HZPPOCAMPUS, VOL. 2, NO. 1, PAGES 81-88, JANUARY 1992
Preserved Configural Learning and Spatial Learning Impairment in Rats With Hippocampal Damage Michela Gallagher* and Peter C. Hollandt *Department of Psychology, University of North Carolina, Chapel Hill, NC 27599 U.S.A. and ?Department of Psychology, Duke University, Durham, NC 27706 U.S.A.
ABSTRACT This study was undertaken to compare the effect of hippocampal neurotoxic lesions in rats on two behavioral tasks, one a test of spatial learning, and the other an operant discrimination task that is acquired by forming nonspatial configural associations. Lesions of the hippocampus were made with microinjections of ibotenic acid. After postoperative recovery, rats were trained initially to locate a camouflaged escape platform in a water maze using distal spatial cues. Rats also were trained in the maze apparatus with a visible escape platform under conditions in which spatial information was made irrelevant to performance, i.e., cue learning. In an operant task, the same rats were then trained on a discrimination that included simultaneous feature positive and feature negative components (trial types XA + , A - , XB - , B +). After completion of this nonspatial configural learning task, rats received additional training in the water maze using a new platform location for spatial learning. To the extent that proficient performance in both the maze and operant tasks depends on a common function of the hippocampus, i.e., configural learning, the expectation was that hippocampal lesions would prove equally detrimental to performance in both tasks. Contrary to this expectation, lesioned rats were severely impaired in spatial learning but readily acquired the operant discrimination, even exhibiting some evidence of enhanced performance on this nonspatial configural learning task. Performance of the lesioned rats during cue training in the water maze was also enhanced relative to the control group. These results demonstrate that a spatial learning deficit in rats with hippocampal damage does not necessarily occur in the setting of a general configural learning impairment, as originally suggested by Sutherland and Rudy (1989, Psychobiology 17: 129- 144). Key words: hippocampus, neurotoxic lesions, ibotenic acid, maze learning, discrimination learning
Rats with damage to the hippocampus have difficulty learning spatial tasks (Morris et al., 1982; 1990; Sutherland et al., 1982; Jarrard, 1983; Aggleton et al., 1986; DiMattia and Kesner, 1988). Although this finding is commonly reported, there are many different accounts of the cause underlying this deficit. By one account, the hippocampus is a neural system used for mapping spatial information (O’Keefe and Nadel, 1978). According to this view, spatial tasks are sensitive to hippocampal damage due to the specific nature of the information upon which learning depends in these tasks. The fact that rats with hippocampal damage are capable of learning many nonspatial tasks is often taken as further support of this view. Other proposed hippocampal functions gain support from demonstrations of learning impairments after hippocampal damage on certain nonspatial tasks. For example, a recent theory proposed by Sutherland and Rudy (1989) is intended Correspondence and reprint requests to Michela Gallagher, Department of Psychology, CB 3270, Davie Hall, University of North Carolina, Chapel Hill, NC 27599 U. S. A.
to address the effects of hippocampal damage in both spatial and nonspatial learning. By their account, the spatial learning impairment is viewed within a broader framework that emphasizes the nature of the processing performed by the hippocampus as distinct from the specific content of the information elaborated by it. Specifically, they propose that the hippocampus is specialized for configural learning, in which the various spatial, temporal, or other attributes of a compound stimulus are combined or configured to form a single functional unit distinct from its constituent elements. Thus, the spatial learning deficit produced by hippocampal damage is considered a specific instance of a more general impairment-one that should extend to nonspatial tasks that critically depend on configural learning. In support of Sutherland and Rudy’s (1989) notion that the integrity of the hippocampus is important for configural learning, they found that neurotoxic damage of the hippocampus impaired the performance of rats on an operant discrimination task commonly referred to as a negative patterning problem (Rudy and Sutherland, 1989). Rats were presented with three
81
