Impairment of Radial Maze Delayed Nonmatching After Lesions of Anterior Thalamus and Parahippocampal Cortex

Transcription

1 Behavioral Neuroscience Copyright 2003 by the American Psychological Association, Inc. 2003, Vol. 117, No. 3, /03/$12.00 DOI: / Impairment of Radial Maze Delayed Nonmatching After Lesions of Anterior Thalamus and Parahippocampal Cortex Robert G. Mair, Joshua A. Burk, and M. Christine Porter University of New Hampshire This study compared the effects of lesions damaging hippocampus-related pathways in anterior thalamus (AT) and parahippocampal (PH) cortex on allocentric spatial memory. Rats were trained to perform radial maze delayed nonmatching (DNM) with random selection of arms to prevent egocentric solutions. After experimental treatment (control, excitotoxic AT, radiofrequency PH, or combined AT PH lesions), rats were retrained for 30 sessions from 2 to 8 weeks after surgery. Results showed comparable impairments for AT and PH lesions that added without interaction in the combined AT PH group. During chronic recovery, the AT PH group exhibited delay-dependent deficits comparable to previous results for hippocampal lesions. Thus, AT and PH lesions appear to have separate effects that together disrupt hippocampus-dependent spatial memory. Medial thalamus is one of the core areas of the brain in which localized lesions can cause global amnesia (Graff-Radford, Tranel, van Hoesen, & Brandt, 1990; Partlow, del Carpio-O Donovan, Melanson, & Peters, 1992; Squire, Amaral, Zola-Morgan, Kritchevsky, & Press, 1989; Van der Werf, Witter, Ulyings, & Jolles, 2000; Victor, Adams, & Collins, 1989; von Cramon, Hebel, & Schuri, 1985). The neurological basis of thalamic amnesia is not established. Some investigators have emphasized the frequent involvement of the medial mammillary nuclei and the mammillothalamic tract and argued that amnesia is caused by disruption of hippocampal functions mediated by the anterior thalamic (AT) nuclei (Aggleton & Brown, 1999; Bentivoglio, Aggleton, & Mishkin, 1997; Delay, Brion, & Elissalde, 1958). Alternatively, thalamic amnesia has been ascribed to lesions of the mediodorsal, midline, or intralaminar nuclei (Mair, 1994; Mair, Warrington, & Weiskrantz, 1979; Markowitsch, 1982; Partlow et al., 1992; Victor et al., 1989), presumably involving pathways related to frontal cortex and striatum. In a recent meta-analysis of thalamic infarction cases, Van der Werf et al. (2000) argued that damage to the mammillothalamic tract is associated with a medial temporal lobe-like amnesia (p. 613), whereas lesions of the mediodorsal, midline, or intralaminar nuclei are associated with impairments in attentional, cognitive, and executive functions mediated by frontal cortex. Experimental studies have confirmed the importance of the AT and intralaminar nuclei but have not clearly differentiated between their effects on memory and thus their potential contributions to Robert G. Mair, Joshua A. Burk, and M. Christine Porter, Department of Psychology, University of New Hampshire. This research was supported by National Institute of Neurological Disorders and Stroke Grant NS We thank Jennifer Koch, Jim Howard, Melissa Rose, Meg Toupin, Monique Teran, Tim Carlone, and Jessica Maehr for technical assistance. Correspondence concerning this article should be addressed to Robert G. Mair, Department of Psychology, University of New Hampshire, Durham, New Hampshire rgm@unh.edu thalamic amnesia. Experimental evidence linking the medial mammillary nuclei or the mammillothalamic tract, the actual sites of pathology in clinical cases of amnesia, to memory is more problematic (Sziklas & Petrides, 1998; Thomas & Gash, 1985). Several reports have indicated that AT lesions impair spatial memory, although it remains uncertain whether the deficits produced by these lesions are comparable to the effects of fimbria fornix or hippocampal lesions (Aggleton, Neave, Nagle, & Hunt, 1995; Aggleton & Sahgal, 1993; Byatt & Dalrymple-Alford, 1996; Sutherland & Rodriguez, 1989; Warburton, Baird, & Aggleton, 1997). The effects of AT lesions on spatial memory have been found to be exacerbated when they involve adjacent areas of thalamus, including mediodorsal, midline, and intralaminar nuclei (Warburton, Morgan, Baird, Muir, & Aggleton, 1999). Other studies have shown that lesions of the intralaminar nuclei impair measures of spatial memory (Burk & Mair, 1998; Mair, Burk, & Porter, 1998; Mair & Lacourse, 1992; Young, Stevens, Converse, & Mair, 1996) that are either unaffected or minimally affected by fimbria-fornix or hippocampal lesions. These results suggest that damage to the AT and intralaminar nuclei affect distinct aspects of memory and that both may contribute to impairments observed with large AT lesions. We studied the effects of AT lesions on varying-choice delayed nonmatching (DNM) trained in a radial maze. This is a measure of allocentric spatial memory that has proven to be differentially sensitive to the effects of intralaminar and hippocampal damage. Hippocampal lesions produce delay-dependent, and intralaminar lesions delay-independent, impairments of this task. Prefrontal cortical lesions spare varying-choice DNM but impair recurring choice DNM, a closely related task designed to measure egocentric aspects of spatial memory (Mair et al., 1998; Porter, Burk, & Mair, 2000; Porter & Mair, 1997). To the extent that AT lesions disrupt hippocampal function, we expected to find delay-dependent impairments comparable to those produced by hippocampal lesions. To maximize the effects on hippocampal-related projections in the anterior nuclear complex (Jones, 1985) we aimed AT lesions at the laterodorsal and anterior nuclei, while minimizing damage to the intralaminar and mediodorsal nuclei as much as possible. 596

