long bias, indicating that they were timing from the onset of the magazine light until the entry of levers into the chamber. This occurred regardless of whether intertrial intervals and delay intervals were dark or illuminated. On test trials in which the sample was omitted, rats responded as if the short sample had been presented. In Experiment 2, the rats received extensive training with dark and illuminated variable delay intervals (1–4 s). However, they continued to exhibit a tendency to time from the onset of the magazine light until entry of the levers into the chamber. Although the use of stationary/uncued and moving/cued levers as response alternatives enhanced the speed of acquisition of the event duration discrimination in rats, additional procedural modifications will be necessary to prevent rats from timing during the delay interval. © 2007 Elsevier B.V. All rights reserved. Keywords: Rats; Memory for time; Time discrimination; Delayed matching-to-sample
1. Introduction Studies of memory for event duration in pigeons have typically employed a delayed symbolic matching-to-sample task in which a pigeon is trained to peck one comparison stimulus (e.g., green) on trials initiated by a 2-s sample stimulus and to peck a different comparison stimulus (e.g., red) on trials initiated by an 8-s sample stimulus. Following training in this task, the delay interval (DI) between the end of the sample stimulus and the presentation of the comparison stimuli is varied. Numerous studies have shown that at extended delays pigeons respond with high accuracy on trials initiated by the short sample, while accuracy on trials initiated by the long sample drops below 50% correct (Fetterman, 1995; Gaitan and Wixted, 2000; Grant, 1993; Grant and Kelly, 1996, 1998; Grant and Spetch, 1991, 1993, 1994; Grant et al., 1997; Kelly and Spetch, 2000; Kraemer et al., 1985; Santi et al., 1993; Santi et al., 1992; Sherburne et al., 1998; Spetch, 1987; Spetch and Rusak, 1989, 1992; Spetch and
∗
Corresponding author. Tel.: +1 519 884 0710x3087; fax: +1 519 746 7605. E-mail address: [email protected] (A. Santi).
0376-6357/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.beproc.2007.11.012
Wilkie, 1983). Several different explanations have been provided for this choose-short effect in pigeons (Dorrance et al., 2000; Gaitan and Wixted, 2000; Kraemer et al., 1985; Spetch and Rusak, 1992; Spetch and Wilkie, 1983; Ward and Odum, 2007). The study of memory for event duration in rats has not produced consistent evidence of a choose-short effect. While a choose-short effect has been found in some studies (Church, 1980, Experiment 3A; Al-Zahrani et al., 1997; Harper and Bizo, 2000; Leblanc and Soffie, 1999, 2001; Santi et al., 1995a,b; Soffie et al., 1999), in other studies a choose-long bias has been obtained (Berz et al., 1992a,b; Church, 1980, Experiment 3B; Keen and Church, 2003; Meck et al., 1984; Roberts, 1982; Santi et al., 1995a, 1997). In part, this may reflect the lack of similarity in the procedures used in these experiments. This procedural variability is reflected in the temporal signals used, the response alternatives used, and the procedures for delay testing. The temporal signals have included an increase in illumination (Al-Zahrani et al., 1997; Berz et al., 1992a,b; Harper and Bizo, 2000; Roberts, 1982; Santi et al., 1995a,b, 1997), a decrease in illumination (Church, 1980), presentation of an auditory signal (Meck et al., 1984; Roberts, 1982; Santi
2
P. Van Rooyen et al. / Behavioural Processes 78 (2008) 1–9
et al., 1995a), and presentation of a bimodal signal (Leblanc and Soffie, 1999, 2001; Soffie et al., 1999). The response alternatives used include spatial locations (Church, 1980; Keen and Church, 2003; Leblanc and Soffie, 1999, 2001; Meck et al., 1984; Roberts, 1982; Soffie et al., 1999), spatial locations with a nose poke/lever press requirement during the delay interval (Al-Zahrani et al., 1997; Berz et al., 1992a,b; Harper and Bizo, 2000), nonspatial–visual response cues (Santi et al., 1995a,b), and nonspatial–auditory response cues (Santi et al., 1997). At this time, no single procedural variable or combination of variables appears to be consistently related to the occurrence of a choose-short, or a choose-long bias during delay testing. One of the difficulties associated with studying memory for event duration in rats is that rats have a difficult time acquiring event duration discriminations when nonspatial response alternatives are used. Nonspatial response alternatives are methodologically better than spatial response alternatives because the use of nonspatial response alternatives makes it more difficult for rats to use spatial positioning during the DI to mediate responding to the correct lever location. When either nonspatial visually cued response alternatives (Santi et al., 1995a), or nonspatial auditory-cued response alternatives (Santi et al., 1997) have been used, rats acquire the discriminations very slowly (between 9500 and 15,700 trials), and their level of accuracy is modest (between 65% and 75% correct). Furthermore, the response bias exhibited during delay testing has not been consistent in these studies. In the tone signal–visual choice cue group of Santi et al. (1995a) and the light signal-tone choice cue group of Santi et al. (1997), rats exhibited a choose-long bias. The rats appeared to use the onset of the choice cue to stop timing rather than offset of the to-be-timed signal. On the other hand, when the signal-cue modalities were the same, such as in the visual signal–visual choice group of the Santi et al. (1995a,b) studies, rats exhibited a choose-short effect. The rats appeared to use the offset of the signal itself to stop the clock. Santi et al. (1995a,b) are the only studies reporting a choose-short effect in rats when nonspatial response alternatives are used. Research by Meck et al. (1985) suggested that a temporal discrimination between 2- and 4 s of white-noise could be rapidly acquired by rats when stationary and moving levers are used as nonspatial response alternatives. Half of their subjects were trained to press a stationary lever following the 2-s signal and a moving lever following the 4-s signal while the remaining rats were trained with the opposite contingency. Meck et al. (1985) did not report the number of trials per session or the mean accuracy for each signal during the acquisition phase of their study. However, they reported that only 10 sessions of training were required before the start of psychophysical testing. Recently, Santi and Van Rooyen (2007) made use of the Meck et al. (1985) procedure in training rats to discriminate sample stimuli that consisted of sequences of tone bursts. The rats were trained to discriminate two tone bursts in 4 s (2b/4 s) or eight tone bursts in 4 s (8b/4 s) by responding to either a stationary lever or a moving lever with a cue light illuminated above it. The rats were able to discriminate the sequences with accuracy above 75% correct within approximately 4800 training trials. This success in using stationary and moving levers as nonspatial response alternatives
in the discrimination of sequences of tone bursts led to the current study in which rats were required to discriminate different durations of magazine light. 2. Experiment 1 The main purpose of the current experiment was to replicate the findings of Meck et al. (1985) by demonstrating that using stationary and moving levers as nonspatial response alternatives would produce faster acquisition and a higher level of temporal discrimination accuracy than that reported in previous research using different nonspatial response alternatives (Santi et al., 1995a, 1997). However, Meck et al. (1985) did not conduct delay testing in their study. Consequently, a second purpose of the current experiment was to conduct delay tests and determine whether similarity in the ambient illumination conditions of the intertrial interval (ITI) and DI would affect retention functions for event duration in rats in the same way that it affects retention functions in pigeons. Rats were trained to discriminate 2 versus 8 s of magazine light illumination by responding to either a stationary lever or a moving lever with a cue light illuminated above it. The ambient illumination condition during the ITI was manipulated between groups. One group of the rats was trained with an illuminated ITI (Group Light) and the other group was trained with a dark ITI (Group Dark). According to the instructional ambiguity hypothesis (Cohen and Njegovan, 1999; Dorrance et al., 2000; Sherburne et al., 1998; Zentall, 1997, 1999, 2005), if the ambient stimulus conditions during the ITI and delay are the same, a novel DI may be confused with the ITI. This was proposed as an alternative explanation for the choose-short response bias which occurs when memory for duration samples is tested in pigeons. Because pigeons confuse novel delay intervals with the ITI, the presentation of comparisons at the end of the delay may be perceived as a trial without a duration sample, and nosample would likely be judged as more like a short sample than a long sample. Consequently, pigeons increasingly respond to the comparison that was associated with the short sample during training. As a result, a choose-short effect occurs, but primarily when the stimulus conditions during the ITI and delay are the same (i.e., confusable). If a similar process occurs for rats, they should increasingly respond to the response alternative associated with the short sample at extended delays when the stimulus conditions during the ITI and delay are the same (i.e., light ITI – light DI, dark ITI – dark DI), but not when they are different (i.e., light ITI – dark DI, dark ITI – light DI). The current experiment will be the first to examine whether the stimulus conditions during the ITI and delay affect retention functions for event duration in the rat. 2.1. Method 2.1.1. Subjects Eight experimentally na¨ıve Long-Evans hooded male rats were used as subjects. The 58–63-days old rats were obtained from Charles River Canada (St. Constant, PQ). They were individually housed in clear Plexiglas cages in a temperature- and humidity-controlled holding room under an alternating 12 h
P. Van Rooyen et al. / Behavioural Processes 78 (2008) 1–9
light–dark cycle. Experimental sessions were given during the light phase. The rats had free access to water but their diet was restricted to maintain approximately 85% of their free-feeding bodyweight. 2.1.2. Apparatus Four Coulbourn modular operant test cages (Model #E1010), housed individually in isolation cubicles (Model #E10-20) and located in the same room were used. The cubicles were equipped with a ventilation fan and baffled air intake exhaust system. Each test cage was equipped with a 45-mg pellet dispenser (Model #E14-06), which was mounted in the center of the front wall of each test cage just above the steel-grid flooring. Two retractable levers (Model #E23-07) were mounted 2.5 cm above the floor on either side of the pellet feeder and an opaque cue-light was positioned above each lever. A single Sonalert tone module (Model # E12-02) was mounted 2.5 cm above each cue light, but they were not used in the current experiments. A houselight was also installed 6.5 cm above the pellet feeder and positioned so that the light was facing upward and reflected off of the ceiling of the test cage (Coulbourn Model #E11-01 with bulb #SL1819X). All of the experimental events and responses were controlled by a microcomputer located in the same room as the four test cages. 2.1.3. Procedure Each rat received several sessions of combined magazine and lever training. The rats were placed in the operant chamber with both the right and left levers retracted. Each trial began with the entry of the left or right lever into the chamber. The lever remained extended until it was pressed or 60 s had elapsed, whichever occurred first. Either event resulted in delivery of a 45-mg food pellet and retraction of the lever. Pellet delivery produced an audible “click” and the light in the magazine was illuminated for 0.5 s. The houselight was illuminated during all of the preliminary training sessions. The sessions were terminated when the rat pressed the lever 60 times or 60 m had passed. After the rats had acquired the lever press to a stationary lever, they were trained to respond to both a stationary/uncued lever and a moving/cued lever randomly presented on both the right and the left side of the test chamber. When the stationary/uncued lever was presented, it entered the chamber without the opaque cue light above it being illuminated. It remained extended until a lever press occurred or 6 s elapsed without a lever press, whichever occurred first. If no response occurred the stationary/uncued lever was retracted and it was presented again after 5 s had elapsed. When the moving/cued lever was presented, it entered the chamber with the cue light above it on and it was inserted and retracted at 0.6 s intervals. These 1.2-s cycles continued for a maximum of five repetitions (i.e., 6 s). If no response occurred within 6 s, the moving/cued lever was retracted and the cue light above it was terminated for 5 s, before another set of 5 cycles of lever insertion/retraction was presented. Responses to either the stationary/uncued or moving/cued lever within the 6 s presentation interval resulted in immediate retraction of the lever, termination of the cue light if it had been illuminated, and delivery of a food pel-
3
let. Preliminary training continued until all of the rats were reliably pressing both the stationary/uncued and moving/cued lever. After preliminary training, the rats were trained to discriminate between 2 and 8 s of magazine light illumination. Trials were separated by a randomly selected ITI of 8, 16, 32, or 64 s. Four rats were trained with the houselight on during the ITI (Group Light) and the remaining four rats were trained with the houselight off during the ITI (Group Dark). Each trial commenced with a sample presentation in which the light in the food magazine was illuminated for 2 or 8 s. Following presentation of the sample, the rats were presented with either the stationary/uncued lever on the left and the moving/cued lever on the right, or the moving/cued lever on the left and the stationary/uncued lever on the right. The moving lever was retracted and inserted at 0.6-s intervals and these repeating 1.2-s cycles continued until the first lever press after a 2.4-s delay. The 2.4-s delay between the insertion of the levers and the recording of the choice response was used so that the animal had enough time to determine which lever was moving before making its response (Meck et al., 1985). This will be referred to as a 0-s delay condition. If a rat did not press one of the levers between 2.4 and 6 s after lever insertion, the levers were retracted, the cue light turned off, and the same sample duration and comparison lever configuration was presented again after a 5-s delay. Four rats were trained to press the stationary/uncued lever following the short sample, and the moving/cued lever following the long sample (i.e., two from Group Light and two from Group Dark). The remaining four rats were trained similarly with the opposite contingencies. For all rats, a response to either lever retracted both levers and turned off the cue light. A correct response resulted in delivery of a food pellet, while an incorrect response resulted in a 5-s delay period and re-presentation of the same sample duration and configuration of comparison levers. Only the choice response on the initial (noncorrection) trial was used to calculate response accuracy. After three consecutive incorrect responses had been made on any given trial, the ITI was presented and the next trial was initiated. Each session of training consisted of 120 trials. Each combination of sample type and comparison test condition occurred once in each block of four trials. The order of trial presentation was randomized individually for each rat in each session. After 25 sessions of discrimination training, the rats received 15 sessions of delay testing. During these sessions, the delay between end of the sample and the entry of the comparison levers into the chamber was 0, 1, 2, 4 and 8 s. Within each session, 24 trials for each sample (short and long) occurred at the 0s baseline delay, and four trials for each sample occurred at each of the other delays (1, 2, 4, and 8 s). For one half of the trials involving the 1-, 2-, 4-, 8-s delays, the entire delay was spent in darkness. On the other half of the trials, the houselight was illuminated for the entire delay. During delay testing, the correction procedure remained in effect only for 0-s delay trials. Errors on all other delay trials were followed by an illuminated (Group Light) or a dark (Group Dark) ITI and the presentation of the next sample duration. All other parameters remained the same as those described previously.