82 HIPPOCAMPUS VOL. 2, NO. 1, JANUARY 1992 types of trials in which the stimulus elements consisted of a Table 1. Coordinates for Hippocampal Lesions light and a tone. Responding was reinforced (A+ or B +) in Anterior (rnm) Lateral (rnrn) Ventral (mrn) trials consisting of either a light or tone presentation alone, - 3.3 2 1.8 + 6.2 but responding was not reinforced (AB -) when the light and -3.5 23.0 +5.4 tone were presented together in compound form. Control rats -3.5 +4.8 23.8 learned to respond less to the nonreinforced compound -1.5 2 5.0 +4.2 (AB -) than to the elements that signaled reinforcement (A+ - 4.7 +4.2 2 5.0 - 7.0 24.5 +3.5 and B + ) . Clearly, responding to the compound did not - 4.5 24.5 + 3.5 merely reflect the sum of associative strengths conditioned -7.0 1-4.8 +2.9 to each of the elements. Rather, the rats learned to withhold -4.5 2 4.8 + 2.9 responding to a cue unique to the AB - configuration. Rudy - 6.0 24.8 +2.2 and Sutherland (1989) reported that this discrimination was -4.3 24.8 +2.2 not acquired by rats with hippocampal damage; in contrast All coordinates are given with reference to stereotaxic zero acto the control rats, lesioned animals did not withhold reto the Paxinos and Watson (1986) rat brain atlas. cording sponding in the presence of the nonreinforced compound. The possibility that this deficit might reflect a general inability to withhold responding was not supported by additional data, Surgery indicating that a simple discrimination ( A + , B - ) was acBilateral lesions of the hippocampus were made by injecquired normally by rats with the same hippocampal neurotion of ibotenic acid (Sigma Chemical Company) while the toxic damage. The present study was undertaken to compare the effect rats were anesthetized with Nembutal (50 mg/kg). The sterof hippocampal neurotoxic lesions in rats on two different eotaxic coordinates, with the head positioned level between tasks, one a test of spatial learning using the Morris water bregma and lambda, were adopted from the Paxinos and Watmaze, and the other an operant discrimination task that is son (1986) atlas, as shown in Table 1. Ibotenic acid (10 acquired by forming nonspatial configural associations. In the mg/mL) was administered using a sterile phosphate buffered latter task, rats are trained using four types of trials (XA , saline vehicle (pH 7.4) in a 0.2-pL volume at each site. Rats A- , B + , and XB -). Rats acquire this discrimination by in the operated control group (N = 7) received injections of learning to respond on B + alone trials in the presence of a the vehicle alone. The animals were allowed 7 days of postcue unique to the XA+ (lightitone) compound, and by learn- operative recovery before the beginning of the experiment. ing to withhold responding on A - alone trials and in the presence of the XB - (lightlwhite noise) compound. Consid- Apparatus erable behavioral evidence (Holland, 1991; Holland and For assessment of spatial learning, the test apparatus conReeve, 199I ) indicates that this discrimination is based on the sisted of a circular galvanized steel tank (depth, 0.58 m; diability to abstract cues unique to the stimulus compounds and ameter, 1.83 m), which served as a water maze. The interior to associate independently element and compound cues with surface of the maze was painted white. The maze was filled reinforcement and nonreinforcement, according to the dis- to a depth of 35.5 cm with lukewarm water (27”C), which was crimination contingencies. made cloudy by the addition of 0.9 kg of powdered milk. The present experiment used a within-subject design, in During “place” training, a camouflaged escape platform which the same groups of lesioned and control rats were (height, 34.5 cm; diameter, 10.2 cm) was submerged 1 cm tested in both the spatial and nonspatial tasks. To the extent below the water surface in the center of one quadrant of the that proficient performance in both tasks depends on a com- maze. A black platform of identical construction, but meamon function of the hippocampus, i.e., configural learning, suring 37.0 cm in height, was used for “cue” training; this the expectation was that hippocampal lesions would prove platform was visible approximately 1 cm above the water surequally detrimental to perform in both tasks. Contrary to this face. Black patterns were affixed to white curtains that surexpectation, lesioned rats were severely impaired in spatial rounded the maze during the initial training phase. Additional learning but readily acquired the operant discrimination, even training, which occurred after the configural discrimination exhibiting some evidence of enhanced performance on this task, took place in the presence of different spatial cues, i.e., white patterns affixed to black curtains. A Panasonic WVnonspatial configural learning task. 241 video camera (wide-angle lens) was placed above the cenMATERIALS A N D METHODS ter of the water maze to record performance. Data were collected and analyzed by a video tracking system (HVS Image Subjects Analyzer VP-112 and IBM PC compatible computer), with Eighteen male Long-Evans hooded rats, obtained from software developed by San Diego Instruments, Inc. An Charles River Laboratories (Raleigh, NC), were 300-325 g Andex Color monitor situated outside the curtained area was at the beginning of the experiment. The animals were housed used by the experimenter to locate the animal’s position in individually under a 12-hour light-dark cycle (lights on 07:00), the maze. and were provided with ad libitum food and water, except For configural discrimination learning, there were 11 idenwhere noted in the behavioral protocol. tical experimental chambers, each 22.9 x 20.3 x 20.3 cm.