2 ANTERIOR THALAMIC AND PARAHIPPOCAMPAL LESIONS 597 One reason that AT lesions may not produce impairments comparable to those caused by hippocampal lesions is that surviving connections with parahippocampal areas (PH) of cortex may be sufficient to mediate some of the contributions of hippocampus to memory. There are a number of reports indicating that hippocampus-related projections to AT and to perirhinal cortex play distinct roles in remembering (Aggleton & Brown, 1999; Eichenbaum, 1997; Gaffan, Eacott, & Simpson, 2000; Squire, Knowlton, & Mussen, 1993). To examine the relative contributions of AT and parahippocampal projections, we compared the effects of AT lesions in rats with and without radiofrequency (RF) lesions designed to produce extensive damage to hippocampusrelated projections in PH cortex. Method Subjects Subjects were 41 male Long Evans rats (Charles River Breeding Labs, Wilmington, MA) that were 2 months old at the beginning of presurgical training, months old at surgery, and months old during postsurgical training. Rats were caged singly on a 12-hr light dark cycle, with training occurring during the light cycle. Access to water was restricted so that it could be used as a reinforcer. On days when they were trained, rats received water during training sessions and for a period of 30 min of free access at the end of the light phase. On days when they were not trained, the free-access period was increased to 60 min. Surgery Rats were matched for percent correct over the last three presurgical sessions and assigned by block randomization to one of four treatments: control, AT lesion, PH lesion, or combined AT PH lesion (n 9 per group). Rats were anesthetized (85 mg/kg ketamine and 8.5 mg/kg xylazine im), placed in a stereotaxic instrument (David Kopf Instruments, Tujunga, CA), and the skull was opened with aseptic procedures. Control rats were given an incision that was sutured shut without opening the skull. Stereotaxic coordinates were measured in millimeters relative to the interaural line for anterior posterior and dorsoventral and relative to the sagittal suture for midline. AT lesions were aimed at the anteromedial, anterodorsal, anteroventral, and laterodorsal nuclei. These lesions were made by infusing 0.1 l of a 150 mm N-methyl-D-aspartate (NMDA) solution in four sites in each hemisphere: AP 7.6, ML 1.0, DV 3.6; AP 7.6, ML 1.5, DV 4.6; AP 6.6, ML 1.4, DV 5.2; and AP 6.6, ML 2.4, DV 5.2. Infusions were made at 0.2 l/min through a 26-gauge cannula with a Kopf 5000 microinjection unit, and cannulas were left in place for 1 min after each infusion. We used excitotoxic procedures for lesioning thalamus to avoid damaging fiber tracts that pass through this area of thalamus. PH lesions were designed to isolate hippocampus from association areas of neocortex by damaging the full extent of perirhinal cortex (as defined by Paxinos & Watson, 1998). We made these lesions by heating the tip of a Radionics (Burlington, MA) TCZ electrode to 70 C for 30 s with RF current from a Radionics RFG-4 lesion generator. This was done at seven sites in each hemisphere: AP 7.2, ML 6.2, DV 3.0; AP 6.2, ML 6.4, DV 2.8; AP 5.2, ML 6.4, DV 2.6; AP 4.2, ML 7.0, DV 2.8; AP 3.2, ML 6.6, DV 2.8; AP 2.2, ML 6.5, DV 3.0; AP 1.2, ML 5.8, DV 4.4. We used radiofrequency rather than excitotoxic procedures to damage fibers and cell bodies and to avoid the possibility of producing remote lesions by activating glutamatergic terminals within hippocampus (Collins & Olney, 1982; Winn, 1991). Apparatus We trained rats in radial mazes (MED Associates, Georgia, VT) described by Porter and Mair (1997). Each maze had an octagonal hub (30.0 cm in diameter and 33.0 cm high) with 8 arms (60.0 cm long, 17.5 cm wide, and 20.0 cm high) with motorized gates to control access to and egress from each of the arms. There were wells (3.0 cm in diameter, 0.5 cm deep) milled into the floors, centered 4.5 cm from the end of each arm, in which water was delivered as a reinforcement by activation of a miniature solenoid valve (LFAA H, The Lee Co., Essex, CT). There were two photocells in each arm, both positioned 4.0 cm off the floor, with one 4.5 cm from the hub to detect arm entries and the other 4.5 cm from the end to detect when rats reached the well. The floor of the maze was made of white polycarbonate and the walls of the arms and ceilings of clear polycarbonate. The mazes were located in a windowless lab room and remotely controlled through an interface (MED Associates) connected to a 486 computer. Behavioral Training Rats were shaped to perform the varying-choice DNM task by means of the procedures described for Experiment 2 in Porter et al. (2000). This task was designed as a test of allocentric spatial memory. Sessions began with the rat in the central compartment and a gate (randomly selected from the eight alternatives) opening to allow access to the sample arm for the first trial. When rats responded to the well at the end of the sample arm, reinforcement was delivered (two 0.1-s pulses of water 2.0 s apart) and the gate closed. After the reinforcement was delivered, the gate to the sample arm was reopened along with the gate to a second (holding) arm (randomly selected from the seven alternatives). When rats responded to the well at the end of the holding arm, reinforcement was delivered and both gates closed, trapping rats in the holding arm for the duration of the retention interval (randomly selected on a trial-by-trial basis to be 1.0, 1.6, 2.5, 4.0, 6.3, 10.0, 15.8, or 25.1 s). At the end of the retention interval, the choice response began. The gates to the sample and holding arms were reopened along with the gate to a third (reinforced [S ] or nonmatching) arm that was randomly selected from the six remaining alternatives. When rats responded to the well at the end of the S arm, reinforcement was delivered and gates were closed until the start of the next trial. When the S was the first arm entered, the trial was scored as correct. Rats were trained on DNM in a continuous fashion. At the start of subsequent trials, rats were trapped in the S arm from the preceding trial (which then served as the sample arm for the new trial). Trials began with arms opening to allow access to a new holding arm (randomly selected from the seven available alternatives), where rats were reinforced and held for the duration of a newly selected retention interval. Gates then opened to allow a choice between the sample and a new S arm, following the procedures described above for the initial trial in a session. Training sessions lasted until 40 trials were completed or 75 min elapsed. Prior to surgery, rats were trained for a minimum of 25 sessions until they completed all trials at a criterion of 85% correct averaged across retention intervals. Two weeks after surgery, water deprivation was reestablished and rats were trained for 30 daily sessions with the same procedures as before surgery. The first 10 of these sessions were conducted from 2 to 4 weeks after surgery. The second 10 were conducted from 4 to 6 weeks, and the third 10 from 6 to 8 weeks, after surgery. Histology On completion of behavioral studies, rats were killed under deep anesthesia (100 mg/kg ketamine, 10 mg/kg xylazine im) by transcardiac perfusion of physiological saline followed by 5% (vol/vol) neutral buffered Formalin. Brains were then cryoprotected by immersion in solutions of 10% glycerin 4% neutral buffered Formalin and then 20% glycerin 4% neutral buffered Formalin (both vol/vol). Tissue was sectioned frozen in the coronal plane at 30 m, and every fifth section was mounted and stained with Cresyl violet for histological examination of the lesions.