4
P. Van Rooyen et al. / Behavioural Processes 78 (2008) 1–9
Following delay testing, the rats received three sessions of baseline training, which were identical to those that preceded delay testing. Following these sessions, the rats received five sessions of sample-omission testing. During sample-omission test sessions, 96 of the 120 trials were baseline trials identical to those that preceded delay testing. On the remaining 24 trials, the sample stimulus was not presented. One no-sample trial occurred randomly within each block of five trials. On no-sample trials, in both Groups, the ITI was followed by a 4-s interval of darkness prior to entry of the levers into the chamber. On these test trials, a response to either the stationary or the moving lever was randomly followed by delivery of a food pellet with a probability of .50. All of the other parameters during testing were the same as those described for baseline training. In all the statistical analyses reported in this article, the rejection region was p < .05. 2.2. Results and discussion Fig. 1 presents the acquisition data for Group Light in the top panel and those for Group Dark in the bottom panel. Both groups were able to discriminate the event durations at a high level of accuracy within 25 sessions of training. An ANOVA was conducted on these data with group as a between-subjects factor and sample type (short and long) and blocks of five sessions as within-subjects factors. There was a significant main effect
Fig. 2. The mean percentage of correct responding on short- and long-sample trials for Group Light (top panel) and for Group Dark (bottom panel) during delay testing as a function of delay interval length and delay interval illumination. Error bars represent the standard error of the mean.
Fig. 1. The mean percentage of correct responding on short- and long-sample trials during acquisition in Experiment 1 for the group trained with the houselight on during the ITI (Group Light) and the group trained with the houselight off during the ITI (Group Dark). Error bars represent the standard error of the mean.
of blocks, F(4, 24) = 77.32, as well as a group × blocks interaction, F(4, 24) = 4.32. While accuracy increased across blocks for both Group Light and Group Dark, during Blocks 3 and 4 accuracy was significantly greater for Group Dark than for Group Light, F(1, 6) = 5.89 and 12.45 respectively. However, in the final block of training, there was no significant difference in accuracy between the two groups F < 1. No other main effects or interaction effects were significant. The mean percentage of correct responding during delay testing sessions is shown in Fig. 2. The data for Group Light is presented in the top panel and the data for Group Dark in the bottom panel. One half of all 0-s delays were randomly designated as dark delays, and the other half were randomly designated as illuminated delays. At the 0-s delay, accuracy was slightly better for short samples than for long samples in both groups. However, at extended delays, for both groups, accuracy was much higher for long samples than for short samples regardless of the illumination condition during the delay. An ANOVA was conducted on the delay test data with group as a betweensubjects factor, and sample type (short and long), DI illumination (light and dark), and DI length (0, 1, 2, 4, 8 s) as within-subjects factors. The analysis revealed a significant main effect of sample type, F(1, 6) = 19.20, and DI, F(4, 24) = 31.94. In addition, there was a significant sample type × DI length interaction, F(4, 24) = 109.95. At the 0-s delay, accuracy was significantly
P. Van Rooyen et al. / Behavioural Processes 78 (2008) 1–9
greater on the short-sample trials than on the long-sample trials, F(1, 6) = 19.12. There was no difference in accuracy between short- and long-sample trials at the 1-s delay. However, at each of the remaining delays, accuracy was significantly greater for long-sample than for short-sample trials, F(1, 6) = 9.14, 64.62, and 54.31 respectively. There was no significant main effect for group or significant interactions involving the group factor. The mean percentage of long responses during sampleomission test sessions was analyzed with group as a between-subjects factor and sample type (short, long and nosample) as a within-factor. Only the main effect of sample type was statistically significant, F(2, 12) = 65.22. The mean percentage of long responses was approximately 85% following the long samples, 19% following the short sample, and 25% on the no-sample trials. This result indicates that in the absence of a sample, rats respond as if the sample was short. A similar finding has been frequently reported in pigeons (see Spetch and Wilkie, 1983). Rats acquired the delayed symbolic matching-to-sample task more quickly and to a higher level of accuracy with nonspatialstationary/moving response alternatives than in previous studies which used either nonspatial visually-cued response alternatives (Santi et al., 1995a), or nonspatial auditory-cued response alternatives (Santi et al., 1997). In the previous Santi et al. studies (1995a, 1997) acquisition occurred very slowly (between 9500 and 15,700 trials) and accuracy levels were modest (65–75%). In the current experiment, rats learned to discriminate event durations at a high level of accuracy (above 85%) after 25 sessions of training (approximately 3000 trials). This result is consistent with the study by Meck et al. (1985) in which psychophysical testing was undertaken following only 10 sessions of training. As noted in Section 1, Meck et al. (1985) did not report the number of trials per session or the mean level of accuracy at the end of training. Nevertheless, the use of stationary versus moving levers appears to be an effective procedure to establish temporal discriminations with nonspatial response alternatives. In the current study, a cue light was illuminated above the moving lever but not above the stationary lever. Meck et al. (1985) did not employ a cue light in their study. It is unclear whether this procedural variation in the current study facilitated acquisition relative to what would have occurred without the cue light. In addition, Meck et al. (1985) did not report whether there was a preference for the stationary or moving lever in their study. In the current study, the overall mean percent choice of the stationary lever versus moving lever during acquisition was 47.8% versus 51.3%. The use of a correction procedure in the current study presumably discouraged the development of a lever preference. Delay tests following acquisition revealed an increasing tendency to respond to the lever correct for the long sample as the DI was extended. Rats apparently learned to time from the onset of magazine light to the insertion of levers into the chamber. When the insertion of the levers was delayed, rats continued to time and as a result exhibited a choose-long bias at extended delays. This tendency to time through the delay was not affected by the ambient illumination conditions during
5
the ITI and DI. Although the choose-long bias obtained with rats is different from the choose-short bias normally obtained in pigeons, both species exhibit a bias to respond to the comparison alternative associated with the short sample on sample-omission trials. 3. Experiment 2 Several studies have examined memory for event duration in pigeons after training with a constant nonzero delay (Grant and Kelly, 1998; Kelly and Spetch, 2000; Spetch, 1987; Spetch and Rusak, 1989, 1992). In all of these studies, pigeons exhibited a choose-short bias at extended delays. However, several recent studies have indicated that following training with variable delays, pigeons do not show a significant choose-short effect (Dorrance et al., 2000; Grant and Talarico, 2004; Talarico and Grant, 2006). Grant (2006) provided the first within-experiment comparison of the effects of nonzero training delays and variability in those training delays on the magnitude of the choose-short effect in pigeons. His results indicated that when different keylight durations were used as samples, training with either a constant or a variable nonzero delay reduced the size of the choose-short effect. In addition, the amount of reduction was greater at the longer training delays. As noted by Grant (2006) the main procedural variable that most reliably differentiates between those studies that demonstrate a sensitivity of the choose-short effect to nonzero delay training and those that do not is the nature of the event that provides the duration information. Studies that demonstrate the sensitivity have used different durations of keylight (Dorrance et al., 2000; Grant, 2006; Grant and Talarico, 2004; Talarico and Grant, 2006), while those that do not have used different durations of houselight or access to food (Kelly and Spetch, 2000; Spetch, 1987; Spetch and Rusak, 1989, 1992). Consistent with this analysis, Grant (2007) employed a within-experiment comparison to demonstrate that training delays reduce the choose-short effect with keylight, but not with food, event duration samples in pigeons. Thus far, no study has examined memory for event durations in rats following training with nonzero delays. Since rats exhibited a choose-long bias in Experiment 1 following 0-s delay training, it would be worthwhile to determine if they continue to do so following training with variable delays of 1–4 s duration which were equally often either dark or illuminated. In Experiment 1, the rats may have learned to time from onset of the magazine light until presentation of the levers, because the offset of the magazine light was not as salient a stimulus as the entry of the levers into the chamber. Training with dark and illuminated variable nonzero delays may make the onset of the delay more salient and result in the rats being more likely to stop timing at the start of the delay. On the other hand, it is possible that rats may continue timing during the delay because with samples of 2 and 8 s and training delays of 1, 2 and 4 s, the total time from onset of the magazine light to the entry of the levers into the chamber would still provide a reliable cue for correct responding (i.e., combined sample-delay durations of 3, 4, or 6 s for “short” and 9, 10, or 12 s for “long”).