+
CONFIGURAL AND SPATIAL LEARNING / Gallagher and Holland
The front and back walls of each chamber were aluminum; the side walls and top were clear acrylic. A food cup was recessed behind a 5 x 5-cm opening in the front wall; the bottom of the opening was 2 cm from the floor, and its center was 2 cm to the right of the center of the front wall. Sucrose solution (the reinforcer) was dispensed to this cup via a gravity-feed solenoid valve. A 6-W jeweled lamp (panel light) was centered on the front wall, 4 cm above the food cup opening. Four centimeters left of the food cup opening was a 2 x 2cm lever, mounted 3 cm above the floor. The floor of the chamber was constructed of 0.48-cm stainless steel rods spaced 1.9 cm apart. Each of the chambers was enclosed in a sound-attenuating shell. A speaker for delivering the auditory stimuli was positioned even with the top of the chamber and 10 cm from one side wall. Finally, each shell was enclosed in another sound-attenuating box. Constant background noise (62.5-dB) was provided by a ventilating fan on each box.
83
addition, the spatial cues surrounding the maze were altered (black curtains with white patterns affixed). During the first 3 days of transfer training, rats were given six trialdday to locate the new position of the escape platform. On day 4, a 30-second free swim was given to assess the search pattern for the new location of the escape platform.
Configural discrimination task All sessions were 60 minutes in duration. First, the rats were trained to press the lever. In the initial session, the rats received 20 response-independent 0.3-mL deliveries of 0.2 M sucrose (the reinforcer used throughout these experiments) on a variable-time 1-minute schedule. Lever presses were reinforced during that period and during the remaining 40 minutes of the session. In the second session, each rat was allowed to remain in its chamber until it had made about 50 lever presses. Lever presses were reinforced, but there were no response-independent sucrose presentations in that session. The next five sessions were designed to establish lever Procedures pressing during the two reinforced stimuli, P T + and N + . Spatial learning task During each of these sessions, there were 30 15-second preAll animals were handled (3 midday) for 7 days before the sentations of a 73-dB white noise (N) and 30 15-second prestart of behavioral testing. The initial assessment of spatial sentations of a compound that comprised a 74 dB 1,500-Hz learning occurred on 5 consecutive days before training in the tone and the illumination of the panel light (PT). In the first configural discrimination task. An additional 4 days of testing two sessions, each lever press made during one of these cues in the spatial task (transfer training) occurred after animals was followed by sucrose delivery. In the remaining sessions (of both this and subsequent phases), reinforcement was completed the configural discrimination task. On day 1 in the initial training phase, each animal was available only during the final 5 seconds of each reinforced habituated to the water maze. On this habituation trial, each cue. During all sessions throughout this experiment, trial seanimal was placed in the maze for 90 seconds without an quences were generated randomly (without constraint) for escape platform present and then returned to its home cage. each session. Intertrial intervals were randomized daily, with The animals received place training for the next 3 consecutive the constraint that the range of intervals has from 0.5 to 2.0 days (6 trialdday). During each of these sessions, the sub- times the mean interval. Next, discrimination training began, in which illumination mergedkamouflaged escape platform was positioned in the maze and remained in the same location. The animal's start- of the panel light (P) indicated the availability of reinforceing position, however, varied. The animals were introduced ment during the tone (T) and the nonavailability of reinforceinto the maze at one of four equally spaced locations around ment during the noise (N). All rats received four kinds of trials the perimeter of the tank, with the provision that every four in each of the 20 conditioning sessions. Reinforced PT+ and trials included varying permutations of the four starting lo- N + trials were identical to those received previously. In adcations. On each trial, the rat was given 90 seconds to find dition, there were 15-second presentations of the tone alone the escape platform. If the rat did not locate the platform in (T-), and of a compound of the panel light and the noise 90 seconds, the experimenter placed the rat on the platform (PN-). In each of sessions 1-10 there were 15 of each trial for 20 seconds prior to placement in a holding cage for a 60- type, randomly intermixed, and in each of sessions 11-20 there were 10 N + , 10 PT + , 20 PN - , and 20 T - trials. second intertrial interval. In order to assess conditioning established to the panelOn day 5 each rat received a free swim followed by cue training. For the free swim, each rat was placed in the maze light feature, the discrimination training phase was followed for 30 seconds with the escape platform removed. The rat's by two test sessions that included presentations of the feature swim pattern