3 598 MAIR, BURK, AND PORTER Histological Analyses Results All rats in the PH and combined AT PH lesion groups were found to have bilateral damage to the areas of cortex surrounding the rhinal fissure. As is apparent in Figures 1 and 2, these lesions produced extensive damage throughout the rostrocaudal extent of perirhinal cortex, as defined by Paxinos and Watson (1998). Postrhinal cortex, as defined by Burwell and Amaral (1998a, 1998b) and Burwell (2001), was at least partially damaged in all cases. In four of the PH and three of the AT PH groups, this damage involved the full anterior posterior extent of postrhinal cortex. These lesions tended to extend into other areas adjacent to PH cortex. In all cases with PH lesions, tracks of electrodes were visible in overlying areas of temporal, parietal, and occipital cortex that were penetrated during surgical procedures. In 4 cases in the PH group (3 bilateral and 1 unilateral) and 4 cases in the AT PH group (all bilateral), there was more extensive damage in parts of these overlying areas. Dorsal portions of lateral entorhinal cortex Figure 2. Photomicrographs in the coronal plane showing representative lesions in the parahippocampal cortex. Lesions are from the left (A, C, and E) and right (B, D, and F) hemispheres in anterior (A and B) and posterior (C and D) perirhinal cortex and in postrhinal cortex (E and F). The calibration mark in D represents 1 mm for A, B, C, and D. The bar in F represents 1 mm in E and F. Figure 1. Reconstruction of one of the smallest (black) and one of the largest (gray) lesions of the parahippocampal cortex. Sections are 6.2, 4.2, 2.2, and 1.2 mm anterior to the interaural line in Paxinos and Watson (1998). Reprinted from The Rat Brain in Stereotaxic Coordinates, 4th ed., G. Paxinos and C. Watson, Figures 31, 39, 47, and 51, Copyright 1998, with permission from Elsevier Science. adjacent to the perirhinal area were involved in all rats with PH lesions. Two rats in the PH group and 3 in the AT PH group had lesions extending into ventral areas of lateral entorhinal cortex. It was not possible to identify unequivocally 1 rat as having the largest or the smallest amount of damage to PH cortex. Figure 1 shows drawings indicating the area of tissue damage produced by one of the largest (gray) and one of the smallest (black) of these lesions. Figure 2 shows photomicrographs of a representative lesion in anterior (Panels A and B) and posterior (Panels C and D) areas of

4 ANTERIOR THALAMIC AND PARAHIPPOCAMPAL LESIONS 599 perirhinal cortex and still more posterior areas of postrhinal cortex (lower power in Panels E and F) on the left (Panels A, C, and E) and right (Panels B, D, and F) sides of the brain. Lesions were characterized by areas of gliosis lacking neurons and tissue loss, including cavitation and the loss of cortical layers that normally separate the rhinal fissure from the striatum and lateral ventricle in more anterior areas and ventral hippocampus in more posterior areas of perirhinal cortex (Figure 2). Lesions to AT were associated with bilateral damage to the anteromedial, anterodorsal, anteroventral, and laterodorsal nuclei for all rats in the AT and AT PH groups. Medial portions of the interanteromedial nuclei appeared to be spared in all cases. Figure 3 shows drawings indicating the extent of damage produced by one of the largest (gray) and one of the smallest (black) thalamic lesions. AT lesions tended to encroach on adjacent portions of the intralaminar and (to a lesser extent) the mediodorsal nuclei in anterior areas. In more posterior regions, these nuclei were spared apart from dorsal areas of the centrolateral nucleus adjacent to the laterodorsal nucleus. The NMDA lesions were characterized by a necrotic core adjacent to the sites of infusion, surrounded by areas of gliosis containing few neurons (see Figure 4). For all rats in the AT and AT PH groups, damage was visible in more dorsal areas of hippocampus, along the tracks made by cannulas used to lesion thalamus. In 4 cases in the AT group (3 bilateral and 1 unilateral) and 6 in the AT PH group (3 bilateral and 3 unilateral), this damage extended beyond the cannula tracks within dorsal anterior hippocampus. Behavioral Analyses Postsurgical training took place in three blocks of 10 sessions. The first block of 10 sessions occurred from 3 to 4 weeks after surgery, the second block from 5 to 6, and the third block from 7 Figure 3. Reconstruction of one of the smallest (black) and one of the largest (gray) lesions of the anterior thalamic nuclei. Sections are 7.6 and 6.7 mm anterior to the interaural line in Paxinos and Watson (1998). Reprinted from The Rat Brain in Stereotaxic Coordinates, 4th ed., G. Paxinos and C. Watson, Figures 24 and 29, Copyright 1998, with permission from Elsevier Science. to 8 weeks after surgery. In each of these blocks, all groups exhibited a decrease in percent correct at longer retention intervals (see Figure 5). The performance of the control group was stable throughout the three blocks, varying from about 90% correct at short delays to about 80% correct at the longest delays, averaging 85.9% correct across delays during the first block of 10 sessions, 86.5% correct during the second block, and 86.4% correct during the third block. The performances of the three lesion groups tended to improve across blocks, particularly for responses made at short retention intervals (Figure 5). During the last two blocks, each of the lesions produced impairments that appeared to be delay-dependent, increasing in severity as the length of the retention interval was increased. There were no apparent interactions between the impairments produced by the AT and PH lesions. The effects of AT lesions alone (AT vs. sham control) were comparable to the effects of AT lesions in rats with PH lesions (AT PH vs. PH; see Figure 6). Likewise, the effects of PH lesions alone (PH vs. sham control) were comparable to the effects of PH lesions in rats with AT lesions (AT PH vs. AT). Results were analyzed for all 30 postsurgical sessions with a four-factor, mixed model analysis of variance (ANOVA) with AT (AT vs. no AT lesion) and PH (PH vs. no PH lesion) as betweengroup variables and delay (retention interval) and block (first, second, or third) of 10 postsurgical sessions as within-subject variables. There were significant effects of AT, F(1, 32) , p.01, and PH, F(1, 32) , p.01, lesions that did not interact (F 1). Thus, both lesions had significant effects that added without interacting. There was a significant effect of delay, F(7, 224) , p.01, that interacted with both the AT, F(7, 224) 2.323, p.03, and PH, F(7, 224) 3.094, p.01, lesions. Thus, the temporal decay at longer retention intervals was significant and was affected significantly by both the AT and PH lesions (Figure 5). The three-way interaction between AT lesion, PH lesion, and delay was not significant, F(7, 224) 1.076, p.37. There was a significant effect of block, F(2, 64) 9.944, p.01, that interacted significantly with the effects of the AT, F(2, 64) 3.986, p.03, but not the PH (F 1) lesion. The interaction between block and delay was only marginally significant, F(14, 448) 1.623, p.07. None of the interactions between block and combinations of AT, PH, and delay approached significance. We performed separate ANOVAs for Sessions 1 10 to characterize acute recovery and for Sessions to characterize impairments during more chronic phases of recovery. During acute recovery (Sessions 1 10), there were significant effects of AT, F(1, 32) , p.01, and PH, F(1, 32) 7.422, p.02, that did not interact (F 1). There was also a significant effect of delay, F(7, 224) , p.01, that did not interact with AT, F(7, 224) 1.297, p.25; PH (F 1); or the combination of AT and PH (F 1). During more chronic recovery (Sessions 11 30), there were significant effects of AT, F(1, 32) 6.340, p.02, and PH, F(1, 32) 9.405, p.01, that did not interact (F 1), as well as a significant effect of delay, F(7, 224) , p.01. In chronic recovery, there was a significant interaction between PH and delay, F(7, 224) 2.932, p.01, but not between AT and delay F(7, 224) 1.663, p.12. To explore the influence of delay on the effects of AT and PH lesions during acute and chronic recovery, we analyzed simple main effects of these lesions at each retention interval for Sessions