6
P. Van Rooyen et al. / Behavioural Processes 78 (2008) 1–9
3.1. Method 3.1.1. Subjects and apparatus The subjects and apparatus used in Experiment 1 were also used in Experiment 2. 3.1.2. Procedure The rats received three sessions of baseline training which were identical to those that preceded testing in Experiment 1. This was followed by 60 sessions of training with variable delays of 1, 2, and 4 s. The delays were equally often either dark or illuminated by the houselight. Each combination of DI illumination and DI length condition occurred equally often for each of the four combinations of sample type and configuration of comparison levers. A session consisted of 120 trials. A correction procedure was used on all trials. Consequently, a correct response resulted in delivery of a food pellet, while an incorrect response resulted in a 5-s delay period and re-presentation of the same sample, DI condition, and configuration of comparison levers. Only the choice response on the initial (noncorrection) trial was used to calculate response accuracy. After three consecutive incorrect responses had been made on any given trial, the ITI was presented and the next trial was initiated. All other aspects of discrimination training remained the same as previously described. Following this training, rats received 10 sessions of testing at extended delays. During these sessions, the delay between end of the sample and the entry of the comparison levers into the chamber was 1, 2, 4, 8, and 16 s. For one half of the trials, the DI was dark, while for the remaining trials it was illuminated by the houselight. Each combination of DI illumination and DI length condition occurred equally often for each of the four combinations of sample type and configuration of comparison levers. During delay testing, the correction procedure remained in effect only for 1-, 2- and 4-s delay trials. Errors on the 8- and 16-s delay trials were followed by an illuminated (Group Light) or a dark (Group Dark) ITI and the presentation of the next trial. All other parameters remained the same as those described previously. 3.2. Results and discussion Overall accuracy remained the same between the first 5 sessions of training with variable delays of 1, 2, and 4 s and the last 5 sessions of training. Mean accuracy was 75.9% over the first 5 sessions of training and it was 75.8% over sessions 55–60. Fig. 3 shows accuracy averaged over the last 5 sessions of training as a function of sample duration and DI condition. The data for Group Light (n = 4) is presented in the top panel and the data for Group Dark (n = 4) in the bottom panel. At the 1-s delay, Group Light exhibited higher accuracy for short samples than for long samples regardless of whether the delay was dark or illuminated. As the delay was extended, accuracy decreased for the short sample, but it increased for the long sample. Group Dark did not exhibit an overall difference in accuracy for short and long samples at the 1-s delay, however similarly to Group Light, they exhibited increasing accuracy for long samples and decreasing accuracy for short samples as the DI was lengthened.
Fig. 3. The mean percentage of correct responding on short- and long-sample trials for Group Light (top panel) and for Group Dark (bottom panel) averaged over the last five sessions of variable delay training as a function of delay interval length and delay interval illumination. Error bars represent the standard error of the mean.
At the longest delay of 4 s, both groups exhibited higher accuracy for long samples than for short samples. These data suggest that both groups continued to time during the DI regardless of whether it was dark or illuminated. An ANOVA was conducted on these data with group as a between-subjects factor, and sample type (short and long), DI illumination (light and dark), and DI length as within-subjects factors. The analysis revealed a significant main effect of DI length, F(2, 12) = 12.20. In addition, there was a significant sample type × DI length interaction, F(2, 12) = 36.30. At the 1-s delay, performance was significantly better on short-sample trials than on the long-sample trials, F(1, 6) = 12.26. There was no significant difference between short and long samples at the 2-s delay. However, at the 4-s delay, performance was significantly better on long-sample trials than on short-sample trials, F(1, 6) = 18.87. In addition, there was a significant sample type × group interaction, F(1, 6) = 6.44, and a DI illumination × sample type × group interaction, F(1, 6) = 16.22. The three-way interaction was the result of a significant DI illumination × sample type interaction for Group Dark, F(1, 6) = 15.88, but not for Group Light. For Group Dark, when the sample was short, accuracy was significantly higher if the DI was dark than if it was illuminated, F(1, 6) = 8.91. However, when the sample was long, the presence or absence of illumination during the DI did not significantly affect accuracy.
P. Van Rooyen et al. / Behavioural Processes 78 (2008) 1–9
Fig. 4. The mean percentage of correct responding on short- and long-sample trials for Group Light (top panel) and for Group Dark (bottom panel) during the delay tests which followed variable delay training. The data are presented as a function of delay interval length and delay interval illumination. Error bars represent the standard error of the mean.
Despite having received training with variable delays of 1, 2 and 4 s for 60 sessions, the rats continued to exhibit a tendency to time during the DI. Additional evidence of this was obtained during testing with extended delays as can been seen in Fig. 4. At the extended delays of 8 and 16 s, both groups exhibited above-chance accuracy following long samples, but below-chance accuracy following short samples. An ANOVA was conducted with group as a between-subjects factor, and sample type (short and long), DI illumination (light and dark), and DI length as within-subjects factors. The analysis revealed a significant main effect of sample type, F(1, 6) = 23.69, and DI length, F(4, 24) = 27.94. In addition, there was a significant sample type × DI length interaction, F(4, 24) = 45.42. At the 1-s delay, accuracy was significantly greater on the short-sample trials than on the long-sample trials, F(1, 6) = 9.14. There was no significant overall difference in accuracy between short- and long-samples at the 2- or 4-s delay. However, at the 8- and 16-s delay, accuracy was significantly greater on long-sample trials than on short-sample trials, F(1, 6) = 63.82 and 46.84 respectively. 4. General discussion Rats acquired the delayed symbolic matching-to-sample task more quickly and to a higher level of accuracy with nonspatial-
7
stationary/moving response alternatives than in previous studies which used either nonspatial visually-cued response alternatives (Santi et al., 1995a), or nonspatial auditory-cued response alternatives (Santi et al., 1997). This result is consistent with the study by Meck et al. (1985) in which psychophysical testing was undertaken following only 10 sessions of training. In the current study, a cue light was illuminated above the moving lever but not above the stationary lever. Meck et al. (1985) did not employ a cue light in their study and they did not mention whether their rats exhibited a preference for either the stationary or moving lever. In the current study, there was no evidence of a stationary or moving lever preference, but it is likely that any preference would have been eliminated by the use of a correction procedure. Group Dark initially acquired the discrimination somewhat faster than Group Light. This was probably due to the greater salience of the onset of the magazine light sample following a dark ITI than an illuminated ITI. However, during the last block of training, accuracy was equivalent for Group Dark and Group Light. As the DI was extended, accuracy remained well above chance on trials initiated by the long sample, but it declined to well below chance on trials initiated by the short sample. This bias to respond to the response alternative correct for the long sample was not affected by either the ITI or the DI illumination condition. It appears that the rats learned to time from onset of the signal duration until presentation of the comparisons. This may have occurred because the offset of the signal (i.e., magazine light) was not as salient a stimulus as the entry of the levers into the chamber during the choice phase of a trial. In Experiment 2, rats were trained with both dark and illuminated variable nonzero delays in an attempt to make the onset of the DI more salient and to discourage them from continuing to time during the DI. Despite the use of a correction procedure on all trials throughout the 60 sessions of variable delay training, rats continued to time during the delay. With samples of 2 and 8 s and training delays of 1, 2 and 4 s, the total time from onset of the magazine light to the entry of the levers into the chamber still provided a reliable cue for correct responding and reinforcement (3, 4, or 6 s for “short” and 9, 10, or 12 s for “long”). Subsequent testing at delays greater than 4 s resulted in a strong bias to respond to the response alternative correct for long-sample trials. Experiment 2 might have been more successful if the correct response contingencies had been reversed from those used in Experiment 1, or if na¨ıve rats had been used. Another possibility would be to employ shorter signal durations and/or longer delay intervals in training to produce more of an overlap between the onset of the sample and the entry of the response levers on short- and long-sample trials. Although many studies of temporal memory in pigeons have reliably reported a choose-short effect, there are a few studies in which a choose-long effect has been reported. For example, Dorrance et al. (2000) observed a choose-long effect in a group of pigeons trained with a bright houselight during the ITI, a dimmer keylight as the duration sample, and a dark DI. Pigeons timed from the offset of the houselight until onset of the comparisons in this study. Choose-long effects in pigeons have also been reported when auditory duration samples are used (Miki and
8
P. Van Rooyen et al. / Behavioural Processes 78 (2008) 1–9
Santi, 2001; Santi et al., 1998a,b), or empty intervals bound by brief light or tone markers are used (Santi et al., 1999, 2003). Several previous studies in rats have reported a choose-long effect when the DI is increased (Berz et al., 1992a,b; Church, 1980; Meck et al., 1984; Santi et al., 1997; Roberts, 1982). In all of these studies, except Santi et al. (1997), the discriminative response was based on the spatial location of the levers (left/right). Fetterman et al. (1998) presented evidence that temporal judgements by rats performing a spatially-differentiated response were mediated by collateral behaviors. Behavioral mediation could account for the development of a choose-long bias as the delay is lengthened in a timing task with spatial response alternatives. Studies which minimize the development of a response strategy during the delay by requiring a response (e.g., nose poke or lever press) at a central location (Al-Zahrani et al., 1997; Harper and Bizo, 2000) or by using nonspatial comparison response alternatives (Santi et al., 1995a) have been somewhat more successful in obtaining a choose-short effect in rats. However, as the present findings indicate, the use of nonspatial comparison response alternatives in rats does not always result in choose-short effects. Santi et al. (1997) trained rats to discriminate light durations by responding to levers with an auditory cue above them on (cued lever) or off (uncued lever). They observed a choose-long bias when the interval between offset of the sample and onset of the auditory cue was increased. It also appears that a choose-long bias can occur even when a spatial response alternative with a “nose poke” response requirement during the DI is used (Berz et al., 1992a,b). In addition, Leblanc and Soffie (1999, 2001) and Soffie et al. (1999) used spatial response alternatives without a “nose poke” response requirement and they reported choose-short effects. The tendency of the rats to time through the DI regardless whether the ITI and the DI was dark or illuminated is an interesting result when considered relative to the effects of gaps and distracters in the peak interval procedure (PI). Buhusi and Meck (2006) recently presented evidence that gaps and distracters disrupt timing in rats in proportion to the salience of these extraneous events. Their findings were consistent with a time-sharing hypothesis which predicts that gaps or distracters cause a reallocation of resources from timing with the result that working memory for time counts begins to decay. The extent to which this occurs according to the time-sharing hypothesis is a function of the salience of the gap and/or the distracter. It was suggested that the salience of gaps and distracters depends on their dissimilarity from the to-be-timed signal rather than on event-ITI similarity (Buhusi et al., 2005, 2006). Gaps or distracters could be similar or dissimilar to the ITI and in both cases they could be dissimilar from the to-be-timed signal, and as a result salient enough to disrupt timing. Although the current procedure was not a PI procedure, the rats were timing the interval from the onset of the magazine light until entry of the levers into the chamber. The DI inserted prior to the entry of levers into the chamber could be viewed as being similar to a gap introduced in a PI procedure, and the houselight illumination during the DI could be viewed as a distracter. In the case of Group Light, a dark DI would be a salient extraneous event because it was dissimilar to both the to-be-timed signal (i.e., magazine light)
and the ITI (i.e., houselight illumination). On the other hand, an illuminated DI would be a much less salient event because it was similar to both the to-be-timed signal (i.e., magazine light) and the ITI (i.e., houselight illumination). According to the timesharing hypothesis, for Group Light, a dark DI should have a produced greater reallocation of memory resources away from timing than an illuminated ITI. Consequently, the choose-long effect should have been smaller when Group Light was tested with a dark ITI than with an illuminated ITI. The time-sharing hypothesis does not make as clear a prediction between the dark and illuminated DI functions in the case of Group Dark. For this group, an illuminated DI was similar to the to-be-timed signal (i.e., magazine light), but dissimilar from ITI (i.e., darkness). On the other hand, the dark DI was dissimilar from the to-be-timed signal (i.e., magazine light), but very similar to the ITI (i.e., darkness). However, DI functions for Group Dark should have been intermediate to those obtained in Group Light. That is, both the dark and illuminated DI test condition in Group Dark should have produced a larger choose-long effect than the dark DI test condition in Group Light, but a smaller choose-long effect than the illuminated DI test condition in Group Light. Neither of these expectations derived from the time-sharing hypothesis were confirmed in the data. While the present study was not designed to directly test the time-sharing hypothesis, it nevertheless provides data which suggests that a difference in the salience of gaps and distracters does not necessarily result in differential reallocation of memory for time under all conditions of timing. In conclusion, while the procedures used in the present study indicate that rats can readily learn to discriminate event durations at a high level of accuracy by responding to nonspatial comparison alternatives consisting of a stationary/uncued and moving/cued lever, additional procedural modifications will be necessary to prevent rats from timing during the delay interval. Acknowledgements This research was supported by Grant OGPOOD6378 from the Natural Sciences and Engineering Research Council of Canada to A.S. The authors thank Kelley Putzu and Kristin Lukashal for their animal care assistance. References Al-Zahrani, S.S.A., Ho, M.Y., Al-Ruwaitea, A.S.A., Bradshaw, C.M., Szabadi, E., 1997. Effect of destruction of the 5-hydroxytryptaminergic pathways on temporal memory: quantitative analysis with a delayed interval bisection task. Psychopharmacology 129, 48–55. Berz, S., Battig, K., Welzl, H., 1992a. The effects of anticholinergic drugs on delayed time discrimination performance in rats. Physiol. Behav. 51, 493–499. Berz, S., Battig, K., Welzl, H., 1992b. Effects of CGS 19755 and dizocilpine (MK 801) on delayed time discrimination performance. Behav. Brain Res. 51, 185–192. Buhusi, C.V., Meck, W.H., 2006. Interval timing with gaps and distracters: evaluation of the ambiguity, switch, and time-sharing hypothesis. J. Exp. Psychol. Anim. Behav. Process. 32, 329–338. Buhusi, C.V., Paskalis, J.G., Cerutti, D.T., 2006. Time-sharing in pigeons: independent effects of gap duration, position and discriminability from the timed signal. Behav. Process. 71, 116–125.