on this trial was used to analyze the accuracy alone (P-). In each of these sessions, there were 12 presenof the search for the escape platform. Immediately after the tations of each of the trial types PT + , T - , N + , PN + , and free swim, rats received six cue training trials. On each of P- (described previously). these trials the rats were permitted to escape to the visible platform, which was positioned in different locations from Histology After the completion of all behavioral testing, rats were trial to trial. The rats were allotted 60 seconds to locate the deeply anesthetized with choral hydrate and perfused transplatform on these trials. After the completion of the configural discrimination task cardially with 0.9% saline followed by a 10% formol-saline described below, rats received an additional 4 days of training solution. Brains were removed and stored in 10% formalin in the water maze. The escape platform was repositioned in for 1 week. Brains were then sectioned (30 Fm) in a cryostat, the middle of the quadrant 180" from its original location. In mounted on slides and Nissl-stained. Histologicd examina-
84
HZPPOCAMPUS VOL. 2, NO. 1, JANUARY 1992
tion was performed with the aid of the Paxinos and Watson (1986) atlas using an Olympus BH-2 microscope.
Data analysis Analysis of variance was used to assess the performance of rats during training trials in the spatial learning task (latencies and distance traversed to locate the escape platform). Measures of spatial bias obtained from the free-swim trials were also analyzed using these statistical methods. In the configural discrimination task, we recorded the rate of lever pressing, the percentage of trials on which at least one response occurred, and the latency to the first response, during each 5-second interval of the CS presentations and the 5-second pre-CS interval. However, response rates did not provide an appropriate measure of discrimination because the rats did not press for several seconds after sucrose was delivered; thus, response rates were artificially depressed during reinforced trials. Neither the latency nor percentage of trials with a response was affected in this manner. Those two measures were very closely related throughout the experiment and showed similar patterns of statistical significance. For the sake of brevity, only the percentages of trials with a response are reported here. Because there were occasional severe violations of homogeneity of variance assumptions with that measure, distribution-free inferential statistics were used to analyze the results of the operant phases. The level of significance throughout the study was defined as 0.05.
RESULTS Histology An image-analyzing system (MCID) was used to estimate the extent of hippocampal lesions. The total length of the principal cell layer was measured in histological sections from the control brains and from brains with hippocampal damage. This measurement included the pyramidal cell layer in CAI and CA3 regions and the granule cell layer in the dentate gyrus. The cell field was rated as spared even when neurons appeared sparser than normal. Thus this method provided a conservative estimate of the severity of hippocampal damage. This analysis revealed that eight of the rats in the lesioned group had very extensive bilateral damage to the hippocampus (>60%, bilaterally). In all but two cases, bilateral damage was in excess of 70%. Figure 1 illustrates the appearance of a lesion in which there was extensive loss of hippocampal neurons. Little damage was detectible to structures outside the hippocampus. However, in several cases (N = 3), some damage in the ventral hippocampus extended into the subicular complex. The rats in the lesioned group that were excluded from the data analysis had significant unilateral sparing of hippocampus (>50%). Thus the data from eight lesioned and seven operated control rats were analyzed.
Spatial learning Performance on the habituation trial provided an assessment of swimming performance prior to training. Analysis of the total distance traversed in the maze during the 90-second interval revealed no difference between the two groups
Fig. 1 . Coronal hemisections from brains representative of the control group (on the left) and the lesioned group (on the right). Extensive cell loss is evident in ventral hippocampus of the lesioned brain. Some sparing is evident in the hippocampus at the dorsal level in this section.
(3.70k.26 and 3.84t.18 m for the control and lesioned groups, respectively). Hippocampal damage, however, caused a reliable impairment in place learning, as reflected in performance during the training trials and the free swim. Both latencies and distance traversed in locating the escape platform were significantly increased for the lesioned group relative to the control group over the course of training. A two-way analysis of variance (group X blocks of six training trials) performed on the latency data yielded a main effect for group [F(1,13)=4.55, P
Preserved Configural Learning and Spatial Learning Impairment in Rats With Hippocampal Damage Michela Gallagher* and Peter C. Hollandt *Department of Psychology, University of North Carolina, Chapel Hill, NC 27599 U.S.A. and ?Department of Psychology, Duke University, Durham, NC 27706 U.S.A.