5 600 MAIR, BURK, AND PORTER Figure 4. Photomicrographs in the coronal plane showing representative lesions of the anterior thalamus. Areas damaged by the lesion are outlined in black in low-power photomicrographs showing typical patterns of damage in anterior (A) and posterior (B) levels. Higher power photomicrographs (C and D) show the transition between areas damaged and areas spared. The locations of the higher powered photomicrographs are indicated by the arrows in the figures immediately above. The lesions were characterized by a necrotic core adjacent to the track of the cannula used to infuse the excitotoxin, surrounded by areas of gliosis containing few neurons having fairly sharp boundaries with areas that appeared undamaged by the treatment and The results of these analyses (see Table 1) showed evidence of delay-dependent impairment for both lesions during chronic, but not acute, recovery. During chronic recovery, AT lesions had significant effects only at the two longest delays. PH lesions spared performances at the two shortest delays (1.0, 1.6 s) while producing significant impairments at the three longest (10.0, 15.8, 25.1 s). Thus, although only PH lesions interacted significantly with delay during chronic recovery, both were associated with a pattern of impairment (deficits at long, but not short, delays) consistent with a delay-dependent memory impairment. Both AT and PH lesions were associated with damage to overlying parts of the brain that were penetrated by cannulas or electrodes during surgery (see Histological Analyses above). To assess the effects of this damage, we compared overall percent correct for rats with and without bilateral damage extending beyond cannula or electrode tracks. For AT lesions, rats with bilateral hippocampal damage performed slightly worse than rats without bilateral damage in the AT group (80.6% vs. 82.6% correct), however, this trend was reversed for the AT PH group (80.1% vs. 79.2% correct). For PH lesions, rats with bilateral damage to overlying cortex performed slightly better than rats without this damage for the PH group (80.3% vs. 83.9% correct), but again, this trend was reversed for the AT PH group (80.0% vs. 79.1% correct). Compared directly by t test, none of these differences approached significance ( ps.24). For purposes of comparison, the control group averaged 86.3% correct overall. Discussion Effects of AT Lesions on Allocentric Memory Lesions to AT that involved the anterior and laterodorsal nuclei interfered with the ability to perform varying-choice DNM. Initially, this impairment was delay-independent, but it changed to a delay-dependent pattern by the second block of 10 sessions (5 6 weeks after surgery) as the effects of AT lesions recovered to normal at short retention intervals (see Figure 5 and Table 1). These results provide evidence that AT lesions have lasting effects on the ability to remember information beyond the span of imme-

6 ANTERIOR THALAMIC AND PARAHIPPOCAMPAL LESIONS 601 lesions had more widespread effects at first, affecting other areas of thalamus that recovered over the first few weeks after surgery. Lesions involving adjacent areas of the intralaminar nuclei have been found to produce persistent delay-independent impairments of this task affecting performance equivalently at delays ranging from 0 to 20 s long (Mair et al., 1998). The varying-choice impairment is consistent with other evidence that the AT complex is necessary for allocentric memory, or the ability to remember location relative to external stimuli independent of body orientation (Vann, Brown, & Aggleton, 2000; Warburton & Aggleton, 1999; Warburton, Baird, Morgan, Muir, & Aggleton, 2001; Warburton et al., 1999). Egocentric and allocentric cues are confounded in tasks like the T maze, in which there is an invariant relationship between the locations of the choice arms and the directions animals must turn to reach them. Experimental studies have provided evidence that rats rely on both types of cues to perform T maze DNM (Dudchenko, 2001). In varyingchoice DNM, the locations and configuration of arms change randomly from trial to trial. Unlike the strategy required in a T maze, the direction of a correct response in varying-choice DNM cannot be determined until after the retention interval ends and gates open to reveal which of 6 possible arms is the S for that trial. Hippocampal lesions have been found to produce delaydependent impairments of varying-choice DNM, qualitatively similar (although larger in magnitude) to AT lesions (Figure 6). Thus hippocampal lesions reduced DNM accuracy by an average of 8.0%, increasing from 2.0% at the 1.0-s delay to 20.4% at the 25.1-s delay (Porter et al., 2000). Using identical procedures, we found that AT lesions reduced accuracy by an average of 4.2%, increasing from 0.9% at the 1.0-s delay to 7.7% at the 25.1-s delay. The hippocampus was damaged by cannulas used to produce AT lesions. Comparison of rats with and without bilateral hippocampal pathology showed little indication that this damage contributed importantly to the impairment of the AT or AT PH groups (see Figure 5. Percent correct delayed nonmatching as a function of retention interval for the control (Cnt), anterior thalamic (AT), parahippocampal (PH), and combined (ATPH) lesion groups. Results are plotted separately for Sessions 1 10 (3 4 weeks after surgery), (5 6 weeks after surgery), and (7 8 weeks after surgery). Error bars indicate the standard error of the mean. diate memory. It is not certain why AT lesions affected performance at short delays during early postsurgical sessions. It is possible that these rats had difficulty at the outset performing procedural aspects of the task, even though they had received extensive training prior to surgery. It is also possible that AT Figure 6. Effects of different hippocampal system lesions on varyingchoice delayed nonmatching plotted as percentage of control performance. Results from the present study are plotted as anterior thalamic (AT), parahippocampal (PH), and combined (ATPH) to show lesion effects as a percentage of unlesioned controls. Results are plotted as AT* (ATPH as a percentage of the PH group) and PH* (ATPH as a percentage of the AT group) to compare effects of lesions in the combined lesion group. Results from previous studies are plotted as percentage of unlesioned control performance for hippocampal lesions (H1: Mair et al., 1998; H2: Porter et al., 2000). Error bars indicate standard error of the mean.