P. Van Rooyen et al. / Behavioural Processes 78 (2008) 1–9 Buhusi, C.V., Perera, D., Meck, W.H., 2005. Memory for timing visual and auditory signals in albino and pigmented rats. J. Exp. Psychol. Anim. Behav. Process. 31, 18–30. Church, R.M., 1980. Short-term memory for time intervals. Learn. Motiv. 11, 208–219. Cohen, J.S., Njegovan, M., 1999. The role of trial tracking in rats’ working memory. Anim. Learn. Behav. 27, 211–220. Dorrance, B.R., Kaiser, D.H., Zentall, T.R., 2000. Event duration discrimination by pigeons: the choose-short effect may result from retention-test novelty. Anim. Learn. Behav. 28, 344–353. Fetterman, J.G., 1995. The psychophysics of remembered duration. Anim. Learn. Behav. 23, 49–62. Fetterman, J.G., Killeen, P.R., Hall, S., 1998. Watching the clock. Behav. Process. 44, 211–224. Gaitan, S.C., Wixted, J.T., 2000. The role of “nothing” in memory for event duration in pigeons. Anim. Learn. Behav. 28, 147–161. Grant, D.S., 1993. Coding processes in pigeons. In: Zentall, T.R. (Ed.), Animal cognition: A tribute to Donald A. Riley. Erlbaum, Hillsdale, NJ, pp. 193–216. Grant, D.S., 2006. Training delays reduce the choose-short effect with keylight duration samples in pigeons. Behav. Process. 72, 139–148. Grant, D.S., 2007. Training delays reduce the choose-short effect with keylight, but not with food, duration samples in pigeons. Behav. Process. 74, 209–216. Grant, D.S., Kelly, R., 1996. The role of minimum wait time and sample discriminability in the coding of event duration in pigeons. Learn. Motiv. 27, 243–259. Grant, D.S., Kelly, R., 1998. The effect of variable-delay training on coding of event duration in pigeons. Learn. Motiv. 29, 49–67. Grant, D.S., Spetch, M.L., 1991. Pigeons’ memory for event duration: differences between choice and successive matching tasks. Learn. Motiv. 22, 180–199. Grant, D.S., Spetch, M.L., 1993. Analogical and nonanalogical coding of samples differing in duration in a choice-matching task in pigeons. J. Exp. Psychol. Anim. Behav. Process. 19, 15–25. Grant, D.S., Spetch, M.L., 1994. The role of asymmetrical coding of duration samples in producing the choose-short effect in pigeons. Learn. Motiv. 25, 413–430. Grant, D.S., Spetch, M.L., Kelly, R., 1997. Pigeons’ coding of event duration in delayed matching-to-sample. In: Bradshaw, C.M., Szabaldi, E. (Eds.), Time and Behaviour: Psychological and Neurobehavioural Analyses. Elsevier, Amsterdam, pp. 217–264. Grant, D.S., Talarico, D., 2004. Processing of empty and filled time intervals in pigeons. Anim. Learn. Behav. 32, 477–490. Harper, D.N., Bizo, L.A., 2000. Mediation of timing accuracy by operant behavior. Behav. Process. 50, 143–154. Keen, R., Church, R.M., 2003. Reexamination of interpretations of memory for temporal intervals. In: Presented at the International Conference on Comparative Cognition, Melbourne, Florida. Kelly, R., Spetch, M.L., 2000. Choice biases in delayed matching-to-sample duration with pigeons: manipulations of ITI and delay illumination. Quart. J. Exp. Psychol. 53B, 309–323. Kraemer, P.J., Mazmanian, D.S., Roberts, W.A., 1985. The choose-short effect in pigeon memory for event duration: Subjective shortening versus coding models. Anim. Learn. Behav. 13, 349–354. Leblanc, P., Soffie, M., 1999. Effects of age on short-term memory for time in rats. Exp. Aging Res. 25, 267–284. Leblanc, P., Soffie, M., 2001. The choose-short effect in rat memory for event duration: the subjective-shortening model. Behav. Process. 56, 31– 40. Meck, W.H., Church, R.M., Gibbon, J., 1985. Temporal integration in duration and number discrimination. J. Exp. Psychol. Anim. Behav. Process. 4, 591–597.
9
Meck, W.H., Church, R.M., Olton, D.S., 1984. Hippocampus, time, and memory. Behav. Neurosci. 98, 3–22. Miki, A., Santi, A., 2001. Pigeons’ timing of an arbitrary and a naturalistic auditory stimulus: tone versus cooing. Behav. Process. 53, 103–111. Roberts, S., 1982. Cross-modal use of an internal clock. J. Exp. Psychol. Anim. Behav. Process. 8, 2–22. Santi, A., Bridson, S., Ducharme, M.J., 1993. Memory codes for temporal and nontemporal stimuli in many-to-one matching by pigeons. Anim. Learn. Behav. 21, 120–130. Santi, A., Coyle, J., Coppa, R., Ross, L., 1998a. The timing of auditory and visual signals by pigeons: differential sensitivity to the intertrial interval duration. Learn. Motiv. 29, 345–365. Santi, A., Ducharme, M.J., Bridson, S., 1992. Differential outcome expectancies and memory for temporal and nontemporal stimuli in pigeons. Learn. Motiv. 23, 156–169. Santi, A., Hornyak, S., Miki, A., 2003. Pigeons’ memory for empty and filled time intervals signaled by light. Learn. Motiv. 34, 282–302. Santi, A., Ross, L., Coppa, R., Coyle, L., 1999. Pigeons’ memory for empty time intervals marked by visual or auditory stimuli. Anim. Learn. Behav. 27, 190–205. Santi, A., Stanford, L., Coyle, J., 1997. Rats’ memory for event duration: differential effects of delaying the discriminative choice cue as opposed to the opportunity to execute the choice response. Behav. Process. 40, 193–199. Santi, A., Stanford, L., Coyle, J., 1998b. Pigeons’ memory for event duration: differences between visual and auditory signals. Anim. Learn. Behav. 26, 163–171. Santi, A., Van Rooyen, P., 2007. Memory for sequences of tone bursts in the rat: reliance on temporal cues and evidence for an instructional ambiguity explanation of the choose-few effect. Learn. Motiv. 38, 295–313. Santi, A., Weise, L., Kuiper, D., 1995a. Memory for event duration in rats. Learn. Motiv. 26, 83–100. Santi, A., Weise, L., Kuiper, D., 1995b. Amphetamine and memory for event duration in rats and pigeons: disruption of attention to temporal samples rather than changes in the speed of the internal clock. Psychobiology 23, 224–232. Sherburne, L.M., Zentall, T.R., Kaiser, D.H., 1998. Timing in pigeons: the choose-short effect may result from pigeons’ “confusion” between delay and intertrial interval. Psychon. Bull. Rev. 5, 516–522. Soffie, M., Hahn, K., Terao, E., Eclancher, F., 1999. Behavioural and glial changes in old rats following environmental enrichment. Behav. Brain Res. 101, 37–49. Spetch, M.L., 1987. Systematic errors in pigeons’ memory for event duration: interaction between training and test delay. Anim. Learn. Behav. 15, 1–5. Spetch, M.L., Rusak, B., 1989. Pigeons’ memory for event duration: intertrial interval and delay effects. Anim. Learn. Behav. 17, 147–156. Spetch, M.L., Rusak, B., 1992. Temporal context effects in pigeons’ memory for event duration. Learn. Motiv. 23, 117–144. Spetch, M.L., Wilkie, D.M., 1983. Subjective shortening: a model of pigeons’ memory for event durations. J. Exp. Psychol. Anim. Behav. Process. 9, 14–30. Talarico, D., Grant, D.S., 2006. Effect of training delays and start and stop markers on the choose-short effect in pigeons. Behav. Process. 71, 98–106. Ward, R.D., Odum, A.L., 2007. Disruption of temporal discrimination and the choose-short effect. Learn. Behav. 35, 60–70. Zentall, T.R., 1997. Animal memory: the role of “instructions”. Learn. Motiv. 28, 280–308. Zentall, T.R., 1999. Support for a theory of memory for event duration must distinguish between test-trial ambiguity and actual memory loss. J. Exp. Anal. Behav. 72, 467–472. Zentall, T.R., 2005. Timing, memory for intervals, and memory for untimed stimuli: the role of instructional ambiguity. Behav. Process. 70, 209–222.