ABSTRACT This study was undertaken to compare the effect of hippocampal neurotoxic lesions in rats on two behavioral tasks, one a test of spatial learning, and the other an operant discrimination task that is acquired by forming nonspatial configural associations. Lesions of the hippocampus were made with microinjections of ibotenic acid. After postoperative recovery, rats were trained initially to locate a camouflaged escape platform in a water maze using distal spatial cues. Rats also were trained in the maze apparatus with a visible escape platform under conditions in which spatial information was made irrelevant to performance, i.e., cue learning. In an operant task, the same rats were then trained on a discrimination that included simultaneous feature positive and feature negative components (trial types XA + , A - , XB - , B +). After completion of this nonspatial configural learning task, rats received additional training in the water maze using a new platform location for spatial learning. To the extent that proficient performance in both the maze and operant tasks depends on a common function of the hippocampus, i.e., configural learning, the expectation was that hippocampal lesions would prove equally detrimental to performance in both tasks. Contrary to this expectation, lesioned rats were severely impaired in spatial learning but readily acquired the operant discrimination, even exhibiting some evidence of enhanced performance on this nonspatial configural learning task. Performance of the lesioned rats during cue training in the water maze was also enhanced relative to the control group. These results demonstrate that a spatial learning deficit in rats with hippocampal damage does not necessarily occur in the setting of a general configural learning impairment, as originally suggested by Sutherland and Rudy (1989, Psychobiology 17: 129- 144). Key words: hippocampus, neurotoxic lesions, ibotenic acid, maze learning, discrimination learning
Rats with damage to the hippocampus have difficulty learning spatial tasks (Morris et al., 1982; 1990; Sutherland et al., 1982; Jarrard, 1983; Aggleton et al., 1986; DiMattia and Kesner, 1988). Although this finding is commonly reported, there are many different accounts of the cause underlying this deficit. By one account, the hippocampus is a neural system used for mapping spatial information (O’Keefe and Nadel, 1978). According to this view, spatial tasks are sensitive to hippocampal damage due to the specific nature of the information upon which learning depends in these tasks. The fact that rats with hippocampal damage are capable of learning many nonspatial tasks is often taken as further support of this view. Other proposed hippocampal functions gain support from demonstrations of learning impairments after hippocampal damage on certain nonspatial tasks. For example, a recent theory proposed by Sutherland and Rudy (1989) is intended Correspondence and reprint requests to Michela Gallagher, Department of Psychology, CB 3270, Davie Hall, University of North Carolina, Chapel Hill, NC 27599 U. S. A.
to address the effects of hippocampal damage in both spatial and nonspatial learning. By their account, the spatial learning impairment is viewed within a broader framework that emphasizes the nature of the processing performed by the hippocampus as distinct from the specific content of the information elaborated by it. Specifically, they propose that the hippocampus is specialized for configural learning, in which the various spatial, temporal, or other attributes of a compound stimulus are combined or configured to form a single functional unit distinct from its constituent elements. Thus, the spatial learning deficit produced by hippocampal damage is considered a specific instance of a more general impairment-one that should extend to nonspatial tasks that critically depend on configural learning. In support of Sutherland and Rudy’s (1989) notion that the integrity of the hippocampus is important for configural learning, they found that neurotoxic damage of the hippocampus impaired the performance of rats on an operant discrimination task commonly referred to as a negative patterning problem (Rudy and Sutherland, 1989). Rats were presented with three
81