7 602 MAIR, BURK, AND PORTER Table 1 Simple Main Effects at Each Retention Interval During Acute and Chronic Phases of Recovery Anterior thalamic lesion Parahippocampal lesion RI F(1, 32) p Power F(1, 32) p Power Acute recovery (Sessions 1 10) * * * * * * * * * Chronic recovery (Sessions 11 30) * * * * * *.934 Note. RI retention interval (in seconds). *p.05. Behavioral Analyses above). Our results suggest that AT lesions partially disrupt the contributions of hippocampus to spatial memory. Other studies have shown that AT lesions impair aspects of spatial memory mediated by the hippocampus (Aggleton & Brown, 1999; Warburton et al., 2001), although there is disagreement over whether the effects of AT lesions are smaller than (Aggleton et al., 1995; Byatt & Dalrymple-Alford, 1996; Sutherland & Rodriguez, 1989) or comparable in magnitude to (Aggleton & Sahgal, 1993; Warburton & Aggleton, 1999; Warburton et al., 1997, 1999) fimbria fornix or hippocampal lesions. There are several factors that may have limited the extent of impairment associated with AT lesions in the present study. First, we trained rats prior to surgery to avoid confounding impairments related to learning the DNM task with those that reflect an inability to remember spatial information over the span of working memory. Sutherland and Rodriguez (1989) found that presurgical training can eliminate the effects of AT lesions on a water maze task, although this finding was questioned in a later study by Warburton et al. (1999). The normal performance of all lesion groups at short retention intervals during chronic phases of testing confirms their ability to perform the present DNM task when mnemonic demands were reduced. Second, we trained rats for a relatively large number of sessions over defined periods of time after surgery to differentiate between acute and chronic effects of lesions. We found that the delay-dependent effects of AT lesions persisted throughout training, whereas deficits at short retention intervals disappeared after the first block of 10 sessions (Figure 5). Had the experiment been stopped after the first block of sessions, our measures of AT impairment would have been exacerbated by these more acute effects. Third, the effects of AT lesions may have been limited by the range of retention intervals tested. Some reports have indicated that the effects of AT lesions are only apparent at longer retention intervals (Beracochea & Jaffard, 1994; Beracochea, Jaffard, & Jarrard, 1989). It should be noted, however, that given comparable training, rats with hippocampal lesions performed varying-choice DNM at a level approaching a floor effect (62.8% correct) at the 25.1-s delay (Porter et al., 2000). Thus, the range of retention intervals tested was sufficient to demonstrate substantial hippocampal impairment. Effects of Parahippocampal Lesions In rats with AT lesions, hippocampus remains connected to association areas of neocortex by projections involving PH cortex (Burwell & Amaral, 1998a, 1998b). To the extent that hippocampal function is spared by AT lesions, it seems likely that pathways involving PH cortex play an important role in this function. We found modest effects of PH lesions alone (Figures 5 and 6), consistent with reports that PH lesions have limited effects on allocentric spatial memory compared with hippocampal lesions (Aggleton, Vann, Oswald, & Good, 2000; Liu & Bilkey, 1998, 2001). When combined, the effects of the AT and PH lesions added without interaction to produce deficits comparable to those produced by complete hippocampal lesions (Figure 6). These results suggest that AT and PH areas make separate contributions essential for hippocampus-dependent spatial memory. The possibility remains, however, that this deafferentation had secondary effects on internal hippocampal circuits that may have contributed to these impairments. Aggleton and his colleagues have argued that hippocampusrelated projections to AT mediate aspects of memory that are distinct from those of perirhinal cortex (Aggleton & Brown, 1999;

8 ANTERIOR THALAMIC AND PARAHIPPOCAMPAL LESIONS 603 Bussey, Duck, Muir, & Aggleton, 2000; Vann, Brown, Erichsen, & Aggleton, 2000; Wan, Aggleton, & Brown, 1999). Our finding that the effects of AT and PH lesions add without interaction are consistent with this view, although our lesions affected more than perirhinal cortex. There is controversy over the importance of perirhinal cortex for spatial memory (Bussey, Muir, & Aggleton, 1999; Bussey et al., 2000; Liu & Bilkey, 1998, 1999, 2001; Machin, Vann, Muir, & Aggleton, 2002; Wiig & Bilkey, 1994; Wiig & Burwell, 1998). However, there is more general agreement on the importance of entorhinal and postrhinal areas also affected by PH lesions (Aggleton et al., 2000; Good & Honey, 1997; Nagahara, Otto, & Gallagher, 1995; Otto, Wolf, & Walsh, 1997). Lesions to PH were associated with damage to overlying areas of cortex through which electrodes were lowered during surgery. There are several reasons to discount the possibility that this unintended damage contributed importantly to the impairments produced by PH lesions. First, the impairment of the PH group was limited and (as noted above), consistent with previous results for parahippocampal lesions. Second, PH-lesioned rats were able to perform the DNM task when memory demands were minimized at short delays and exhibited a hippocampal-like impairment when delays were longer. Third, the combined AT PH lesion produced an impairment that approached, but did not exceed, previous results for hippocampal lesions. Thus, the impairment observed can be explained entirely by hippocampal dysfunction produced by surgical isolation from PH cortex. Fourth, there were no clear-cut differences in performance