Behavioural Processes 78 (2008) 1–9
Rats’ memory for event duration in delayed matching-to-sample with nonspatial comparison response alternatives Patrick Van Rooyen, Neil McMillan, Angelo Santi ∗ Department of Psychology, Wilfrid Laurier University, Waterloo, Ontario, Canada N2L 3C5 Received 8 June 2007; received in revised form 24 October 2007; accepted 29 November 2007
Abstract Previous research has suggested that using stationary and moving levers as nonspatial response alternatives can significantly enhance the speed of acquiring a temporal discrimination in rats. In Experiment 1, rats were trained to discriminate 2 and 8 s of magazine light illumination by responding to either a stationary lever or a moving lever with a cue light illuminated above it. Rats learned to discriminate event durations at a high level of accuracy after 25 sessions of training. During subsequent delay tests, rats exhibited a strong choose-long bias, indicating that they were timing from the onset of the magazine light until the entry of levers into the chamber. This occurred regardless of whether intertrial intervals and delay intervals were dark or illuminated. On test trials in which the sample was omitted, rats responded as if the short sample had been presented. In Experiment 2, the rats received extensive training with dark and illuminated variable delay intervals (1–4 s). However, they continued to exhibit a tendency to time from the onset of the magazine light until entry of the levers into the chamber. Although the use of stationary/uncued and moving/cued levers as response alternatives enhanced the speed of acquisition of the event duration discrimination in rats, additional procedural modifications will be necessary to prevent rats from timing during the delay interval. © 2007 Elsevier B.V. All rights reserved. Keywords: Rats; Memory for time; Time discrimination; Delayed matching-to-sample
1. Introduction Studies of memory for event duration in pigeons have typically employed a delayed symbolic matching-to-sample task in which a pigeon is trained to peck one comparison stimulus (e.g., green) on trials initiated by a 2-s sample stimulus and to peck a different comparison stimulus (e.g., red) on trials initiated by an 8-s sample stimulus. Following training in this task, the delay interval (DI) between the end of the sample stimulus and the presentation of the comparison stimuli is varied. Numerous studies have shown that at extended delays pigeons respond with high accuracy on trials initiated by the short sample, while accuracy on trials initiated by the long sample drops below 50% correct (Fetterman, 1995; Gaitan and Wixted, 2000; Grant, 1993; Grant and Kelly, 1996, 1998; Grant and Spetch, 1991, 1993, 1994; Grant et al., 1997; Kelly and Spetch, 2000; Kraemer et al., 1985; Santi et al., 1993; Santi et al., 1992; Sherburne et al., 1998; Spetch, 1987; Spetch and Rusak, 1989, 1992; Spetch and
∗
Corresponding author. Tel.: +1 519 884 0710x3087; fax: +1 519 746 7605. E-mail address: [email protected] (A. Santi).
0376-6357/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.beproc.2007.11.012
Wilkie, 1983). Several different explanations have been provided for this choose-short effect in pigeons (Dorrance et al., 2000; Gaitan and Wixted, 2000; Kraemer et al., 1985; Spetch and Rusak, 1992; Spetch and Wilkie, 1983; Ward and Odum, 2007). The study of memory for event duration in rats has not produced consistent evidence of a choose-short effect. While a choose-short effect has been found in some studies (Church, 1980, Experiment 3A; Al-Zahrani et al., 1997; Harper and Bizo, 2000; Leblanc and Soffie, 1999, 2001; Santi et al., 1995a,b; Soffie et al., 1999), in other studies a choose-long bias has been obtained (Berz et al., 1992a,b; Church, 1980, Experiment 3B; Keen and Church, 2003; Meck et al., 1984; Roberts, 1982; Santi et al., 1995a, 1997). In part, this may reflect the lack of similarity in the procedures used in these experiments. This procedural variability is reflected in the temporal signals used, the response alternatives used, and the procedures for delay testing. The temporal signals have included an increase in illumination (Al-Zahrani et al., 1997; Berz et al., 1992a,b; Harper and Bizo, 2000; Roberts, 1982; Santi et al., 1995a,b, 1997), a decrease in illumination (Church, 1980), presentation of an auditory signal (Meck et al., 1984; Roberts, 1982; Santi
2
P. Van Rooyen et al. / Behavioural Processes 78 (2008) 1–9
et al., 1995a), and presentation of a bimodal signal (Leblanc and Soffie, 1999, 2001; Soffie et al., 1999). The response alternatives used include spatial locations (Church, 1980; Keen and Church, 2003; Leblanc and Soffie, 1999, 2001; Meck et al., 1984; Roberts, 1982; Soffie et al., 1999), spatial locations with a nose poke/lever press requirement during the delay interval (Al-Zahrani et al., 1997; Berz et al., 1992a,b; Harper and Bizo, 2000), nonspatial–visual response cues (Santi et al., 1995a,b), and nonspatial–auditory response cues (Santi et al., 1997). At this time, no single procedural variable or combination of variables appears to be consistently related to the occurrence of a choose-short, or a choose-long bias during delay testing. One of the difficulties associated with studying memory for event duration in rats is that rats have a difficult time acquiring event duration discriminations when nonspatial response alternatives are used. Nonspatial response alternatives are methodologically better than spatial response alternatives because the use of nonspatial response alternatives makes it more difficult for rats to use spatial positioning during the DI to mediate responding to the correct lever location. When either nonspatial visually cued response alternatives (Santi et al., 1995a), or nonspatial auditory-cued response alternatives (Santi et al., 1997) have been used, rats acquire the discriminations very slowly (between 9500 and 15,700 trials), and their level of accuracy is modest (between 65% and 75% correct). Furthermore, the response bias exhibited during delay testing has not been consistent in these studies. In the tone signal–visual choice cue group of Santi et al. (1995a) and the light signal-tone choice cue group of Santi et al. (1997), rats exhibited a choose-long bias. The rats appeared to use the onset of the choice cue to stop timing rather than offset of the to-be-timed signal. On the other hand, when the signal-cue modalities were the same, such as in the visual signal–visual choice group of the Santi et al. (1995a,b) studies, rats exhibited a choose-short effect. The rats appeared to use the offset of the signal itself to stop the clock. Santi et al. (1995a,b) are the only studies reporting a choose-short effect in rats when nonspatial response alternatives are used. Research by Meck et al. (1985) suggested that a temporal discrimination between 2- and 4 s of white-noise could be rapidly acquired by rats when stationary and moving levers are used as nonspatial response alternatives. Half of their subjects were trained to press a stationary lever following the 2-s signal and a moving lever following the 4-s signal while the remaining rats were trained with the opposite contingency. Meck et al. (1985) did not report the number of trials per session or the mean accuracy for each signal during the acquisition phase of their study. However, they reported that only 10 sessions of training were required before the start of psychophysical testing. Recently, Santi and Van Rooyen (2007) made use of the Meck et al. (1985) procedure in training rats to discriminate sample stimuli that consisted of sequences of tone bursts. The rats were trained to discriminate two tone bursts in 4 s (2b/4 s) or eight tone bursts in 4 s (8b/4 s) by responding to either a stationary lever or a moving lever with a cue light illuminated above it. The rats were able to discriminate the sequences with accuracy above 75% correct within approximately 4800 training trials. This success in using stationary and moving levers as nonspatial response alternatives
in the discrimination of sequences of tone bursts led to the current study in which rats were required to discriminate different durations of magazine light. 2. Experiment 1 The main purpose of the current experiment was to replicate the findings of Meck et al. (1985) by demonstrating that using stationary and moving levers as nonspatial response alternatives would produce faster acquisition and a higher level of temporal discrimination accuracy than that reported in previous research using different nonspatial response alternatives (Santi et al., 1995a, 1997). However, Meck et al. (1985) did not conduct delay testing in their study. Consequently, a second purpose of the current experiment was to conduct delay tests and determine whether similarity in the ambient illumination conditions of the intertrial interval (ITI) and DI would affect retention functions for event duration in rats in the same way that it affects retention functions in pigeons. Rats were trained to discriminate 2 versus 8 s of magazine light illumination by responding to either a stationary lever or a moving lever with a cue light illuminated above it. The ambient illumination condition during the ITI was manipulated between groups. One group of the rats was trained with an illuminated ITI (Group Light) and the other group was trained with a dark ITI (Group Dark). According to the instructional ambiguity hypothesis (Cohen and Njegovan, 1999; Dorrance et al., 2000; Sherburne et al., 1998; Zentall, 1997, 1999, 2005), if the ambient stimulus conditions during the ITI and delay are the same, a novel DI may be confused with the ITI. This was proposed as an alternative explanation for the choose-short response bias which occurs when memory for duration samples is tested in pigeons. Because pigeons confuse novel delay intervals with the ITI, the presentation of comparisons at the end of the delay may be perceived as a trial without a duration sample, and nosample would likely be judged as more like a short sample than a long sample. Consequently, pigeons increasingly respond to the comparison that was associated with the short sample during training. As a result, a choose-short effect occurs, but primarily when the stimulus conditions during the ITI and delay are the same (i.e., confusable). If a similar process occurs for rats, they should increasingly respond to the response alternative associated with the short sample at extended delays when the stimulus conditions during the ITI and delay are the same (i.e., light ITI – light DI, dark ITI – dark DI), but not when they are different (i.e., light ITI – dark DI, dark ITI – light DI). The current experiment will be the first to examine whether the stimulus conditions during the ITI and delay affect retention functions for event duration in the rat. 2.1. Method 2.1.1. Subjects Eight experimentally na¨ıve Long-Evans hooded male rats were used as subjects. The 58–63-days old rats were obtained from Charles River Canada (St. Constant, PQ). They were individually housed in clear Plexiglas cages in a temperature- and humidity-controlled holding room under an alternating 12 h
P. Van Rooyen et al. / Behavioural Processes 78 (2008) 1–9
light–dark cycle. Experimental sessions were given during the light phase. The rats had free access to water but their diet was restricted to maintain approximately 85% of their free-feeding bodyweight. 2.1.2. Apparatus Four Coulbourn modular operant test cages (Model #E1010), housed individually in isolation cubicles (Model #E10-20) and located in the same room were used. The cubicles were equipped with a ventilation fan and baffled air intake exhaust system. Each test cage was equipped with a 45-mg pellet dispenser (Model #E14-06), which was mounted in the center of the front wall of each test cage just above the steel-grid flooring. Two retractable levers (Model #E23-07) were mounted 2.5 cm above the floor on either side of the pellet feeder and an opaque cue-light was positioned above each lever. A single Sonalert tone module (Model # E12-02) was mounted 2.5 cm above each cue light, but they were not used in the current experiments. A houselight was also installed 6.5 cm above the pellet feeder and positioned so that the light was facing upward and reflected off of the ceiling of the test cage (Coulbourn Model #E11-01 with bulb #SL1819X). All of the experimental events and responses were controlled by a microcomputer located in the same room as the four test cages. 2.1.3. Procedure Each rat received several sessions of combined magazine and lever training. The rats were placed in the operant chamber with both the right and left levers retracted. Each trial began with the entry of the left or right lever into the chamber. The lever remained extended until it was pressed or 60 s had elapsed, whichever occurred first. Either event resulted in delivery of a 45-mg food pellet and retraction of the lever. Pellet delivery produced an audible “click” and the light in the magazine was illuminated for 0.5 s. The houselight was illuminated during all of the preliminary training sessions. The sessions were terminated when the rat pressed the lever 60 times or 60 m had passed. After the rats had acquired the lever press to a stationary lever, they were trained to respond to both a stationary/uncued lever and a moving/cued lever randomly presented on both the right and the left side of the test chamber. When the stationary/uncued lever was presented, it entered the chamber without the opaque cue light above it being illuminated. It remained extended until a lever press occurred or 6 s elapsed without a lever press, whichever occurred first. If no response occurred the stationary/uncued lever was retracted and it was presented again after 5 s had elapsed. When the moving/cued lever was presented, it entered the chamber with the cue light above it on and it was inserted and retracted at 0.6 s intervals. These 1.2-s cycles continued for a maximum of five repetitions (i.e., 6 s). If no response occurred within 6 s, the moving/cued lever was retracted and the cue light above it was terminated for 5 s, before another set of 5 cycles of lever insertion/retraction was presented. Responses to either the stationary/uncued or moving/cued lever within the 6 s presentation interval resulted in immediate retraction of the lever, termination of the cue light if it had been illuminated, and delivery of a food pel-
3
let. Preliminary training continued until all of the rats were reliably pressing both the stationary/uncued and moving/cued lever. After preliminary training, the rats were trained to discriminate between 2 and 8 s of magazine light illumination. Trials were separated by a randomly selected ITI of 8, 16, 32, or 64 s. Four rats were trained with the houselight on during the ITI (Group Light) and the remaining four rats were trained with the houselight off during the ITI (Group Dark). Each trial commenced with a sample presentation in which the light in the food magazine was illuminated for 2 or 8 s. Following presentation of the sample, the rats were presented with either the stationary/uncued lever on the left and the moving/cued lever on the right, or the moving/cued lever on the left and the stationary/uncued lever on the right. The moving lever was retracted and inserted at 0.6-s intervals and these repeating 1.2-s cycles continued until the first lever press after a 2.4-s delay. The 2.4-s delay between the insertion of the levers and the recording of the choice response was used so that the animal had enough time to determine which lever was moving before making its response (Meck et al., 1985). This will be referred to as a 0-s delay condition. If a rat did not press one of the levers between 2.4 and 6 s after lever insertion, the levers were retracted, the cue light turned off, and the same sample duration and comparison lever configuration was presented again after a 5-s delay. Four rats were trained to press the stationary/uncued lever following the short sample, and the moving/cued lever following the long sample (i.e., two from Group Light and two from Group Dark). The remaining four rats were trained similarly with the opposite contingencies. For all rats, a response to either lever retracted both levers and turned off the cue light. A correct response resulted in delivery of a food pellet, while an incorrect response resulted in a 5-s delay period and re-presentation of the same sample duration and configuration of comparison levers. Only the choice response on the initial (noncorrection) trial was used to calculate response accuracy. After three consecutive incorrect responses had been made on any given trial, the ITI was presented and the next trial was initiated. Each session of training consisted of 120 trials. Each combination of sample type and comparison test condition occurred once in each block of four trials. The order of trial presentation was randomized individually for each rat in each session. After 25 sessions of discrimination training, the rats received 15 sessions of delay testing. During these sessions, the delay between end of the sample and the entry of the comparison levers into the chamber was 0, 1, 2, 4 and 8 s. Within each session, 24 trials for each sample (short and long) occurred at the 0s baseline delay, and four trials for each sample occurred at each of the other delays (1, 2, 4, and 8 s). For one half of the trials involving the 1-, 2-, 4-, 8-s delays, the entire delay was spent in darkness. On the other half of the trials, the houselight was illuminated for the entire delay. During delay testing, the correction procedure remained in effect only for 0-s delay trials. Errors on all other delay trials were followed by an illuminated (Group Light) or a dark (Group Dark) ITI and the presentation of the next sample duration. All other parameters remained the same as those described previously.