82 HIPPOCAMPUS VOL. 2, NO. 1, JANUARY 1992 types of trials in which the stimulus elements consisted of a Table 1. Coordinates for Hippocampal Lesions light and a tone. Responding was reinforced (A+ or B +) in Anterior (rnm) Lateral (rnrn) Ventral (mrn) trials consisting of either a light or tone presentation alone, - 3.3 2 1.8 + 6.2 but responding was not reinforced (AB -) when the light and -3.5 23.0 +5.4 tone were presented together in compound form. Control rats -3.5 +4.8 23.8 learned to respond less to the nonreinforced compound -1.5 2 5.0 +4.2 (AB -) than to the elements that signaled reinforcement (A+ - 4.7 +4.2 2 5.0 - 7.0 24.5 +3.5 and B + ) . Clearly, responding to the compound did not - 4.5 24.5 + 3.5 merely reflect the sum of associative strengths conditioned -7.0 1-4.8 +2.9 to each of the elements. Rather, the rats learned to withhold -4.5 2 4.8 + 2.9 responding to a cue unique to the AB - configuration. Rudy - 6.0 24.8 +2.2 and Sutherland (1989) reported that this discrimination was -4.3 24.8 +2.2 not acquired by rats with hippocampal damage; in contrast All coordinates are given with reference to stereotaxic zero acto the control rats, lesioned animals did not withhold reto the Paxinos and Watson (1986) rat brain atlas. cording sponding in the presence of the nonreinforced compound. The possibility that this deficit might reflect a general inability to withhold responding was not supported by additional data, Surgery indicating that a simple discrimination ( A + , B - ) was acBilateral lesions of the hippocampus were made by injecquired normally by rats with the same hippocampal neurotion of ibotenic acid (Sigma Chemical Company) while the toxic damage. The present study was undertaken to compare the effect rats were anesthetized with Nembutal (50 mg/kg). The sterof hippocampal neurotoxic lesions in rats on two different eotaxic coordinates, with the head positioned level between tasks, one a test of spatial learning using the Morris water bregma and lambda, were adopted from the Paxinos and Watmaze, and the other an operant discrimination task that is son (1986) atlas, as shown in Table 1. Ibotenic acid (10 acquired by forming nonspatial configural associations. In the mg/mL) was administered using a sterile phosphate buffered latter task, rats are trained using four types of trials (XA , saline vehicle (pH 7.4) in a 0.2-pL volume at each site. Rats A- , B + , and XB -). Rats acquire this discrimination by in the operated control group (N = 7) received injections of learning to respond on B + alone trials in the presence of a the vehicle alone. The animals were allowed 7 days of postcue unique to the XA+ (lightitone) compound, and by learn- operative recovery before the beginning of the experiment. ing to withhold responding on A - alone trials and in the presence of the XB - (lightlwhite noise) compound. Consid- Apparatus erable behavioral evidence (Holland, 1991; Holland and For assessment of spatial learning, the test apparatus conReeve, 199I ) indicates that this discrimination is based on the sisted of a circular galvanized steel tank (depth, 0.58 m; diability to abstract cues unique to the stimulus compounds and ameter, 1.83 m), which served as a water maze. The interior to associate independently element and compound cues with surface of the maze was painted white. The maze was filled reinforcement and nonreinforcement, according to the dis- to a depth of 35.5 cm with lukewarm water (27”C), which was crimination contingencies. made cloudy by the addition of 0.9 kg of powdered milk. The present experiment used a within-subject design, in During “place” training, a camouflaged escape platform which the same groups of lesioned and control rats were (height, 34.5 cm; diameter, 10.2 cm) was submerged 1 cm tested in both the spatial and nonspatial tasks. To the extent below the water surface in the center of one quadrant of the that proficient performance in both tasks depends on a com- maze. A black platform of identical construction, but meamon function of the hippocampus, i.e., configural learning, suring 37.0 cm in height, was used for “cue” training; this the expectation was that hippocampal lesions would prove platform was visible approximately 1 cm above the water surequally detrimental to perform in both tasks. Contrary to this face. Black patterns were affixed to white curtains that surexpectation, lesioned rats were severely impaired in spatial rounded the maze during the initial training phase. Additional learning but readily acquired the operant discrimination, even training, which occurred after the configural discrimination exhibiting some evidence of enhanced performance on this task, took place in the presence of different spatial cues, i.e., white patterns affixed to black curtains. A Panasonic WVnonspatial configural learning task. 241 video camera (wide-angle lens) was placed above the cenMATERIALS A N D METHODS ter of the water maze to record performance. Data were collected and analyzed by a video tracking system (HVS Image Subjects Analyzer VP-112 and IBM PC compatible computer), with Eighteen male Long-Evans hooded rats, obtained from software developed by San Diego Instruments, Inc. An Charles River Laboratories (Raleigh, NC), were 300-325 g Andex Color monitor situated outside the curtained area was at the beginning of the experiment. The animals were housed used by the experimenter to locate the animal’s position in individually under a 12-hour light-dark cycle (lights on 07:00), the maze. and were provided with ad libitum food and water, except For configural discrimination learning, there were 11 idenwhere noted in the behavioral protocol. tical experimental chambers, each 22.9 x 20.3 x 20.3 cm.