associated with bilateral damage to overlying cortex for rats in the PH and AT PH groups (see Behavioral Analyses above). Clinical Implications We found that AT lesions, including the lateral dorsal nucleus, produced DNM impairments that persisted through 30 training sessions and 8 weeks of recovery from surgery. AT lesions alone had relatively modest effects. Combined with PH lesions, they produced a more substantial impairment comparable to earlier results for hippocampal lesions. Although these findings demonstrate a significant role for AT, they also suggest that AT lesions only partially disrupt the contributions of hippocampus to spatial memory. Clinical studies have shown that diencephalic amnesia can produce global memory impairments comparable in severity to medial temporal lobe amnesia (Butters & Cermak, 1980; Graff- Radford et al., 1990; Mair, 1994; Mair et al., 1979; Partlow et al., 1992; Talland, 1965; von Cramon et al., 1985). The limited effects of AT lesions in the present study coupled with evidence that AT lesions spare other aspects of memory, such as object recognition (Aggleton & Brown, 1999; Aggleton et al., 1995; Wilton, Baird, Muir, Honey, & Aggleton, 2001), make it difficult to reconcile the global memory impairments observed in clinical cases with the hypothesis that hippocampus-related projections to the AT represent the main site of pathology in diencephalic amnesia. The AT nuclei themselves are not a consistent site of pathology in diencephalic amnesia. They have been implicated indirectly by evidence that the medial mammillary nuclei are damaged consistently in Korsakoff s disease (Mair et al., 1979; Victor et al., 1989) and that the area of the mammillothalamic tract is often involved in patients with memory loss after a thalamic infarct (Graff- Radford et al., 1990; Van der Werf et al., 2000; von Cramon et al., 1985). Experimental studies have not shown lesions in either of these sites to have substantial and lasting effects on memory. Lesions restricted to the medial mammillary nuclei have been found to have little effect on measures of spatial memory compared with lesions involving the AT or fornix (Aggleton, Keith, & Sahgal, 1991; Aggleton et al., 1995; Aggleton & Sahgal, 1993; Harper, Dalrymple-Alford, & McLean, 1993; Harper, McLean, & Dalrymple-Alford, 1994; Jarrard, Okaichi, Steward, & Goldschmidt, 1984; Mair & Lacourse, 1992; Sziklas & Petrides, 1998). Thomas and Gash (1985) reported that complete mammillothalamic tract lesions produce severe but temporary impairments of spatial DNM trained in a T maze that recovered in most cases within 60 trials of training after surgery. Partial mammillothalamic tract lesions were without effect on this task. Thus, although lesions of the AT nuclei can produce persistent signs of spatial memory impairment, their implication in amnesia rests on the unsubstantiated assumption that lesions restricted to the medial mammillary nuclei or the mammillothalamic tract would have comparable effects. The midline and intralaminar nuclei provide an alternative account for thalamic amnesia. The intralaminar nuclei are a part of the ascending reticular formation that has been implicated in the activation of functions in cortex (Llinas, Leznik, & Urbano, 2002; Steriade, Contreras, Amzica, & Timofeev, 1996; Sukov & Barth, 2001) and striatum (Matsumoto, Minamimoto, Graybiel, & Kimura, 2001; Minamimoto & Kimura, 2002). Clinical studies have associated intralaminar lesions with impairments in memory, attention, and intentional motor function (Exner, Weniger, & Irle, 2001; Mair, 1994; Purpura & Schiff, 1997; Van der Werf et al., 2000). Experimental studies suggest two main differences between the effects of intralaminar and hippocampal lesions on memory: Intralaminar lesions affect a broader range of memory tasks and are associated with delay-independent impairments of radial maze DNM (Mair et al., 1998; Zhang, Burk, Glode, & Mair, 1998). The present results are consistent with this last distinction. Although the interaction of AT lesions with retention interval was not significant, the pattern of impairment during chronic recovery (deficits at long, but not short, intervals) is indicative of delaydependent impairment consistent with hippocampal dysfunction. Nevertheless, the limited extent of impairment and the lack of a significant interaction argue for caution in considering the possible contributions of intralaminar damage to the effects of AT lesions in the present study. References Aggleton, J. P., & Brown, M. W. (1999). Episodic memory, amnesia, and the hippocampal-anterior thalamic axis. Behavioral and Brain Sciences, 22, Aggleton, J. P., Keith, A. B., & Sahgal, A. (1991). Both fornix and anterior thalamic, but not mammillary, lesions disrupt delayed non-matching-toposition memory in rats. Behavioural Brain Research, 44, Aggleton, J. P., Neave, N., Nagle, S., & Hunt, P. R. (1995). A comparison of the effects of anterior thalamic, mamillary body and fornix lesions on reinforced spatial alternation. Behavioural Brain Research, 68, Aggleton, J. P., & Sahgal, A. (1993). The contribution of the anterior thalamic nuclei to anterograde amnesia. Neuropsychologia, 31, Aggleton, J. P., Vann, S. D., Oswald, C. J., & Good, M. (2000). Identifying cortical inputs to the rat hippocampus that subserve allocentric spatial