4
P. Van Rooyen et al. / Behavioural Processes 78 (2008) 1–9
Following delay testing, the rats received three sessions of baseline training, which were identical to those that preceded delay testing. Following these sessions, the rats received five sessions of sample-omission testing. During sample-omission test sessions, 96 of the 120 trials were baseline trials identical to those that preceded delay testing. On the remaining 24 trials, the sample stimulus was not presented. One no-sample trial occurred randomly within each block of five trials. On no-sample trials, in both Groups, the ITI was followed by a 4-s interval of darkness prior to entry of the levers into the chamber. On these test trials, a response to either the stationary or the moving lever was randomly followed by delivery of a food pellet with a probability of .50. All of the other parameters during testing were the same as those described for baseline training. In all the statistical analyses reported in this article, the rejection region was p < .05. 2.2. Results and discussion Fig. 1 presents the acquisition data for Group Light in the top panel and those for Group Dark in the bottom panel. Both groups were able to discriminate the event durations at a high level of accuracy within 25 sessions of training. An ANOVA was conducted on these data with group as a between-subjects factor and sample type (short and long) and blocks of five sessions as within-subjects factors. There was a significant main effect
Fig. 2. The mean percentage of correct responding on short- and long-sample trials for Group Light (top panel) and for Group Dark (bottom panel) during delay testing as a function of delay interval length and delay interval illumination. Error bars represent the standard error of the mean.
Fig. 1. The mean percentage of correct responding on short- and long-sample trials during acquisition in Experiment 1 for the group trained with the houselight on during the ITI (Group Light) and the group trained with the houselight off during the ITI (Group Dark). Error bars represent the standard error of the mean.
of blocks, F(4, 24) = 77.32, as well as a group × blocks interaction, F(4, 24) = 4.32. While accuracy increased across blocks for both Group Light and Group Dark, during Blocks 3 and 4 accuracy was significantly greater for Group Dark than for Group Light, F(1, 6) = 5.89 and 12.45 respectively. However, in the final block of training, there was no significant difference in accuracy between the two groups F < 1. No other main effects or interaction effects were significant. The mean percentage of correct responding during delay testing sessions is shown in Fig. 2. The data for Group Light is presented in the top panel and the data for Group Dark in the bottom panel. One half of all 0-s delays were randomly designated as dark delays, and the other half were randomly designated as illuminated delays. At the 0-s delay, accuracy was slightly better for short samples than for long samples in both groups. However, at extended delays, for both groups, accuracy was much higher for long samples than for short samples regardless of the illumination condition during the delay. An ANOVA was conducted on the delay test data with group as a betweensubjects factor, and sample type (short and long), DI illumination (light and dark), and DI length (0, 1, 2, 4, 8 s) as within-subjects factors. The analysis revealed a significant main effect of sample type, F(1, 6) = 19.20, and DI, F(4, 24) = 31.94. In addition, there was a significant sample type × DI length interaction, F(4, 24) = 109.95. At the 0-s delay, accuracy was significantly
P. Van Rooyen et al. / Behavioural Processes 78 (2008) 1–9
greater on the short-sample trials than on the long-sample trials, F(1, 6) = 19.12. There was no difference in accuracy between short- and long-sample trials at the 1-s delay. However, at each of the remaining delays, accuracy was significantly greater for long-sample than for short-sample trials, F(1, 6) = 9.14, 64.62, and 54.31 respectively. There was no significant main effect for group or significant interactions involving the group factor. The mean percentage of long responses during sampleomission test sessions was analyzed with group as a between-subjects factor and sample type (short, long and nosample) as a within-factor. Only the main effect of sample type was statistically significant, F(2, 12) = 65.22. The mean percentage of long responses was approximately 85% following the long samples, 19% following the short sample, and 25% on the no-sample trials. This result indicates that in the absence of a sample, rats respond as if the sample was short. A similar finding has been frequently reported in pigeons (see Spetch and Wilkie, 1983). Rats acquired the delayed symbolic matching-to-sample task more quickly and to a higher level of accuracy with nonspatialstationary/moving response alternatives than in previous studies which used either nonspatial visually-cued response alternatives (Santi et al., 1995a), or nonspatial auditory-cued response alternatives (Santi et al., 1997). In the previous Santi et al. studies (1995a, 1997) acquisition occurred very slowly (between 9500 and 15,700 trials) and accuracy levels were modest (65–75%). In the current experiment, rats learned to discriminate event durations at a high level of accuracy (above 85%) after 25 sessions of training (approximately 3000 trials). This result is consistent with the study by Meck et al. (1985) in which psychophysical testing was undertaken following only 10 sessions of training. As noted in Section 1, Meck et al. (1985) did not report the number of trials per session or the mean level of accuracy at the end of training. Nevertheless, the use of stationary versus moving levers appears to be an effective procedure to establish temporal discriminations with nonspatial response alternatives. In the current study, a cue light was illuminated above the moving lever but not above the stationary lever. Meck et al. (1985) did not employ a cue light in their study. It is unclear whether this procedural variation in the current study facilitated acquisition relative to what would have occurred without the cue light. In addition, Meck et al. (1985) did not report whether there was a preference for the stationary or moving lever in their study. In the current study, the overall mean percent choice of the stationary lever versus moving lever during acquisition was 47.8% versus 51.3%. The use of a correction procedure in the current study presumably discouraged the development of a lever preference. Delay tests following acquisition revealed an increasing tendency to respond to the lever correct for the long sample as the DI was extended. Rats apparently learned to time from the onset of magazine light to the insertion of levers into the chamber. When the insertion of the levers was delayed, rats continued to time and as a result exhibited a choose-long bias at extended delays. This tendency to time through the delay was not affected by the ambient illumination conditions during
5
the ITI and DI. Although the choose-long bias obtained with rats is different from the choose-short bias normally obtained in pigeons, both species exhibit a bias to respond to the comparison alternative associated with the short sample on sample-omission trials. 3. Experiment 2 Several studies have examined memory for event duration in pigeons after training with a constant nonzero delay (Grant and Kelly, 1998; Kelly and Spetch, 2000; Spetch, 1987; Spetch and Rusak, 1989, 1992). In all of these studies, pigeons exhibited a choose-short bias at extended delays. However, several recent studies have indicated that following training with variable delays, pigeons do not show a significant choose-short effect (Dorrance et al., 2000; Grant and Talarico, 2004; Talarico and Grant, 2006). Grant (2006) provided the first within-experiment comparison of the effects of nonzero training delays and variability in those training delays on the magnitude of the choose-short effect in pigeons. His results indicated that when different keylight durations were used as samples, training with either a constant or a variable nonzero delay reduced the size of the choose-short effect. In addition, the amount of reduction was greater at the longer training delays. As noted by Grant (2006) the main procedural variable that most reliably differentiates between those studies that demonstrate a sensitivity of the choose-short effect to nonzero delay training and those that do not is the nature of the event that provides the duration information. Studies that demonstrate the sensitivity have used different durations of keylight (Dorrance et al., 2000; Grant, 2006; Grant and Talarico, 2004; Talarico and Grant, 2006), while those that do not have used different durations of houselight or access to food (Kelly and Spetch, 2000; Spetch, 1987; Spetch and Rusak, 1989, 1992). Consistent with this analysis, Grant (2007) employed a within-experiment comparison to demonstrate that training delays reduce the choose-short effect with keylight, but not with food, event duration samples in pigeons. Thus far, no study has examined memory for event durations in rats following training with nonzero delays. Since rats exhibited a choose-long bias in Experiment 1 following 0-s delay training, it would be worthwhile to determine if they continue to do so following training with variable delays of 1–4 s duration which were equally often either dark or illuminated. In Experiment 1, the rats may have learned to time from onset of the magazine light until presentation of the levers, because the offset of the magazine light was not as salient a stimulus as the entry of the levers into the chamber. Training with dark and illuminated variable nonzero delays may make the onset of the delay more salient and result in the rats being more likely to stop timing at the start of the delay. On the other hand, it is possible that rats may continue timing during the delay because with samples of 2 and 8 s and training delays of 1, 2 and 4 s, the total time from onset of the magazine light to the entry of the levers into the chamber would still provide a reliable cue for correct responding (i.e., combined sample-delay durations of 3, 4, or 6 s for “short” and 9, 10, or 12 s for “long”).