+
CONFIGURAL AND SPATIAL LEARNING / Gallagher and Holland
The front and back walls of each chamber were aluminum; the side walls and top were clear acrylic. A food cup was recessed behind a 5 x 5-cm opening in the front wall; the bottom of the opening was 2 cm from the floor, and its center was 2 cm to the right of the center of the front wall. Sucrose solution (the reinforcer) was dispensed to this cup via a gravity-feed solenoid valve. A 6-W jeweled lamp (panel light) was centered on the front wall, 4 cm above the food cup opening. Four centimeters left of the food cup opening was a 2 x 2cm lever, mounted 3 cm above the floor. The floor of the chamber was constructed of 0.48-cm stainless steel rods spaced 1.9 cm apart. Each of the chambers was enclosed in a sound-attenuating shell. A speaker for delivering the auditory stimuli was positioned even with the top of the chamber and 10 cm from one side wall. Finally, each shell was enclosed in another sound-attenuating box. Constant background noise (62.5-dB) was provided by a ventilating fan on each box.
83
addition, the spatial cues surrounding the maze were altered (black curtains with white patterns affixed). During the first 3 days of transfer training, rats were given six trialdday to locate the new position of the escape platform. On day 4, a 30-second free swim was given to assess the search pattern for the new location of the escape platform.
Configural discrimination task All sessions were 60 minutes in duration. First, the rats were trained to press the lever. In the initial session, the rats received 20 response-independent 0.3-mL deliveries of 0.2 M sucrose (the reinforcer used throughout these experiments) on a variable-time 1-minute schedule. Lever presses were reinforced during that period and during the remaining 40 minutes of the session. In the second session, each rat was allowed to remain in its chamber until it had made about 50 lever presses. Lever presses were reinforced, but there were no response-independent sucrose presentations in that session. The next five sessions were designed to establish lever Procedures pressing during the two reinforced stimuli, P T + and N + . Spatial learning task During each of these sessions, there were 30 15-second preAll animals were handled (3 midday) for 7 days before the sentations of a 73-dB white noise (N) and 30 15-second prestart of behavioral testing. The initial assessment of spatial sentations of a compound that comprised a 74 dB 1,500-Hz learning occurred on 5 consecutive days before training in the tone and the illumination of the panel light (PT). In the first configural discrimination task. An additional 4 days of testing two sessions, each lever press made during one of these cues in the spatial task (transfer training) occurred after animals was followed by sucrose delivery. In the remaining sessions (of both this and subsequent phases), reinforcement was completed the configural discrimination task. On day 1 in the initial training phase, each animal was available only during the final 5 seconds of each reinforced habituated to the water maze. On this habituation trial, each cue. During all sessions throughout this experiment, trial seanimal was placed in the maze for 90 seconds without an quences were generated randomly (without constraint) for escape platform present and then returned to its home cage. each session. Intertrial intervals were randomized daily, with The animals received place training for the next 3 consecutive the constraint that the range of intervals has from 0.5 to 2.0 days (6 trialdday). During each of these sessions, the sub- times the mean interval. Next, discrimination training began, in which illumination mergedkamouflaged escape platform was positioned in the maze and remained in the same location. The animal's start- of the panel light (P) indicated the availability of reinforceing position, however, varied. The animals were introduced ment during the tone (T) and the nonavailability of reinforceinto the maze at one of four equally spaced locations around ment during the noise (N). All rats received four kinds of trials the perimeter of the tank, with the provision that every four in each of the 20 conditioning sessions. Reinforced PT+ and trials included varying permutations of the four starting lo- N + trials were identical to those received previously. In adcations. On each trial, the rat was given 90 seconds to find dition, there were 15-second presentations of the tone alone the escape platform. If the rat did not locate the platform in (T-), and of a compound of the panel light and the noise 90 seconds, the experimenter placed the rat on the platform (PN-). In each of sessions 1-10 there were 15 of each trial for 20 seconds prior to placement in a holding cage for a 60- type, randomly intermixed, and in each of sessions 11-20 there were 10 N + , 10 PT + , 20 PN - , and 20 T - trials. second intertrial interval. In order to assess conditioning established to the panelOn day 5 each rat received a free swim followed by cue training. For the free swim, each rat was placed in the maze light feature, the discrimination training phase