9 604 MAIR, BURK, AND PORTER processes: A simple problem with a complex answer. Hippocampus, 10, Bentivoglio, M., Aggleton, J. P., & Mishkin, M. (1997). The thalamus and memory formation. In M. Steriade, E. G. Jones, & D. A. McCormick (Eds.), Thalamus (Vol II, pp ). Oxford, UK: Elsevier. Beracochea, D. J., & Jaffard, R. (1994). Effects of anterior thalamic lesions on spatial memory in mice. NeuroReport, 5, Beracochea, D. J., Jaffard, R., & Jarrard, L. E. (1989). Effects of anterior or dorsomedial thalamic ibotenic lesions on learning and memory in rats. Behavioral and Neural Biology, 51, Burk, J. A., & Mair, R. G. (1998). Thalamic amnesia reconsidered: Excitotoxic lesions of the intralaminar nuclei, but not the mediodorsal nucleus disrupt place DMTS performance in the rat (Rattus norvegicus). Behavioral Neuroscience, 112, Burwell, R. D. (2001). Borders and cytoarchitecture of the perirhinal and postrhinal cortices in the rat. Journal of Comparative Neurology, 437, Burwell R. D., & Amaral, D. G. (1998a). Cortical afferents of the perirhinal, postrhinal, and entorhinal cortices of the rat. Journal of Comparative Neurology, 398, Burwell R. D., & Amaral, D. G. (1998b). Perirhinal and postrhinal cortices of the rat: Interconnectivity and connections with the entorhinal cortex. Journal of Comparative Neurology, 398, Bussey, T. J., Duck, J., Muir, J. L., & Aggleton, J. P. (2000). Distinct patterns of behavioural impairments resulting from fornix transection or neurotoxic lesions of the perirhinal and postrhinal cortices in the rat. Behavioural Brain Research, 111, Bussey, T. J., Muir, J. L., & Aggleton, J. P. (1999). Functionally dissociating aspects of event memory: The effects of combined perirhinal and postrhinal cortex lesions on object and place memory in the rat. Journal of Neuroscience, 19, Butters, N., & Cermak, L. S. (1980). Alcoholic Korsakoff s syndrome. New York: Academic Press. Byatt, G., & Dalrymple-Alford, J. C. (1996). Both anteromedial and anteroventral thalamic lesions impair radial-maze learning in rats. Behavioral Neuroscience, 110, Collins, R. C., & Olney, J. W. (1982, October 8). Focal cortical seizures cause distant thalamic lesions. Science, 218, Delay, J., Brion, S., & Elissalde, B. (1958). Corps mamillaires et syndrome Korsakoff: Etude anatomique de huit cas de syndrome Korsakoff d origine alcoholique sans alterations significative du cortex cerebral [Mammillary bodies and Korsakoff s syndrome: Anatomical study of eight cases of alcoholic Korsakoff s syndrome without significant alteration of the cerebral cortex]. Press Medicale, 66, , Dudchenko, P. A. (2001). How do animals actually solve the T maze? Behavioral Neuroscience, 115, Eichenbaum, H. (1997). Declarative memory: Insights from cognitive neurobiology. Annual Review of Psychology, 48, Exner, C., Weniger, G., & Irle, E. (2001). Implicit and explicit memory after focal thalamic lesions. Neurology 57, Gaffan, D., Eacott, M. J., & Simpson, E. L. (2000). Perirhinal cortex ablation in rats selectively impairs object identification in a simultaneous visual comparison task. Behavioral Neuroscience, 114, Good, M., & Honey, R. C. (1997). Dissociable effects of selective lesions to hippocampal subsystems on exploratory behavior, contextual learning, and spatial learning. Behavioral Neuroscience, 111, Graff-Radford, N. R., Tranel, D., van Hoesen, G. W., & Brandt, J. P. (1990). Diencephalic amnesia. Brain 113, Harper, D. N., Dalrymple-Alford, J. C., & McLean, A. P. (1993). The effect of medial septal and mammillary body lesions on the serial position curve in rats. Psychobiology, 21, Harper, D. N., McLean, A. P., & Dalrymple-Alford, J. C. (1994). Forgetting in rats following medial septum or mammillary body damage. Behavioral Neuroscience, 108, Jarrard, L. E., Okaichi, H., Steward, O., & Goldschmidt, R. B. (1984). On the role of hippocampal connections in the performance of place and cue tasks: Comparisons with damage to hippocampus. Behavioral Neuroscience, 98, Jones, E. G. (1985). The thalamus (pp ). New York: Plenum Press. Liu, P., & Bilkey, D. K. (1998). Excitotoxic lesions centered on perirhinal cortex produce delay-dependent deficits in a test of spatial memory. Behavioral Neuroscience, 112, Liu, P., & Bilkey, D. K. (1999). The effect of excitotoxic lesions centered on the perirhinal cortex in two versions of the radial arm maze task. Behavioral Neuroscience, 113, Liu, P., & Bilkey, D. K. (2001). The effect of excitotoxic lesions centered on the hippocampus or perirhinal cortex in object recognition and spatial memory tasks. Behavioral Neuroscience, 115, Llinas, R. R., Leznik, E., & Urbano, F. J. (2002). Temporal binding via cortical coincidence detection of specific and nonspecific thalamocortical inputs: A voltage-dependent dye-imaging study in mouse brain slices. Proceedings of the National Academy of Sciences, USA, 99, Machin, P., Vann, S. D., Muir, J. L., & Aggleton, J. P. (2002). Neurotoxic lesions of the rat perirhinal cortex fail to disrupt the acquisition or performance of tests of allocentric spatial memory. Behavioral Neuroscience, 116, Mair, R. G. (1994). On the role of thalamic pathology in diencephalic amnesia. Reviews in the Neurosciences, 5, Mair, R. G., Burk, J. A., & Porter, M. C. (1998). In the rat, lesions of frontal cortex, hippocampus, and the intralaminar thalamic nuclei have distinct effects on remembering. Behavioral Neuroscience, 112, Mair, R. G., & Lacourse, D. M. (1992). Radiofrequency lesions of thalamus produce delayed non-matching to sample impairments comparable to pyrithiamine-induced thiamine deficiency. Behavioral Neuroscience, 106, Mair, W. P. G., Warrington, E. K., & Weiskrantz, L. (1979). Memory disorders in Korsakoff s psychosis: A neuropathological and neuropsychological investigation of two cases. Brain, 102, Markowitsch, H. J. (1982) Thalamic mediodorsal nucleus and memory: A critical evaluation of studies in animals and man. Neuroscience and Biobehavioral Reviews, 6, Matsumoto, N., Minamimoto, T., Graybiel, A. M., & Kimura, M. (2001). Neurons in the thalamic CM-Pf complex supply striatal neurons with information about behaviorally significant sensory events. Journal of Neurophysiology, 