6
P. Van Rooyen et al. / Behavioural Processes 78 (2008) 1–9
3.1. Method 3.1.1. Subjects and apparatus The subjects and apparatus used in Experiment 1 were also used in Experiment 2. 3.1.2. Procedure The rats received three sessions of baseline training which were identical to those that preceded testing in Experiment 1. This was followed by 60 sessions of training with variable delays of 1, 2, and 4 s. The delays were equally often either dark or illuminated by the houselight. Each combination of DI illumination and DI length condition occurred equally often for each of the four combinations of sample type and configuration of comparison levers. A session consisted of 120 trials. A correction procedure was used on all trials. Consequently, a correct response resulted in delivery of a food pellet, while an incorrect response resulted in a 5-s delay period and re-presentation of the same sample, DI condition, and configuration of comparison levers. Only the choice response on the initial (noncorrection) trial was used to calculate response accuracy. After three consecutive incorrect responses had been made on any given trial, the ITI was presented and the next trial was initiated. All other aspects of discrimination training remained the same as previously described. Following this training, rats received 10 sessions of testing at extended delays. During these sessions, the delay between end of the sample and the entry of the comparison levers into the chamber was 1, 2, 4, 8, and 16 s. For one half of the trials, the DI was dark, while for the remaining trials it was illuminated by the houselight. Each combination of DI illumination and DI length condition occurred equally often for each of the four combinations of sample type and configuration of comparison levers. During delay testing, the correction procedure remained in effect only for 1-, 2- and 4-s delay trials. Errors on the 8- and 16-s delay trials were followed by an illuminated (Group Light) or a dark (Group Dark) ITI and the presentation of the next trial. All other parameters remained the same as those described previously. 3.2. Results and discussion Overall accuracy remained the same between the first 5 sessions of training with variable delays of 1, 2, and 4 s and the last 5 sessions of training. Mean accuracy was 75.9% over the first 5 sessions of training and it was 75.8% over sessions 55–60. Fig. 3 shows accuracy averaged over the last 5 sessions of training as a function of sample duration and DI condition. The data for Group Light (n = 4) is presented in the top panel and the data for Group Dark (n = 4) in the bottom panel. At the 1-s delay, Group Light exhibited higher accuracy for short samples than for long samples regardless of whether the delay was dark or illuminated. As the delay was extended, accuracy decreased for the short sample, but it increased for the long sample. Group Dark did not exhibit an overall difference in accuracy for short and long samples at the 1-s delay, however similarly to Group Light, they exhibited increasing accuracy for long samples and decreasing accuracy for short samples as the DI was lengthened.
Fig. 3. The mean percentage of correct responding on short- and long-sample trials for Group Light (top panel) and for Group Dark (bottom panel) averaged over the last five sessions of variable delay training as a function of delay interval length and delay interval illumination. Error bars represent the standard error of the mean.
At the longest delay of 4 s, both groups exhibited higher accuracy for long samples than for short samples. These data suggest that both groups continued to time during the DI regardless of whether it was dark or illuminated. An ANOVA was conducted on these data with group as a between-subjects factor, and sample type (short and long), DI illumination (light and dark), and DI length as within-subjects factors. The analysis revealed a significant main effect of DI length, F(2, 12) = 12.20. In addition, there was a significant sample type × DI length interaction, F(2, 12) = 36.30. At the 1-s delay, performance was significantly better on short-sample trials than on the long-sample trials, F(1, 6) = 12.26. There was no significant difference between short and long samples at the 2-s delay. However, at the 4-s delay, performance was significantly better on long-sample trials than on short-sample trials, F(1, 6) = 18.87. In addition, there was a significant sample type × group interaction, F(1, 6) = 6.44, and a DI illumination × sample type × group interaction, F(1, 6) = 16.22. The three-way interaction was the result of a significant DI illumination × sample type interaction for Group Dark, F(1, 6) = 15.88, but not for Group Light. For Group Dark, when the sample was short, accuracy was significantly higher if the DI was dark than if it was illuminated, F(1, 6) = 8.91. However, when the sample was long, the presence or absence of illumination during the DI did not significantly affect accuracy.
P. Van Rooyen et al. / Behavioural Processes 78 (2008) 1–9
Fig. 4. The mean percentage of correct responding on short- and long-sample trials for Group Light (top panel) and for Group Dark (bottom panel) during the delay tests which followed variable delay training. The data are presented as a function of delay interval length and delay interval illumination. Error bars represent the standard error of the mean.
Despite having received training with variable delays of 1, 2 and 4 s for 60 sessions, the rats continued to exhibit a tendency to time during the DI. Additional evidence of this was obtained during testing with extended delays as can been seen in Fig. 4. At the extended delays of 8 and 16 s, both groups exhibited above-chance accuracy following long samples, but below-chance accuracy following short samples. An ANOVA was conducted with group as a between-subjects factor, and sample type (short and long), DI illumination (light and dark), and DI length as within-subjects factors. The analysis revealed a significant main effect of sample type, F(1, 6) = 23.69, and DI length, F(4, 24) = 27.94. In addition, there was a significant sample type × DI length interaction, F(4, 24) = 45.42. At the 1-s delay, accuracy was significantly greater on the short-sample trials than on the long-sample trials, F(1, 6) = 9.14. There was no significant overall difference in accuracy between short- and long-samples at the 2- or 4-s delay. However, at the 8- and 16-s delay, accuracy was significantly greater on long-sample trials than on short-sample trials, F(1, 6) = 63.82 and 46.84 respectively. 4. General discussion Rats acquired the delayed symbolic matching-to-sample task more quickly and to a higher level of accuracy with nonspatial-
7
stationary/moving response alternatives than in previous studies which used either nonspatial visually-cued response alternatives (Santi et al., 1995a), or nonspatial auditory-cued response alternatives (Santi et al., 1997). This result is consistent with the study by Meck et al. (1985) in which psychophysical testing was undertaken following only 10 sessions of training. In the current study, a cue light was illuminated above the moving lever but not above the stationary lever. Meck et al. (1985) did not employ a cue light in their study and they did not mention whether their rats exhibited a preference for either the stationary or moving lever. In the current study, there was no evidence of a stationary or moving lever preference, but it is likely that any preference would have been eliminated by the use of a correction procedure. Group Dark initially acquired the discrimination somewhat faster than Group Light. This was probably due to the greater salience of the onset of the magazine light sample following a dark ITI than an illuminated ITI. However, during the last block of training, accuracy was equivalent for Group Dark and Group Light. As the DI was extended, accuracy remained well above chance on trials initiated by the long sample, but it declined to well below chance on trials initiated by the short sample. This bias to respond to the response alternative correct for the long sample was not affected by either the ITI or the DI illumination condition. It appears that the rats learned to time from onset of the signal duration until presentation of the comparisons. This may have occurred because the offset of the signal (i.e., magazine light) was not as salient a stimulus as the entry of the levers into the chamber during the choice phase of a trial. In Experiment 2, rats were trained with both dark and illuminated variable nonzero delays in an attempt to make the onset of the DI more salient and to discourage them from continuing to time during the DI. Despite the use of a correction procedure on all trials throughout the 60 sessions of variable delay training, rats continued to time during the delay. With samples of 2 and 8 s and training delays of 1, 2 and 4 s, the total time from onset of the magazine light to the entry of the levers into the chamber still provided a reliable cue for correct responding and reinforcement (3, 4, or 6 s for “short” and 9, 10, or 12 s for “long”). Subsequent testing at delays greater than 4 s resulted in a strong bias to respond to the response alternative correct for long-sample trials. Experiment 2 might have been more successful if the correct response contingencies had been reversed from those used in Experiment 1, or if na¨ıve rats had been used. Another possibility would be to employ shorter signal durations and/or longer delay intervals in training to produce more of an overlap between the onset of the sample and the entry of the response levers on short- and long-sample trials. Although many studies of temporal memory in pigeons have reliably reported a choose-short effect, there are a few studies in which a choose-long effect has been reported. For example, Dorrance et al. (2000) observed a choose-long effect in a group of pigeons trained with a bright houselight during the ITI, a dimmer keylight as the duration sample, and a dark DI. Pigeons timed from the offset of the houselight until onset of the comparisons in this study. Choose-long effects in pigeons have also been reported when auditory duration samples are used (Miki and
8
P. Van Rooyen et al. / Behavioural Processes 78 (2008) 1–9
Santi, 2001; Santi et al., 1998a,b), or empty intervals bound by brief light or tone markers are used (Santi et al., 1999, 2003). Several previous studies in rats have reported a choose-long effect when the DI is increased (Berz et al., 1992a,b; Church, 1980; Meck et al., 1984; Santi et al., 1997; Roberts, 1982). In all of these studies, except Santi et al. (1997), the discriminative response was based on the spatial location of the levers (left/right). Fetterman et al. (1998) presented evidence that temporal judgements by rats performing a spatially-differentiated response were mediated by collateral behaviors. Behavioral mediation could account for the development of a choose-long bias as the delay is lengthened in a timing task with spatial response alternatives. Studies which minimize the development of a response strategy during the delay by requiring a response (e.g., nose poke or lever press) at a central location (Al-Zahrani et al., 1997; Harper and Bizo, 2000) or by using nonspatial comparison response alternatives (Santi et al., 1995a) have been somewhat more successful in obtaining a choose-short effect in rats. However, as the present findings indicate, the use of nonspatial comparison response alternatives in rats does not always result in choose-short effects. Santi et al. (1997) trained rats to discriminate light durations by responding to levers with an auditory cue above them on (cued lever) or off (uncued lever). They observed a choose-long bias when the interval between offset of the sample and onset of the auditory cue was increased. It also appears that a choose-long bias can occur even when a spatial response alternative with a “nose poke” response requirement during the DI is used (Berz et al., 1992a,b). In addition, Leblanc and Soffie (1999, 2001) and Soffie et al. (1999) used spatial response alternatives without a “nose poke” response requirement and they reported choose-short effects. The tendency of the rats to time through the DI regardless whether the ITI and the DI was dark or illuminated is an interesting result when considered relative to the effects of gaps and distracters in the peak interval procedure (PI). Buhusi and Meck (2006) recently presented evidence that gaps and distracters disrupt timing in rats in proportion to the salience of these extraneous events. Their findings were consistent with a time-sharing hypothesis which predicts that gaps or distracters cause a reallocation of resources from timing with the result that working memory for time counts begins to decay. The extent to which this occurs according to the time-sharing hypothesis is a function of the salience of the gap and/or the distracter. It was suggested that the salience of gaps and distracters depends on their dissimilarity from the to-be-timed signal rather than on event-ITI similarity (Buhusi et al., 2005, 2006). Gaps or distracters could be similar or dissimilar to the ITI and in both cases they could be dissimilar from the to-be-timed signal, and as a result salient enough to disrupt timing. Although the current procedure was not a PI procedure, the rats were timing the interval from the onset of the magazine light until entry of the levers into the chamber. The DI inserted prior to the entry of levers into the chamber could be viewed as being similar to a gap introduced in a PI procedure, and the houselight illumination during the DI could be viewed as a distracter. In the case of Group Light, a dark DI would be a salient extraneous event because it was dissimilar to both the to-be-timed signal (i.e., magazine light)
and the ITI (i.e., houselight illumination). On the other hand, an illuminated DI would be a much less salient event because it was similar to both the to-be-timed signal (i.e., magazine light) and the ITI (i.e., houselight illumination). According to the timesharing hypothesis, for Group Light, a dark DI should have a produced greater reallocation of memory resources away from timing than an illuminated ITI. Consequently, the choose-long effect should have been smaller when Group Light was tested with a dark ITI than with an illuminated ITI. The time-sharing hypothesis does not make as clear a prediction between the dark and illuminated DI functions in the case of Group Dark. For this group, an illuminated DI was similar to the to-be-timed signal (i.e., magazine light), but dissimilar from ITI (i.e., darkness). On the other hand, the dark DI was dissimilar from the to-be-timed signal (i.e., magazine light), but very similar to the ITI (i.e., darkness). However, DI functions for Group Dark should have been intermediate to those obtained in Group Light. That is, both the dark and illuminated DI test condition in Group Dark should have produced a larger choose-long effect than the dark DI test condition in Group Light, but a smaller choose-long effect than the illuminated DI test condition in Group Light. Neither of these expectations derived from the time-sharing hypothesis were confirmed in the data. While the present study was not designed to directly test the time-sharing hypothesis, it nevertheless provides data which suggests that a difference in the salience of gaps and distracters does not necessarily result in differential reallocation of memory for time under all conditions of timing. In conclusion, while the procedures used in the present study indicate that rats can readily learn to discriminate event durations at a high level of accuracy by responding to nonspatial comparison alternatives consisting of a stationary/uncued and moving/cued lever, additional procedural modifications will be necessary to prevent rats from timing during the delay interval. Acknowledgements This research was supported by Grant OGPOOD6378 from the Natural Sciences and Engineering Research Council of Canada to A.S. The authors thank Kelley Putzu and Kristin Lukashal for their animal care assistance. References Al-Zahrani, S.S.A., Ho, M.Y., Al-Ruwaitea, A.S.A., Bradshaw, C.M., Szabadi, E., 1997. Effect of destruction of the 5-hydroxytryptaminergic pathways on temporal memory: quantitative analysis with a delayed interval bisection task. Psychopharmacology 129, 48–55. Berz, S., Battig, K., Welzl, H., 1992a. The effects of anticholinergic drugs on delayed time discrimination performance in rats. Physiol. Behav. 51, 493–499. Berz, S., Battig, K., Welzl, H., 1992b. Effects of CGS 19755 and dizocilpine (MK 801) on delayed time discrimination performance. Behav. Brain Res. 51, 185–192. Buhusi, C.V., Meck, W.H., 2006. Interval timing with gaps and distracters: evaluation of the ambiguity, switch, and time-sharing hypothesis. J. Exp. Psychol. Anim. Behav. Process. 32, 329–338. Buhusi, C.V., Paskalis, J.G., Cerutti, D.T., 2006. Time-sharing in pigeons: independent effects of gap duration, position and discriminability from the timed signal. Behav. Process. 71, 116–125.