was followed for 30 seconds with the escape platform removed. The rat's by two test sessions that included presentations of the feature swim pattern on this trial was used to analyze the accuracy alone (P-). In each of these sessions, there were 12 presenof the search for the escape platform. Immediately after the tations of each of the trial types PT + , T - , N + , PN + , and free swim, rats received six cue training trials. On each of P- (described previously). these trials the rats were permitted to escape to the visible platform, which was positioned in different locations from Histology After the completion of all behavioral testing, rats were trial to trial. The rats were allotted 60 seconds to locate the deeply anesthetized with choral hydrate and perfused transplatform on these trials. After the completion of the configural discrimination task cardially with 0.9% saline followed by a 10% formol-saline described below, rats received an additional 4 days of training solution. Brains were removed and stored in 10% formalin in the water maze. The escape platform was repositioned in for 1 week. Brains were then sectioned (30 Fm) in a cryostat, the middle of the quadrant 180" from its original location. In mounted on slides and Nissl-stained. Histologicd examina-
84
HZPPOCAMPUS VOL. 2, NO. 1, JANUARY 1992
tion was performed with the aid of the Paxinos and Watson (1986) atlas using an Olympus BH-2 microscope.
Data analysis Analysis of variance was used to assess the performance of rats during training trials in the spatial learning task (latencies and distance traversed to locate the escape platform). Measures of spatial bias obtained from the free-swim trials were also analyzed using these statistical methods. In the configural discrimination task, we recorded the rate of lever pressing, the percentage of trials on which at least one response occurred, and the latency to the first response, during each 5-second interval of the CS presentations and the 5-second pre-CS interval. However, response rates did not provide an appropriate measure of discrimination because the rats did not press for several seconds after sucrose was delivered; thus, response rates were artificially depressed during reinforced trials. Neither the latency nor percentage of trials with a response was affected in this manner. Those two measures were very closely related throughout the experiment and showed similar patterns of statistical significance. For the sake of brevity, only the percentages of trials with a response are reported here. Because there were occasional severe violations of homogeneity of variance assumptions with that measure, distribution-free inferential statistics were used to analyze the results of the operant phases. The level of significance throughout the study was defined as 0.05.
RESULTS Histology An image-analyzing system (MCID) was used to estimate the extent of hippocampal lesions. The total length of the principal cell layer was measured in histological sections from the control brains and from brains with hippocampal damage. This measurement included the pyramidal cell layer in CAI and CA3 regions and the granule cell layer in the dentate gyrus. The cell field was rated as spared even when neurons appeared sparser than normal. Thus this method provided a conservative estimate of the severity of hippocampal damage. This analysis revealed that eight of the rats in the lesioned group had very extensive bilateral damage to the hippocampus (>60%, bilaterally). In all but two cases, bilateral damage was in excess of 70%. Figure 1 illustrates the appearance of a lesion in which there was extensive loss of hippocampal neurons. Little damage was detectible to structures outside the hippocampus. However, in several cases (N = 3), some damage in the ventral hippocampus extended into the subicular complex. The rats in the lesioned group that were excluded from the data analysis had significant unilateral sparing of hippocampus (>50%). Thus the data from eight lesioned and seven operated control rats were analyzed.
Spatial learning Performance on the habituation trial provided an assessment of swimming performance prior to training. Analysis of the total distance traversed in the maze during the 90-second interval revealed no difference between the two groups
Fig. 1 . Coronal hemisections from brains representative of the control group (on the left) and the lesioned group (on the right). Extensive cell loss is evident in ventral hippocampus of the lesioned brain. Some sparing is evident in the hippocampus at the dorsal level in this section.
(3.70k.26 and 3.84t.18 m for the control and lesioned groups, respectively). Hippocampal damage, however, caused a reliable impairment in place learning, as reflected in performance during the training trials and the free swim. Both latencies and distance traversed in locating the escape platform were significantly increased for the lesioned group relative to the control group over the course of training. A two-way analysis of variance (group X blocks of six training trials) performed on the latency data yielded a main effect for group [F(1,13)=4.55, P