85, Minamimoto, T., & Kimura, M. (2002). Participation of the thalamic CM-Pf complex in attentional orienting. Journal of Neurophysiology, 87, Nagahara, A. H., Otto, T., & Gallagher, M. (1995). Entorhinal-perirhinal lesions impair performance of rats on two versions of place learning in the Morris water maze. Behavioral Neuroscience, 109, 3 9. Otto, T., Wolf, D., & Walsh, T. J. (1997). Combined lesions of perirhinal and entorhinal cortex impair rat s performance in two versions of the spatially guided radial-arm maze. Neurobiology of Learning and Memory, 68, Partlow, G. D., del Carpio-O Donovan, R., Melanson, D., & Peters, T. M. (1992). Bilateral thalamic glioma: Review of eight cases with personality change and mental deterioration. American Journal of Neuroradiology, 13, Paxinos, G., & Watson, C. (1998). The rat brain in stereotaxic coordinates (4th ed.). San Diego, CA: Academic Press. Porter, M. C., Burk, J. A., & Mair, R. G. (2000). A comparison of the effects of hippocampal or prefrontal cortical lesions on three versions of

10 ANTERIOR THALAMIC AND PARAHIPPOCAMPAL LESIONS 605 delayed non-matching-to-sample based on positional or spatial cues. Behavioural Brain Research, 109, Porter, M. C., & Mair, R. G. (1997). The effects of frontal cortical lesions on remembering depend on the procedural demands of tasks performed in the radial arm maze. Behavioural Brain Research, 87, Purpura, K. P., & Schiff, N. D. (1997). The thalamic intralaminar nuclei: A role in visual awareness. The Neuroscientist, 3, Squire, L. R., Amaral, D. G., Zola-Morgan, S., Kritchevsky, M., & Press, G. (1989). Description of brain injury in the amnesic patient N. A. based on magnetic resonance imaging. Experimental Neurology, 105, Squire, L. R., Knowlton, B., & Mussen, G. (1993). The structure and organization of memory. Annual Review of Psychology, 44, Steriade, M., Contreras, D., Amzica, F., & Timofeev, I. (1996). Synchronization of fast (30 40 Hz) spontaneous oscillations in intrathalamic and thalamocortical networks. Journal of Neuroscience, 16, Sukov, W., & Barth, D. S. (2001). Cellular mechanisms of thalamically evoked gamma oscillations in auditory cortex. Journal of Neurophysiology, 85, Sutherland, R. J., & Rodriguez, A. J. (1989). The role of the fornix/fimbria and some related subcortical structures in place learning and memory. Behavioural Brain Research, 32, Sziklas, V., & Petrides, M. (1998). Memory and the region of the mammillary bodies. Progress in Neurobiology, 54, Talland, G. A. (1965). Deranged memory. New York: Academic Press. Thomas, G. J., & Gash, D. M. (1985). Mammillothalamic tracts and representational memory. Behavioral Neuroscience, 99, Van der Werf, Y. D., Witter, M. P., Ulyings, H. B., & Jolles, J. (2000). Neuropsychology of infarctions in the thalamus: A review. Neuropsychologia, 38, Vann, S. D., Brown, M. W., & Aggleton, J. P. (2000). Fos expression in the rostral thalamic nuclei and associated cortical regions in response to different spatial memory tests. Neuroscience, 101, Vann, S. D., Brown, M. W., Erichsen, J. T., & Aggleton, J. P. (2000). Fos imaging reveals differential patterns of hippocampal and parahippocampal subfield activation in rats in response to different spatial memory tests. Journal of Neuroscience, 20, Victor, M., Adams, R. D., & Collins, G. H. (1989). The Wernicke- Korsakoff syndrome. Philadelphia: F. A. Davis. von Cramon, D. Y., Hebel, N., & Schuri, U. (1985). A contribution to the anatomical basis of thalamic amnesia. Brain, 108, Wan, H., Aggleton, J. P., & Brown, M. W. (1999). Different contributions of the hippocampus and perirhinal cortex to recognition memory. Journal of Neuroscience, 19, Warburton, E. C., & Aggleton, J. P. (1999). Differential deficits in the Morris water maze following cytotoxic lesions of the anterior thalamus and fornix transection. Behavioural Brain Research, 98, Warburton, E. C., Baird, A. L., & Aggleton, J. P. (1997). Assessing the magnitude of the allocentric spatial deficit associated with complete loss of the anterior thalamic nuclei in rats. Behavioural Brain Research, 87, Warburton, E. C., Baird, A., Morgan, A., Muir, J. L., & Aggleton, J. P. (2001). The conjoint importance of the hippocampus and anterior thalamic nuclei for allocentric spatial learning: Evidence from a disconnection study in the rat. Journal of Neuroscience, 21, Warburton, E. C., Morgan, A., Baird, A. L., Muir, J. L., & Aggleton, J. P. (1999). Does pretraining spare the spatial deficit associated with anterior thalamic damage in rats? Behavioral Neuroscience, 113, Wiig, K. A., & Bilkey, D. K. (1994). Perirhinal cortex lesions in rats disrupt performance in a spatial DNMS task. NeuroReport, 5, Wiig, K. A., & Burwell, R. D. (1998). Memory impairment on a delayed non-matching-to-position task after lesions of the perirhinal cortex in the rat. Behavioral Neuroscience, 112, Wilton, L. A., Baird, A. L., Muir, J. L., Honey, R. C., & Aggleton, J. P. (2001). Loss of thalamic nuclei for head direction impairs performance on spatial memory tasks in rats. Behavioral Neuroscience, 115, Winn, P. (1991). Excitotoxins as tools for producing brain lesions. In P. M. Conn (Ed.), Lesions and transplantation (pp ). San Diego, CA: Academic Press. Young, H. L., Stevens, A. A., Converse, E., & Mair, R. G. (1996). A comparison of temporal decay in place memory tasks in rats with lesions of thalamus, frontal cortex, and the hippocampal system. Behavioral Neuroscience, 110, Zhang, Y. P., Burk, J. A., Glode, B. M., & Mair, R. G. (1998). The effects of thalamic and olfactory cortical lesions on continuous olfactory DNMTS and olfactory discrimination in the rat (Rattus norvegicus). Behavioral Neuroscience, 112, Received June 25, 2002 Revision received November 6, 2002 Accepted December 30, 2002