P. Van Rooyen et al. / Behavioural Processes 78 (2008) 1–9 Buhusi, C.V., Perera, D., Meck, W.H., 2005. Memory for timing visual and auditory signals in albino and pigmented rats. J. Exp. Psychol. Anim. Behav. Process. 31, 18–30. Church, R.M., 1980. Short-term memory for time intervals. Learn. Motiv. 11, 208–219. Cohen, J.S., Njegovan, M., 1999. The role of trial tracking in rats’ working memory. Anim. Learn. Behav. 27, 211–220. Dorrance, B.R., Kaiser, D.H., Zentall, T.R., 2000. Event duration discrimination by pigeons: the choose-short effect may result from retention-test novelty. Anim. Learn. Behav. 28, 344–353. Fetterman, J.G., 1995. The psychophysics of remembered duration. Anim. Learn. Behav. 23, 49–62. Fetterman, J.G., Killeen, P.R., Hall, S., 1998. Watching the clock. Behav. Process. 44, 211–224. Gaitan, S.C., Wixted, J.T., 2000. The role of “nothing” in memory for event duration in pigeons. Anim. Learn. Behav. 28, 147–161. Grant, D.S., 1993. Coding processes in pigeons. In: Zentall, T.R. (Ed.), Animal cognition: A tribute to Donald A. Riley. Erlbaum, Hillsdale, NJ, pp. 193–216. Grant, D.S., 2006. Training delays reduce the choose-short effect with keylight duration samples in pigeons. Behav. Process. 72, 139–148. Grant, D.S., 2007. Training delays reduce the choose-short effect with keylight, but not with food, duration samples in pigeons. Behav. Process. 74, 209–216. Grant, D.S., Kelly, R., 1996. The role of minimum wait time and sample discriminability in the coding of event duration in pigeons. Learn. Motiv. 27, 243–259. Grant, D.S., Kelly, R., 1998. The effect of variable-delay training on coding of event duration in pigeons. Learn. Motiv. 29, 49–67. Grant, D.S., Spetch, M.L., 1991. Pigeons’ memory for event duration: differences between choice and successive matching tasks. Learn. Motiv. 22, 180–199. Grant, D.S., Spetch, M.L., 1993. Analogical and nonanalogical coding of samples differing in duration in a choice-matching task in pigeons. J. Exp. Psychol. Anim. Behav. Process. 19, 15–25. Grant, D.S., Spetch, M.L., 1994. The role of asymmetrical coding of duration samples in producing the choose-short effect in pigeons. Learn. Motiv. 25, 413–430. Grant, D.S., Spetch, M.L., Kelly, R., 1997. Pigeons’ coding of event duration in delayed matching-to-sample. In: Bradshaw, C.M., Szabaldi, E. (Eds.), Time and Behaviour: Psychological and Neurobehavioural Analyses. Elsevier, Amsterdam, pp. 217–264. Grant, D.S., Talarico, D., 2004. Processing of empty and filled time intervals in pigeons. Anim. Learn. Behav. 32, 477–490. Harper, D.N., Bizo, L.A., 2000. Mediation of timing accuracy by operant behavior. Behav. Process. 50, 143–154. Keen, R., Church, R.M., 2003. Reexamination of interpretations of memory for temporal intervals. In: Presented at the International Conference on Comparative Cognition, Melbourne, Florida. Kelly, R., Spetch, M.L., 2000. Choice biases in delayed matching-to-sample duration with pigeons: manipulations of ITI and delay illumination. Quart. J. Exp. Psychol. 53B, 309–323. Kraemer, P.J., Mazmanian, D.S., Roberts, W.A., 1985. The choose-short effect in pigeon memory for event duration: Subjective shortening versus coding models. Anim. Learn. Behav. 13, 349–354. Leblanc, P., Soffie, M., 1999. Effects of age on short-term memory for time in rats. Exp. Aging Res. 25, 267–284. Leblanc, P., Soffie, M., 2001. The choose-short effect in rat memory for event duration: the subjective-shortening model. Behav. Process. 56, 31– 40. Meck, W.H., Church, R.M., Gibbon, J., 1985. Temporal integration in duration and number discrimination. J. Exp. Psychol. Anim. Behav. Process. 4, 591–597.
9
Meck, W.H., Church, R.M., Olton, D.S., 1984. Hippocampus, time, and memory. Behav. Neurosci. 98, 3–22. Miki, A., Santi, A., 2001. Pigeons’ timing of an arbitrary and a naturalistic auditory stimulus: tone versus cooing. Behav. Process. 53, 103–111. Roberts, S., 1982. Cross-modal use of an internal clock. J. Exp. Psychol. Anim. Behav. Process. 8, 2–22. Santi, A., Bridson, S., Ducharme, M.J., 1993. Memory codes for temporal and nontemporal stimuli in many-to-one matching by pigeons. Anim. Learn. Behav. 21, 120–130. Santi, A., Coyle, J., Coppa, R., Ross, L., 1998a. The timing of auditory and visual signals by pigeons: differential sensitivity to the intertrial interval duration. Learn. Motiv. 29, 345–365. Santi, A., Ducharme, M.J., Bridson, S., 1992. Differential outcome expectancies and memory for temporal and nontemporal stimuli in pigeons. Learn. Motiv. 23, 156–169. Santi, A., Hornyak, S., Miki, A., 2003. Pigeons’ memory for empty and filled time intervals signaled by light. Learn. Motiv. 34, 282–302. Santi, A., Ross, L., Coppa, R., Coyle, L., 1999. Pigeons’ memory for empty time intervals marked by visual or auditory stimuli. Anim. Learn. Behav. 27, 190–205. Santi, A., Stanford, L., Coyle, J., 1997. Rats’ memory for event duration: differential effects of delaying the discriminative choice cue as opposed to the opportunity to execute the choice response. Behav. Process. 40, 193–199. Santi, A., Stanford, L., Coyle, J., 1998b. Pigeons’ memory for event duration: differences between visual and auditory signals. Anim. Learn. Behav. 26, 163–171. Santi, A., Van Rooyen, P., 2007. Memory for sequences of tone bursts in the rat: reliance on temporal cues and evidence for an instructional ambiguity explanation of the choose-few effect. Learn. Motiv. 38, 295–313. Santi, A., Weise, L., Kuiper, D., 1995a. Memory for event duration in rats. Learn. Motiv. 26, 83–100. Santi, A., Weise, L., Kuiper, D., 1995b. Amphetamine and memory for event duration in rats and pigeons: disruption of attention to temporal samples rather than changes in the speed of the internal clock. Psychobiology 23, 224–232. Sherburne, L.M., Zentall, T.R., Kaiser, D.H., 1998. Timing in pigeons: the choose-short effect may result from pigeons’ “confusion” between delay and intertrial interval. Psychon. Bull. Rev. 5, 516–522. Soffie, M., Hahn, K., Terao, E., Eclancher, F., 1999. Behavioural and glial changes in old rats following environmental enrichment. Behav. Brain Res. 101, 37–49. Spetch, M.L., 1987. Systematic errors in pigeons’ memory for event duration: interaction between training and test delay. Anim. Learn. Behav. 15, 1–5. Spetch, M.L., Rusak, B., 1989. Pigeons’ memory for event duration: intertrial interval and delay effects. Anim. Learn. Behav. 17, 147–156. Spetch, M.L., Rusak, B., 1992. Temporal context effects in pigeons’ memory for event duration. Learn. Motiv. 23, 117–144. Spetch, M.L., Wilkie, D.M., 1983. Subjective shortening: a model of pigeons’ memory for event durations. J. Exp. Psychol. Anim. Behav. Process. 9, 14–30. Talarico, D., Grant, D.S., 2006. Effect of training delays and start and stop markers on the choose-short effect in pigeons. Behav. Process. 71, 98–106. Ward, R.D., Odum, A.L., 2007. Disruption of temporal discrimination and the choose-short effect. Learn. Behav. 35, 60–70. Zentall, T.R., 1997. Animal memory: the role of “instructions”. Learn. Motiv. 28, 280–308. Zentall, T.R., 1999. Support for a theory of memory for event duration must distinguish between test-trial ambiguity and actual memory loss. J. Exp. Anal. Behav. 72, 467–472. Zentall, T.R., 2005. Timing, memory for intervals, and memory for untimed stimuli: the role of instructional ambiguity. Behav. Process. 70, 209–222.